Advances in Insect Physiology, Volume 25

Advances in Insect Physiology

Volume 25

This Page Intentionally Left Blank

Advances in Insect Physiology edited by

P. D. EVANS Department of Zoology, The University Cambridge, England

Volume 25

ACADEMIC PRESS Harcourt Brace & Company, Publishers London San Diego New York Sydney Toronto Tokyo

Boston

ACADEMIC PRESS LIMITED 24-28 Oval Road London NW17DX United States Edition published by ACADEMIC PRESS INC. San Diego, CA 92101 This book is printed on acid-free paper Copyright 01994 by ACADEMIC PRESS LIMITED

All Rights Reserved

No part of this book may be reproduced in any form by photostat, microfilm, or any other means, without written permission from the publishers A catalogue record for this book is available from the British Library ISBN 0-1 2-0242257 Typeset by Keyset Composition, Colchester, Essex Printed and bound in Great Britain by TJ Press Ltd, Padstow, Cornwall

Contents Contributors Advances in Insect Virology L. A. KING, R. D. POSSEE, D. S. HUGHES, A. E. ATKINSON, C. P. PALMER, S. A. MARLOW, J. M. PICKERING, K. A. JOYCE, A. M. LAWRIE, D. P. MILLER, D. J. BEADLE Genetic Mechanisms of Early Neurogenesis in Drosophila melanogaster J. A. CAMPOS-ORTEGA

vii

1

75

Molecular Biology of the Honeybee

R. F. A. MORITZ

105

Information Processing in the Insect Ocellar System: Comparative Approaches t o the Evolution of Visual Processing and Neural Circuits M. MlZUNAMl

151

Allatostatins: Identification, Primary Structures, Functions and Distribution 6. STAY, S. S. TOBE, W. G. BENDENA

267

Index

339

Color Plates are located between pages 152 and 153.

This Page Intentionally Left Blank

Contributors A. E. Atkinson

School of Biological and Molecular Sciences, Oxford Brookes University, Gipsy Lane Campus, Oxford OX3 OBP, U K D. J. Beadle

School of Biological and Molecular Sciences, Oxford Brookes University, Gipsy Lane Campus, Oxford OX3 OBP, UK

W. G. Bendena Department of Biology, University of Iowa, Iowa City, ZA 52242-1324, USA J. A. Carnpos-Ortega

Institut fur Entwicklungsbiologie, Universitat zu Koln, 0-50931 Koln, Germany D. S. Hughes

School of Biological and Molecular Sciences, Oxford Brookes University, Gipsy Lane Campus, Oxford OX3 OBP, UK K. A. Joyce

School of Biological and Molecular Sciences, Oxford Brookes University, Gipsy Lane Campus, Oxford OX3 OBP, U K L. A. King

School of Biological and Molecular Sciences, Oxford Brookes University, Gipsy Lane Campus, Oxford OX3 OBP, UK A. M. Lawrie

School of Biological and Molecular Sciences, Oxford Brookes University, Gipsy Lane Campus, Oxford OX3 OBP, U K

S. A. Marlow

School of Biological and Molecular Sciences, Oxford Brookes University, Gipsy Lane Campus, Oxford OX3 OBP, U K Current address: Royal Free Medical School, UK D. P. Miller

School of Biological and Molecular Sciences, Oxford Brookes University, Gipsy Lane Campus, Oxford OX3 OBP, U K M . Mizunami

Laboratory of Neuro-Cybernetics, Research Institute for Electronic Science, Hokkaido University, Sapporo 060, Japan R. F. A. Moritz

Institut fur Biologie, Technische Universitat Berlin, Franklinstr 28/29, 10587 Berlin, Germany C. P. Palmer

School of Biological and Molecular Sciences, Oxford Brookes University, Gipsy Lane Campus, Oxford OX3 OBP, U K Current address: Department of Microbiology, University of Reading, UK J. M. Pickering

School of Biological and Molecular Sciences, Oxford Brookes University, Gipsy Lane Campus, Oxford OX3 OBP, UK R. D. Possee

NERC Institute of Virology & Environmental Microbiology, Mansfield Road, Oxford, U K B. Stay

Department of Biology, University of Iowa, Iowa City, IA 52242-1324, U S A S. S. Tobe

Department of Zoology, University of Toronto, Toronto, Canada

Advances in Insect Virology Linda A. King,a Robert D. Possee! David S. Hughes: Allan E. Atkinson,c Christopher P. Palmer,d Susan A. Marlow,8 Jason M . Pickering,f Kirsti A. Joyce,a Alison M. Lawrie,a Davin P. MilleP and David J. Beadlea a School

of Biological and Molecular Sciences, Oxford Brookes University, Oxford, UK NERC Institute of Virology and Environmental Microbiology, Mansfield Road, Oxford, UK Current address: Department of Biology, University of California, San Diego, California, USA Current address: Department of Microbiology, University of Reading, Reading, UK Current address: Royal Free Hospital Medical School, London, UK 'Current address: St Mary's Hospital Medical School, London, UK Correspondence to L. A. King, School of Biological and Molecular Science, Oxford Brookes University, Gipsy Lane Campus, Oxford OX3 OBe UK

1 Introduction 2 2 Baculoviruses 2 2.1 Introduction 2 2.2 Isolation and host range 2 2.3 Structure and classification 3 2.4 Baculovirus replication in vivo 5 2.5 Transmission of baculoviruses between hosts 6 2.6 Baculovirus replication in vitro 9 2.7 Biological control of insect pests 15 2.8 Baculovirus expresssion vectors 22 3 Entomopoxviruses 29 3.1 Isolation and host range 29 3.2 Structure and classification 30 3.3 Replication cycle in insects 33 3.4 Molecular studies 34 3.5 Replication in vitro 36 3.6 Biological control 38 4 Iridescent viruses 38 4.1 Classification, isolation and host range 38 4.2 Virus structure 39 4.3 Replication cycle 40 4.4 Molecular studies 42 4.5 Biological control 42 5 RNA viruses of insects 43 5.1 Introduction 43 ADVANCES IN INSECT PHYSIOLOGY VOL 25 ISBN &124)24225-7

0

Copyright 1994 Academic Press Limited All rights ofreproducnon in any form reserved

2

L. A. KING e t a / .

5.2 Nodaviridae 43 5.3 Tetraviridae (Nudaurelia p virus group) 48 5.4 Picornaviridae 50 References 54

1 Introduction

This chapter describes current advances that have been made in a number of aspects of insect virology in the past few years. As in all such reviews, the baculoviruses feature most dominantly, as it is this group of insect viruses that forms the focus of attention for the majority of insect virologists and molecular biologists. However, a number of recent advances have been made in our understanding of the replication of other insect viruses, notably the entomopoxviruses. This chapter does not aim to provide a comprehensive treatise on every aspect of insect virology, and at the beginning of each section the reader is referred to other recent review articles for further information as necessary. 2 Baculoviruses

2.1

INTRODUCTION

This section provides an overview of baculovirus biology and molecular biology, and focuses on progress in these areas. In particular, we describe advances in the use of baculoviruses as expression vectors of foreign genes, in the development of genetically modified insecticides, vertical transmission, and aspects of interactions with host cells including the phenomenon of apoptosis. Other recent reviews provide detailed accounts of the use of baculoviruses as insecticides (Entwistle and Evans, 1985; Podgewaite, 1985; Bilimoria, 1986; Huber, 1986; Possee et al., 1993; Vlak, 1993a,b), of their biology (Granados, 1980; Mazzone, 1985; Granados and Williams, 1986; Kelly, 1987; Volkman and Keddie, 1990; Cory, 1993), molecular biology (Blissard and Rohrmann, 1990) and their use as expression vectors (Luckow and Summers, 1988; Miller, 1988; Maeda, 1989a; Atkinson et al., 1990; Bishop and Possee, 1990; Fraser, 1989; King and Possee, 1992; O’Reilly er al., 1992). 2.2

ISOLATION AND HOST RANGE

Baculoviruses have only been isolated from invertebrates. Most examples have been found in insect species, but there are some reports of baculoviruses which are pathogenic for crustacea (Anderson and Prior, 1992). Baculovirus infections have been described in over 600 species of insects

ADVANCES IN INSECT VIROLOGY

3

including Lepidoptera (butterflies and moths), Hymenoptera (sawflies and wasps), Diptera (flies), Coleoptera (beetles) and Trichoptera (caddis flies) (Martignoni and lwai, 1986). The diseases that they can produce in insect pest populations have made them an obvious choice for use as biological insecticides (Entwistle and Evans, 1985). It is only in the last 10 years that the development of the baculovirus expression vector system has overshadowed, but not replaced, the original reason for studying these viruses. 2.3

STRUCTURE AND CLASSIFICATION

Baculoviruses have a large, double-stranded, covalently closed, circular DNA genome of between 88 and 200 kbp (Arif, 1986). This is associated with a highly basic (arginine-rich) protein of 6.5 kDa (Tweeten et al., 1980; Kelly et al., 1983; Wilson et al., 1987; Russell and Rohrmann, 1990b; Maeda et al., 1991). The DNA-protein complex is contained by a rod-shaped nucleocapsid comprising a 39 kDa capsid protein (Pearson et a f . , 1988; Blissard et al., 1989; Thiem and Miller, 1989) and an 87 kDa capsid protein (Mueller et al., 1990). Other structural components almost certainly remain to be identified. The size of the virus genome determines the length of the nucleocapsid, which may be 200-400 nm. The width remains constant at about 36 nm (Fraser, 1986). One or more nucleocapsids are further packaged within a single lipoprotein envelope to form the virus particle or virion. These structures may be occluded within a crystalline matrix, referred to as a polyhedron or granule (see below). Baculoviruses used to be classified into three subgroups (Matthews, 1982). This system has recently been superseded by classification into two subfamilies, the Eubaculovirinae and the Nudibaculovirinae (Table 1) (Francki et al., 1991). The former contains the nuclear polyhedrosis virus (NPV) and granulosis virus (GV) genera. The NPV genus is further subdivided into two subgenera: multiple nucleocapsids per envelope (MNPV) and single nucleocapsids per envelope (SNPV). The Nudibaculovirinae comprises one genus, the non-occluded baculoviruses, which do not produce a crystalline protein matrix around the virus particles. The Autographa californica (Ac) NPV is the type species of the subgenus MNPV and the Bombyx mori (Bm) NPV is the type species of the subgenus SNPV. In each case, several multiply or singly enveloped virions are embedded in a proteinaceous occlusion body or polyhedron. Polyhedra consist largely of a single protein (polyhedrin) of about 30 kDa (Hooft van Iddekinge et al., 1983; reviewed by Rohrmann, 1986, 1992), and form in the nucleus of infected cells. Polyhedra are large structures, ranging in size from 1 to 1 5 p m in diameter with an outer envelope which appears to confer additional strength and protection (Vlak et a f . , 1988; Williams et al., 1989). It was suggested that the polyhedrin envelope (PE) was carbohydrate in

L. A. KING et a/.

4

TABLE 1 Classification of the Baculoviridae ~

English vernacular name Baculoviruses

Nuclear polyhedrosis virus (NPV) Multiple nucleocapsids per envelope (MNPV) Single nucleocapsids per envelope Granulosis viruses (GV) Non-occluded baculoviruses Non-occluded baculoviruses (NOB)

Taxonomic status (international name)

~

Type species

Family (Baculoviridae) Subfamily (Eubaculovirinae) Genus Subgenus

Autographa californica (Ac) MNPV

Subgenus

Bombyx mori (Bm) SNPV

Genus Subfamily (Nudibuculovirinae) Genus

Plodia interpunctella (Pi) G V Heliothis zea (Hz) NOB

nature (Minion et al., 1979). More recent evidence supports the view that there is a protein component in the PE (Whitt and Manning, 1988). Gombart et al. (1989) mapped a gene encoding the PE protein in the AcMNPV and Orgyia pseudotsugata OpMNPV genomes. Studies with immunoelectron microscopy have shown that the PE protein is a major component of the PE (Russell and Rohrmann, 1990a). The PE is sensitive to protease, suggesting that protein forms a major part of the structure (Russell and Rohrmann, 1990a). Virions that have been released from polyhedra are known as polyhedra-derived virus (PDV), whereas virions that are released from cells without occlusion are called extracellular virus (ECV). The type species of the granulosis virus genus is the Plodia interpunctella GV. Granulosis viruses contain one virion (singly enveloped nucleocapsid) per virus occlusion body or granule. Granulin, the major granule protein, is similar in function to polyhedrin (Longworth et al., 1972; Akiyoshi et al., 1985; Chakerian et al., 1985). The type species of the non-occluded baculoviruses (NOB) is Heliothis zea NOB. These viruses are composed of singly enveloped nucleocapsids, but as the name suggests, they are not further packaged into occlusion bodies. Baculoviruses are usually named after the host from which they were isolated. While this svstem is convenient, it ignores the genetic relatedness

ADVANCES IN INSECT VIROLOGY

5

of viruses from different species. For example, the baculovirus isolated from the alfalfa looper was designated Aurugruphu culifurnica (Ac) MNPV (Vail et al., 1973); however, baculoviruses which are almost identical to AcMNPV have been found in Trichuplusia ni, Galleria mellunella (Smith and Summers, 1979) and Rachiplusia uu (Summers and Anderson, 1973). 2.4

BACULOVIRUS REPLICATIONI N V I V O

The larval or caterpillar stage of the insect life cycle is the most susceptible to infection with polyhedra or granules. Polyhedra are ingested when the insect feeds on contaminated diet and dissolve in the alkaline environment of the mid-gut to release the virus particles. After negotiating the peritrophic membrane lining the gut, the virus lipoprotein envelope then fuses with the plasma membrane of the gut wall cells and liberates nucleocapsids into the cytoplasm. The nucleocapsids serve to transport the virus DNA to the nucleus of the cell; it is unclear whether the nucleocapsid enters the nucleus or merely ‘injects’ the virus DNA via a nuclear pore. The processes of virus gene expression have been most thoroughly examined using cultures of insect cells maintained in vitro and infected with NPVs. What was apparent from the earliest studies on virus replication in vivo was the fact that baculoviruses produce two distinct structural forms in a biphasic replication cycle. In the infected gut cells, nucleocapsids are formed by about 8 h post-infection (hpi) and begin to bud through the nuclear membrane by 12 hpi, thus acquiring a lipid envelope. This membrane appears to be ‘lost’ in the cytoplasm, but the nucleocapsid gains another as it buds through the plasma membrane. In the course of this latter process, it also acquires a virus-encoded glycoprotein of 67 kDa (gp67; Blissard and Rohrmann, 1989; Whitford et al., 1989). This protein most probably serves to attach the budded virus to other susceptible cells within the insect; in cell culture the budded virus is 1000-fold more infectious than virus particles released from polyhedra, which lack gp67 (Volkman et al., 1976; Keddie and Volkman, 1985). The budded, or extracellular virus (ECV) is released into the haemolymph to infect other cells and disseminate infection throughout the insect; affected tissues include fat bodies, nerve cells and haemocytes. The cells infected in the second round of virus replication in the insect larva also produce ECV, but in addition occlude virus particles within polyhedra, in the nucleus. The virus particles occluded within polyhedra, which are genetically identical with the ECV, obtain their lipid envelope de nuvu within the nucleus and lack the gp67 found in the budded virus phenotype. The accumulation of polyhedra within the insect proceeds until the host consists almost entirely of a bag of virus. In the terminal stages of infection the insect liquifies and thus releases polyhedra which can infect other insects. Volkrnan and Keddie (1990) aptly described the infected insect at

L. A. KING et a/.

6

this stage as an ‘amorphous puddle’. Recent studies have demonstrated that a virus-encoded chitinase has a role in this process (Hawtin, 1993). Deletion of the chitinase gene from AcMNPV abolished the liquefaction process seen in the latter stages of virus infection. However, to date, the chitinase gene has only been found in AcMNPV or closely related viruses. It remains to be seen how baculoviruses lacking the chitinase gene can also cause liquefaction of virus-infected larvae.

2.5

TRANSMISSION OF BACULOVIRUSES BETWEEN HOSTS

Baculoviruses must infect a succession of susceptible host insects to ensure the maintenance of the virus population. In the most simple case, occluded virus released from a dead insect is almost immediately consumed by an uninfected insect larva and the replication cycle is reinitiated. Inevitably, the process is disrupted when host insects are unavailable. This may occur when the numbers of insects in an area decline due to the effects of the virus outbreak or, as in temperate climates, the host species overwinters as pupae (e.g. Panolis flammea) or eggs (e.g. Neodiprion sertifer). Occluded virus can overwinter on plant surfaces and in the soil. Carruthers et a f . (1988) showed that during fieldwork studies on the pine beauty moth P. Pammea, significant quantities of viable NPV inclusion bodies persisted on pine foliage, after a spray-initiated epizootic, and were still present at the hatch of the following year’s generation. Virus can also remain associated with the insects throughout the winter period. In a continuation of the study reported by Carruthers et a f . (1988), further work at the same site, 1 year after initial spraying, indicated a low-level persistent NPV infection in both the P. jlammea larvae and in the resulting overwintering pupae (Cory and Entwistle, 1990). Vertical transmission of baculoviruses, from parents to offspring, is probably a vital factor in the host-virus relationship. Fuxa et a f . (1992) have shown that polyhedra could be observed in larvae, pupae and adults of Spodoptera frugiperda whose parents had survived exposure to the S . frugiperda NPV. These polyhedra were isolated from the F, insects and fed to first instars. The polyhedra isolated from F1 adult insects did not cause further infection in the first instar larvae. Electron microscopy indicated that the non-infectious polyhedra did not contain virions. From these observations it could be seen that when the S. frugiperda NPV is vertically transmitted, the infections in the F1 adults are non-lethal and are manifested as non-infectious, empty polyhedra. Transmission of the virus between generations is thought to involve both infectious virus remaining viable on the outside of the egg and also the viral genome inside the egg existing in a persistent state (Longworth and Cunningham, 1968; Shapiro and Robertson, 1987). Recent work by Murray and Elkington (1989) has suggested that Lymanfriu dispar NPV is transmit-

ADVANCES IN INSECT VIROLOGY

7

ted between generations primarily via external contamination of eggs. Burand et af. (1992) reported the use of the polymerase chain reaction and short-wave UV irridiation to detect baculovirus DNA sequences from viral occlusion bodies contaminating the surface of gypsy moth eggs. They found that virus contaminating the surface of eggs is the primary source of inoculum for newly hatched larvae. When these larvae die, virus is released, providing a source of inoculum for subsequent infections (Murray and Elkington, 1989). There have been many reports of sporadic outbreaks of NPV infections in natural populations of insects that appear to be caused by the activation of latent or occult NPV or cytoplasmic (C)PV infections (Longworth and Cunningham, 1968; Evans and Harrap, 1982). Latency provides another mechanism for the transgeneration transmission of virus from one generation to another. It has been suggested that such infections can be activated by stress factors (Smith, 1963; Longworth and Cunningham, 1968) such as overcrowding and the consequent lack of food (Steinhaus and Dineen, 1960). This phenomenon occurs frequently in nature, as in the case of mass infestations of the gypsy moth, Porthetria dispar. The CPV of the alfalfa caterpillar, Cofias eurytheme, was discovered in insects subjected to the stress of overcrowding (Steinhaus and Dineen, 1960). Increased temperature and UV radiation may also activate occult virus. Hukuhara (1962) showed the effect of increased temperature on the induction of polyhedrosis by the CPV of the silkworm, Bombyx mori. Unsuitable diet, or an abrupt change of food plant, has in some situations, caused the activation of occult viruses. David and Gardiner (1966) reported the onset of polyhedrosis in Pieris brassicae by a granulosis virus following changes of population density, food and temperature. Biever and Wilkinson (1978) showed that when larvae of P. rapae were reared at 25°C on dehydrated diet (ca 50% moisture), the granulosis virus of P . rapae was activated. Himeno et af. (1973) investigated the effect of various temperature treatments on the infectivity of the nuclear polyhedrosis virus of silkworm larvae. From their results they suggest the virulent virus is bound with biochemical substances which change the virus into an inactive, non-infectious form. These substances, it is suggested, should be present in sufficient quantity to maintain a physiological inactive state. Such virus complexes would be sensitive to thermal shock. By keeping insects at low temperature, the viruses become free from the complexes in the insect, and can replicate. The ingestion of foreign NPV has also been shown to activate latent CPV and NPV infections (Smith, 1963). McKinley et af. (1981) demonstrated that feeding insect larvae (Spodoptera fittorufis)with heterologous NPV (Spodoptera exempta NPV, Hefiothis armigera NPV and S . frugiperda NPV) resulted in the activation of a latent virus resembling the homologous S. fittorafis NPV, rather than replication of the virus used as the inoculum. In another study, the polyhedra of Adoxophyes orana (SNPV) and Barathra

8

L. A. KING et a/.

brassicae (MNPV) were used in cross-inoculation experiments of A . orana and B. brassicae larvae, which were each suspected to harbour latent NPV infections (Ponsen and de Jong, 1964). Subsequent analysis of the DNA obtained from virus harvested from infected insects indicated that the two NPVs were not cross-infective but had activated latent virus in each host which resembled the homologous pathogen (Jurkovicova, 1979). A recent report by Hughes et al. (1993) found that a culture of Mamestra brassicae insects was found to harbour a latent baculovirus infection. The latent virus was activated by feeding the M . brassicae larvae with either the closely related P . fIarnmea NPV, or the distantly related AcMNPV. Restriction fragment profiles of the activated virus DNA showed that it was very closely related, if not identical, to M . brassicae NPV. Further studies using polymerase chain reaction (PCR) amplification of polyhedrin gene sequences demonstrated that the latent virus was present throughout the life cycle of the insect; eggs, larvae, pupae and adults. Using PCR analysis of DNA isolated from dissected tissues of fourth instar larvae, latent virus sequences were only detected in the fat body. Cell lines established from the isolated fat body tissue were also shown to harbour the latent virus sequences. Preliminary experiments using these latent virus-containing cell lines have demonstrated a use in further studies to elucidate the mechanisms of baculovirus latency and virus activation. There have also been reports of baculovirus-like particles in the reproductive tracts of female braconid wasps. Virus-like particles have been found in specific regions of the reproductive tracts of seven different species of wasp, all parasitoids of the tobacco hornworm, Heliothis virescens. The particles are nuclear in origin and are suggested to be related to baculoviruses on the basis of structural homologies (Stoltz et a[., 1976). Other baculovirus-like particles have been found persistently infecting an insect cell line derived from Heliothis zea (Granados et al., 1978; Kelly et al., 1981). Attempts to infect these cells with other baculoviruses induced the replication of the baculovirus-like particle. Despite these reports, the state of the latent or occult virus and the mechanism of activation remain unknown. The role of foreign polyhedra in the activation of latent virus is interesting, as previous studies by Grace (1962) and Longworth and Cunningham (1968) have suggested that polyhedra or polyhedrin itself, rather than the virons contained within, may be the important factor in the activation process. These suggestions were supported by Kelly et al. (1981) who found that the baculovirus-like particles persistently infecting an H . zea cell line could be induced by inoculation with inactivated baculovirus preparations. These reports suggest that polyhedra may have a more significant role than merely acting as a protective coat for the virus between hosts. Activation of the latent A . orana NPV with polyhedra of B. brassicae CPV has also indicated that the activating agent does not have to be a component of an NPV (Jurkovicova, 1979).

ADVANCES IN INSECT VIROLOGY

2.6

9

BACULOVIRUS REPLICATION IN VITRO

Cell lines which support the replication of AcMNPV have been derived from S. frugiperda (Sf) (Fall army worm) pupal ovarian tissue (Vaughn et a f . , 1977) or ovaries from adult Trichoplusia ni (cabbage looper) (Hink, 1970) and M. brassicae (cabbage moth) (King et al., 1991). Although other baculoviruses replicate in cell culture, few match the efficiency of the AcMNPV-S. frugiperda cell combination. The study of baculovirus replication in vitro has greatly simplified experiments to understand the kinetics of virus gene expression and replication. It should be noted, however, that after repeated passage in cell culture, baculoviruses may suffer insertions of host cell DNA and transposable elements within its genome (reviewed by Blissard and Rohrmann, 1990). A consequence of one such insertion within a gene encoding a 25 kDa protein, is the production of viruses which yield few polyhedra (FP phenotype) (Beames and Summers, 1988, 1989, 1990). Clearly, baculovirus genomes, in common with those of other viruses, are subject to mutations which may produce altered phenotypes. The biphasic production of ECV and polyhedra observed in insect larvae is also found in cell culture and, in general, the processes are similar. In this section more details are provided of the molecular events accompanying virus replication. The budded or ECV form of AcMNPV enters insect cells in culture by the process of adsorptive endocytosis. The nucleocapsids serve to translocate the DNA to the nucleus of the cell where virus replication is initiated. Baculovirus genes are expressed in a regulated fashion in infected insect cells. For convenience, and to match the effects of inhibitors of virus replication, virus gene expression is divided into three phases, early, late and very late (Friesen and Miller, 1986). In general, the expression level attained in each succeeding phase is higher than that of the preceding one. 2.6.1 Early gene expression Early gene expression is sometimes further subdivided into immediate-early and delayed-early phases. Immediate-early (IE) genes have been defined as those genes which can be transcribed in the presence of inhibitors of protein synthesis (e.g. cycloheximide) (Kelly and Lescott, 1981; Guarino and Summers, 1986a,b), indicating that other virus proteins are not required for their activation. Several examples have been described for AcMNPV. These include IE-1 (Guarino and Summers, 1986a), IE-N (Carson et a f . , 1988), PE38 (Krappa and Knebel-Morsdorf, 1991; Krappa et al., 1992) and ME53 (Knebel-Morsdorf et al., 1993). A spliced version of IE-1 has also been detected very early in AcMNPV-infected cells (Chisolm and Henner, 1988), and has been designated IE-0. Their classification as immediate-early genes was supported by the fact that copies of these genes, inserted into plasmids, were transcriptionally active after transfection into uninfected insect cells (Guarino and Summers, 1986a; Carson et al., 1988, 1991).

10

L. A. KING e t a / .

Delayed early genes were originally identified using cycloheximide and other inhibitors of protein synthesis in virus-infected cells (Kelly and Lescott, 1981). When cells were treated with these inhibitors and then returned to normal growth conditions, a distinctive pattern of protein expression was observed. Some proteins were expressed immediately after the inhibitor was removed, while others were produced after a delay; hence the division between immediate-early and delayed-early gene products. These results were extended in more recent experiments where immediateearly gene products have been found to transactivate the expression of delayed-early genes after transfection of uninfected insect cells (Guarino and Summers, 1986a; Carson et al., 1988). The 39K delayed-early gene is first detected in infected cells between 3 and 6 hpi. The chloramphenicol acetyl transferase (CAT) gene coding sequences were inserted, in frame, with a truncated AcMNPV 39K gene in a recombinant plasmid. Transfection of insect cells with this construct did not result in detectable CAT activity. When a plasmid containing the IE-1 gene was co-transfected with the 39K-CAT construct, however, significant quantities of CAT enzyme activity were produced (Guarino and Summers, 1986a). It was concluded from these experiments that the immediate-early class of gene products is required in the virus infection to transactivate the delayed-early genes. It has also been demonstrated that IE-N can augment the transactivation of 39K in combination with IE-1, but not when transfected with 39K alone (Carson et al., 1988). More recent results, however, do not support the immediate-early/ delayed-early subdivision. Studies with uninfected insect cell nuclear extracts confirm that both immediate-early and delayed-early genes are transcribed in vitro (Hoopes and Rohrmann, 1991; Glocker et al., 1992). This suggests that virus-encoded transcription factors are not required for the transcription of genes such as 39K. 2.6.2 Late genes The second major class of virus genes expressed in infected cells coincides with the onset of virus DNA replication at about 6hpi. If virus DNA synthesis is inhibited with aphidicolin, the late genes are not transcribed (Miller et al., 1981; Wang and Kelly, 1983). The virus encodes a DNA polymerase gene, the transcription of which is also inhibited by aphidicolin (Tomalski et al., 1988). Virus genes which are expressed during this phase include those encoding structural elements of the virus particles, e.g. the basic protein (Wilson e f al., 1987), the 39K capsid protein (Blissard et al., 1989; Thiem and Miller, 1989), the p87 capsid protein (Mueller et al., 1990) and the virus membrane glycoprotein (gp67) (Whitford et al., 1989; Blissard and Rohrmann, 1991). Transcription of these genes is probably mediated by the action of an unusual RNA polymerase which is a-amanitin resistant and

ADVANCES IN INSECT VIROLOGY

11

induced in cells late in infection (Grula et al., 1981; Fuchs ef al., 1983). It is unclear whether this enzyme is virus encoded or simply a modification of a host cell RNA polymerase. It does, however, have a different subunit composition to host RNA polymerases (Yang et al., 1991). Interestingly, the delayed-early 39K gene promoter is also active in the late phase of baculovirus gene expression. By virtue of a shift in the position of transcription initiation to a late gene promoter start site about 25 nucleotides upstream from the early mRNA start site, the gene is expressed at least until 18 hpi (Guarino and Summers, 1986a). 2.6.3 Very late genes There is some debate as to whether this class of genes should be regarded as a division separate from the preceding late genes. The principal reason for designating a third class is that they are transcribed in the period when the virus is assembling occlusion bodies within the nucleus of the infected cell, from about 15 hpi. The very late gene products include the polyhedrin protein, which forms the matrix of the occlusion body, and the p10 protein, which most probably has a role in polyhedra formation (Vlak et al., 1988; Williams et al., 1989; Russell et al., 1987). The p10 protein forms a crystalline matrix in the infected cell nucleus that is associated with polyhedra formation, although it does not form part of the mature polyhedron. Another justification might be to define very late genes as those which play no role in the formation of infectious virus particles; the polyhedrin and p10 genes may be deleted from the virus genome without affecting virion (ECV) production (Smith et al., 1983b; Vlak et al., 1988). These two very late genes have certainly been the major focus for development of baculovirus expression vectors, since their promoters are extremely efficient and can result in their combined proteins accounting for up to 50% of the total cell protein mass in the terminal stages of infection. 2.6.4 Baculovirus gene promoters Our increasing knowledge of baculovirus gene promoters has paralleled the development of the expression vector system. Many studies which have reported new expression vectors have added information concerning the nature of the promoters. The interest in the use of the very late gene promoters (polyhedrin and p10) as expression systems has focused attention on these structures. Transcription is initiated at a TAAG motif located 50 (AcMNPV polyhedrin; Howard er al., 1986) or 70 (AcMNPV p10; Kuzio et al., 1984) nucleotides upstream from the respective translation initiation codons. This motif is found in all of the baculovirus late and verv late gene

12

L. A. KING et a/.

promoters identified to date (Rohrmann, 1986; Blissard and Rohrmann, 1990). The importance of the 5' non-coding leader sequence of the polyhedrin gene was suggested by early results from Smith et al. (1985), which showed that removing part of this sequence decreased the level of expression. These results were extended in a more detailed study (Matsuura et al., 1987) where progressive deletions were made between the translation initiation codon and the transcription start site (i.e. 3' to 5'). These data showed that the length of the 5' non-coding leader could be correlated directly with the level of expression. The seven nucleotides before the ATG codon were found to be particularly important for maximum promoter activity. A series of linker-scan mutations in the 5' non-coding leader also confirmed the importance of this region (Rankin et al., 1988; Ooi et a l . , 1989). Furthermore, these latter studies also demonstrated that replacing eight nucleotides spanning the T A A G motif with a synthetic oligonucleotide linker resulted in a 2000-fold decrease in promoter activity. The transcription start site was further dissected in a study by Gearing and Possee (1990), where point mutations were introduced into the 11 nucleotides spanning the T A A G motif. Alteration of the TAAG sequence abolished promoter activity, while changes in the flanking regions only resulted in a small decrease in expression. The T A A G motif appears to be a universal feature of all the late and very late baculovirus gene promoters; it also serves as the late transcription start site in the delayed early 39K gene promoter (for review, see Blissard and Rohrmann, 1990). The region upstream from the transcription start site has also been analysed using deletion mutants (5' to 3') (Possee and Howard, 1987). It was concluded that between 7 and 20 nucleotides were required for maximum promoter activity. Interestingly, Ooi et al. (1989) reported that inserting a synthetic linker between 12 and 22 nucleotides upstream from the TAAG sequence increased the levels of steady-state mRNA by up to 50%. Both studies confirm that the sequences upstream from the mRNA start site are relatively unimportant. Similar results have been obtained with the AcMNPV p10 gene promoter (Weyer and Possee, 1988, 1989). The conclusion from these studies is generally the same as with the studies of the polyhedrin promoter. The p10 promoter consists of about 100 nucleotides extending upstream from the translation initiation codon. There is an absolute requirement for the 5' non-translated leader sequence and about 30 nucleotides upstream from the transcription start site. The p10 promoter is fully functional when located in a heterologous location within the virus genome, namely upstream of the polyhedrin gene (Weyer et af., 1990). Recently, advances have been made in our understanding of the transcriptional control mechanisms involved in the regulation of the late and very late genes. Glocker et al. (1993) have shown that nuclear extracts prepared

ADVANCES IN INSECT VIROLOGY

13

from AcMNPV-infected cells can support accurate transcription from several late and very late gene promoters including p39, 39K, p10 and polyhedrin. In vitro transcription of the 39K promoter was resistant to high concentrations of a-amanitin and tagetitoxin, which suggested that neither RNA polymerase I1 nor 111 was responsible for baculovirus late gene transcription. Furthermore, Passarelli and Miller (1993a) have identified three genes that play a crucial role in late (p39) and very late (polyhedrin) gene transcription. Using a method based on subtraction of clones from an AcMNPV genomic library, which is able to transactivate promoters of reporter plasmids in transient expression assays, the three genes identified were IE-1, IE-N and LEF-2. IE-1 was found to be necessary but insufficient for expression from the p39 and polyhedrin gene promoters. The presence of IE-N increased expression from both gene promoters but the third gene, LEF-2 (late gene expression factor 2), was found to be specifically required for expression from the late and very late gene promoters. Two other late gene expression factors have since been identified and characterized, LEF-1 (Passarelli and Miller, 1993b) and LEF-3 (Li et al., 1993). Other virus gene promoters have been less well characterized at the primary sequence level. The IE-1 gene is transcribed from two mRNA start sites, implying that it has two promoters (Chisholm and Henner, 1988). The first of these structures appears to be functional between 0 and 2 hpi and produces a spliced transcript of 2.1 kb. The 5' end of this transcript maps to a position about 4 k b p upstream of the mRNA start site of the second, 1.9 kb transcript. The 1.9 kb transcript reaches its steady-state level 30 min after the virus adsorption period and maintains this level throughout the infection (Chisholm and Henner, 1988). The 39K gene promoter also utilizes two mRNA start sites in the delayed-early and late phases of gene expression (see above). The DNA polymerase gene also has two start sites for mRNA initiation, but these are both active in the early phase of gene expression (Tomalski et a l . , 1988). Another interesting feature of immediate-earlyldelayed-early gene expression is the role of enhancer elements. In AcMNPV these consist of five regions of homologous repeats (hr 1-hr5) containing repeated sequences with multiple EcoRI sites (Cochran and Faulkner, 1983; Guarino et al., 1986). When linked in cis with immediate-early and delayed-early genes, these sequences may enhance transcription by up to 3000-fold (Guarino and Summers, 1986b; Nissen and Friesen, 1989). The role, if any, of these enhancer elements in late and very late gene transcription remains to be elucidated. 2.6.5 DNA replication Relatively little is understood about the DNA replication of AcMNPV, although the structure and functional organization of the genome has been

14

L. A. KING et a/.

well characterized (reviewed by Kool and Vlak, 1993). A few genes have been identified which are thought to be involved in the process, including a DNA polymerase (Tomalski et al., 1988), a proliferating cell nuclear antigen-like protein (O’Reilly et al., 1989), and a helicase (Lu and Carstens, 1991). Attempts to isolate putative origins of DNA replication have proved difficult (Blinov, 1984). However, more recently there has been some progress in this area and several groups have reported the identification of putative origins of replication on the AcMNPV genome (Pearson et al., 1992; Kool et al., 1993). In the report by Kool and colleagues, use was made of defective AcMNPV genomes that lacked considerable portions of the genome (up to 43%), yet still retained the ability to replicate in cells in vitro. Three separate regions were retained in the mutant genomes and two of these were identified as containing putative origins of replication. The two regions were mapped to the HindIIIB fragment between map units 50.1 and 53.2, and to the HindIIIQ fragment between map units 87.2 and 88.9. Transfection of AcMNPV-infected S. frugiperda cells with plasmids containing these sequences resulted in the amplification of the plasmids, as demonstrated by DpnI sensitivity assays. 2.6.6 Apoptosis Apoptosis or programmed cell death is recognized in vertebrate cells. A similar phenomenon has been reported in certain insect cells infected with an AcMNPV mutant (vAcAnh)(Clem et al., 1991; Crook et al., 1993). This virus was isolated because of its production of small plaques lacking polyhedra in S. frugiperda cells. Subsequent tests in three cell lines demonstrated that it caused premature death of S. frugiperda, and B. mori but not T. ni cells. The mutation in vAcAnh was mapped within the virus genome and the appropriate region of virus DNA sequenced. When compared with parental virus DNA sequence, a deletion of 754 nucleotides was noted, which would foreshorten the native p35 protein by 132 amino acids. Inactivation of p35 in the wild-type virus confirmed its role in preventing apoptosis. These data suggest that apoptosis may enable certain insects to overcome virus infection. Kamita et al. (1993) have also described a similar phenomenon in cells infected with the BmNPV. More recently, Clem and Miller (1993) have shown that viral gene expression is abnormal in S. frugiperda cells infected with the p35-mutant AcMNPV, with a delay in the expression of both early and late genes and a lack of expression of the very late genes. The infectivity of the p35-mutant for S. frugiperda larvae was about 1000-fold lower than wild type or revertant viruses. In contrast, the replication and infectivity of the p35-mutant in T. ni cells or larvae was equivalent to wild-type or revertant viruses (Clem et al., 1991). Thus it would appear that a host apoptotic response provides protection against

ADVANCES IN INSECT VIROLOGY

15

viral infection at the organismal level, and that the p35 gene may constitute a host range determinant for AcMNPV infection (Clem and Miller, 1993). Crook er al. (1993) have identified a gene from the Cydia pomonella granulosis virus that was able to rescue wild-type infection from S. frugiperda cells infected with the p35-mutant AcMNPV. This gene, named iap (inhibitor of apoptosis), had no significant homology to the AcMNPV p35 gene but contained a zinc finger-like motif which has been found in other genes with the potential to regulate apoptosis. Both the p35 and iap genes were able to block apoptosis induced by actinomycin D , indicating that these genes act by blocking cellular apoptosis rather than by preventing viral stimulation of apoptosis. 2.7

BIOLOGICALCONTROL OF INSECT PESTS

2.7.1 History offield control The detailed studies of baculovirus biology and molecular genetics described above were predated by numerous examples of the use of baculoviruses as biological control agents of insect pests. It is outside the scope of this section to provide comprehensive descriptions of each one; the reader is referred to reviews by Podgewaite (1985), Entwistle and Evans (1985), Huber (1986) and Evans and Entwistle (1987) for detailed accounts of this topic. One of the most spectacular successes has been the control of the velvetbean caterpillar, Anticarsia gemmatalis, with the A . gemmuralis (Ag) NPV (Moscardi and Correa Ferreira, 1985; Moscardi, 1989; P. Zanotta, personal communication). A long-term programme to use this virus in insect control programmes has resulted in the annual spraying of 2 million hectares of soybeans (P. Zanotto, personal communication). An accurate description of insect pest control with baculoviruses is also complicated by the fact that many field programmes are not documented in the scientific literature (J. Cory, personal communication). In developing countries, farmers and other agriculturalists are more concerned with controlling the insect pests than conducting and reporting Western-style field trials. 2.7.2 Advantages and disadvantages The principal advantage of baculoviruses as insecticides is their narrow host range. The Baculoviridae are specific for invertebrates. Each virus isolated from insects has a narrow host range enabling the targeting of particular pest species. Conversely, this may be considered a disadvantage, since more than one virus may need to be applied to a crop to effect control of different insect pests. Chemicals can be used to control multiple insect species, although beneficial species may also be affected by these agents. Despite the lack of a direct effect, it has been suggested that the decimation of a pest

16

L. A. KING et ar

population with baculoviruses could seriously affect dependent predator! such as insect parasitoids (Hochberg, 1991). In subsequent years, then would be insufficient predators to provide natural control of the pest insect permitting an explosion in their numbers with inevitable crop damage While it is difficult to refute this argument, the same hypothesis holds gooc for chemical insecticide application. The use of chemical insecticides is frequently associated with the development of resistance in the target species. To date, the same phenomenon has not been documented in situations where baculoviruses have been used for insect pest control. Although it is considered that mas5 resistance to baculoviruses is unlikely to develop (Kirschbaum, 1985), this conclusion has yet to stand the test of time and it would be unwise to assume that insect populations will remain unresponsive to severe selection pressure. Although baculovirus infection of insect larvae often results in death, the virus can coexist with the host without producing symptoms (see below). Chemical insecticides are cheap and easy to produce. Baculoviruses require a biological system, the insect larva, to provide large quantities of infectious material for field application. Large-scale production by this method is labour intensive, which in developed countries is expensive. The maintenance of virus-free insect cultures is also problematic unless one is careful to separate the virus production facility from the stock culture. Baculoviruses may be stored as frozen, or freeze-dried preparations. The latter is particularly advantageous in areas lacking facilities for chilled storage. The most frequently cited disadvantage of baculoviruses as insecticides is the long period, relative to chemical agents, after application of the virus and death of the target insect. This may be as long as 7-10 days, depending on temperature and initial virus dose. In this period, insect larvae continue to feed, resulting in damage to the crop. In contrast, chemicals are usually effective after a matter of hours and almost instantly stop the insects from feeding. Before baculoviruses can seriously challenge the market domination of chemical insecticides their speed of action must be improved. 2.7.3 Genetic modification of baculovirus insecticides

The most promising approach to solving the problem of delay in killing the insect pests with baculovirus insecticides involves the insertion of foreign DNA encoding insect-specific hormones, enzymes or toxins into the virus genome, under the control of a strong virus gene promoter. Several examples have been reported in the literature, each describing varying degrees of success. Water balance in insects is influenced by diuretic and anti-diuretic

ADVANCES IN INSECT VIROLOGY

17

hormones (Maddrell, 1986). To test whether this balance might be affected by elevated levels of a diuretic hormone (DH), Maeda (1989b), inserted a sequence encoding the Manduca sexta (tobacco hornworm) diuretic hormone into the B. mori NPV genome, under the control of the polyhedrin gene promoter. A signal peptide coding region from the Drosophila melanogaster (fruit fly) cuticle protein was fused in frame with the DH coding sequence to effect secretion from virus-infected cells. The polyhedrin-negative virus was used to infect B. mori larvae by injection directly into the haemocoel. It was reported that after infection, a 30% decrease in haemolymph volume was noted compared with mock-infected or unmodified virus-infected larvae. Insect larvae infected with the recombinant virus died one day earlier than controls. The production of D H in insect larvae was inferred from analysis of mRNA isolated from virusinfected insects. A reverse phase column was used to isolate D H from virus-infected insects; this material was assayed using newly emerged Pieris r a p e (small white butterfly) adults (Kataoka et al., 1989). The small size of the D H and complicated assay procedures required to detect biologically active material, make it difficult to assess the effectiveness of inserting such coding regions into the virus genome. Another complex physiological process amenable to disruption by genetically modified baculoviruses is that of metamorphosis. Large amounts of juvenile hormone esterase (JHE) are produced by the insect larva in the last instar, prior to pupation. This enzyme hydrolyses the chemically stable, conjugated methyl ester of juvenile hormone (JH) to the J H acid (Hammock, 1985). A decrease in JH titre in the last larval instar is associated with a cessation of feeding and metamorphosis (de Kort and Granger, 1981). These data, together with the fact that inhibition of J H E produces overgrown insect larvae (Sparks and Hammock, 1980), suggested that elevated levels of JHE in early larval instars would inhibit feeding and cause premature pupation (Hammock et al., 1990). A cDNA clone encoding J H E was prepared from fat body mRNA isolated from H. virescens (tobacco budworm) (Hanzlik et al., 1989). This was inserted into a baculovirus transfer vector, under the control of the AcNPV polyhedrin gene promoter, and used to produce a polyhedrin-negative recombinant virus (Hammock et al., 1990). The virus produced active JHE in cell culture. In virus-infected neonate T. ni, active J H E was also produced and was accompanied by a reduction in feeding. The feeding inhibition was not observed when later larval instars were used. This may have been due to the production of insufficient JHE. The levels of J H E recorded in the haemolymph of virus-infected insects were only 10% of those normally measured in the last larval instar, prior to ecdysis. The virus-induced J H E may have been insufficient to overcome J H biosynthesis, or other control mechanisms as yet undefined. The enzyme is also very unstable in vivo when produced by the recombinant virus, or in its natural form. The baculovirus also produces an

18

L. A. KING e t a / .

ecdysteroid UDP-glucosyl transferase (O'Reilly and Miller, 1989) which may reduce the effects of JHE. The lack of effect of the recombinant J H E on later larval instars is a disadvantage, since it is unlikely that neonates could be targeted consistently. However, it is likely that the effectiveness of JHE will be improved as more is elucidated of its function in the natural insect. The most significant improvements in the effectiveness of baculoviruses have been achieved by inserting insect-specific toxin genes into the virus genome. In two similar studies (Martens er al., 1990; Merryweather et al., 1990) the Bacillus thuringiensis delta endotoxin gene was inserted into the baculovirus genome. This bacterial toxin is produced as a 130 kDa protoxin, in a crystalline form. When ingested by larvae, the protoxin dissolves in the alkaline midgut and is cleaved to an active 62 kDa component (reviewed by Hofte and Whiteley, 1989). This toxin binds to cells lining the gut wall and is thought to introduce pores into the plasma membranes, causing disruption of the osmotic balance and cell lysis. The immediate effect of this process is a cessation of feeding (Heimpel and Angus, 1959). The DNA encoding the B. thuringiensis delta endotoxin was inserted into AcMNPV in lieu of the polyhedrin gene coding region, but under the control of the polyhedrin gene promoter (Martens et al., 1990; Merryweather et al., 1990). A polyhedrin-positive AcMNPV containing the bacterial sequence was produced by inserting the endotoxin coding region under the control of the virus p10 gene promoter (Merryweather et al., 1990). Polyhedrin-negative and polyhedrin-positive viruses produced biologically active endotoxin, as evidenced by a reduction in feeding when diet was contaminated with recombinant virus-infected cell extracts. Although functional toxin was produced in virus-infected cells, this was not accompanied by a reduction in lethal dose (LD,,) when purified recombinant virus polyhedra were assayed in T. ni larvae. Further tests showed that toxin was produced in the virus-infected insects, but it is likely that this material could not access the midgut for proteolytic cleavage and conversion to the active form. Provision of a signal peptide sequence to a truncated form of the toxin does result in some secretion of toxin in virus-infected cells in culture, but has no effect on the biological activity of the virus (A. T . Merryweather and R. D. Possee, unpublished data). While the use of the B. thuringiensis delta endotoxin remains attractive, its production by baculoviruses in insect larvae has yet to show any significant advantage over the unmodified virus. This is unfortunate, since the B. thuringiensis itself is used as a biological control agent and is widely accepted as a safe formulation. The most successful examples of genetic modification of baculoviruses have utilized insect-specific neurotoxin genes. Stewart et al. (1991) inserted a copy of the Androctonus australis (North African scorpion) insect-specific neurotoxin coding region into a polyhedrin-positive AcMNPV, under the control of the p10 gene promoter. This toxin affects the sodium conductance

ADVANCES IN INSECT VIROLOGY

19

of neurones, producing a presynaptic excitatory event, leading to paralysis and death (Walter et al., 1976; Teitelbaum et al., 1979). When produced by the recombinant virus in T. ni larvae, the insects consumed 50% less diet in comparison with unmodified virus-infected controls (Stewart et al., 1991). In other tests, the LDSowas reduced slightly and there was a 25% decrease in the survival time (ST,,). These effects were only observed if the neurotoxin was produced as a fusion protein with the AcNPV glycoprotein (gp) 67 coding sequence (Whitford et al., 1989). Clearly, the neurotoxin must enter the secretory pathway of the cell to attain full biological activity. Other studies using the same scorpion toxin have reported essentially similar results (Maeda et al., 1991; McCutchen et al., 1991). Another insect-specific neurotoxin gene used to enhance the effectiveness of baculoviruses was isolated from Pyernotes tritici (mite) by Tomalski and Miller (1991). This was inserted into the AcMNPV genome under the control of the polyhedrin promoter to derive a polyhedrin-negative virus (Tomalski and Miller, 1991). A polyhedrin-positive virus was also produced by placing the neurotoxin coding region upstream of the native polyhedrin gene, under the control of a hybrid gene promoter comprising elements of both late and very late transcription units (Tomalski and Miller, 1992). Both reports recorded that feeding the recombinant viruses to T. ni larvae resulted in significantly earlier mortality, reduced feeding and weight gains and paralysis. Undoubtedly, the isolation of other insect-specific toxins will lead to the development of recombinant baculoviruses with even better insecticidal activities than those reported by Stewart et al. (1991) and Tomalski and Miller (1991, 1992). It is worth considering, however, that as recombinant viruses which have very rapid effects on the insect host are produced, it will become problematic to amplify those same viruses to levels sufficient for field application. Stewart et a[. (1991) recorded a decrease in virus yield from insects infected with the AcNPV containing the scorpion toxin gene. This decrease was not sufficient to prevent production of enough virus for laboratory use and a future small-scale field trial but it signalled that attention should be directed to this phenomenon. It may be necessary to regulate the production of the foreign protein in the virus amplification stage and then release the control in the field application. Alternatively, insect hosts which are resistant to the effects of the toxin may be used as a virus production facility. 2.7.4

Other techniques for improving baculovirus insecticides

Before the introduction of techniques for the genetic modification of baculoviruses various strategies were evoked to improve the effectiveness of the virus preparation. Despite the protection afforded the virus particles by the occlusion body, solar ultraviolet light can rapidly inactivate virus

20

L. A. KING e t a / .

infectivity, particularly in tropical climates. Brassel and Benz (1979) selected a strain of the codling moth GV with improved resistance to ultraviolet light. In laboratory tests this virus was 5.6 times more resistant to artificial ultraviolet irradiation and retained infectivity twice as long in the field. Wood et al. (1981) conducted studies to increase the virulence of AcMNPV by replication of wild-type virus in the presence of 2-aminopurine. The lethal time 50 of a mutant, designated HOB, was significantly shorter than the parental virus. Fifth instar larvae infected with HOB gained weight at a lower rate than the unmodified virus. Hughes et al. (1983) compared the time-mortality response of H . zea to 14 isolates of H . zea NPV. Those isolates with genomes producing identical profiles in agarose gels after digestion with restriction endonucleases, did not differ significantly in the time required to kill the insect host. Isolates producing dissimilar restriction endonuclease cleavage patterns had significantly different STSovalues. That baculovirus isolates do not consist of genetically homogeneous populations has been confirmed in various studies. For example, Smith and Crook (1988) were able to separate eight distinct genotypes from Arfogeia rapae larvae by low mortality dose infections and three genotypes by the same method using L. dispar larvae. The biological activity of these viruses was not assessed. Weitzman et al. (1992) characterized two variants (PfNPV(A) and (B)) of P . flammea NPV isolated from a population of wild-type virus. These variants had similar restriction endonuclease cleavage patterns, but displayed interesting properties when propagated together in two different insect hosts. In P . flammea larvae, PfNPV(B) formed at least 80% of the virus population. In M . brassicae larvae, PfNPV(A) predominated and attained 80% of the population after only three successive passages. This suggests that, according to the host insect used to amplify the viruses, a ceratin genotype has an advantage. The significance of this for field control programmes is that propagation of a mixed genotype virus population in one host insect for later application to control another species in the field may not derive the most efficient variant. 2.7.5 Assessing the safety of genetically modified baculovirus insecticides Empirically, there is no difference in assessing the safety of unmodified or genetically modified baculovirus insecticides. Inevitably, the proposal to use the altered virus insecticides has provoked additional concerns for various parties. These issues have been discussed at length in previous publications (Bishop et al., 1988; Bishop, 1989; Possee et al., 1992, 1993; Vlak, 1993a,b) and will not be considered in the same detail here. The question most often raised is whether genetically modified baculoviruses have an altered (expanded) host range. Our knowledge of what determines baculovirus host range is very limited and so this must remain an important consideration. Prior to field release experiments, those insect

ADVANCES IN INSECT VIROLOGY

21

species most likely to be present at the site of the trial should be challenged with the virus in the laboratory to determine the outcome. These data may not be so easy to interpret, since it is unclear whether insects requiring very high doses of virus to initiate infection would encounter the same levels of virus in the field. A further concern is whether genetically modified baculoviruses might recombine with indigenous baculoviruses after co-infection of insect larvae and transfer the foreign gene. For this to occur, the two viruses would have to be present in the same cell within the insect. Recombination frequencies are affected by DNA similarity, so co-infection of insect cells would not guarantee the generation of stable virus recombinants. Recombination has been demonstrated to occur between recombinant and unmodified AcMNPVs after coinfection of T. ni larvae (A. T. Merryweather and R. D. Possee, unpublished data). The long-term persistence of the modified baculovirus in the environment has also to be considered. The baculovirus occlusion bodies may persist in soil for many years (Evans and Harrap, 1982). On leaf surfaces, the virus is more susceptible to inactivation by solar ultraviolet radiation. In the short term, field experiments should probably include the option of disinfection of the soil to prevent the formation of a reservoir of infectious viruses. 2.7.6 Past field release experiments The considerable number of laboratory studies to assess the feasibility of improving baculovirus insecticides using genetic engineering techniques has not been matched by subsequent field trials to monitor their effectiveness in the environment. This is understandable, considering the cost of field trials and the trepidation with which such experiments are viewed by scientists and the public alike. Each field trial involves several years of planning, only one experiment can usually be accomplished in a season and the results require careful interpretation afterwards. The first field trial with a genetically modified baculovirus was conducted in 1986 in Oxford (Bishop, 1986; Bishop et al., 1988). A genetic marker was inserted into a non-coding region of the AcMNPV genome, downstream of the polyhedrin coding region. This marker served to distinguish the virus from parental virus and other, unrelated baculoviruses. The virus was used to infect insect larvae in the laboratory. These individuals were then placed onto sugar beet plants within a netted field facility to examine persistence of the genetically modified virus. The virus-infected larvae died within a 7-day period. Marked virus could be recovered from the site up to 6 months later, whereupon the plants were cleared and the soil disinfected. Later experiments, between 1987 and 1989, used AcNPV mutants lacking the polyhedrin gene, but containing another unique genetic marker or a functional lac2 gene from Ewherichia coli. These experiments served to demonstrate that

22

L. A. KING et a/.

viruses lacking the protection of the polyhedrin-based occlusion body cannot persist in the environment for significant periods. They suggest that if a baculovirus insecticide could be modified to package virus particles within polyhedra in the laboratory, but only produced polyhedrin-negative viruses in the field, this would constitute a very safe biological insecticide. The only other field trial with a genetically modified baculovirus has also made use of an AcMNPV mutant deficient in polyhedra production. Hamblin et al. (1990) described the process of co-occlusion of a polyhedrinnegative virus by an unmodified AcMNPV after co-infection of insect cells in culture or in larvae with the two viruses. After several passages in insects, the polyhedrin-negative virus was gradually lost from the population. This study has recently been extended to the field, with spray application of virus to a cabbage plot artificially infested with T. ni larvae (H. A. Wood, personal communication). Monitoring of the site is continuing over a number of years to determine the levels of polyhedrin-negative virus. 2.7.7 Future experiments The field release experiments with genetically modified baculoviruses performed to date have utilized viruses harbouring relatively simple, innocuous changes to the virus genome. These changes were not intended to improve the efficacy of the virus insecticide. Future field tests will involve the use of viruses containing foreign genes encloding insect-specific hormones, enzymes or toxins (see Section 2.7.3). These viruses are known to kill the insect host more rapidly and to reduce the feeding damage to the plant. Laboratory tests must be confirmed by authentic field trials if the early promise shown by these agents is to be confirmed. 2.8

BACULOVIRUS EXPRESSION VECTORS

The expression of foreign genes in insect cells using baculovirus expression vectors is now a well-established technology that is widely used by many investigators (reviewed by Luckow and Summers, 1988; Miller, 1988, 1993; Cameron et al., 1989; Maeda, 1989a; Atkinson et al., 1990; Bishop and Possee, 1990; Possee et al., 1990; King e f al., 1992; King and Possee, 1992; O’Reilly et al., 1992). There are several advantages in the use of baculoviruses instead of other eukaryotic expression systems: the expression vectors do not require a helper virus system; the viruses are non-pathogenic to vertebrates and plants; and no oncogenic or transforming elements are employed. It has also been demonstrated that insect cells will perform a wide range of post-translational processing events that are often required for biological activity of a recombinant protein. One other advantage which has not been well documented is the application of the baculovirus expression vector system to the study of insect

ADVANCES I N INSECT VIROLOGY

23

proteins. This increases the possibility that the protein of interest will retain the range of properties associated with its native counterpart. This review will describe some of the more recent advances in baculovirus expression vector technology.

2.8.1 The development of baculovirus expression vectors Most baculovirus expression vectors make use of the very late polyhedrin gene promoter of AcMNPV to drive the expression of the foreign gene in virus-infected S. frugiperda cells (Smith et al., 1983b). The polyhedrin gene promoter is very strong and can result in the production of very large quantities of the foreign protein in virus-infected cells. An alternative expression vector system is based on the homologous gene in the B. mori (silkworm) NPV, but this has not found such wide application as AcMNPV vector system (Maeda et al., 1985; Maeda, 1989a; Iatrou and Meidinger, 1990). The polyhedrin gene of AcMNPV has been mapped to the EcoRI ‘I’ fragment of the virus genome (Smith ei al., 1983a) and sequenced (Hooft van Iddekinge et al., 1983). It was subsequently demonstrated that the polyhedrin gene could be deleted, preventing formation of virus occlusion bodies, without affecting the production of ECV (Smith et al., 1983b). This also provided a useful phenotypic marker for recognizing recombinant viruses in plaque assays. These features of the polyhedrin gene, including a strong promoter, redundant function of the protein and easily recognizable phenotype, made it an attractive proposition for development as a high level eukaryotic expression system. A complication, however, was the fact that the AcMNPV genome is quite large (approx. 133 kbp), and difficult to modify by conventional ligation with foreign DNA. Consequently, indirect methods are used to modify the baculovirus genome. These involve construction of bacterial plasmids, or transfer vectors, containing portions of the AcMNPV genome, which encompass and flank the polyhedrin gene. The EcoRI ‘I’ fragment has formed the basis for most AcMNPV transfer vectors (Possee et al., 1991). Within a transfer vector, the polyhedrin gene coding sequences are either partially or completely deleted and replaced with a recognition site for a restriction endonuclease. The polyhedrin gene promoter and transcription termination signals are left intact. The foreign gene coding region, with its own translation initiation signal, is inserted into the transfer vector into the correct orientation. The recombinant virus is produced by co-transfection of insect cells with the transfer vector and purified ‘wild-type’ AcMNPV DNA. The foreign gene is inserted into the AcMNPV genome by homologous recombination between identical sequences flanking the native polyhedrin gene in the virus and the foreign DNA in the transfer vector. The net result is the production of a virus which is able to replicate in insect cells but lacks the ability to

24

L. A. KING et a/.

produce polyhedra. The virus progeny from the co-transfection are then titrated in a plaque assay in S. frugiperda cells and polyhedrin-negative plaques identified by visual inspection with the aid of a microscope. The success of the AcMNPV expression vector system has resulted in the derivation of many alternative transfer vectors for introducing foreign coding regions into the baculovirus genome. For the inexperienced user of the system it can be difficult to choose an appropriate plasmid. It is advisable to consult the two practical guides to the baculovirus expression system currently available (King and Possee, 1992; O’Reilly et al., 1992) for guidance in selecting a transfer vector. Briefly, transfer vectors may have varying amounts of virus sequence flanking the foreign coding region; it is often easier to make insertions in the smaller vectors. The bacterial plasmids may have the M13 intergenic region, facilitating production of singlestranded DNA in bacteria for sequencing and rapid mutagenesis. Finally, more than one restriction enzyme site may be available for the insertion of the foreign DNA. This renders some transfer vectors compatible with DNA fragments possessing a variety of cohesive or blunt ends. Although most foreign genes are expressed using the polyhedrin genebased vectors, other baculovirus gene promoters may be used. The p10 gene coding region may also be removed from the virus genome without affecting either ECV or polyhedra production (Vlak et al., 1988; Williams er al., 1989; Weyer et al., 1990). The p10 gene product is, like polyhedrin, non-essential in the production of virus particles (Vlak et al., 1988). Transfer vectors analogous in structure to the polyhedrin gene-based vectors described above are constructed, prior to the insertion of the foreign coding region under the control of the p10 gene promoter. Recombinant viruses are produced by co-transfection of insect cells with plasmid and virus DNA. A complication when using the p10 gene locus to insert foreign DNA into the virus genome is that, unlike polyhedrin, the p10 protein does not produce a recognizable phenotype in virus-infected cells. Fortunately, recent advances in recombinant virus selection have solved this problem (see Section 2.8.3). Foreign genes are inserted directly under the p10 promoter by co-transfection of insect cells with a p10 transfer vector and purified wild-type AcMNPV DNA. Other AcMNPV gene promoters may also be used for the expression of foreign genes in virus-infected cells. Several late gene promoters, while not as active as the very late polyhedrin and p10 gene promoters, can produce useful quantities of recombinant material. One problem encountered when using the late gene promoters is that they are normally associated with genes responsible for producing virus particle structural proteins. In consequence, replacement of the native virus coding sequence with a foreign coding region is unlikely to result in the production of a viable recombinant virus. For example, the AcMNPV basic or arginine-rich protein is associated with virus DNA within the nucleocapsid. In order to use this virus gene promoter as an

ADVANCES IN INSECT VIROLOGY

25

expression system, it was necessary to insert a copy of the promoter in lieu of the polyhedrin gene promoter (Hill-Perkins and Possee, 1990). Production of bacterial P-galactosidase by the basic protein gene expression vectors was significantly lower than when the polyhedrin gene promoter was used, but initiated from about 8 hpi. This was 6 h before the polyhedrin gene promoter attained a reasonable level of activity. Other late gene promoters used in a similar manner include the 39K capsid protein gene promoter (Thiem and Miller, 1989) and the gp67 gene promoter (A. T. Merryweather and R. D. Possee, unpublished data). The earlier production of foreign proteins in virus-infected cells may be advantageous if they require extensive post-translational modifications (see Section 2.8.4).

2.8.2 The development of multiple expression vectors More than one foreign protein may be produced in insect cells simply by co-infection with two or more recombinant viruses. This approach was used to express simultaneously three influenza virus proteins (St Angelo et al., 1987). A more efficient and reproducible way to achieve co-expression, however, is to insert each foreign gene into the same recombinant virus. Emery and Bishop (1987) inserted a copy of the polyhedrin gene promoter and putative transcription termination signals, upstream of the native polyhedrin gene. This enabled the expression of a foreign gene in addition to the virus polyhedrin. Subsequent modification of this vector permitted the co-expression of the hepatitis B virus surface and core antigens in baculovirus-infected cells (Takehara et al., 1988). Each foreign sequence was placed under the control of the native or duplicated polyhedrin gene promoter. Similar expression vectors were derived by using a combination of the polyhedrin and p10 gene promoters (Weyer and Possee, 1991). A copy of the p10 gene promoter was inserted upstream of the polyhedrin gene promoter. The influenza virus haernagglutinin o r neuraminidase gene was placed under the control of each promoter and co-synthesis achieved in recombinant virus-infected cells. Baculovirus expression vectors are not limited to the production of two foreign proteins in insect cells. The synthesis of non-infectious virus-like particles of bluetongue virus by the simultaneous expression of four structural proteins, by the co-infection of two dual recombinant baculoviruses has been described (French et al., 1990). Five bluetongue virus structural proteins have been co-expressed within the same cell by coinfection of two dual recombinants and one single recombinant virus (Loudon and Roy, 1991). The baculovirus expression system, therefore, is ideally suited to the study of protein-protein interactions.

26

L. A. KING eta/.

2.8.3 Improved methods for the selection of recombinant baculoviruses Foreign genes have been inserted into the AcMNPV genome at the polyhedrin locus by co-transfection of insect cells with the appropriate transfer vector and purified wild-type AcMNPV DNA. Recombinant virus is separated from parental virus by titration in a plaque assay and selection of polyhedrin-negative plaques after visual inspection. This method, while very simple, has presented problems to many users of the system. Polyhedrinnegative plaques can be difficult to identify, necessitating many hours of tedious searching. Various methods have been developed to simplify the recombinant virus selection procedure. The virus progeny from a co-transfection may be titrated in a microtitre plate containing insect cells. After a suitable period, the medium is removed from each well and retained. DNA is extracted from the cells remaining in the well and the presence of the foreign gene sequences monitored using simple dot blot hybridization techniques. Those wells eliciting positive results provide the means to select the stored medium samples for further analysis. These samples usually contain enhanced numbers of polyhedrinnegative viruses, which may be more easily identified in subsequent plaque assays. Vialard et al. (1990a) produced polyhedrin gene-based transfer vectors with a copy of the bacterial P-galactosidase gene, under the control of a baculovirus gene promoter, inserted upstream of the polyhedrin gene promoter used for foreign gene expression. Recombination between the transfer vector and AcMNPV DNA produced recombinant viruses with both P-galactosidase and the foreign gene, facilitating selection of blue plaques in a titration stained with X-gal. A similar approach has been used to effect recombinant virus selection when inserting foreign genes at the p10 gene locus (Vlak et al., 1988). In a radically different approach, the poor infectivity of linearized virus DNA has been used as a way of selecting recombinant viruses. Kitts ef af. (1990) observed that when AcMNPV DNA, linearized at the polyhedrin gene locus (using Bsu36I), was co-transfected with a plasmid transfer vector, 30% of the virus progeny contained the foreign gene. The transfer vector appeared to ‘rescue’ the linear virus DNA by recircularization. Fortuitously, the E. coli P-galactosidase coding region contains the appropriate restriction enzyme site (Bsu361). Recombinant virus DNA containing this sequence at the polyhedrin or p10 gene loci may be linearized with Bsu361, mixed with the appropriate transfer vector and used to co-transfect insect cells. Progeny virus, containing the foreign DNA in place of the P-galactosidase sequences, may be selected as colourless plaques in the presence of X-gal at a frequency of about 30%. This system, using linearized virus DNA, has been improved by inserting additional Bsrr36I sites in the virus DNA harbouring the P-galactosidase

ADVANCES IN INSECT VIROLOGY

27

coding region under the control of the polyhedrin gene promoter. The first was inserted within a non-essential virus gene of unknown function upstream of the polyhedrin gene promoter (ORF 603; Gearing and Possee, 1990). The second was inserted within an essential gene, also of unknown function (ORF 1629; Possee et a f . , 1991), downstream of the P-galactosidase coding region. The sequence of the second Bsu36I recognition site was designed to preserve the coding region of ORF 1629. This virus was designated BacPAK6 (Kitts and Possee, 1993). Digestion of the virus DNA with Bsu36I removes part of ORF 603, the polyhedrin gene promoter, the P-galactosidase coding region and part of ORF 1629. When the Bsu361digested BacPAK6 virus DNA was used to transfect insect cells, very low recoveries of infectious virus were attained. Recircularization of linear BacPAK6 DNA produces a crippled virus which is unable to produce infectious virus in insect cells. If the same DNA was co-transfected with a plasmid transfer vector, however, virus yields were enhanced considerably with nearly 100% recovery of recombinant virus. Very few parental (BacPAK6) viruses are evident in the first round of plaque purification, enabling rapid isolation of recombinant viruses. The co-transfection of linear BacPAK6 DNA with the transfer vector serves to repair the partial deletion in ORF 1629 and permit the production of infectious virus.

2.8.4 Post-translational processing in insect cells Post-translational processing plays an important role in determining the biological activity of a recombinant protein. The range of post-translational processing events in insect cells have been widely documented. These include glycosylation, phosphorylation, proteolysis, ADP-ribosylation, sulphation, acylation and disulphide bond formation (reviewed by Miller, 1988; Luckow and Summers, 1988; Maeda, 1989a, Atkinson et af., 1990; Bishop and Possee, 1990; King et al., 1992; King and Possee, 1992; O’Reilly et a f . , 1992). Of the post-translational processing events listed, insect cells appear to possess a different glycosylaction pattern to that associated with other systems. It is known that the glycosylation sites utilized in insect cells are the same as in mammalian cells (Hsieh and Robbins, 1984), i.e. asparagine residues, and that the most common type of glycosylation is N-linked and can be inhibited by tunicamycin (Kelly, 1982). The principal difference is in the nature of the oligosaccharides added to these sites. Insect cells appear to lack galactose and sialic acid transferases and trim the oligosaccharide to a short core containing mannose, whereas in mammalian cells the central core is extensively modified (Kuroda et al., 1990). Jarvis et al. (1990a) performed a study on the role of glycosylation in the transport of recombinant glycoproteins through the secretory pathway of insect cells following treatment with tunicamycin or castanospermine (an inhibitor of the initial steps of N-linked oligosaccharide processing). They demonstrated that

28

L. A. KING eta/.

tunicamycin treatment inhibits the secretion or cell-surface expression of some but not all glycoproteins. Castanospermine treatment did not inhibit the secretion or cell-surface expression of any of the glycoproteins tested. This suggests the possible role of N-glycosylation but not oligosaccharide processing in the transport of some recombinant proteins through the secretory pathway of insect cells. The differences in glycosylation appear to have no effect on the overall function of the recombinant protein. It has been shown that foreign proteins synthesized in insect cells can be targeted to the nucleus (Forstova et a f . , 1989), the cytoplasm (Jeang et al., 1987), to the cell surface (Possee, 1986; Greenfield et al., 1988; Atkinson et al., 1992) or for secretion (Jarvis and Summers, 1989). The baculovirus expression vector system has been used to express proteins from viral, fungal, plant, protozoan, invertebrate and vertebrate origins. Historically, two main cell lines have been used for the baculovirus expression system, S. frugiperda IPLB (Sf21) (Vaughn ei a f . , 1977) and Sf9 (Smith et a f . , 1983b). More recently, it has been documented that other insect cell lines may also be useful. These include T. ni (Tn368) and T, ni High Five (TnSBl-4) (Invitrogen), M . brassicae (Mb) (King et a f . , 1991) and Estigmene acrea (A. L. Lawrie and L. A. King, unpublished data). In some situations, e.g. the expression of recombinant neurotransmitter receptors, Sf9 cells are used in preference to Sf21 cells, as the electrophysiological techniques employed to study these receptors proved difficult using Sf21 cells (King et al., 1992). As mentioned above, the processing of glycoproteins in insect cells results in the addition of mainly mannose-rich side chains which are not further trimmed to form complex oligosaccharides (Kuroda et al., 1990). This observation produces recurrent criticism of the baculovirus expression system because the recombinant material is not identical to the native protein. The glycosylation patterns of influenza haemagglutinin have been compared in Sf9, Tn368 and Esiigmene acrea (Ea) cells (Klenk ei al., 1992). It was shown that the majority of the side chains attached to the haemagglutinin in the Sf9 and Tn368 cells were processed from oligomannosidic to truncated trimannosyl cores. These results were consistent with those of Kuroda et a f . (1990). However, in the Ea cells it appeared that the trimannosyl cores were elongated by the addition of N-acetylglucosamine. Such processing of complex oligosaccharides has not been reported in any other insect cell line to date. The secretion of recombinant human urokinase has also been compared in different insect cell lines: Sf21, Mb, Ea and Tn368 (A. M. Lawrie and L. A. King, unpublished data). In this study the Tn368 and Mb cell lines produced significantly higher yields of intracellular and secreted urokinase than did the Sf21 or Ea cells. Interestingly, secretion of urokinase in the Ea cells was not detected until 48 hpi, whereas in the other cell lines tested, it was apparent in the culture medium after 18-24 hpi.

ADVANCES IN INSECT VIROLOGY

29

2.8,6 Expression of foreign genes in insect cell lines Baculovirus vectors can only produce transient expression of foreign genes in insect cells because the virus infection results in the death of the host cells, after about 72 hpi. This may partly account for the reason that membrane-targeted and secreted proteins are relatively poorly synthesized in recombinant virus-infected cells, as undoubtedly the secretory pathway is compromised very late in the infection cycle. A recent study involved the expression of the bovine GABAA receptor a1- and pl-subunits in insect cells and their subsequent analysis using electrophysiological techniques (Atkinson et al., 1992). It was found that the efficiency of patch clamping to the insect cell plasma membrane was greatly reduced following virus infection. Therefore, in some cases it may be advantageous to use a stable expression system utilizing non-infected cells. It has been demonstrated that the insect cell can be made to continuously express foreign genes by making use of the AcMNPV IE-1 gene promoter incorporated into the insect cell genome (Jarvis et al., 1990b). Since the IE-1 gene promoter is transcriptionally active in the absence of any other viral gene product (see Section 2.6.1), it can be used to express foreign genes in transformed insect cell lines, thus allowing the foreign gene to be continuously synthesized. The stable and continuous expression of tissue plasminogen activator and E. coli /3-galactosidase (Jarvis et al., 1990b) have been described, although the levels of foreign protein synthesized were much reduced compared with those obtained via recombinant virus-infected cells. The expression of functional, GABA-gated homo-oligomeric GABAA receptors, by the integration of the bovine GABAA receptor pl-subunit cDNA under control of the IE-1 gene promoter, has also been described (Joyce et al., 1993).

3 Entomopoxviruses 3.1

ISOLATION A N D HOST R A N G E

Poxviruses in insects were originally described by Vago (1963) and since then entomopoxviruses (EPVs) have been isolated from over 60 different insect species in widespread geographical locations (Arif, 1984). EPVs have been found in all four economically important insect orders, Lepidoptera, Coleoptera, Diptera, and Orthoptera. Compared with several other groups of insect viruses, very little is known about their biochemical and biophysical nature or about the details of their replication cycle (Arif and Kurstak, 1991). In general, EPV infections of Coleopteran species exhibit an exceptionally prolonged course of disease development which may be as long as 30-40

L. A. KING et a/.

30

weeks depending on the temperature (Hurpin and Vago, 1963; Milner and Lutton, 1975). Development of the disease in the cockchafer Mefolontha melofontha (Mm) (Scarabaeidae) is slow, with larvae dying 5-6 months after inoculation. The external symptoms are not very marked; affected larvae are sluggish and lose turgidity. Internally, the adipose tissue and blood cells are primarily affected (Amargier et a f . , 1964; Bergoin et a f . , 1969). All larval stages, particularly the first two instars, are susceptible to infection (Roberts, 1970), and normal development of the insect during the pupal and imaginal moults is arrested. In Lepidopteran EPV infections, the course of disease is relatively short, usually less than 3 weeks. Infected larvae may become lethargic and lose coordination and mobility during the late stages of infection. Death of E. acrea larvae infected with Amsacta moorei (Am) EPV is frequently preceded by paralysis of the abdomen and by regurgitation or defecation of fluid containing virus. The primary site of virus multiplication is in the cytoplasm of fat body cells, although mid-intestinal cells, hypodermis, muscle cells, tracheoblasts, haemocytes and ganglion connective tissue are also affected (Roberts and Granados, 1968). The infection of Lepidopteran pupae by some EPVs has been reported (Retnakaran and Bird, 1972; Sutter, 1972), however, infections of the adult stage are not known. EPV-infected grasshoppers (e .g. Melanoguin sanguinipes, Ms) exhibit a general torpor, take longer to develop and show a high rate of mortality. The bodies of heavily infected nymphs are frequently distended with protruding cervical membranes due to the accumulation of virus in the fat body. The fat body of M . sanguinipes appears to be the only tissue affected by MsEPV. The outward signs of disease in the midge Chironomus luridus (CI) are striking; the entire body exhibits irregular whitish spots caused by massive accumulations of virus occlusion bodies in the fat tissue. Infected fourth instar larvae are of normal size and as active as healthy larvae, but usually die before the next moult (Gotz et al., 1969). The ClEPV appears to be polytrophic, affecting various tissues including fat bodies, haemocytes, epidermis, oenocytes, imaginal discs of the legs and genital organs, muscles, nerve cells, and the intestinal tract (Huger et a f . , 1970). Other chironomid EPVs show a high degree of tissue specificity and are known to infect only fat body and haemocytes (Stoltz and Summers, 1972). Entomopoxvirus-like particles have also been found in three species of bumble bee (Clark, 1982), mosquitoes and other water-borne insects (Lebdeva and Zelenko, 1972). 3.2

STRUCTURE A N D CLASSIFICATION

The EPVs present morphological and physicochemical characteristics that are similar to those of the orthopoxviruses of vertebrates, however they have one distinguishing feature, the virus particles are occluded in a

ADVANCES IN INSECT VIROLOGY

31

TABLE 2 Classification of the entomopoxviruses English vernacular name En tomopoxviruses (EPV) Poxvirus of Coleoptera Poxvirus of Lepidoptera and Orthoptera Poxvirus of Diptera

Taxonomic status (international name) Subfamily (Entomopoxvirinae) Probable Genus (Entomopoxvirus A) Probable Genus (Entomopoxvirus B) Probable Genus (Entomopoxvirus C)

Type species

Melolontha melolontha ( M m ) EPV Amsacta moorei (Am) EPV Chironomus luridus (CI) EPV

paracrystalline matrix, the spheroid. Spheroids appear to be analogous to the baculovirus polyhedra. Because of these similarities, the Entomopoxvirinae have been classified as a subfamily in the Foxviridae family (Francki et al., 1991). The Entomopoxvirinae has been subdivided into three probable genera, as described in Table 2. Historically, an EPV has been named after the host insect from which it was originally isolated. Like the orthopoxviruses, the EPV virion is brick-shaped or oval with sizes ranging from 150 to 470nm long, and from 165 to 300nm wide. Negatively stained virus particles exhibit a folded outer membrane to give the appearance of a mulberry-like surface (Westwood et al., 1964). These spherical folds vary in size depending on the virus species and measure approximately 40 nm for AmEPV and 22 nm for MmEPV (Granados and Roberts, 1970; Bergoin et al., 1971). In cross-section or when the negative stain penetrates the particle, the virion is shown to contain an electrondense core surrounded by a multilayer membrane. EPVs from Orthopteran and Lepidopteran hosts generally contain a cylindrical core and two lateral bodies, while those infecting Dipteran hosts contain a biconcave core and two well-developed lateral bodies. EPVs from Coleoptera contain a unilaterally concave core and one lateral body located in the cavity of the core (Granados and Roberts, 1970; Bergoin and Dales, 1971; Stoltz and Summers, 1972; Granados, 1973a). Three forms of EPV have been identified in virus-infected cells: non-occluded intracellular virus, extracellular released virus and occluded virus. These forms are described in more detail in Section 3.5. Virus occlusion bodies, or spheroids, are spherical or oval in shape and have a paracrystalline matrix which is primarily composed of a single polypeptide, spheroidin. Spheroids and spheroidin appear to be analogous to the polyhedra and polyhedrin protein, respectively, of the baculoviruses (see Section 2.1). Spheroids vary in size from 5 to 24 pm in diameter and contain embedded virus particles arranged in a random or radial orientation. The number of virions occluded is variable, and depends

32

L. A. KING e t a / .

on the host species and even the cell type within the same host. The size, form and distribution of the spheroids also depends on the species of EPV. They range in shape from nearly spherical (Caleopteran hosts) to nearly ellipsoidal (Lepidopteran hosts). Generally, in the infected tissues of Coleoptera, one or more large spheroids develop per cell but in the Lepidoptera, numerous smaller spheroids are found in each cell. The amino acid content of the spheroids of several different EPVs (including AmEPV, E. acrea (Ea)EPV, MmEPV and MsEPV) has been determined (Bergoin et af., 1970; Langridge and Roberts, 1982). It was demonstrated that the acidic amino acids (aspartic and glutamic acids) and the basic amino acids (lysine and arginine) were present in approximately equimolar amounts. The sulphur-containing amino acids, cysteine and methionine, in AmEPV, EaEPV and MsEPV constituted 996, 8.1% and 3.7% of the total amino acids, respectively. From these data, it is likely that the need for disulphide bond-reducing agents during the alkali dissolution of spheroids is due to the formation of disulphide bonds between sulphurcontaining amino acids in the paracrystalline matrix of the spheroid structure. Virions have been reported to contain between 24 and 40 polypeptides, depending on the virus isolate (Bergoin and Roberts, 1971; Langridge and Roberts, 1982), with sizes ranging from 12 to 250 kDa (Bilimoria and Arif, 1980). This represents approximately 38% of the total coding capacity of the genome. Virion cores of a Choristoneura sp. EPV, isolated by treatment with the non-ionic detergent NP-40, have been shown to contain only one major protein VP59 (Bilimoria and Arif, 1980). Four enzymatic activities have been associated with the virion particles of AmEPV: a nucleotide phosphohydrolase (Pogo et af. 1971), acidic and neutral DNAases (Pogo et al., 1971), and a DNA-dependent RNA polymerase (McCarthy et a f . , 1974). An endogenous alkaline proteolytic activity has also been reported to be associated with occlusion bodies isolated from insect larvae. Spheroidin was degraded from a 102 kDa protein to a 52 kDa polypeptide and eventually into smaller polypeptides when the spheroids were dissolved in alkali (Bilimoria and Arif, 1979). Tissue culture-derived spheroids did not exhibit any alkaline protease activity (Langridge and Roberts, 1982). The alkaline protease appears to be similar to the enzyme associated with the matrix protein of baculovirus polyhedra (Epstein and Thoma, 1975; Zummer and Faulkner, 1979). The EPV genome consists of a linear, dsDNA molecule (Gotz et a f . , 1969; Granados and Roberts, 1970; McCarthy et af., 1974) which constitutes about 5% of the viral particle. The genome of AmEPV has been shown to have terminal hairpin loops (Hall and Hink, 1990), that are similar to those found in the orthopoxviruses, and is reported to be 225 kbp by pulse field gel electrophoresis (Hall and Moyer, 1991). The G + C content of EPV DNA is low, varying from 17% to 27% (Arif,

ADVANCES IN INSECT VIROLOGY

33

1976; Langridge et al., 1977), when compared with the G + C content of orthopoxviruses, which ranges from 32.5% to 39% (Joklik, 1962). The significance of this low G + C content is not known. 3.3

REPLICATION CYCLE IN INSECTS

Relatively detailed electron microscopic studies on the replication of EPVs in larvae have been carried out. Since these studies were the result of asynchronous infections, however, the sequence of morphogenic events has been deduced using the vertebrate poxviruses as a model. Most studies have been carried out with AmEPV infections of E. acrea larvae. Infection begins when larvae ingest viral occlusion bodies which dissolve in the alkaline environment of the gut to release virus particles. Granados (1973b) observed that AmEPV was first detected in the gut lumen 1-2 h after per 0s inoculation of E. acrea larvae. The viral envelope then fuses with the plasma membrane of microvilli and subsequently the viral core and lateral bodies enter the cell cytoplasm. This appears to be the normal mechanism of entry into cells, however viropexis has also been demonstrated when larvae receive an intrahaemocoelic injection of virus (Devauchelle er al., 1971). Following uncoating and a period of latency, cytoplasmic foci consisting of either electron-dense amorphous material (type I viroplasm), or aggregates of granular material interspersed with spherical vesicles (type I1 viroplasm) begin to appear in infected cells. The first recognizable viral structures are incomplete crescent-like shells or membranes appearing at the periphery of the virogenic stroma. These membranes develop and eventually enclose a mass of electron-dense material. Electron micrographs have shown that these immature particles consist of an inner trilaminar structure of unit membrane and a spicule coat (Stoltz and Summers, 1972). In type I1 viroplasms, crescent or arch-like envelopes are present in association with fibrillar material containing a large number of vesicles (Bergoin et a f . , 1969). The incomplete viral envelopes progressively close, and in the process, engulf the granulated material which eventually condenses to give the appearance of immature particles found with type I viroplasms. The material inside the particles begins to differentiate and a viral nucleoid forms as a condensed mass. As the nucleoid structure differentiates further into a mature core surrounded by three-layered membrane, the particle begins to assume a more rectangular shape with a concomitant loss of the outer layer of spicules. The lateral bodies also assume a more recognizable structural form. Later the outer membrane is modified by folding to give the appearance of a beaded mulberry-like structure (Devauchelle er al., 1971; Stoltz and Summers, 1972; Granados, 1973a; Bird, 1974; Bergoin et al., 1968a,b, 1969, 1971). The most predominant cytopathic feature of EPV-infected insects is the

L. A. KING e t a / .

34

formation of the large, ovoid occlusion bodies or spheroids within the cytoplasm of infected cells (see Section 3.2). It is likely that the spheroids serve a similar function to the baculovirus polyhedra in protecting virion infectivity during horizontal transmission outside the insect. In some EPV-infected cells, other homogeneous protein structures (spindles), which show a crystalline lattice in thin sections, have been identified. Spindles are smaller than spheroids, measure between 0.5 and 12 p m (Vago and Bergoin, 1968), and are devoid of virions; their function is unknown. The spindle protein is antigenically distinct from that of the virions or spheroids, and in MmEPV the spindle protein has been estimated to have a molecular mass of 50 kDa species (Gauthier et a f . , 1992). 3.4

MOLECULAR STUDIES

A restriction enzyme map has been generated for the AmEPV genome using the enzymes HindIII, BamHI and EcoRI (Hall and Hink, 1990), and showed no similarity to restriction enzyme maps of the vertebrate poxviruses. Extensive genomic heterogeneity was detected in the restriction endonuclease cleavage patterns of DNA from five EPVs (EaEPV, MsEPV, Othnonius batesi (Ob)EPV, CbEPV and AmEPV) and vaccinia virus, strain WR (Langridge, 1984). The first gene to be identified and sequenced from an EPV was the putative spheroidin gene from CbEPV (Yuen et a f . , 1990). The CbEPV spheroidin protein has been reported to be 100 kDa by SDS-PAGE analysis, however the predicted size of the protein following sequencing of the gene was found to be 37 kDa. The authors explained this disparity by suggesting that the spheroidin protein might be glycosylated to form a 50 kDa protein which forms dimers to give the expected molecular mass of 100 kDa. Analysis of the gene sequence suggested that a signal peptide was present at the 5' end and that the gene possessed homology to a baculovirus protein associated with occlusion bodies (Vialard et al., 1990b). A 100 bp sequence upstream of the putative transcription start site was inserted into a vaccinia virus expression vector and the sequence was found to act as an efficient promoter in a recombinant virus (Pearson et a f . , 1991). The spheroidin gene of the AmEPV was sequenced by Hall and Moyer (1991) and surprisingly had little significant homology with the sequence of the CbEPV spheroidin gene. The AmEPV spheroidin gene was identified as a 3.0kbp open reading frame (ORF) potentially encoding a protein of 114.8 kDa, which agreed well with the predicted molecular mass of 115 kDa from SDS-PAGE of purified occlusion bodies. The third spheroidin gene to be sequenced was from MmEPV (Sanz et al., 1992), and this study identified an O R F encoding 942 amino acids, corresponding to a potential polypeptide of 109 kDa. This agreed with the molecular weight of approximately 100 kDa from SDS-PAGE. This gene showed more than 40% amino acid

ADVANCES IN INSECT VIROLOGY

35

homology with the AmEPV, but no homology to the reported CbEPV spheroidin gene sequences. More recently, the spindle (fusolin) gene of MmEPV has been sequenced (Gauthier et al., 1992) and had an unexpected 55.6% amino acid homology to the CbEPV ‘spheroidin’ gene. In a recent study, Hall and Moyer (1993) reported that amino acid sequence data, generated from peptide fragments of purified spheroidin from CbEPV occlusion bodies, differed from that predicted from the reported DNA sequence (Yuen et al., 1990). The new data, however, showed over 80% homology to the predicted amino acid sequence of the AmEPV spheroidin gene (Hall and Moyer, 1993). The sequence data indicate that upstream of the AmEPV and MmEPV spheroidin gene ATG initiation codons is a very A + T rich region containing a TAAATG consensus motif, similar to that found in many late promoters of the vertebrate poxviruses. This sequence was shown by primer extension analysis to be the site of transcription initiation. Primer extension studies also demonstrated that the spheroidin gene mRNA contained 5‘ poiy(A) sequences typical of vertebrate poxvirus late transcripts (Hall and Moyer, 1991). The spindle (fusolin) gene of MmEPV has been identified and sequenced (Gauthier et al., 1992). PAGE analysis of purified spindles indicated a molecular mass of 50 kDa. The spindle gene was shown to encode an O R F consisting of 401 codons. The 5’ region of the gene was shown to act as an early promoter when inserted into a recombinant vaccinia virus. Antibodies generated to the MmEPV spindle protein did not bind to protein extracted from E. acrea insects infected with AmEPV, however they did bind to a 37 kDa protein isolated from CbEPV-infected insects. This suggests that while CbEPV and MmEPV both produce spindle proteins, they are not found in AmEPV-infected insects or cells (Gauthier et al., 1992). A thymidine kinase (TK) gene from AmEPV has been identified and sequenced (Gruidl et al., 1992). Analysis of the data revealed an ORF of 182 amino acids, encoding a polypeptide of 21.2 kDa. Amino acid homology comparisons indicated that the gene was most closely related to the TK genes of poxviruses (45%) and less so to the TK genes of vertebrates (40%). The TK from African swine fever virus (ASF) showed the least homology (31.4%) to the AmEPV TK gene, suggesting that these two viruses are not closely related, although ASF shares some biological features with poxviruses, and both ASF and AmEPV can replicate within arthropod hosts. More recently, Lytvyn et a f . (1992) reported the identification and sequencing of TK genes from C. furniferana (Cf)EPV, CbEPV and AmEPV. The three EPV TK genes were shown to be related, exhibiting 63.2% identity and 9.9% similarity at the amino acid level; only 36.7% identity and 13.6% similarity was observed when the EPV sequences were compared with the TK gene of vaccinia virus. Other partial sequence data for the AmEPV genome have identified a

L. A. KING e t a / .

36

potential nucleoside triphosphate phosphohydrolase I (NPH I) gene which exhibits close homology to the vaccinia NTPase and to the NPH I of CbEPV (Yuen et a f . , 1991). Immediately upstream of the AmEPV spheroidin gene, an ORF has been identified that has significant homology to the capripoxvirus HM3 ORF (Hall and Moyer, 1991). 3.5

REPLICATION I N VITRRO

At present only three EPVs have been reported to replicate in continuous insect cell lines; all have been isolated from Lepidopteran species and, therefore, no cell lines currently exist for the study of the replication of EPVs from Orthopteran, Coleopteran and Dipteran hosts. The ArnEPV has been shown to replicate in a number of cell lines, including E. acrea (EAA-BTI), L. dispar (LD-SDZl), L . dispar (LD-65Z), S . frugiperda (IPLB-Sf21), H . zea (IPLB-1075) and B. mori (Quiot et al., 1975; Granados, 1981; Langridge, 1983a,c; Goodwin et af., 1990; Marlow et al., 1992, 1993). The EAA-BTI cell line was established from primary cultures of E. acrea haemocytes (Granados and Naughton, 1975). Various strains of the L. dispar cell line IPLB-LD-65 (listed in Goodwin et a f . , 1978) have been shown to be capable of supporting the complete replication cycle of AmEPV. Of the different IPLD-65 strains, the IPLB-LD-652 line was reported to contain the highest percentage of cells supporting virus replication. A complex, serum-free medium has been developed for the replication of AmEPV in this cell line. Interestingly, it was found that the sterol concentrations used in serum-free media developed for baculovirus replication studies were not sufficient to support the full replication of AmEPV (Goodwin et a f . , 1990), suggesting that EPV replication makes greater nutritional demands on infected host cells during the replication than do baculoviruses. In addition to the studies on AmEPV replication in v i m , Pseudafetia separata (Ps)EPV has been demonstrated to replicate in two Lepidopteran cell lines isolated from P. separata and B. mori larvae (Hukuhara et af., 1990). A tentative sequence of viral replication events was proposed that was similar to that made by Granados and Roberts (1970) and Devauchelle et af. (1971) for ArnEPV. Adoxphyes orana (Ao)EPV has been shown to replicate in a cell line derived from newborn larvae of A . orana fasciata and Hornona magnanirna (Sato, 1989). Most of the studies on EPV replication in vitro have, however, been performed with AmEPV in either the EAA-BTI or LD-652 cell lines. In EAA-BTI cells, AmEPV DNA synthesis can first be detected between 6 and 12 hpi, with a period of rapid DNA synthesis from 12 to 24 hpi (Langridge, 1983a). The synthesis of viral structural proteins begins at about

ADVANCES IN INSECT VIROLOGY

37

18 hpi with progeny virus detectable after 20 hpi. Extracellular virus and occluded virus are first detected at about 18 hpi and the biosynthesis of virus structural proteins increases rapidly from 18 to 34 hpi. The viral replication cycle resembles that of other poxviruses, except for the appearance of occluded virus late in the infection. Virus-induced rounding of EAA-BTI cells is one of the early features of AmEPV infection, and suggests that virus infection results in a reorganization of the host cell cytoskeleton. In a recent study using immunofluorescent staining of the tubulin and f-actin components of the cytoskeleton, Marlow et al. (1992) demonstrated that the microtubules begin to depofymerize between 12 and 24 hpi, and by 48-72 hpi have further contracted to form a reduced network around the main cell body. At the same time, the f-actin components become rearranged to form distinct foci and microspikes. By 96 hpi, both the tubulin and f-actin are detected in areas of the cytoplasm associated with virus assembly, and by 120 hpi, after the formation of spheroids, the tubulin and f-actin are reduced to sparse patches over the cell surface. Depolymerization of the microtubules by colchicine corresponded to the virus-mediated effects, and virus replication was shown to be unimpeded by colchicineinduced depolymerization. Treatment of the cells with aphidicolin and cycloheximide indicated that the effects on the cytoskeleton in virus-infected cells may have been mediated by both early and late genes (Marlow et al., 1992). Observations on the replication of AmEPV in LD-652 cells revealed similar cytoplasmic events as those observed in virus-infected E. acrea larvae (see Section 3.3) and in the BTI-EAA cell line (Goodwin et al., 1990). However, an absence of type I viroplasms was noted and this may be due to the fact that L. dispar (Lymantriidae) is an alien host species (e.g. the larvae of this species are only susceptible by intrahaemocoelic inoculation), the normal host being A . moorei and the experimental host being E. acrea (both Arctiidae). In EAA-BTI cells we found that mature virus particles were predominantly occluded into spheroids or were found as the non-occluded, intraceilular form; very little virus was released into the culture media. The latter observation probably explains the difficulty in obtaining reliable plaque assays with the EAA-BTI cell line (S. A. Marlow and L. A. King, unpublished observations). A few, limited experiments have been undertaken to discern the ability of EPVs to replicate in vertebrate cells. In murine L-929 cells inoculated with AmEPV, no virus-induced proteins were detected at 37°C using [35S]methionine pulse-labelling. However, in E. acre0 (EAA-BTI) cells inoculated with vaccinia virus, an increase in protein production was detected by ELISA using antiserum raised against purified vaccinia virus (Langridge, 1983b).

L. A. KING eta/.

38

3.6

BIOLOGICALCONTROL

Because of their similarities to the vertebrate poxviruses, few serious studies have reported the use of EPVs as biological control agents. In one study, three viruses, Wiseana nuclear polyhedrosis virus (WNPV) , Wiseana EPV (WEPV) and Wiseana granulosis virus (WGV) that infect the insects W. cervinafa, W. umbraculafa and W. signafa respectively, were compared for their effectiveness as control agents. WNPV, and to a lesser extent WEPV, were found to be effective in controlling their respective hosts to subeconomic levels (Crawford and Kalmakoff, 1975). Aerial application of an EPV and NPV against the spruce budworm at Chapleau, Ontario (Cunningham and McPhee, 1973) indicated that the level of insect control was greater than that achieved by just applying the NPV.

4 Iridescent viruses

The Iridoviridae are a group of large, icosahedral dsDNA viruses that replicate in the cytoplasmic compartment of infected cells (Vaughn, 1979). The viruses infect both vertebrates (amphibians and pleuronect fish) and invertebrates (insects, nematodes and crustaceans) (Kelly, 1985); this review will be confined to current advances made in our understanding of the insect-specific iridoviruses (IVs). 4.1

CLASSIFICATION,ISOLATIONA N D HOST RANGE

The invertebrate IVs are divided into two genera, the small (120-140 nm) IVs or Iridovirus genus and the large (180-200nm) IVs or Chloriridovirus genus (Table 3 ) . The type species of the small IV genus is Chilo iridescent virus type 6 (CIV) which infects the rice stem borer (Chilo suppressah) (Fukaya and Nasu, 1966; Devauchelle et al., 1985), and that of the large IV genus is mosquito iridescent virus type 3 (MIV) which was first isolated from Aedes faeniorhynchus (Clark ef al., 1965). Insect IV infections are characterized by a blue-green or lilac iridescence caused by the formation of crystalline arrays of virus particles in the host tissue (Smith, 1967). Historically, it is this opalesence that has been used as a diagnostic feature of IV infections (Tinsley and Kelly, 1970). In common with most insect viruses, IVs have been named according to the insect host of origin and the sequence of isolation, thus Tipulu IV type 1 (TIV) was the first IV to be discovered (Xeros, 1954; Smith, 1955), and to date about 32 different IVs have been isolated from insects in at least three different orders: Lepidoptera, Diptera and Cdeoptera (Kelly, 1985; Ward and Kalmakoff, 1991;

ADVANCES IN INSECT VIROLOGY

39

TABLE 3 Classification of the Iridoviridae English vernacular name Small iridescent insect viruses Large iridescent insect viruses Frog viruses

Taxonomic status (international name) Genus (Iridovirus) Genus (Chloriridovirus Genus (Ranavirus)

Members (type species first) Chilo iridescent virus (CIV) Insect IVs 1, 2, 6, 9, 10, 16-32 Mosquito iridescent virus (MIV) Insect IVs 3-5, 7, 8, 11-15 Frog virus 3 Frog viruses 1. 2 , 5-24, L2. L4, L5 Tadpole edema virus LT1-4, T6-T20 (Newts) T21 (Xenopus)

Francki et al., 1991). No in vivo replication of insect IVs has been demonstrated in vertebrates (Kelly and Robertson, 1973). Often just one, or a few, obviously infected insects have been recovered from populations of many thousands of apparently healthy individuals (Kelly, 1985). This low frequency of overt infection in the environment contrasts with the high infectivity of virus particles when injected into the haemocoel, and has led to the suggestion that natural transmission of the virus may occur through wounds, by cannibalism or via parasitic nematodes (Ward and Kalmakoff, 1991).

4.2

VIRUS STRUCTURE

Much of the information on the structure of IVs has come from studies on two members, CIV and TIV, and more recently from studies on IV22 (isolated from blackflies) (Devauchelle el al., 1985; Tajbakhsh and Seligy, 1989). Virus particles have an electron-dense core surrounded by an internal lipid-protein envelope and an outer icosahedral shell (Williams and Smith, 1958; Stoltz, 1973; Vaughn, 1979; Orange-Balange and Devauchelle, 1982a,b). This outer protein shell renders the insect IVs resistant to ether (Francki et al., 1991). The dense core contains the dsDNA genome, which in the case of CIV, is packed in a chromatin-like structure with six DNA binding proteins (Cerutti and Devauchelle, 1985). The DNA comprises 11-18% of the virus particle by weight (Vaughn, 1979). The internal lipid membrane is distinct in composition from host cell membranes, is approximately 4 n m thick, and is closely associated with the outer capsid shell (Kelly and Vance, 1973; Stoltz, 1973; Orange-Balange and Devauchelle,

L. A. KING eta/.

40

1982a,b; Kelly, 1985). Protein complexes have been shown to extend throughout the single unit membrane, attaching the core polypeptides to those in the outer capsid, thus probably contributing to the stable nature of these viruses (stable at pH 3-10 and at 4°C for several years) (Klump et al., 1983; Cerutti and Devauchelle, 1985). The outer capsid is icosahedral in shape (Williams and Smith, 1958) and is composed of a lattice of hexagonally packed subunits (Wrigley, 1969). Surface fibres of 3-5 nm in diameter and up to 150nm in length (or microfibrillar fringe) have been reported on the surface of some IVs, including MIV, CIV and IV29 (Willison and Cocking, 1972; Kelly, 1985). It has been suggested that these fibres might be responsible for the interplanar spacings in IV crystals (Kelly and Robertson, 1973). The number of polypeptides associated with the virus particle appears to vary with the method used to prepare purified virions. For CIV, the number has been estimated at 19, ranging from 10 to 213kDa, with the major 65 kDa capsid polypeptide (Kelly and Tinsley, 1972) accounting for 40% of the total virion mass (Moore and Kelly, 1980). The IV22 and TIV capsid proteins have an estimated mass of 48-50 kDa, again comprising about 40% of the total virion mass (Kelly, 1985; Cameron, 1990). Studies by Krell and Lee (1974) have indicated that none of the viral structural proteins are glycosylated. Several enzyme activities have been associated with IV particles including RNA polymerase, nucleotide phosphohydrolase, protein kinase and an alkaline protease (Kelly and Tinsley, 1973; Monnier and Devauchelle, 1980; Farara and Attias, 1983, 1986). The latter may be a larval contaminant as it has not been detected in virus propagated in v i m (Farara and Attias, 1983, 1986). The virus genome consists of one molecule of linear dsDNA with a molecular weight ranging from 100 to 160x lo6 Da (small IVs; Bellet and Inman, 1967; Delius et al., 1984; Ward and Kalmakoff, 1991) and from 160 to 185 x lo6 Da (large IVs; Kelly, 1981). The DNA of two insect IVs (CIV and IVS), in common with the vertebrate IVs, have been shown to be circularly permuted and to have direct terminal repeats accounting for up to 12% (CIV) and 25% (WIV) of the total genome (Delius et al., 1984; Ward and Kalmakoff, 1991). Tajbakhsh and Seligy (1989) have also reported that the TIV genome consists of the major linear dsDNA component and up to three other smaller DNA fragments; the significance of these smaller fragments is uncertain. 4.3

REPLICATION CYCLE

Very little is known about the cellular events that occur in insect IV-infected cells, and surprisingly little is known about the pathway of infection in insect larvae. The latter is probably because of the difficulties in establishing the mechanism of entry of the virus; the initial site of a natural infection is still

ADVANCES IN INSECT VIROLOGY

41

unclear (Kelly, 1985; Ward and Kalmakoff, 1991). After initial infection, the virus spreads throughout the insect tissues (especially the epidermis and fat body) and results in a massive accumulation of virus; 25% by weight of a larva infected with TIV (Williams and Smith, 1957). Within 7-10 days of infection larvae generally become flaccid, with death occurring soon afterwards. Much of what is known about the insect IV replication cycle has come from in vitro studies from either TIV infections in cultured E. acrea or Galleria mellonella haemocytes (Yule and Lee, 1973; Krell and Lee, 1974), or from CIV infections of Choristoneura fumiferana or Trichoplusia ni cells (McIntosh and Kimura, 1974; Cerutti et al., 1981). The gross cytopathological effects of IV infection are characterized by syncytia formation, the marked shut-off of host cell macromolecular (DNA, RNA and protein) synthesis and an increase in cellular DNA and RNA polymerases and thymidine kinase activities (Cerutti and Devauchelle, 1980, 1982; Devauchelle et al., 1985). A virogenic stroma appears in the cytoplasm that contains fine fibrils of viral DNA and proteins (Yule and Lee, 1973). More recently, Cameron (1990) identified three putative phases of protein synthesis in Spodoptera frugiperda cells infected with IV22; these were designated immediate early (until 3 hpi), early (3-9 hpi) and late (from 12hpi). One late polypeptide was shown to migrate with the IV22 major capsid protein. Very little has been published on the molecular aspects of insect IV replication. Instead, comparisons are always drawn with the more extensively studied replication cycle of the vertebrate IV, frog virus 3 (FV3); however, the relevance of the FV3 replication cycle to the insect IVs is not known (Kelly, 1985; Ward and Kalmakoff, 1991). The replication of FV3 has been recently reviewed and will only be briefly summarized here (Murti et al., 1985; Francki et al., 1991). Virus uptake into cells is by pinocytosis and incorporation into phagocytic vesicles, in which the virus particle is uncoated. Replication of FV3 requires a functional host cell nucleus (this requirement has not been conclusively ascertained for insect IVs) in which initial virus DNA replication occurs. After replication, the new DNA molecules migrate to the cytoplasm where further DNA copies are made. The switch to cytoplasmic DNA replication coincides with the synthesis of a virus-encoded protein. In the cytoplasm, the DNA molecules form long concatemers which are necessary for the regeneration of the DNA ends. Early mRNA transcripts are also synthesized in the nucleus using a virus-modified RNA polymerase, leading to early protein synthesis in the first 5 h after infection. The late proteins are produced after 8 h infection and are thought to be associated with DNA replication. Following DNA replication, in the later stages of infection virus release is by budding or lysis. Virus that buds from the cell surface acquires a plasma- or endoplasmic reticulum-derived envelope, although most virus appears to remain cell associated and unenveloped virions are infectious.

42

4.4

L. A. KING et a/. MOLECULAR STUDIES

As mentioned above, there have been few studies on the insect IVs at the molecular level. Physical (restriction endonuclease) maps have been generated for the DNAs of CIV (Soltau et al., 1987) and IV9 (Wiseana spp., WIV) (Ward and Kalmakoff, 1987, 1991), and Fischer et al. (1989) have reported the formation of a complete genome library for CIV, which should form the basis for further studies to examine the structural and functional properties of the IV genome. In addition, DNA-DNA hybridization techniques are now being used for the detection of new IV isolates (Ward and Kalmakoff, 1991) and comparison of restriction endonuclease (RE) profiles can be used to separate and distinguish different IV strains, as has been extensively documented for other insect DNA viruses such as the baculoviruses. The use of these techniques should improve the taxonomy of the IVs, since at present the typing of these viruses is based only upon iridescence, host of origin and date of isolation (Xeros, 1954; Fowler and Robertson, 1972; Batson et al., 1976). Hybridization techniques may also be used in future to study the routes of infection, secondary hosts and virus transmission. This should provide a more reliable indication of infection than the present methods which rely primarily on visual iridescence of infected insects. Three insect IV capsid proteins genes have been identified and sequenced: IV22 (Cameron, 1990), TIV (Tajbakhsh et al., 1990) and CIV (Smith et al., 1993). The predicted molecular mass of the capsid protein of IV22 from the sequence data, 51.9 kDa, is slightly higher than that predicted from SDS-PAGE analyses (49 kDa). In an earlier study, Moore and Kelly (1980) showed that the major capsid protein of three other insect IVs had an N-terminal proline and, therefore, it has been suggested that the capsid of IV22 may be post-translationally processed to give the apparent molecular mass identified on gels (Cameron, 1990). Comparison of the amino acid sequences of the capsid proteins of TIV and IV22 showed that 442 of the first 451 amino acids were identical, however, because of an insertion of an additional 44 bp in the IV22 sequence, the C-terminal sequences had no similarities (Cameron, 1990). A high degree (64.7%) of amino acid sequence identity was reported between the capsid protein of CIV and TIV/IV22 (Smith et al., 1993). It will be interesting to determine if other, more distantly related IVs, also have a conserved capsid protein gene.

4.5

BIOLOGICAL CONTROL

Insect IVs have not been generally considered as viable biological control agents, although they have attracted some interest as a potential means to control some Dipteran pests, an insect group not usually susceptible to the more commonly used baculovirus insecticides. In particular, there has been

ADVANCES IN INSECT VIROLOGY

43

interest in IV22, isolated from the larvae of a blackfly (Sirnufiurn spp.), as blackflies are major vectors of human disease in tropical countries, notably river blindness (T. Williams, personal communication). Part of the reluctance to consider IVs as control agents is the lack of fundamental understanding of their life cycle, in particular how they infect their hosts. Some studies have shown that infection rates in the laboratory may reach levels of 70-90% (Chapman et af., 1971) and it is difficult to correlate this with the proposed mechanisms that exist in the field (i.e. wounding, cannibalism or parasitic inoculation). Use of DNA hybridization techniques to identify virus-infected insects, rather than visual iridescence, has recently been used in one study to show that 100% of G. melfoneffalarvae were susceptible to infection with WIV; none of the larvae showed signs of iridescence (Ward and Kalmakoff, 1991). Thus, IV infections in natural populations may be higher than suggested in the early literature, where infection rates were based solely on the identification of iridescence. Other studies have shown that massive virus proliferation is not necessary for IVs to cause insect death (Sieburth and Carner, 1987), which may also help to explain the lack of iridescence in many infected larvae. These studies, together with the advances that are being made in the molecular characterization of the IV genome, which may permit strain selection or genetic manipulation, suggest that the IVs may in future be considered as serious candidates for the biological control of pest species.

5 RNA viruses of insects 5.1

INTRODUCTION

There have recently been a number of excellent reviews on the three families of small insect viruses that have an ssRNA genome: Nodaviridue (Garzon and Charpentier, 1991; Hendry, 1991), Tetraviridae (Moore, 1991a; Reinganum, 1991), and Picornaviridae (Moore, 1991b; Moore and Eley, 1991). This section will, therefore, include the general classification and structure of these viruses (see Table 4), but will be primarily confined to a discussion of the more recent advances in our understanding of the replication of these groups of viruses. In addition, there is one group of segmented dsRNA viruses that infect insects; these are the cytoplasmic polyhedrosis viruses. 5.2

NODAVlRIDAE

The Nodaviridue are non-enveloped, icosahedral insect pathogenic viruses, which contain a bipartite RNA genome encapsidated within a single virion (Matthews, 1982). The single-stranded RNAs are messenger-sense and both

44

L. A. KING e t a / .

are required for infectivity (Newman and Brown, 1973; Gallagher et al., 1983). The virus particles have a diameter of 3 0 n m and the sizes of the major coat protein and the two genomic RNA species, RNAl and RNA2, are 40 kDa, 1.1 x 10' Da and 0.47 x 10' Da, respectively (Garzon and Charpentier, 1991; Hendry, 1991). As a family they have been more thoroughly studied than any of the other small insect riboviruses, and their initial popularity probably arose because their type member, Nodamura virus, was found to infect mammals as well as insects (Scherer and Hurlbut, 1967). Interest eventually spread to other nodaviruses when it was discovered they were easier to amplify in tissue culture, and were as a consequence more amenable to molecular analyses. A number of significant advances in nodavirus research have previously been reviewed (Garzon and Charpentier, 1991; Hendry, 1991) including the respective roles of RNAs 1 and 2 together with their complete nucleotide sequences (Friesen and Rueckert, 1981, 1982, 1984; Gallagher et al., 1983; Dasgupta et al., 1984); the first production of an infectious mRNA transcript from a cloned cDNA (of insect or animal RNA virus) (Dasmahapatra et al., 1986); and the elucidation of the atomic structure of the viral capsid at 3.0 A resolution. 5.2.1 Isolation and host range Nodamura virus was first isolated from mosquitoes and was initially classified as an arbovirus. Later, on the basis of physicochemical properties, this was changed to membership of the picornaviruses (Murphy et al., 1970). However, when it was demonstrated that the virus had a bipartite genome, the virus was classified in a new family, the Nodaviruses (Newman and Brown, 1973, 1977). Six viruses have been classified as members of the Nodaviridae (Table 4); all have been isolated from insects (Lepidoptera, Coleoptera and Diptera) but none appear to be host specific. Nodamura virus is the only virus known to infect other animals (e.g. mice) as well as insects. Little is known about the spread of nodavirus infections in natural insect populations, as each virus was isolated from individual insects in a single location (Hendry, 1991). For example, Nodamura virus was isolated from female Culex tritaeniorhynchus mosquitoes trapped at Nodamura, near Tokyo, Japan in 1956; black beetle virus was isolated from Heteronychus arator collected near Helensville in New Zealand in 1974 (Longworth and Archibald, 1975); and Flock House virus was isolated from the New Zealand grass grub, Costelytra zealandica, collected at Flock House in New Zealand in 1980 (Dearing et al., 1980). 5.2.2 Structure and classification The six viruses which have been classified by the International Committee on Taxonomy for Viruses (ICTV) as belonging to the family Nodaviridae

ADVANCES IN INSECT VIROLOGY

45

TABLE 4 R N A viruses of insects" Family (Genus)

Species

Nodaviridae

Nodarnuara virus (NMV) Black beetle virus (BBV) Flock House virus (FHV) Boolarra virus (BoV) Gypsy moth virus (GMV) Manawata virus (MwV)

Tetraviridae (Nudaurelia p virus)

Nudaurelia p virus (NdpV) Antheraea /3 virus (AnPV) Darna p virus (DrpV) Thosea G , virus (ThPV) Philosamia /3 virus (PhPV) Dasychira p virus (DspV) Trichoplusia /3 virus (TrPV)

Picornaviridae

Cricket paralysis virus (CrPV) Drosophila C virus (DCV) Gonometu virus (GV)

Reoviridae (Cytoplasmic polyhedrosis virus, CPV)

Bombyx mori CPV (type species; type 1) lnachis io CPV (type 2 ) S. exempta CPV (type 3 ) Actias selene CPV (type 4) T. ni CPV (type 5 ) Biston betularia CPV (type 6) Triphena pronuba NPV (type 7) Abraxas grossulariata CPV (type 8) Agrotis segetum CPV (type 9) Aporophytla iutulenta CPV (type 10) S. exigua CPV (type 11) S. exempta CPV (type 12)

"Francki et ul 1991: Garzon and Charpentier. 1991; Hendry, 1991; Moore. 1991a,b; Moore and Eley, 1991; Reinganum. 1991.

L. A. KING eta/.

46

are Nodamura virus (NV), black beetle virus (BBV) (Longworth and Carey, 1976), Flock House virus (FHV) (Dearing et al., 1980); Boolarra virus (BoV) (Reinganum et a f . , 1985), gypsy moth virus (GMV) (Reavy et al., 1982), and Manawata virus (MwV) (see Table 4). Serology has been the most convenient method for identifying nodaviruses, although none of the six identified in Table 4 cross-react with all the others. In addition, nodaviruses can be diagnosed by their particle diameter (about 30 nm), size of the major coat protein (about 40 kDa) and the presence of two RNA species (about 0.5 and 1.0 X lo6 Da). On the basis of such comparisons of morphological and biochemical properties, Mori et al. (1992) have proposed that a new virus, designated Striped Jack Nervous Necrosis Virus (SJNNV), be included in the Nodaviridae. SJNNV is non-enveloped with a bipartite, positive-sense genome and has a virion diameter of approximately 25 nm. The genomic RNAs have molecular masses of 1.01 X lo6 Da (RNA1) and 0.49 x lo6 Da (RNA2), and neither have a poly(A) tail at the 3' terminus. There are two viral structural proteins with molecular masses of 42 and 40 kDa, and RNAs 1 and 2 direct the synthesis of proteins of 100 kDa and 42 kDa, respectively, in reticulocyte Iysates. The physicochemical properties of all members of the nodavirus family have been summarized by Hendry (1991). Briefly, virions are 29-31 nm in diameter based on electron microscopy measurements. BBV has also been measured by small-angle X-ray scattering (Hosur et al., 1984) and this has given an accurate sizin of 31.2 nm. The structure of the BBV capsid has been determined at 3 resolution using X-ray diffraction of virus crystals (Hosur et al., 1984). The data obtained from this study have been reviewed (Hendry, 1991) and indicated that the BBV capsid is composed of 180 identical protomers arranged in a T = 3 icosahedral symmetry (Hosur et al., 1984). Each protomer was shown to consist of 407 amino acids; either as the alpha precursor polypeptide or as the two cleavage products, beta and gamma (see below). Using the information obtained from sequencing of the genome and predicted protein sizes, the virion particle weight for BBV is 9.4 X lo6 Da. The same data indicate that, for BBV, the RNA content comprises 16% of the virion mass. Virions are resistant to both ether and chloroform, suggesting that they do not contain any lipid. Neither FHV, BBV nor NV are affected by acidic pH treatments but heating to 5040°C appears to result in loss of infectivity of NV and FHV (Murphy et al., 1970; Scotti et al., 1983). Using BBV as the best studied example, the nodavirus capsid consists of a major (beta, 39 kDa) and two minor protein species (gamma, 4.5 kDa; and a precursor, alpha, 44kDa) (Friesen and Rueckert, 1981). The alpha precursor is cleaved between an asparagine residue at position 363 and an alanine residue at 364, to generate beta and gamma (Dasgupta et al., 1984). There is considerable variation in the reported number and sizes of the minor proteins of the other nodaviruses. It is only in the case of BBV, which

1

ADVANCES IN INSECT VIROLOGY

47

has been sequenced, that the relationship between alpha, beta and gamma has been firmly established (Dasgupta et al., 1984).

5.2.3 The bipartite RNA genome Both nodaviral RNA species are messenger-sense, single-stranded and both are necessary for infectivity (Newman and Brown, 1973; Gallagher et al., 1983). The BBV RNAs have been shown to have typical methylated guanine cap structures at t h e 5' terminus, but the 3' terminus is thought to be modified so as to prevent polyadenylation. The larger of the two RNAs (RNA1) has a molecular mass of about 1.1 X 10' (3106 bases (b) for BBV; Kaesberg, 1987) and encodes two viral proteins, one of which is known to be a viral replicase. The smaller RNA2 molecule has a molecular mass of about 0.47 X lo6 (1399 b for BBV; Kaesberg, 1987) and encodes the viral coat protein precursor (alpha). Computer analysis using the nucleotide sequence data of BBV predicts a stable and distinctive secondary structure at both the 3' and 5' ends of RNA2 (Dasgupta et a!., 1984). The mechanism by which each of the genomic RNAs is selected and packaged into the same virion is unknown. Recently, however, a report by Zhong et a f . (1992) indicated that a 32-base region of the RNA2 (bases 186-217) of FHV appears to be important in this process. Analysis predicts that the RNA is folded in this region into a stem-loop structure with a 5-base loop and a 13 base-pair bulged stem.

5.2.4 Molecular studies The demonstration that RNA transcripts of BBV cDNA are infectious in cultured Drosophilu cells (Dasmahapatra et al., 1986) allows the possibility of modifying the nodaviral genome for analysis by recombinant DNA techniques. Dasmahapatra er al. (1987) devised a cell-free expression system by constructing a plasmid containing a translation initiation signal from the 5' non-coding region of BBV. They inserted the 3C protease coding region of Coxsackievirus B3 (CVB3) next to the ribosome binding sequences and initiator A U G site of the nodavirus. Transcripts of this plasmid directed the efficient synthesis of an active protease. More recently, Dasmahapatra et al. (1991) demonstrated that using this system, a biologically active protease is synthesized which possesses both cis and trans processing capabilities. This in v i m synthesized protease is analogous to the native 3C produced by CVB3-infected HeLa cells, and antibody prepared against the native protease cross reacts with the in vitro protease. Using the translational initiation signal from BBV RNA1, the authors have also expressed the CVB3 capsid precursor and part of the P2 region in vitro. Additionally, they report that the capsid precursor is cleaved, between IC (VP3) and 1D (VPI), by the proteolytic activity of in vitro synthesized 3C in trans.

,

L. A. KING e t a / .

48

The amplification of RNA that is a result of RNA replication has been found to occur naturally only in RNA viruses. Nodavirus cDNA sequences have recently been used in an attempt to harness this power for the amplification of heterologous mRNAs (Ball, 1992). The authors have expressed both an RNA replicase and its corresponding RNA templates in functional form, using a vaccinia virus-bacteriophage T7 RNA polymerase vector. Plasmids were constructed which contained in 5' to 3' order: a T7 promoter; a full length cDNA encoding either the RNA replicase (RNAl) or the coat protein (RNA2) of FHV; a cDNA encoding the self-cleaving ribozyme of satellite tobacco ringspot virus; and a T7 transcriptional terminator. RNAs were produced, both in vivo and in vitro, by T7 RNA polymerase with sizes that closely resembled those of the two authentic genomic RNAs (RNA1 and 2). In baby hamster kidney (BHK) cells that expressed authentic FHV RNA replicase, the RNA2 transcripts were accurately replicated. More importantly, the RNAl transcripts directed the synthesis of an enzyme that could replicate not only authentic virion-derived FHV RNA, but also the plasmid-derived transcripts themselves. Under the latter conditions, replicative amplification of the RNA transcripts ensued and resulted in a high rate of synthesis of the encoded products. This successful expression from a DNA vector of the complex biological process of RNA replication will greatly facilitate studies of its mechanism and is a major step towards the goal of achieving RNA replication for mRNA amplification.

5.3

P

T E T R A V I R I D A E ( N C J D A C J R E L I A VIRUS GROUP)

The Tetraviridae, also referred to as the Nudaurefia P virus family after the type member of the group, are relatively important small insect riboviruses that naturally regulate a number of pest species of Lepidoptera. There is, however, no tissue culture system available to support their replication, and so they are not widely studied. Past work on these viruses has been restricted to physicochemical, in vivo pathogenicity and replication studies using in vitro translation systems (Hendry, 1991; Moore, 1991a).

5.3.1 Classijication, isolation and host range This family of insect viruses comprises seven members (see Table 4) that have all been isolated from pest species of Lepidoptera. All members replicate exclusively in insect hosts. The type member is Nudaurefia capensis p virus (NPV) which was originally identified from the pine emperor moth, N . cytharea capensis (Struthers and Hendry, 1974; du Plessis et a f . , 1991). Using antisera raised against purified virions, five other members of this family have been identified, as shown in Table 4. The final member,

ADVANCES IN INSECT VIROLOGY

49

Nudaurelia capensis virus (NV), was also identified from the pine emperor moth, but showed no serological relationship to NPV.

5.3.2 Virion structure The virions of the NPV are non-enveloped, icosahedrons with a diameter of 3540 nm and contain a single-stranded RNA genome with a molecular mass of about 1.8 x lo6 Da. The virion capsid contains a single protein, which has a molecular mass of 60-68 kDa. The biophysical properties of the individual group members have recently been reviewed by Moore (1991a) and Olson et nl. (1990). Olson et al. (1990) have recently reconstructed the three-dimensional structure of Nudaurelia P virus (NPV), to a resolution of 3.2nm, using images of frozen-hydrated virions. The model of a distinctly icosahedral capsid (with 240 copies of a single 61 kDa subunit) and T = 4 symmetry compares well with what was previously observed with negative staining using electron microscopy. The authors state that analyses of the density maps, volume estimates and model building experiments, indicate that each subunit consists of two domains. Each large cylindrical domain (40 kDa; 4 X 4 nm) associates with two other large domains in neighbouring subunits to form a Y-shaped trimeric aggregate in the outer capsid surface. Four trimers come together to make each of the 20 planar faces of the icosahedron. The small domains (21 kDa; 13-16.5 nm) are presumed to be associated with the inside of the virion to make a contiguous, non-spherical shell. The small ssRNA genome is loosely packed inside the capsid with a low average density. Similar physical characteristics were reported for NV, following virus crystallization and X-ray diffraction studies at 2.8 8, resolution (Cavarelli et al., 1991).

5.3.3 Replication and molecular studies Very little is known about the replication cycle or genetic organization of this group of viruses, mainly due to the lack of suitable cell culture systems capable of supporting virus replication. Electron microscopy and ELISA studies have indicated that replication in vivo occurs, at least initially, in the cytoplasm of fore- and midgut cells (Reavy and Moore, 1982). In vitro replication studies have been confined to translation of the viral RNA species in rabbit reticulocyte lysates. A number of putative polypeptides have been identified in this way, including a potential capsid precursor and polymerase, although it was not possible in these experiments to draw definite conclusions (Reavy and Moore, 1984; King et al., 1984). In a recent study, Agrawal and Johnson (1992) report the sequencing and analysis of the second RNA species of NV. It was found to consist of 2448 b and contained one long ORF encoding a 644 amino acid capsid protein

L. A. KING eta/.

50

precursor (70kDa). The 5' and 3' ends contained 366 and 150 b of non-coding region, respectively. The capsid protein precursor was shown to start at the second AUG initiation codon. The authors identified a putative asparagine/phenylalanine cleavage site in the precursor protein, which yields the previously identified capsid protein of 62 kDa, and a smaller polypeptide of 8 kDa that was also discovered by the authors in mature virus particles (Agrawal and Johnson, 1992). du Plessis et a!. (1991) reported that the replicative form of NPV is a dsRNA molecule. They found that larvae of the pine emperor moth consistently contained a ds species of RNA with the expected size for a ds replicative form. In Northern blots, it hybridized with a '*P-labelled vRNA probe, whereas other smaller dsRNAs did not, and cell extracts from non-infected larvae contain no dsRNA molecules. 5.3.4 Biological control All members of this family of viruses have been isolated from pest species of Lepidoptera and, to date, analyses indicate that replication is confined to insect hosts, making these viruses potential candidates for biological control programmes (Moore, 1991a). Darna trima is an important pest of several crops, including coconut and oil palms in Southeast Asia, and Thoseu asigna is a pest of oil palms in Malaysia. The viruses isolated from N . cytherea capensis and D. trima have been shown to cause large reductions in insect populations, affecting up to 90% of the larvae, and the latter virus has been used successfully in biological control programmes (Moore, 1991a). 5.4

PICORNAVIRIDA E

There are more than 30 insect viruses which have been proposed as members of the Picornuviridae, but only three have been studied in enough detail to be classified as such by the International Committee on Taxonomy for Viruses. The three that have been assigned to this group, although not to specific genera, are cricket paralysis virus (CrPV), Drosophila C virus (DCV) and Gonometa virus (GV). Of these, replicating isolates exist only for CrPV and DCV, and so most of the existing information about insect picornaviruses comes from work performed with these two viruses (Moore, 1991b). 5.4.1 Isolation, classification and host range CrPV was first identified during a mass breeding programme of the Australian field cricket Teleogryllus oceanicus. Early instar nymphs became paralysed and eventually died (Reinganum et al., 1970; Reinganum, 1973). When virus particles were injected into crickets, death occurred within 3

ADVANCES IN INSECT VIROLOGY

51

days. DCV was first identified in a laboratory colony of the fruit fly Drosophila melanogaster (reviewed in Moore, 1991b) and was shown to be pathogenic for the flies. GV was isolated by Harrap et al. (1966) from the insect Gonomefa podocarpi, which at the time was a serious pest of exotic pines in Uganda. GV proved to be an extremely effective biocontrol agent for this insect species (Longworth ef a f . , 1973). One problem with GV is that no laboratory stocks of the virus exist, so it is difficult to include this virus in any classification system (Moore, 1991b). A number of other picorna-like viruses have more recently been isolated from a variety of different insects (Thomas-Orillard, 1988; see Table 4). Kawino virus was isolated from Mansonia uniformis mosquitoes in Kenya (Pudney ef al., 1978). The virus was also shown to replicate in three mosquito cell lines at 28°C and had most of the physicochemical properties associated with picornaviruses, with the exception that the 3' end of the RNA genome appeared not be be polyadenylated. The host ranges of CrPV and DCV have been evaluated by the experimental infection of purified virions into a range of hosts and by the serological identification of related viruses in natural insect populations. The results of these tests have recently been reviewed by Moore (1991b). However, as both procedures are open to criticism, the natural host range of these viruses remains to be determined. Fediere et al. (1990) have described a new picorna-like virus from the oil palm pest Lafoia viridissima (Parasa viridissima]. Virions of this new virus have a diameter of 30 nm which is composed of four proteins, two major (30 and 31 kDa) and two minor. The genome was shown to comprise a single molecule of ssRNA with a molecular mass of 2.9 x lo6 Da. In a comparative in vitro study of picornaviruses in Drosophila cells, Plus (1989) has suggested that three major conclusions can be reached about DCV isolated from Drosophila and CrPV isolated from different Grylid and Lepidopteran populations: they are both true picornaviruses but are unique in their properties; they are serologically related but are appearing increasingly different as research continues; and that DCV isolated from insect populations and screened on Drusophita cell lines form a homogeneous group, whereas CrPV isolates from five different insect species do not. Despite the fact that the latter were screened on Drosophila cell lines, they were found to be divided into three distinct host range/ serological groups. As previously mentioned, there are many picorna-like viruses that are not well enough characterized to be classified as members of the Picornaviridae. Two of these viruses, aphid lethal paralysis virus (ALPV) and Rhopalosiphum padi virus (RhPY), have recently been the focus of a number of studies and may now be closer to being assigned family status. Williamson et al. (1989) published an account of comparisons made between a South African strain of RhPV and an isolate from Illinois, USA. Both viruses were

52

L. A. KING e t a /

serologically related but there were differences in their capsid protein molecular weights and virion buoyant densities. The South African isolate was found to have two more capsid proteins than was previously thought, and the size of the RNA, which contains a poly(A) tract, was found to be 10 kb. The virions were stable at pH 5.0-8.4, degraded quickly below pH 3.0 and the virions lost genomic RNA when heated to 56°C. From this evidence, together with the results of radio-iodination and Western blotting of capsid proteins, the authors proposed that RhPV be classified as a picornavirus. Support for the inclusion of both RhPV and ALPV into this family is given in the report by Williamson and Rybicki (1989), which details a comparative study on the cell-free translation of the genomic RNAs of both viruses. Studies have also been performed, using nucleic acid in situ hybridization, to identify and localize actively replicating ALPV in whole body sections of aphids (Hatfill et al., 1990). Virus was detected in the stomach and intestinal epithelia of infected Rh. padi aphids and, in the advanced stages of infection, virus spread throughout the protocerebrum. In another report, the pathogenicity, host specificity and tissue specificity of RhPV were studied by inoculation of virus-free clones of five aphid species with gradient-purified virus (Gildow and D’Arcy, 1990). Only two out of the five test species of insects became infected by RhPV, and upon ultrastructural examination, virus was only visualized in the posterior region of the midgut and the hindgut. The authors suggested that viral entry was via endocytosis and that infection resulted in progressive loss of cytoplasmic organelles and the development of membrane vesicles. Virions accumulated only in the cytoplasm and were subsequently released into the gut lumen and haemolymph. The results confirm the pathogenicity of RhPV and suggest a high degree of host and tissue specificity for the virus. 5.4.2 Virion structure In common with their mammalian counterparts, insect picornaviruses have non-enveloped, icosahedral virions which contain a single-stranded, messenger-sense RNA genome (2.5-3.0 x lo6 Da). Virus particles are about 27nm in diameter and the vRNA has been shown to possess a genomelinked protein (VpG; approximately 3.9 x lo3 Da) at the 5’ end and a poly(A) tract at the 3‘ end (King et al., 1987; Moore et al., 1987; King and Moore, 1988). The capsid is composed of 60 copies each of four viral proteins (VP1-4, although CrPV appears to have only VP1-3), and these have molecular masses of 31-35, 30-34, 28-30 and 5.5-13.5 kDa, respectively (Moore, 1991b). As with the Enterovirus genus of the Picornaviridue, insect picornavirus particles are acid stable; with the vertebrate viruses this protects the virus in the gut. The significance of this finding for the insect picornaviruses is uncertain as the insect gut is alkaline.

ADVANCES IN INSECT VIROLOGY

53

5.4.3 Virus replication and molecular studies Most of the information on the replication and molecular biology of insect picornaviruses has come from in vitro studies of CrPV infection in insect cell lines. CrPV will replicate in a number of cell lines including D. mefanogasfer, T. ni, L. dispar and A. aegyptii (Reinganum, 1973; Scotti, 1976). Most of the replication studies have utilized Scheider's Drosophifa DL1 cells, as these produce a reliable cytopathic effect which can be used to titrate the virus by plaque assay (Scotti et af., 1981). Both CrPV and DCV replicate to high titres in the DL1 cells (Moore et af., 1981a; Moore and Pullin, 1982), with the CrPV infection being more lytic and DCV more cell associated. About 20 virus-specified polypeptides have been identified in CrPVinfected DLl cells using SDS-PAGE (Moore et a f . , 1982). Moore and co-workers have shown that virion structural protein synthesis, from high molecular weight precursors, is similar to that observed for mammalian picornaviruses (Moore et al., 1980, 1981a). A putative VPO, the precursor of VP2 and VP4 in mammalian picornaviruses, was identified in both CrPVand DCV-infected cells. With DCV, VPO was shown to be the precursor of VP2 and VP4 by pulse-chase experiments, but with CrPV, there was no evidence for VP4 (Moore et al., 1980, 1981b). Pactinomycin mapping studies have demonstrated that the order of capsid proteins in the CrPV polyprotein is VPO, VP3, VPl and that these are encoded at the 5' terminus of the genome (Reavy and Moore, 1982). The CrPV genome has a poly(A) tail and a 5' linked VPg (King and Moore, 1988). Sequence data generated from the 3' terminal 1600 bases of the CrPV genome confirmed the presence of the poly(A) tail, a 3' non-coding region and an ORF that encodes a potential RNA polymerase (King et al., 1987). 5.4.4 Biological control There are relatively few reports of the use of small RNA viruses of insects as biological control agents. As mentioned above, GV has been used successfully in Uganda to control G. podocarpi. There have been two recent reports describing the potential use of the insect picornviruses as control agents. Manousis and Moore (1987) demonstrated that CrPV caused a high mortality when tested against Dacus ofeae, a serious pest of olive plantations, and Fediere et al. (1990) reported that different doses of a picorna-like virus sprayed onto an industrial oil palm plantation infested by P. viridissima gave good control of the pest. After a period of 1 week, a dose-related mortality gradient ranging from 11% to 61% was obtained. Two weeks after spraying, the mortality of larvae in treated plots reached 92% and during the next generation the numbers of larvae were very low.

54

L. A. KING eta/.

References Agrawal, D. K. and Johnson, J . E. (1992). Sequence analysis of the capsid protein of Nudaurelia capensis omega virus, an insect virus with T equals 4 icosahedral symmetry. Virology 190, 806-814. Akiyoshi, D.. Chakerian, R., Rohrmann, G . F., Nesson, M. H. and Beaudrea, G. S. (1985). Cloning and sequencing of the granulin gene from the Trichoplusia ni granulosis virus. Virology 141, 328-332. Amargier, A , , Vago, C. and Meynadier, G. (1964). Etude histopathologique de I’evolution de la virose a fuseaux chez le Coleoptere Melolontha rnelolontha. Mikroskopie 19, 309-315. Anderson, I. G. and Prior, H. C. (1992). Baculovirus infections in the mud crab Scylla serrata, and a freshwater crayfish, Cherax quadricarinolus, from Australia. J . Invert. Pathol. 60, 265-273. Arif, B. M. (1976). Isolation of an entomopoxvirus and characterization of its DNA. Virology 69, 626-634. Arif, B. M. (1984). The entomopoxviruses. Adv. Virus Res. 29, 195-213. Arif, B. M. (1986). The structure of the viral genome. I n “The Molecular Biology of Baculovirmes’‘ (Eds W Doerfler and i? Bohrn), pp. 21-29, Sprri-lger-Verlag, Berlin. Arif, B. M. and Kurstak, E. (1991). The entomopoxviruses. I n “Viruses of Invertebrates” (Ed. E. Kurstak), pp. 179-195, Dekker, New York. Atkinson, A. E., Weitzman, M. D., Obosi, L., Beadle, D. J. and King, L. A. (1990). Baculoviruses as vectors for foreign gene expression in insect cells. Pest. Sci. 28, 215-224. Atkinson, A. E . , Bermudez, I . , Darlison, M. G., Barnard, A. E., Earley, F. P. G., Possee, R. D., Beadle, D. .I.and King, L. A. (1992). Assembly of functional GABAA receptors in insect cells using baculovirus expression vectors. NeuroReport 3, 597400. Ball, L. A. (1992). Cellular expression of a functional nodavirus RNA replicon from vaccinia virus vectors. J. Virol. 66, 2335-2345. Batson, B. S., Johnson, M. R. L . , Arnold, M. K. and Kelly, D. C. (1976). An iridescent virus from Sirnuliurn sp. (Diptera Simuliidae) in Wales. J. Invert. Pathol. 27, 133-135. Beames, B. and Summers, M. D. (1988). Comparisons of host cell DNA insertions and altered transcription at the site of insertions in few polyhedra baculovirus mutants. Virology 162, 206-220. Beames, B. and Summers, M. D. (1989). Location and nucleotide sequence of the 25K protein missing from baculovirus few polyhedra (FP) mutants. Virology 168, 344-353. Beames, B. and Summers, M. D. (1990). Sequence comparison of cellular and viral copies of host cell DNA insertions found in Autographa californica nuclear polyhedrosis virus. Virology 174, 354-363. Bellet, A. J. D. and Inman, R. B. (1967). Some properties of deoxyribonucleic acid preparations from Chifo, Sericesthis and Tipula iridescent viruses. 1. Mol. Biol. 25, 425-532. Bergoin, M. and Dales, S. (1971). Comparative observations on poxviruses of invertebrates and vertebrates. In “Comparative Virology” (Eds K. Maramorosch, and E. Kurstak), pp. 169-205, Academic Press, New York. Bergoin, M. and Roberts, D. W. (1971). Enzymes associated with an insect poxvirus. Virology 43, 306-309. Bergoin, M., Devauchelle, G . , Duthoit, J. L. and Vago, C. (1968a). Etude au

ADVANCES IN INSECT VIROLOGY

55

microscope electronique des inclusions de la virose a fuseaux des coleopteres. C. R. Acad. Sci. 266D, 2126-2128. Bergoin, M., Devauchelle, G . and Vago, C. (1968b). Observations au microscope electronique sur le developpement du virus de la maladie a fuseaux du coleoptere Melolontha melolontha, L. C. R. Acad. Sci. Ser. D267, 382-385. Bergoin, M . , Devauchelle, G. and Vago, C. (1969). Electron microscopy study of the pox-like virus of Melolontha melolontha L.: Virus morphogenesis. Arch. Ges. Virusforsch 28, 285-302. Bergoin, M., Veyrunes, J . C. and Scalia, R. (1970). Isolation and amino acid composition of the inclusions of Melolontha melolontha poxvirus. Virology 40, 760-763. Bergoin, M., Devauchelle, G. and Vago, C. (1971). Electron microscopy study of Melontha poxvirus: The fine structure of occluded virions. Virology 43, 453467. Biever, K . D. and Wilkinson, J . D. (1978). A stress induced granulosis virus of Pieris brassicae. Environ. Entomol. 7, 572-573. Bilimoria, S. L. (1986). Taxonomy and identification of baculoviruses. In “The Biology of Baculoviruses, Vol I , Biological Properties and Molecular Biology” (Eds R. R. Granados and B. A. Federici), pp. 37-59. CRC Press, Boca Raton, Florida. Bilimoria, S. L. and Arif, B. M. (1979). Subunit protein and alkaline protease of entomopoxvirus spheroids. Virology 96, 596-603. Bilimoria, S. L. and Arif, B. M. (1980). Structural polypeptides of Choristoneura biennis entomopoxvirus. Virology 104, 253-257. Bird, E. T. (1974). The development of spindle inclusions of Choristoneura fumiferana (Lepidoptera: Torticidae) infected with entomopox virus. J. Invert. Pathol. 23, 325-332. Bishop, D. H. L. (1986). UK release of a genetically marked virus. Nature, London 323, 496. Bishop, D. H. L. (1989). Genetically engineered viral insecticides: a progress report 1986-1989. Pest. Sci. 27, 173-189. Bishop, D. H. L. and Possee, R. D. (1990). Baculovirus expression vectors. Adv. Gene Technol. 1, 55-72. Bishop, D. H. L., Entwistle, P. F., Cameron, I. R., Allen, C. J. and Possee, R. D. (1988). Field trials with genetically engineered baculovirus insecticides. In “The Release of Genetically Engineered Micro-organisms” (Eds M. Sussman, C. H. Collins, F. A. Skinner and D. E. Stewart-Tull), pp. 143-179, Academic Press, New York and London. Blinov, V. M. (1984). Nucleotide sequence of the Galleria mellonella nuclear polyhedrosis virus origin of DNA replication. FEBS Letters 161, 254-258. Blissard, G. W. and Rohrmann, G. F. (1989). Location, sequence, transcriptional mapping and temporal expression of the gp64 envelope glycoprotein gene of the Orgyia pseudotsugata multicapsid nuclear polyhedrosis virus. Virology 170, 537-555. Blissard, G. W. and Rohrmann, G. F. (1990). Baculovirus diversity and molecular biology. Annu. Rev. Entomol. 35, 127-155. Blissard, G. W. and Rohrmann, G. F. (1991). Baculovirus gp64 gene expression: analysis of sequences modulating early transcription and transactivation by IE-1. J . Virol. 65, 5820-5827. Blissard, G. W., Quant-Russell, R. L., Rohrmann, G. F. and Beaudreau, G. S. (1989). Nucleotide sequence, transcriptional mapping and temporal expression of the gene encoding p39, a major structural protein of the multicapsid nuclear polyhedrosis virus of Orgyia pseudotsugata. Virology 168, 354-362. Brassel, J. and Benz, G. (1979). Selection of a strain of the granulosis virus of the

56

L. A. KING eta/.

codling moth with improved resistance against artificial ultraviolet radiation and sunlight. J. Invert. Pathol. 33, 358-363. Burand, J . P., Horton, H. M., Retnasami, S. and Elkington, J . S. (1992). The use of polymerase chain reaction and shortwave UV irradiation to detect baculovirus DNA on the surface of gypsy moth eggs. J. Virol. Methods 36, 141-150. Cameron, I. R. (1990). Identification and characterisation of the gene encoding the major structural protein of insect iridescent virus type 22. Virology 178, 3542. Cameron, I. R., Possee, R. D. and Bishop, D. H. L. (1989). Insect cell culture technology in baculovirus expression systems. Trends Biotechnol. 7, 66-70. Carruthers, W. R., Cory, J. S. and Entwistle, P. F. (1988). Recovery of pine beauty moth (Panolis flamrnea) nuclear polyhedrosis virus from pine foliage. J. Invert. Pathol. 52, 21-32. Carson, D. D., Guarino, L. A. and Summers, M. D. (1988). Functional mapping of an AcNPV immediate early gene which augments expression of the IE-1 trans-activated 39K gene. Virology 162, 444-451. Carson, D. D., Summers, M. D . and Guarino, L. A. (1991). Transient expression of the Autographa californica nuclear polyhedrosis virus immediate-early gene, IE-N, is regulated by three viral elements. J. Virol. 65, 945-951. Cavarelli, J . , Bomu, W., Liljas, L., Kim, S., Minor, W., Munshi. S., Muchmore, S., Schmidt, J.. Johnson, J. and Hendry, D. A . (1991). Crystallisation and preliminary structure analysis of an insect virus with T equal to 4 quasi-symmetry: Nudaurelia capensis o virus. Acta Crystallography 47( l ) , 23-29. Cerutti, M. and Devauchelle, G. (1980). Inhibition of macromolecular synthesis in cells infected with an invertebrate virus (iridovirus type 6 or CIV). Arch. Virol. 63, 297-303. Cerutti, M. and Devauchelle, G . (1982). Isolation and reconstitution of Chilo iridescent virus membrane. Arch. Virol. 74, 145-155. Cerutti, M. and Devauchelle, G . (1985). Characterisation and localisation of CIV polypeptides. Virology 145, 123-131. Cerutti, M., Guerillon, J., Arella, M. and Devauchelle, G. (1981). La replication de I’iridovirus de type 6 (CIV) dans differentes lignees cellulaires. CR Seances Academy of Sciences [III] 292, 791-802. Chakerian, R.. Rohrmann, G . F. and Beaudreau, G. S. (1985). The nucleotide sequence of the Pieris brassicae granulosis virus granulin gene. J. Gen. Virol. 66, 1263-1269. Chapman, H. C., Clark, T. B., Anthony, D. W. and Glenn, F. E . (1971). An iridescent virus from the larvae of Corethralla brakelyi (Diptera: Chaoboridae) in Louisiana. J . Invert. Pathol. 18, 284-287. Chisholm, G. E. and Henner, D. J . (1988). Multiple early transcripts and splicing of the Aurographa cali,fornica nuclear polyhedrosis virus IE-1 gene. J. Virol. 62, 3 193-3200. Clark, T. B. (1982). Entomopox-like particles in three species of bumble bees. J. Invert. Pathol. 39, 119-122. Clark, T. B., Kellen, W. R. and Lum, P. T. M. (1965). A mosquito iridescent virus (MIV) from Aedes taeniorhynchus (Weidermann). J. Invert. Pathol. 7, 519-524. Clem, R. J. and Miller, L. K. (1993). Apoptosis reduces both the in vitro and the in vivo infectivity of baculoviruses. J. Virol. 67, 313g3738. Clem, R. J . , Fechheimer, M. and Miller, L. K. (1991). Prevention of apoptosis by a baculovirus gene during infection of insect cells. Science 254, 1388-1390. Cochran, M. A. and Faulkner, P. (1983). Location of homologous DNA sequences interspersed at five regions in the baculovirus AcMNPV genome. J. Virol. 45, 961-970. Cory, J. S. (1993). Biology and ecology of baculoviruses. In “Opportunities for

ADVANCES IN INSECT VIROLOGY

57

Molecular Biology in Crop Production” (Eds D. J . Beadle, D. H. L. Bishop, L. G. Copping and D. W. Hollomon), pp. 3-10. BCPC Monograph 55. Cory, J. S. and Entwistle, P. F. (1990). The effect of time of spray application on infection of the pine beauty moth, Panolis frammea (Den. and Schiff.) (Lep., Noctuidae), with nuclear polyhedrosis virus. J . Appl. Entomol. 110, 235-241. Crawford, A. M. and Kalmakoff, J . (197-5). A host-virus interaction in a pasture habitat. J. Invert. Pathol. 29, 81-87. Crook, N. E., Clem, R. J. and Miller, L. K. (1993). An apoptosis-inhibiting baculovirus gene with a zinc finger-like motif. J . Virol. 67, 2168-2174. Cunningham, J. C. and McPhee, J . R. (1973). Aerial application of entomopoxvirus and nuclear polyhedrosis virus against the spruce budworm at Chapleau Ontario. Report of the Insect Pathology Research Institute, Canada Forest Service, Sault Ste. Marie, Ontario. Dasgupta, R., Ghosh, A., Dasmahapatra, B., Guarino, L. A. and Kaesberg, P. (1984). Primary and secondary structure of Black Beetle virus RNA2 the genomic messenger for BBV coat protein precursor. Nucl. Acids Res. 12, 7215-7223. Dasmahapatra, B., Dasgupta, R., Saunders, K . , Selling, B., Gallagher, T. and Kaesburg, P . (1986). Proc. Nad. Acad. Scc. 83, 63-66. Dasmahapatra, B., Rozhon, E. and Schwartz, J. (1987). Cell free expression of Coxsackievirus 3C protease from a cDNA clone which contains the translational initiation signal of an insect virus. In “Proceedings of the VIIth International Congress of Virology”, Edmonton, Alberta, p. 135. Dasmahapatra, B., Rozhon, E., Hart, A , , Cox S., Tracy, S. and Schwartz, J. (1991). Cell-free expression of the coxsackie 3C protease using the translational initiation signal of an insect virus RNA and its characterisation. Virus Res. 20, 237-249. David, W. A. L. and Gardiner, B. 0. C. (1966). Persistence of a granulosis virus of Pierk brassicae on cabbage leaves. J . Invert. Pathol. 8, 180-183. Dearing, S. C., Scotti, P. D., Wigley, P. J . and Dhana, S. D. (1980). A small RNA virus isolated from the grass grub. Costelyfra zealandica (Coleoptera: Scarabaeidae). N. Z.1.Zool. 7, 267-269. de Kort, C.A.D. and Granger, N. A. (1981). Regulation of the juvenile hormone titre. Annu. Rev. Entomol. 26, 1-28. Delius, H . , Darai, G. and Flugel, R. M . (1984). DNA analysis of the insect iridescent virus type 6: evidence for circular permutation and terminal redundancy. J . Virol. 49, 609-614. Devauchelle, G., Bergoin, M. and Vago, C. (1971). J. Ultrastruct. Res. 37, 301-321. Devauchelle, G., Attias, J., Monnier, C., Barray, S., Cerutti, M., Guerillon, J. and Orange-Balange, N. (1985). Chilo iridescent virus. Curr. Top. Microbiol. Immunol. 116, 37-51. duPlessis, D. A., Mokhosi, G. and Hendry, D. A. (1991). Cell free translation and identification and the replicative form of Nudaurelia beta virus RNA. J . Gen. Virol. 12, 267-274. Emery, V. C. and Bishop, D. H. L. (1987). The development of multiple expression vectors for high level synthesis of eukaryotic proteins: expression of LCMV-N and AcNPV polyhedrin protein by a recombinant baculovirus. Protein Eng. 1, 359-366. Entwistle, P. F. and Evans, H. F. (1985). Viral control. In “Comprehensive Insect Physiology, Biochemistry and Pharmacology” (Eds G. A. Kerkut and L. I. Gilbert), pp. 347-412, Pergamon Press, Oxford. Epstein, D. A. and Thoma, J . A. (1975). Alkaline protease activity associated with the matrix protein of a virus infecting the cabbage looper. Biochem. Biophys. Res. Commun. 62. 478-484.

58

L. A. KING eta/.

Evans, H. F. and Entwistle, P. F. (1987). Viral diseases. In “Epizootiology of insect diseases” (Eds J. R. Fuxa and Y. Tanada), pp. 257-322. Wiley, New York. Evans, H. F. and Harrap, K. A. (1982). Persistence of insect viruses. In “Virus Persistence” (Eds B. W. J . Mahy, A. C. Minson and G. K. Darby), pp. 57-96. SGM Symposium, Vol. 33, Cambridge University Press, Cambridge. Farara, T. and Attias, J . (1983). Alkaline protease activity associated with iridescent virus type 6. F E M S Lett. 20, 351-355. Farara, T. and Attias. J . (1986). Further characterisation of an alkaline protease activity associated with iridescent virus type 6. Arch. Virol. 67, 307-314. Fediere, G., Philippe, R., Veyrunes, J. C. and Monsarrat. P. (1990). Biological control of the oil palm pest Latoia viridissima. Entomophaga 35, 347-354. Fischer, M., Schnitzler, P., Delius, H., Rosen-Wolff, A. and Darai, G. (1989). Molecular biology of insect iridescent virus type 6. In “Molecular Biology of Iridoviruses” (Ed. G. Darai), Developments in Molecular Virology. pp. 47-82. Kluwer Academic Publishers. Boston. Forstova, J., Krauzewicz, N. and Griffin, B. E. (1989). Expression of biologically active middle T antigen of polyoma virus from recombinant baculovirus. Nucl. Acids Res. 17, 1427-1443. Fowler, M. and Robertson, J . S. (1972). Iridescent virus infection in field populations of Wiseana cervinata (Lepid0ptera:Hepialidae) and Witlesia sp. (Lepidoptera:Pyralidae) in New Zealand. J. Invert. Pathol. 19, 154-157. Francki, R. I . B., Faquet. C. M., Knudson, D. L. and Brown, F. (1991). Classification and nomenclature of viruses. In “The Fifth ReDort of the International Committee on Taxonomy of Viruses.” Arch. Virol. Suppl. 2. pp. 117-1 23. Fraser, M. J. (1986). Ultrastructural observations of virion maturation in Autographa californica nuclear polyhedrosis virus infected Spodoptera frugiperda cell cultures. J . Ultrastruct. Mol. Struct. Res. 95, 189-1 95. Fraser, M. J . (1989). Expression of eucaryotic genes in insect cell cultures. In Vitro Cell. Dev. Biol. 25, 225-235. French, T. J.. Marshall, J . J . A. and Roy, P. (1990). Assembly of double-shelled, viruslike particles of bluetongue virus by the simultaneous expression of four structural proteins. J. Virol. 64, 5695-5700. Friesen, P. D. and Miller. L. K. (1986). The regulation of baculovirus gene expression. In “The Molecular Biology of Baculoviruses” (Eds W. Doerfler and P. Bohm). pp. 31-49. Springer-Verlag, Berlin. Friesen, P. D. and Rueckert, R. R. (1981). Synthesis of black beetle virus proteins in cultured Drosophilu cells: differential expression of RNAs 1 and 2. J. Virol. 37, 876-886. Friesen, P. D. and Rueckert, R. R. (1982). Black beetle virus: messenger for protein B is a subgenomic vital RNA. J. Virol. 42, 986-995. Friesen, P. D. and Rueckert, R. R. (1984). Early and late functions in a bipartite RNA virus: evidence for translational control by competition between vital RNAs. J. Virol. 49, 116-124. Fuchs, L. Y . , Woods. M. S. and Weaver, R. F. (1983). Viral transcription during Autogropha californica nuclear polyhedrosis virus infection: a novel RNA polymerase induced in infected Spodoptera frugiperda cells. J . Virol. 48, 641-646. Fukaya, M. and Nasu, S. (1966). A Chilo iridescent virus (CIV) from the stem borer, Chilo cervinata (Lepidoptera: Pyralidae). Appl. Entomol. Zool. 1, 69-72. Fuxa, J . R., Weidner. E. H. and Richter, A. R. (1992). Polyhedra without virions in a vertically transmitted nuclear polyhedrosis virus. J. Invert. Pathol. 60, 53-58. Gallagher, T . M., Friesen, P. D. and Rueckert, R. R. (1983). Autonomous

ADVANCES IN INSECT VIROLOGY

59

replication and expression of RNA 1 from Black Beetle virus. J. Virol. 46, 481489. Garzon, S. and Charpentier. G . (1991). Nodaviridae. In “Atlas of Invertebrate Viruses” (Eds J . R. Adams and J. R. Bonami), pp. 351-370, C R C Press, Boca Raton. Gauthier, L., Veyrunes. J. C., Cousserans, F. and Bergoin, M. (1992). Abstracts: 9th International Conference on Poxviruses and Iridoviruses, Les Diablerets, Switzerland. Gearing, K. L. and Possee, R. D . (1990). Functional analysis of a 603 nucleotide open reading frame upstream of the polyhedrin gene of Autographa californica nuclear polyhedrosis virus. J. Gen. Virol. 71, 251-262. Gildow, F. E. and D’Arcy, C. J . (1990). Cytopathology and experimental host range of Rhopalosiphum padi virus a small isometric R N A virus infecting cereal grain aphids. J . Invert. Pathol. 55, 245-257. Glocker, B., Hoopes. R. R. and Rohrmann, G. F. (1992). fn vitro transactivation of baculovirus early genes by nuclear extracts from Arctographa californica nuclear polyhedrosis virus-infected Spodoptera frugiperda cells. J . Virol. 66, 3476-3484. Glocker, B., Hoopes, R . R., Hodges, L. and Rohrmann, G . F. (1993). In vitro transcription from baculovirus late gene promoters: Accurate mRNA initiation by nuclear extracts prepared from infected Spodoptera frugiperdu cells. J . Virol. 67, 377 1-3776. Gombart, A . F., Pearson. M. N., Rohrmann. G . F. and Beaudreau, G . S. (1989). A baculovirus polyhedral envelope-associated protein: genetic location, nucleotide sequence and immunocytochemical characterization. Viro/ogy 169. 182-193. Goodwin, R. H., Tompkins, G . J . and McCauley. P. (1978). Gypsy moth cell lines divergent in viral susceptibility. I, Culture and identification. In Vitro 14, 485-494. Goodwin, R. H . , Adams, J. R. and Shapiro, M. (1990). Replication of the entomopoxvirus from Amsacfa moorei in serum-free cultures of a gypsy moth cell line. J. Invert. Pathol. 56. 19(k205. Gotz, P., Huger, A . M. and Kreig, A . (1969). Uber ein insektenpathogenes virus aus der gruppe der pockenviren. Naturwissenschuften 56, 145. Grace, T. D. C. (1962). The development of a cytoplasmic polyhedrosis in insect cells grown in vitro. Virology 18. 33-42. Granados, R. R. (1973a). Insect poxviruses: Pathology, morphology, and development. Miscellaneous Publications of the Entomology Society of America 9, 73-94. Granados, R. R. (1973b). Entry of an insect poxvirus by fusion of the virus envelope with the host cell membrane. Virology 5 2 , 305-309. Granados, R . R. (1980). Infectivity and mode of action of baculoviruses. Biotechnology and Bioengineering 23, 1377-1405. Granados, R. R. (1981). Entomopoxvirus infections in insects. In “Pathogenesis of Invertebrate Microbial Diseases” (Ed. E. W. Davidson), pp. 101-126. Allanheld, New Jersey. Granados, R. and Roberts, D . (1970). Electron microscopy of a poxlike virus infecting an invertebrate host. Virology 40, 23G243. Granados, R. R. and Naughton, M. (1975). Development of Amsacta moorei entomopoxviruses in ovarian and hemocyte cultures from Estigmene acrea larvae. Intervirology 5 , 62-68. Granados, R. R. and Williams, K. A . (1986). In vitro infection and replication of baculoviruses. In “The Biology of Baculoviruses, Vol I , Biological Properties and Molecular Biology” (Eds R . R. Granados and B. A . Federici). pp. 89-108, C R C Press, Boca Raton. Florida.

60

L. A. KING eta/.

Granados, R. R., Nguyen, T. and Cato, B. (1978). An insect cell line persistently infected with a baculovirus-like particle. lntervirology 10, 309-317. Greenfield, C . , Patel, G . , Clark, S . . Jones, N . and Waterfield, M. D . (1988). Expression of the human EGF receptor with ligand-stimulatable kinase activity in insect cells using a baculovirus vector. EMBO Journal 7 , 139-146. Gruidl, M. E . , Hall, R. L. and Moyer, R . W. (1992). Mapping and molecular characterization of a functional thymidine kinase from Amsacta moorei. Virology 186, 507-516. Grula, M. A . , Buller, P. L. and Weaver, R . F. (1981). a-amanitin-resistant viral R N A synthesis in nuclei isolated from nuclear polyhedrosis virus-infected Heliothis zea larvae and Spodoptera frugiperda cells. J . Virol. 38, 916-921. Guarino, L. A . and Summers, M. D . (1986a). Functional mapping of a tramactivating gene required for expression of a baculovirus delayed-early gene. J. Virol. 57, 563-571. Guarino, L. A . and Summers, M. D . (1986b). Interspersed homologous DNA of Autographa californica nuclear polyhedrosis virus enhances delayed-early gene expression. J . Virol. 60, 215-223. Guarino, L. A . , Gonzalez, M. A . and Summers, M. D. (1986). Complete sequence and enhancer function of the homologous D N A regions of Autographa californica nuclear polyhedrosis virus. J . Virol. 60, 224-229. Hall, R. L. and Hink, W. F. (1990). Physical mapping and field inversion gel electrophoresis of Amsacta moorei entomopoxvirus DNA. Arch. Virol. 110, 77-90, Hall, R . L. and Moyer, R. W. (1991). Identification, cloning, and sequencing of a fragment of Amsacta moorei entomopoxvirus D N A containing the spheroidin gene and three vaccinia virus-related open reading frames. J . Virol. 65, 6516-6527. Hall, R. L. and Moyer, R . W. (1993). Identification of an Amsacta spheroidin-like protein within the occlusion bodies of Choristoneura entomopoxviruses. Virology 192, 179-187. Hamblin, M., van Beek, N . A. M., Hughes, P. R . and Wood, H . A . (1990). Co-occlusion and persistence of a baculovirus mutant lacking the polyhedrin gene. Applied and Environmental Microbiology 56, 3051-3062. Hammock, B. D . (1985). Regulation of the juvenile hormone titre: degradation. In “Comprehensive Insect Physiology, Biochemistry and Pharmacology” (Eds G. A. Kerkut and L. I. Gilbert), pp. 431472. Pergamon Press, New York. Hammock, B. D . , Bonning, B. C., Possee, R. D . , Hanzlik, T. N. and Maeda, S . (1990). Expression and effects of the juvenile hormone esterase in a baculovirus vector. Nature, London 344, 458-461. Hanzlik, T. N . , Abdel-Aal, Y. A . I., Harshman, L. G. and Hammock, B. D . (1989). Isolation and sequencing of the cDNA clones coding for juvenile esterase from Heliothis virescens. J . Biol. Chem. 264, 12419-1 2425. Harrap, K. A . , Longworth. J. F., Tinsley, T. W. and Brown, K . W. (1966). A noninclusion virus of Gonometa podocarpi (Lepidoptera: Lasiocampidae). J. Invert. Pathol. 8, 27c272. Hatfill, S. J., Williamson, C., Kirby, R. and Von Wechmar. M. B. (1990). Identification and localisation of aphid lethal paralysis virus particles in thin tissue sections of the Rhopalosiphum padi aphid by in situ nucleic acid hybridisation. J. Invert. Pathol. 55, 265-271. Hawtin, R. E. (1993). The chitinase if Autographa californica nuclear polyhedrosis virus. PhD Thesis, Oxford Brookes University. Heimpel, A . M. and Angus, T. A . (1959). The site of action of crystalliferous bacteria in lepidopteran larvae. J . Insect Pathol. 1, 152-170.

ADVANCES IN INSECT VIROLOGY

61

Hendry, D. A. (1991). Nodaviridae of Invertebrates. In “Viruses of Invertebrates” (Ed. E. Kurstak), pp. 227-276, Marcel Dekker. New York. Hill-Perkins, M. S. and Possee, R. D. (1990). A baculovirus expression vector derived from the basic protein promoter of Autographa californica nuclear polyhedrosis virus. J . Gen. Virol. 71, 971-976. Himeno, M., Matsubara, F. and Hayashiya, K. (1973). The occult virus of nuclear polyhedrosis of the silkworm larvae. J . Invert. Pathol. 22, 292-295. Hink, W. F. (1970). Established insect cell line from the cabbage looper, Trichoplusia ni. Nature 226, 266467. Hoehberg, M. E. (1991). Extra-host interactions between a braconid endoparasitoid, Apanteles glomeratus, and a baculovirus for larvae of Pieris brassicae. J . Anim. Eco~.60, 65-77. Hofte, H. and Whiteley, H. R. (1989). Insecticidal crystal proteins of Bacillus thuringiensis. Microbiol. Rev. 53, 242-255. Hooft van Iddekinge, B. J . L., Smith, G. E. and Summers, M. D. (1983). Nucleotide sequence of the polyhedrin gene of Autographa californica nuclear polyhedrosis virus. Virology 131, 561-565. Hoopes, R. R. and Rohrmann, G. F. (1991) In vitro transcription of baculovirus immediate early genes: accurate mRNA initiation by nuclear extracts from both insect cells and human cells. Proc. Natl. Acad. Sci. USA 88, 4513-4517. Hosur, M. V . , Schmidt, T., Tucker, C., Johnson, J , E., Selling, B. H. and Rueckert, R. R. (1984). Black beetle virus+-crystallisation and particle symmetry. Virology 133, 119-127. Howard, S. C., Ayres, M. D. and Possee, R. D. (1986). Mapping the 5‘ and 3’ ends of Autographa californica nuclear polyhedrosis virus polyhedrin mRNA. Virus Res. 5 , 109-119. Hsieh, P. and Robbins, P. W. (1984). Regulation of asparagine-linked oligosaccharide processing. J . Biol. Chem. 259, 2375-2382. Huber, J. (1986). Use of baculoviruses in pest management programs. In “The Biology of Baculoviruses Vol. 11: Practical Application for Insect Control” (Eds R. R. Granados and B. A. Federici). pp. 181-202. CRC Press, Boca Raton, Florida. Huger, A. M., Kreig, A , , Emschermann, P. and Gotz, P. (1970). Further studies on Polypoxvirus chironomi, an insect virus of the pox group isolated from the midge Chironomus luridus. J . Invert. Pathol. 15, 253-261. Hughes, D. S . , Possee, R. D. and King, L. A. (1993). Activation and detection of a latent baculovirus resembling Mamestra brassicae nuclear polyhedrosis virus. Virology 194, 608-615. Hughes, P. R., Gettig, R. R. and McCarthy, W. J. (1983). Comparison of the time-mortality response of Heliothis zea to 14 isolates of Heliothis nuclear polyhedrosis virus. J . Invert. Pathol. 41, 246261. Hukuhara, T. (1962). Generation to generation transmission of the cytoplasmic polyhedrosis virus of the silkworm, Bombyx mori (Linnaeus). J . Invert. Pathol. 4, 32-35. Hukuhara, T., Xu, J. and Yano, K . (1990). Replication of an Entomopoxvirus in two Lepidopteran cell lines. J . Invert. Pathol. 56, 222-232. Hurpin, B. and Vago, C. (1963). Une maladie a inclusions cytoplasmiques fusiformes chez le Coleoptere Melolontha melolontha. Rev. Pathol. Vegetale Entomoi. Agr. 42, 115-117. Iatrou, K. and Meidinger, R. G. (1990). Tissue-specific expression of silkmoth chorion genes in vivo using Bombyx mori nuclear polyhedrosis virus as a transducing vector. Proc. Natl Acad. Sci., USA 87, 3650-3654. Jarvis, D. L. and Summers, M D. (1989). Glycosylation and secretion of human

62

L. A. KING eta/.

tissue plasminogen activator in recombinant baculovirus-infected cells. Mol. Cell. Biol. 9, 214-233. Jarvis. D. L., Oker-Blom, C. and Summers, M. D. (1990a). Role of glycosylation in the transport of recombinant glycoproteins through the secretory pathway of Lepidopteran insect cells. J . Cell. Biochem. 42, 181-191. Jarvis, D. L. Fleming, J-A. G. W . , Kovacs, G . R. Summers, M. D. and Guarino, L. A. (1990b). Use of early baculovirus promoters for continuous expression and efficient processing of foreign gene products in stably transformed Lepidopteran cells. BiolTechnology 8, 950-955. Jeang, K-T., Giam, C-Z.. Nerenberg, M. and Khoury, G . (1987). Abundant synthesis of functional human T-cell leukaemia virus type I p40” protein in eucaryotic cells by using a baculovirus expression vector. J . Virol. 61, 708-713. Joklik, W. K. (1962). Some properties of poxvirus deoxyribonucleic acid. J. Mol. Biol. 5, 265-274. Joyce, K. A,. Atkinson, A. E., Bermudez, I . , Beadle. D . J. and King, L. A. (1993). Synthesis of functional GABAA receptors in stable insect cell lines. FEBS Left. 335, 61-64. Jurkovicova, M . (1979). Activation of latent virus infections in larvae of Adoxophyes orana (Lepidoptera: Tortricidae) and Barathra brassicae (Lepidoptera: NOCtuidae) by foreign polyhedra. J. Invert. Pathol. 34, 213-223. Kaesburg. P. (1987). Organisation of pipartite insect virus genomes: the genome of black beetle virus. In “The Molecular Biology of the Positive Strand RNA Viruses” (Eds D. J . Rowlands. M. A. Mayo and B. W . J . Mahy), pp. 207-218. Academic Press. New York. Kamito, S. G., Majima, K. and Maeda. S. (1993). Identification and characterisation of the p35 gene of Bombyx mori nuclear polyhedrosis virus that prevents virus-induced apoptosis. J. Virol. 67, 455-463. Kataoka. H . , Troetschler, R. G.. Li, J. P.. Kramer, S. J., Carney, R. L. and Schooley, D. A. (1989). Isolations and identification of a diuretic hormone from the tobacco hornworm Manduca sexta. Proc. Natl. Acad. Sci. USA 86,2976-2950. Keddie, B. A . and Volkman, L. E. (1985). Infectivity difference between the two phenotypes of Autographa californica nuclear polyhedrosis virus: importance of the 64K envelope glycoprotein. J . Gen. Virol. 66, 1195-1200. Kelly, D. C. (1981). Non-occluded viruses. In “Pathogenesis of Invertebrate Microbial Diseases” (Ed. E. W. Davidson), pp. 3 9 4 0 , Allanheld Osmun. Kelly. D. C. (1982). Baculovirus replication. J. Gen. Virol. 63, 1-13. Kelly. D. C. (1985). Insect iridescent viruses. Curr. Top. Microbiol. Immunol. 116, 23-35. Kelly, D. C. (1987). Baculovirus pathogenesis. In “Molecular Basis of Virus Disease” (Eds W . C. Russell and J. W. Almond), pp. 289-302. Cambridge University Press, Cambridge. Kelly, D. C. and Robertson. J. S. (1973). Icosahedral cytoplasmic deoxyriboviruses. J . Gen. Virol. 20, 17-41. Kelly, D . C. and Tinsley, T. W. (1972). The proteins of iridescent virus types 2 and 6. J. Invert. Pathol. 19, 273-275. Kelly. D. C. and Tinsley, T. W. (1973). Ribonucleic acid polymerase activity associated with particles of iridescent virus types 2 and 6. J. Invert. Pathol. 22, 199-202. Kelly, D. C. and Vance, D. E. (1973). The lipid content of two iridoviruses. J. Gen. Virol. 21, 417-423. Kelly, D. C. and Lescott, T. (1981). Baculovirus replication: protein synthesis in Spodoptera frugiperda cells infected with Trichoplusia ni nuclear polyhedrosis virus. Microbiologica 4, 35-57.

ADVANCES IN INSECT VIROLOGY

63

Kelly, D. C., Lescott, T., Ayres, M. D.. Carey, D., Coutts, A. and Harrap, K. A. (1981). Induction of a non-occluded baculovirus persistently infecting Heliothis zea cells by Heliothis armigera and Trichoplusia ni nuclear polyhedrosis viruses. Virology 112, 174-189. Kelly, D. C., Brown, D. A., Ayres, M. D., Allen, C. J. and Walker, I. 0. (1983). Properties of the major nucleocapsid protein of Heliorhis zea singly enveloped nuclear polyhedrosis virus. J . Gen. Virol. 64,399-408. King, L. A. and Moore, N. F. (1988). Evidence for the presence of a genome linked protein in two insect picornaviruses, Cricket paralysis virus and Drosophila C virus. FEMS Lett. 50, 41-44. King, L. A. and Possee, R. D. (1992). “The Baculovirus Expression System. A Laboratory Guide.” Chapman and Hall, London, UK. King, L. A., Pullin, J. S. K. and Moore, N . F. (1984). Characterisation of a picornavirus isolated from a tumorous blood cell line of Drosophila melanogaster. Microbios 26, 121-127. King, L. A . , Pullin. J. S. K., Stanway, C . , Almond, J. W. and Moore, N. F. (1987). Cloning of the genome of Cricket paralysis virus and sequence of the 3’ end. Virus Res. 6, 331-344. King, L. A., Mann, S. G., Lawrie. A. M. and Mulshaw, S. H. (19Y1). Replication of wild-type and recombinant Autographa californica nuclear polyhedrosis virus in a cell line derived from Mamestra brassicae. Virus Res. 19, 93-104. King, L. A . , Mann, S. G., Marlow, S. A., Bermudez. I., Lawrie, A. M., Obosi, L. A. and Atkinson, A. E. (1992). Alternative insect cell lines for baculovirus expression vectors. In “Baculovirus and Recombinant Protein Production Processes” (Eds J. M . Vlak, E.-J. Schlaeger and A. R. Bernard), pp. 157-165, Editiones Roche. Basel. Kirschbaum, J. B. (1985). Potential implication of genetic engineering and other biotechnologies to insect control. Annu. Rev. Enrornol. 30, 51-70. Kitts, P. A. and Possee, R. D. (1993). A method for producing recombinant baculovirus expression vectors at high frequency. BiolTechniques 14. 810-817. Kitts, P. A , , Ayres. M. A. and Possee, R. D. (1990). Linearization of baculovirus DNA enhances the recovery of recombinant virus expression vectors. Nucl. Acids Res. 18, 566775672, Klenk, H.-D., Cramer, A , , Wagner, R.. Groner, A. and Kretzschmar, E. (1992). Processing of influenza virus hemagglutinin in insect cells: variations depending on the host and on co-expression with other viral proteins. I n “Baculoviruses and Protein Product Processes” (Eds J . M. Vlak. E.-J. Schlaeger and A. R. Bernard), pp. 166174, Editiones Roche, Basel. Klump, H., Beaumais, J. and Devauchelle, G. (1983). Structural and thermodynamic investigation of the Chilo iridescent virus (iridovirus type 6). Arch. Virol. 75, 269-276. Knebel-Morsdorf, D., Kremer, A. and Jahnel, F. (1993). Baculovirus gene ME53, which contains a putative zinc finger motif, is one of the major early-transcribed genes. J . Virol. 67, 753-758. Kool, M. and Vlak. J. M. (1993). The structural and functional organisation of the Aurographa californica nuclear polyhedrosis virus genome. Arch. Virol. 130, 1-16. Kool, M., van den Berg. P. M. M., Tramper, J . , Goldbach, R. W. and Vlak, J . M. (1993). Location of two putative origins of DNA replication of Autographa californica nuclear polyhedrosis virus. Virology 192, 94-101. Krappa, R. and Knebel-Morsdorf. D. (1991). Identification of the very early transcribed baculovirus gene PE-38. J . Virol. 65, 805-812. Krappa, R., Bohn-Krappa. A , , Jahnel, F., Doerfler. W. and Knebel-Morsdorf, D.

64

L. A. KING eta/.

(1992). Differential factor binding at the promoter of early baculovirus gene PE38 during viral infection: GATA motif is recognised by an insect protein. J. Virol. 66. 3494-3503. Krell. P. and Lee, P. E. (1974). Polypeptides in Tipula iridescent virus and in TIV-infected haemocytes of Galleria mellonella (L.) larvae. Virology 60, 315-326. Kuroda, K., Geyer, H.. Geyer, R., Doerfler, W. and Klenk, H-D. (1990). The oligosaccharides of influenza virus haemagglutinin expressed in insect cells by a baculovirus vector. Virology 174. 418-429. Kuzio. J.. Rohel. D. Z., Curry, C. J . , Krebs, A , , Carstens, E. B. and Faulkner, P. (1984). Nucleotide sequence of the p10 polypeptide gene of Autographa Californica nuclear polyhedrosis virus. Virology 139, 414-418. Langridge, W. (19834. Partial characterization of DNA from five Entomopoxviruses. J . invert. Pathol. 42, 369-375. Langridge, W. H. R . (1983b). Detection of Amsacta moorei Entomopoxvirus and Vaccinia virus proteins in cell cultures restrictive for poxvirus multiplication. J. Invert. Pathol. 42, 77-82. Langridge, W. H. R. (1983~).Virus DNA replication and protein synthesis in Amsacta moorei entomopoxvirus-infected Estigmene acrea cells. J . Invert. Pathol. 41,34\-349.

Langridge, W. H. R. (1984). Detection of DNA base sequence homology between entomopoxviruses isolated from Lepidoptera and Orthoptera. J . Invert. Pathol. 43. 4146. Langridge. W. H. R. and Roberts, D. W. (1982). Structural proteins of Amsacta moorei, Euxoa auxiliaris and Melanoplus sanguinpes, entomopoxviruses. J. Invert. Pathol. 39, 346353. Langridge. W. H. R., Bozarth, R. F. and Roberts, D. W. (1977). The base composition of entomopoxvirus DNA. Virology 76, 616-620. Lebdeva, 0. P. and Zelenko. A. P. (1972). Virus-like formations in larvae of Aedes and Culex mosquitoes. Med. Parazitof. Parazit. Bolezni 41, 490492. Li, Y . . Passareli, L. and Miller, L. K. (1993). Identification, sequence and transcriptional mapping of fef-3, a baculovirus gene involved in late and very late gene expression. J . Virol. 67, 5260-5268. Longworth. J . F. and Archibald, R. D. (1975). A virus of black beetle, Heteronychur arator (F.) (Coleoptera: Scarabaeidae). N . Z . J . Zool. 2, 233-236. Longworth, J. F. and Carey. G. P. (1976). A small RNA virus with a divided genome from Heteronychus arator (F.) (Coleoptera: Scarabaeidae). J . Gen. Virol. 33, 3140. Longworth, J . F. and Cunningham, J . C. (1968). The activation of occult nuclear-polyhedrosis viruses b; foreign nuclear polyhedra. J. Invert. Pathol. 10, 36 1-367. Longworth, J . F.. Robertson, J . S. and Payne, C. C. (1972). The purification and properties of inclusion body protein of the granulosis virus of Pieris brassicae. J. invert. Pathol. 19, 42-50. Longworth. J . F., Payne, C. C. and MacLeod, R. (1973). Studies on a virus isolated from Gonometa podocarpi (Lepidoptera: Lasiocampidae). J . Gen. Virol. 18, 119-125. Loudon, P. T. and Roy, P. (1991). Assembly of five bluetongue virus proteins expressed by recombinant baculoviruses: inclusion of the largest protein VP1 in the core and virus-like particles. Virology 180, 798-802. Lu, A. and Carstens. E. B. (1991). Nucleotide sequence of a gene essential for viral DNA replication in the baculovirus Autographa californica nuclear polyhedrosis virus. Virolog,v 181. 336-347.

ADVANCES IN INSECT VIROLOGY

65

Luckow, V. A. and Summers, M. D. (1988). Trends in the development of baculovirus expression vectors. BiolTechnology 6, 47-55. Lytvyn, V., Fortin, Y., Banville, M., Arif. B. and Richardson, C. (1092). Comparison of the thymidine kinase genes from three entomopoxviruses. J . Gen. Virol. 73, 3235-3240. Maddrell, S. (1986). In “Insect Neurochemistry and Neurophysiology” (Eds A. B. Borkovec and D. B. Gelman). pp. 35-74. Marcel Dekker. New York. Maeda, S. (1989a). Expression of foreign genes in insects using baculovirus vectors. Annu. Rev. Entomol. 34, 351-372. Maeda, S . (1989b). Increased insecticidal effect by a recombinant baculovirus carrying a synthetic diuretic hormone. Biochim. Biophys. Actu Res. Comrnun. 165, 1177-1183. Maeda, S., Kawai, T., Obinata, M.. Fujiwara, H . , Horiuchi, T., Saeki, Y . , Sato, Y. and Furusawa, M. (1985). Production of human a-interferon in silkworm using a baculovirus vector. Nature, London 315, 592-594. Maeda, S., Kamita, S. G. and Kataoka. H. (1991). The basic DNA-binding protein of Bombyx mori nuclear polyhedrosis virus: the existence of an additional arginine repeat. Virology 180, 807-810. Manousis, T. and Moore, N. F. (1987). Cricket Paralysis virus, a potential control agent for the olive fruit fly. Dacus oleae. Appl. Environ. Microbiol. 53. 142-. Marlow, S. A . , Palmer. C. P. and King, L. A. (1992). Cytopathic effects of Amsacru moorei entomopoxvirus infection on the cytoskeleton of Estigmene ucrea cells. Virus Res. 26, 41-55. Marlow, S. A., Billam, L. J., Palmer, C. P. and King, L. A. (1993). Replication and morphogenesis of Amsacfa moorei entomopoxvirus in cultured cells of Estigmene ucreu (salt marsh caterpillar). J . Gen. Virol. 74, 1457-1461. Martens, J. W. M.. Honee, G., Zuidema, D., van Lent, J . W. M.. Visser, B. and Vlak, J. M. (1990). Insecticidal activity of a bacterial crystal protein expressed by a recombinant baculovirus in insect cells. Appl. Environ. Microbiol. 56. 27642770. Martignoni, M. E. and Iwai, P. J. (1986) “A Catalogue of Viral Diseases of Insects, Mites and Ticks,” 4th edn. USDA Forest Service PNW-195. Washington, DC. Matsuura, Y . , Possee, R. D., Overton, H. A. and Bishop, D. H. L. (1987). Baculovirus expression vectors: the requirements for high level expression of proteins, including glycoproteins. J . Gen. Virol. 68, 1233-1250. Matthews, R. E. F. (1982). Classification and nomenclature of viruses. Intervirology 17, 1-200. Mazzone, H. M. (1985). Pathology associated with baculovirus infection. In “Viral Insecticides for Biological Control” (Eds K. Maramorosch and K. E. Sherman), pp. 81-120, Academic Press, London. McCarthy, W. J., Granados, R. R. and Roberts, D. W. (1974). Isolation and characterization of entomopox virions from virus-containing inclusions of Amsactu moorei (Lepidoptera: Arctiidae). Virology 59, 5940. McCutchen, B. F., Choudary, P. V., Crenshaw, R., Maddox, D., Kamita, S. G., Palekar, N., Volrath. S., Fowler, E., Hammock, B. D. and Maeda, S. (1991). Biotechnology 9, 848-852. McIntosh, A. H. and Kimura, M. (1974). Replication of the insect Chilo iridescent virus (CIV) in a poikilothermic vertebrate cell line. Intervirology 4, 257-267. McKinley, D. J . Brown, D. A., Payne, C. C. and Harrap. K. A. (1981). Cross infectivity and activation studies with four baculoviruses. Entomophuga 26, 79-90. Merryweather, A. T., Weyer, U., Harris, M. P. G., Hirst, M., Booth, T. and Possee. R . D. (1990). Construction of genetically engineered baculovirus insecti-

66

L. A. KING eta/.

cides containing the Bacillus thuringiensis susp. kurstaki HD-73 delta endotoxin. J. Gen. Virol. 71, 1535-1544. Miller, L. K. (1988). Baculoviruses as gene expression vectors. Annu. Rev. Microbiol. 42, 177-199. Miller, L. K. (1993). Baculoviruses. High-level expression in insect cells. Curr. Opin. Genet. Dev. 3, 97-101. Miller, L. K., Jewell, J . E. and Browne, D. (1981). Baculovirus induction of a DNA polymerase. J. Virol. 46. 275-278. Milner, R. J. and Lutton, G. G. (1975). The pathogenicity of an entomopoxvirus from Othonius batesi and its possible use as a control agent. Entomophaga 20, 2 13-220, Minion, F. C.. Coons, L. B. and Broom, J . R. (1979). Characterization of the polyhedral envelope of the nuclear polyhedrosis virus of Heliothis virescens. J. Invert. Pathol. 34, 303-307. Monnier. C. and Devauchelle. G. (1980). Enzyme activities associated with an invertebrate iridovirus: protein kinase activity associated with iridescent virus type 6 (Chilo iridescent virus). J. Virol. 35, 444-450. Moore, N. F. (1991a). The Nudaurelia p family of insect viruses. In “Viruses of Invertebrates” (Ed. E. Kurstak). pp. 277-286, Marcel Dekker, New York. Moore, N. F. (1991b). Identification, pathology, structure, and replication of insect picornaviruses. In “Viruses of Invertebrates” (Ed. E. Kurstak), pp. 287-300, Marcel Dekker, New York. Moore. N. F. and Eley. S. M. (1991). Picornaviridae: Picornaviruses of Invertebrates. In “Atlas of lnvertebrake Viruses” (Eds J . R. Adams and J . R. Bonami), pp. 371-386. CRC Press, Boca Raton. Moore, N. F. and Kelly, D. C. (1980). A comparative study of the polypeptides of three iridescent viruses by N-terminal analysis, amino acid analysis and surface labelling. J. Invert. Pathol. 36, 415-422. Moore. N . F. and Pullin, J. S. K. (1982). Plaque purification of cricket paralysis virus using an agar overlay on Drosophila cells. .I. Invert. Pathol. 39, 10-14. Moore, N. F., Kearns. A. and Pullin, J. S. K. (1980). Characterisation of Cricket paralysis virus-induced polypeptides in Drosophila cells. J. Virol. 33, 1-9. Moore, N. F., Pullin, J. S. K., Crump, W. A . L., Reavy. B. and Greenwood, L. K. (1981a). Comparison of two insect picornaviruses, Drosophila C and cricket paralysis viruses. Microbiologica 4, 359-370. Moore, N. F., Reavy, B. and Pullin, J. S. K . (1981b). Processingof Cricket paralysis virus induced polypeptides in Drosophila cells: production of high molecular weight polypeptides by treatment with iodoacetamide. Arch. Virol. 68, 1-8. Moore, N. F., Pullin. J. S. K . , Crump, W. A. L. and Plus, N. (1982). The proteins expressed by different isolates of Drosophila C virus. Arch. Virol. 74. 21-30. Moore, N. F., King, L. A. and Pullin. J. S. K. (1987). Insect Picornaviruses. In “The Molecular Biology of the Positive Strand RNA Viruses” (Eds D. Rowlands. M. Mayo and B. Mahy), pp. 67-74. Academic Press. New York. Mori. K. I . , Nakai, T.. Muroga, K.. Arimoto. M., Mushiake, K. and Furusawa, I. (1992). Properties of a new virus belonging to Nodaviridae found in larval striped jack (Pseudocarunx dentex) with nervous necrosis. Virology 187, 368-371. Moscardi. F. (1989). Use of viruses for pest control in Brazil: the case of the nuclear polyhedrosis virus of the soybean caterpillar, Anticarsia gemmatalis. Mem. Inst. Oswaldo Cruz 84, 51-56. Moscardi, F. and Correa Ferreira. B. S. (1988). Biological control of the soybean caterpillar. In “World Research Conference 111: Proceedings” (Ed. R. Shibles), pp. 703-71 1, Westview Press, London. Mueller, R., Pearson. M. N.. Russell, R. I. Q. and Rohrmann. G. F. (1990). A

ADVANCES IN INSECT VIROLOGY

67

capsid-associated protein of the multicapsid nuclear polyhedrosis virus of Orgyia pseudotsugata: genetic location, sequence, transcriptional mapping and immunocytochemical characterisation. Virology 176. 133-144. Murphy, F. A., Scherer, W. F.. Harrison. A . K., Dunne, M. W. and Gary, G. W. (1970). Characterisation of nodamura virus, an arthropod transmissible picornavirus. Virology 40,1008-1021. Murray, K. D. and Elkington, J . S. (1989). Environmental contamination of egg masses as a major component of transgenerational transmission of gypsy moth nuclear polyhedrosis virus (LdMNPV). J . Invert. Pathol. 53, 324-334. Murti, K. G., Goorha, R . and Granoff, A . (1985). An unusual replication strategy of an animal iridovirus. Adv. Virus Res. 30. 1-15. Newman, J. F. E . and Brown, F. (1973). Evidence for a divided genome in Nodamura virus, an arthropod borne picornavirus. J. Gen. Virol. 21. 371-384. Newman, J. F. E. and Brown, F. (1977). Further physicochemical characterisation of nodamura virus. Evidence that the divided genome occurs in a single component.

J. Gen. Virol. 38, 8SY5. Nissen, M . S. and Friesen, P. D . (1989). Molecular analysis of the transcriptional regulatory region of an early baculovirus gene. J . Virol. 63, 493-503. Olson, N. H . , Baker, T. S., Johnson, J . E. and Hendry. D. A . (1990). The three dimensional structure of frozen hydrated Nuduurelia capensis beta virus, a T equals 4 insect virus. J . Struct. B i d . 105, 11 1-122. Ooi, B. G . , Rankin, C . and Miller. L. K. (1989). Downstream sequences augment transcription from the essential initiation site of a baculovirus polyhedrin gene. J . Mol. Biol. 210. 721-736. Orange-Balange. N . and Devauchelle, G . (1982a). Lipid composition of an iridescent virus type 6 (CJV]. Arch. Virol. 73, 36S367. Orange-Balange, N. and Devauchelle, G. (1982b). Effect of infection with iridovirus type 6 on invertebrate cell phospholipids. Annals of Virology (Insfirure Pasteur) 133E, 205-214. O’Reilly, D . R. and Miller, L. K. (1989). A baculovirus blocks insect molting by producing ecdysteroid UDP-glucosyl transferase. Science 245, 1110-1 112. OReilly, D. R . . Crawford. A . M. and Miller, L. K. (1989). Viral proliferating cell nuclear antigen. Nature 337, 606. O’Reilly, D . R.. Miller, L. K. and Luckow. V. A . (1992). “Baculovirus Expression Vectors. A Laboratory Manual.” W. H . Freeman and Company, New York. Passarelli, A . L. and Miller, L. K . (lYY3a). Three baculovirus genes involved in late and very late gene expression: iel, ien and lef2. J . Virol. 67, 2149-2158. Passarelli, A. L. and Miller. L. K. (1093b). Identification and characterisation of left, a baculovirus gene involved in late and very late gene expression. J . Virol. 67, 3481-3488. Pearson, A . , Richardson, C. and Yuen. L. (1991). The 5’ noncoding region sequence of the Choristoneura hiennis Entomopoxvirus spheroidin gene functions as an efficient late promoter in the mammalian vaccinia expression system. Virology 180, 561-566. Pearson, M. N., Russell, R. L., Rohrmann, G . F. and Beaudreau, G . S. (1988). P39, a major baculovirus structural protein: immunocytochemical characterization and genetic location. Virology 167, 407-413. Pearson, M. N . , Bjornson, R.. Pearson, G. and Rohrmann, G . (1992). The Autograph californica baculovirus genome: evidence for multiple replication origins. Science 257, 1382-1384. Plus, N. (1989). Comparative study of picornaviruses in Drosophila cells in vitro. In “Invertebrate Cell System Applications” (Ed. J . Mitsuhashi), pp. 6 3 4 8 . CRC Press, Boca Raton.

68

L. A. KING eta/.

Podgewaite, J. D. (1985). Strategies for field use of baculoviruses. In “Viral Insecticides for Biological Control” (Ed. K. Maramorosch and K. E. Sherman), pp. 775-797. Academic Press, London. Pogo, B. G. T.. Bergoin. M., Dales, S. and Roberts, D. W. (1971). Enzymes associated with an insect poxvirus. Virology 43. 306-309. Ponsen, M. B. and de Jong, D. J. (1964). On a nuclear and cytoplasmic polyhedrosis of Adoxophyes reticulana HB. Entomophaga 9(3), 253-255. Possee, R. D. (1986). Cell-surface expression of influenza virus haemagglutinin in insect cells using a baculovirus vector. Virus Res. 5, 43-59. Possee, R. D. and Howard, S. C. (1987). Analysis of the polyhedrin gene promoter of the Autographa californica nuclear polyhedrosis virus. Nucleic Acids Res. 15, 10233-10248. Possee, R. D., Weyer, U. and King, L. A. (1990). Recombinant antigen production using baculovirus expression vectors. In “Control of Virus Diseases” (Eds N. J. Dimmock, P. D. Griffiths and C. R. Madely), pp. 53-76. Society General Microbiology Symposium Vol. 45. SGM, Cambridge University Press, Cambridge. Possee, R. D., Sun, T.-P., Howard, S. C., Ayres. M. D., Hill-Perkins, M. and Gearing, K. L. (199 1). Nucleotide sequence of the Autographa californica nuclear polyhedrosis virus 9.4 kbp EcoRI-I and -R (polyhedrin gene) region. Virology 185, 229-24 1. Possee, R. D., King, L. A.. Weitzman, M. D., Mann, S. G . , Hughes, D. S., Cameron, I. R., Hirst, M. L. and Bishop D. H. L. (1992). Progress in the genetic modification and field-release of baculovirus insecticides. In “The Release of Genetically Modified Microorganisms-REGEM 2”. (Eds D. E. S. Stewart-Tull and M. Sussman). Plenum Press, New York. Possee, R. D., Hirst, M., Jones, L. D. and Bishop, D. H. L. (1993) i n “Opportunities for Molecular Biology in Crop Production” (Eds D. J . Beadle, D. H. L. Bishop, L. G. Copping and D. W. Hollomon). pp. 23-36. BCPC Monograph 55. Pudney, M., Newman, J. F. E. and Brown, F. (1978). Characterisation of Kawino virus; an entero-like virus isolated from the mosquito Mansonia uniformis (Diptera: Culicidae). J . Gen. Virol. 40, 433441. Quiot, J. M., Bergoin, M. and Vago, C. (1975). Developpement et pathogenese d’un entomopoxvirus en culture cellulaire de lepidoptere. C. R. Acad. Sci., Serte D 280, 2273-2275. Rankin, C.. Ooi, B. G. and Miller, L. K. (1988). Eight base pairs encompassing the transcriptional start point are the major determinant for baculovirus polyhedrin gene expression. Gene 70. 39-49, Reavy, B. and Moore, N . F. (1982). The replication of small RNA-containing viruses of insects. Microbiologica 5, 63-84. Reavy, B. and Moore, N. F. (1984). In vitro translation of the genome of a small RNA virus from Trichoptusia ni. J . invert. Pathol. 44,244-248. Reavy, B., Crump, W. A. L., Greenwood, L. K. and Moore, N. F. (1982). Characterisation of a bipartite genome ribovirus isolated from Lymantria ninayi. Microbiologica 5, 305-319. Reinganum. C. (1973). Studies o n a non-occluded virus of the field crickets Telleogryllus oceanicus and T . comrnodus. MSc Thesis, Monash University, Melbourne, Australia. Reinganum, C. (1991). Tetraviridae. In “Atlas of Invertebrate Viruses’’ (Eds J. R. Adams and J . R. Bonami), pp. 387-392, CRC Press, Boca Raton. Reinganum, C., O’Loughlin, G. T. and Hogan, T. W. (1970). A non-occluded virus

ADVANCES IN INSECT VIROLOGY

69

of field crickets Telleogryllus oceanicus and T. cornmodus (Orthoptera: Gryllidae). J . Invert. Pathol. 16, 216220. Reinganum, C., Bashiruddin, J. B. and Cross, G. F. (1985). Boolarra virus: a member of the Nodaviridae isolated from Oncopera intricoides (Lepidoptera: Hapialidae). Intervirology 24, 10-17. Retnakaran, A. and Bird, F. T. (1972). Apparent hormone imbalance syndrome caused by an insect virus. J . Invert. Pathol. 20, 358-360. Roberts, D. W. (1970). Poxvirus of Arctiid larvae: Infectivity for greater wax moth. Bacteriol. Proc. 70th Meeting 151, 177. Roberts, D. W. and Granados, R. R . (1968). A poxlike virus from Amsacta moorei (Lepidoptera: Arctiidae). J . Invert. Pathol. 12, 141-143. Rohrmann, G. F. (1986). Polyhedrin structure. J . Gen. Virol. 67, 1499-1513. Rohrmann, G. F. (1992). Baculovirus structural proteins. J . Gen. Virol. 73, 749-761. Russell, R. L. Q. and Rohrmann, G. F. (1990a). A baculovirus polyhedron envelope protein: immunogold localization in infected cells and mature polyhedra. Virology 174, 177-184. Russell, R. L. 0. and Rohrmann, G. F. (1990b). The p6.5 gene region of a nuclear polyhedrosis virus of Orgyia pseudotsugata: DNA sequence and transcriptional analysis of four late genes. J . Gen. Virol. 71, 551-560. Russell, R. L. Q . , Pearson, M. N., Rohrmann, G. F. and Beaudreau, G. S. (1987). Characterization of baculovirus p10 synthesis using monoclonal antibodies. Virology 160, 9-19. Sanz, P., Veyrunes, J. C., Cousserans, F. and Bergoin, M. (1992). Abstracts, 9th International Conference on Poxviruses and Iridoviruses, Les Diablerets, Switzerland. Sato, T. (1989). Establishment of eight lines from newborn larvae of tortricids and characteristics including susceptibility to insect viruses. Bulletin of the Fruit Tree Research Station, Series A . (Yatabe) 16, 115-132. Scherer, W. F. and Hurlbut, H. S. (1967). Nodamura virus from Japan: a new and unusual arbovirus resistant to diethyl ether and chloroform. A m . J . Epidemiol. 86, 271-285. Scotti, P. D. (1976). Cricket paralysis virus replicates in cultured Drosophila cells. Intervirology 6, 333-342. Scotti, P. D., Longworth, J. F. E., Plus, N., Croizier, G. and Reinganum, C. (1981). The biology and ecology of strains of an insect small RNA virus complex. A d v . Virus Res. 26, 117-143. Scotti, P. D., Dearing, S. and Mossop, D. W. (1983). Flock house virus: a nodavirus isolated from Costelytra zealandica (White) (Coleoptera: Scarabaeidae). Arch. Virol. 75, 181-189. Shapiro, M. and Robertson, J. L. (1987). Yield and activity of gypsy moth (Lepidoptera: Lymantriidae) nuclear polyhedrosis virus recovered from survivors of viral challenge. J . Econ. Entomol. 80, 901-905. Sieburth, P. J. and Carner, G. R. (1987). Infectivity of an iridescent virus for larvae of Anticarsia gemmatalis (Lepidoptera: Noctuidae). J. Invert. Pathol. 49, 49-51. Smith, G. E. and Summers, M. D. (1979). Restriction maps of five Autographa californica MNPV variants, Trichoplusia ni MNPV, and Galleria mellonella MNPV DNAs with endonucleases SmaI, KpnI, BamHI, S a d , XhoI and EcoRI. J . Virol. 30, 828-838. Smith, G. E., Vlak, J. M. and Summers, M. D. (1983a). Physical analysis of Autogrupha californica nuclear polyhedrosis virus transcripts for polyhedrin and 10,000-molecular-weight protein. J . Virol. 45, 215-225. Smith, G. E., Fraser, M. J. and Summers, M. D. (1983b). Molecular engineering of

70

L. A. KING eta/.

the Autographa californica nuclear polyhedrosis virus genome: deletion mutations within the polyhedrin gene. J . Virol. 46, 584-593. Smith, G. E., Ju, G., Ericson, B. L., Moschera, J . , Lahm, H., Chizzonite, R. and Summers, M. D. (1985). Modification and secretion of human interleukin 2 produced in insect cells infected with a baculovirus expression vector. Proc. Nod. Acad. Sci. USA 82, 8404-8408. Smith, G. E., J u , G., Ericson, B. L.. Moschera, J . , Lahm, H., Chizzonite, R., Stohwasser, R . , Raab. K., Schnitzler, P., Janssen. W. and Darai, G. (1993). Identification of the gene encoding the major capsid protein of insect iridescent virus type 6 by polymerase chain reaction. J . Gen. Virol. 74. 873-879. Smith, I . R. L. and Crook, N . E. (1988). In vivo isolation of baculovirus genotypes. Virology 166, 240-244. Smith, K . M . (1955). What is a virus‘? Nature 175. 12-21. Smith, K. M. (1963). The cytoplasmic virus diseases. In “Insect Pathology: an Advanced Treatise” (Ed. E. A . Steinhaus). Vol. 1 , pp. 457-497. Academic Press, New York. Smith, K. M. (1967). “Insect Virology.” Academic Press, New York. Soltau, J. B., Fischer, M., Schnitzler, P., Scholz. J. and Darai, G. (1987). Characterisation of the genome of insect iridescent virus type 6 by physical mapping. J . Gen. Virol. 68, 2717-2722. Sparks, T . C . and Hammock, B. D. (1980). Comparative inhibition of the juvenile hormone esterase from Trichoplusia ni, Tenebrio monitor, and Musca domestica. Pesticide Biochem. Physiol. 14, 29G302. St Angelo, C., Smith, G. E . , Summers, M. D. and Krug, R. M. (1987). Two of the three influenza viral polymerase proteins expressed by using baculovirus vectors form a complex in insect cells. J . Virol. 61. 361-365. Steinhaus. E . A . and Dineen, J. P. (1960). Observations on the role of stress in a granulosis of the variegated cutworm. J . Insect Pathol. 2, 55-65. Stewart, L. M. D., Hirst, M., Lopez-Ferber, M., Merryweather, A . T., Cayley, P. J. and Possee. R . D. (1991). Construction of an improved baculovirus insecticide containing an insect-specific toxin gene. Nature 352. 85-88. Stoltz, D. B. (1973). The structure of icosahedral cytoplasmic deoxyriboviruses. An alternative model. J . Ultrastruct. Res. 43, 58-74. Stoltz, D . B. and Summers, M. D. (1972). Observations on the morphogenesis and structure of a hemocytic poxvirus in the midge Cliironomus attenuatus. J. Ultrastruct. Res. 40, 58 1-598. Stoltz, D . B . , Vinson, S. B. and MacKinnon, E. A. (1976). Baculovirus-like particles in the reproductive tracts of female parasitoid wasps. Can. J . Microbiol. 22(7), 1013-1 023. Struthers, J. K . and Hendry, D . A . (1974). Studies on the protein and nucleic acid components of Nudaurelia capensis beta virus. J . Gen. Virol. 22, 255-362. Summers, M. D . and Anderson, D. L. (1973). Characterization of nuclear polyhedrosis virus DNAs. J . Virol. 12, 13361346. Sutter, G. (1972). A poxvirus of the army cutworm. J . Invert. Pathol. 19, 375-382. Tajbakhsh, S. and Seligy. V. L. (1989). Molecular biology of Tipula iridescent virus. In “Molecular Biology of Iridoviruses, Developments in Molecular Virology” (Ed. G. Darai), pp. 98-132, Kluwer Academic Publishers, Boston. Tajbakhsh. S . , Lee, P. E., Watson, D . C. and Seligy. V. L. (1990). Molecular cloning. characterisation and expression of the Tipula iridescent virus capsid gene. J . Virol. 64,125-136. Takehara, K.. Ireland, D. and Bishop. D . H. L. (1988). Co-expression of the hepatitis B surface antigen and core antigens using baculovirus multiple expression vectors. J . Gen. Virol. 69. 2763-2777.

ADVANCES IN INSECT VIROLOGY

71

Teitelbaum, Z., Lazarovici, P . and Zlotkin, E. (1979). Selective binding of the scorpion venom insect toxin to insect nervous tissue. Insect Biochem. 9, 343-346. Thiem, S. M. and Miller, L. K. (1989). Identification, sequence and transcriptional mapping of the major capsid protein gene of the baculovirus Autographa californica nuclear polyhedrosis virus. J . Virol. 63, 2008-2018. Thomas-Orillard, M. (1988). Interaction between a picornavirus and a wild population of Drosophila melanogaster. Oecologia 75, 516 5 2 0 . Tinsley, T. W. and Kelly, D. C. (1970). An interim nomenclature system for the iridescent group of insect viruses. J . Invert. Pathol. 16, 470-472. Tomalski, M. D. and Miller, L. K. (1991). Insect paralysis by baculovirus-mediated expression of a mite-neurotoxin gene. Nature 352, 82-85. Tomalski, M. D. and Miller, L. K. (1992). Expression of the paralytic neurotoxin to improve baculoviruses as biopesticides. BioiTechnology 10, 545-549. Tomalski, M. D., Wu, J. and Miller, L. K. (1988). The location, sequence, transcription and regulation of a baculovirus DNA polymerase. Virology 167, 591-600. Tweeten, K. A., Bulla, L. A. and Consigli, R. A. (1980). Characterisation of an extremely basic protein derived from granulosis virus nucleocapsids. J . Virol. 33, 866-876. Vago, C. (1963). A new type of insect virus. J . Invert. Pathol. 5 , 275-276. Vago, C. and Bergoin, M. (1968). Viruses of invertebrates. Adv. Virus Res. 13, 247-303. Vail, P. V., Jay, D. L. and Hink, W. F. (1973). Replication and infectivity of the nuclear polyhedrosis virus of the Alfalfa looper Autographa californica, produced in cells grown in vitro. J . Invert. Pathol. 22, 231-237. Vaughn, J. L. (1979). Virus and rickettsia1 diseases. In “Insect Diseases Vol I” (Ed. G . E. Cantwell), pp. 4 9 4 5 , Marcel Dekker, New York. Vaughn, J. L., Goodwin, R. H., Thompkins, G. J. and McCawley, P. (1977). The establishment of two cell lines from the insect Spodoptera frugiperda (Lepidoptera: Noctuidae). In Vitro 13, 213-217. Vialard, J., Lalumiere, M., Vernet, T., Briedis, D., Alkhatib, G., Henning, D., Levin, D. and Richardson, C. (1990a). Synthesis of the membrane fusion and hemagglutinin proteins of measles virus, using a novel baculovirus vector containing the P-galactosidase gene. J . Virol. 64, 37-50. Vialard. J. E., Yuen. L. and Richardson, C. D. (1990b). Identification and characterization of a baculovirus occlusion body glycoprotein which resembles spheroidin, an entomopoxvirus protein. J . Virol. 64,5804-5811. Vlak, J. (1993a) Genetic engineering of baculoviruses. In “Opportunities for Molecular Biology in Crop Production” (Eds D. J. Beadle, D. H. L. Bishop, L. G. Copping and D. W. Hollomon), pp. 11-22, BCPC Monograph 55. Vlak, J. M. (1993b). Genetic engineering of baculovirus insecticides for insect control. In “Molecular Approaches to Pure and Applied Entomology” (Eds M. J. Whitten and J. G. Oakeshott), pp. 9@127, Springer-Verlag, New York. Vlak, J. M., Klinkenberg, F. A , , Zaal, K. J. M., Usmany, M., Klinge-Roode, E. C., Geervliet, J. B. F., Roosien, J. and van Lent, J. W. M. (1988). Functional studies on the p10 gene of Autographa californica nuclear polyhedrosis virus using a recombinant expressing a pl0-P-galactosidase fusion gene. J . Gen. Virol. 69, 765-776. Volkman, L. E. and Keddie, B. A. (1990). Nuclear polyhedrosis virus pathogenesis. Semin. Virol. I , 249-256. Volkman, L. E., Summers, M. D. and Hsieh, C. H. (1976). Occluded and non-occluded nuclear polyhedrosis virus grown in Trichoplusia ni: comparative neutralization, comparative infectivity, and in vitro growth studies. J . Virol. 19, 82C832

72

L. A. KING eta/.

Walter, C . , Zlotkin. E. and Rathmeyer, W. (1976). Action of different toxins from the scorpion Androctonus australis on a locust nerve-muscle preparation. J. Insect Physiol. 22, 1187-1 194. Wang, X . and Kelly, D. C. (1983). Baculovirus replication: purification and identification of Trichoplusia ni nuclear polyhedrosis virus-induced DNA polymerase. J. Gen. Virol. 64, 2229-2236. Ward, V. K. and Kalmakoff, J. (1987). Physical mapping of the D N A genome of insect iridescent virus type 9 from Wiseana spp. larvae. Virology 160, 507-510. Ward, V. K. and Kalmakoff, J . (1991). Invertebrate iridoviridae. In “Viruses of Invertebrates” (Ed. E. Kurstaki), pp. 197-226, Marcel-Dekker, New York. Weitzman, M. D., Possee, R . D. and King, L. A . (1992). Characterisation of two variants of Panolis flammea multiple nucleocapsid nuclear polyhedrosis virus. J. Gen. Virol. 7, 1881-1886. Westwood, J. C . N., Harris, W. J . , Zwartouw, H. T . , Titmus, D . H. J. and Appleyard, G. (1964). Studies on the structure of vaccinia virus. J . Gen. Microbiol. 34. 67-78. Weyer, U.and Possee, R. D. (1988). Functional analysis of the p10 gene 5’ leader sequence of the Autographa californica nuclear polyhedrosis virus. Nucleic Acids Res. 16, 3635-3653. Weyer, U.and Possee, R. D. (1989). Analysis of the promoter of the Autographa californica nuclear polyhedrosis virus p10 gene. J. Gen. Virol. 70, 203-208. Weyer, U. and Possee, R. D . (1991). A baculovirus dual expression vector derived from the Autographa californica nuclear polyhedrosis virus polyhedrin and p10 promoters: co-expression of two influenza virus genes in insect cells. J . Gen. Virol. 72, 2967-2974. Weyer, U . , Knight, S. and Possee, R. D. (1990). Analysis of very late expression by Autographa californica nuclear polyhedrosis virus and the further development of multiple expression vectors. J. Gen. Virol. 71, 1525-1534. Whitford, M., Stewart, S., Kuzio, J. and Faulkner. P. (1989). Identification and sequence analysis of a gene encoding gp67, an abundant envelope glycoprotein of the baculovirus Autographa californica nuclear polyhedrosis virus. J. Virol. 63, 1393-1 399. Whitt, M. A. and Manning, J. S. (1988). A phosphorylated 34-kDa protein and a subpopulation of polyhedrin are thiol linked to the carbohydrate layer surrounding a baculovirus occlusion body. Virology 163, 33-42. Williams, G. V., Rohel, D. Z . , Kuzio, J. and Faulkner, P. (1989). A cytopathological investigation of Autographa californica nuclear polyhedrosis virus p10 gene function using insertion/deletion mutants. J. Gen. Virol. 70. 187-202. Williams, R. C.-and Smith, K. M. (1957). A crystallizable insect virus. Nature 179, 119-1 20. Williams, R. C . and Smith, K. M. (1958). The polyhedral form of the Tipula iridescent virus. Biochim. Biophys. Acta 28. 464-469. Williamson, C. and Rybicki, E. P. (1989). A comparative study on the cell-free translation of the genomic RNAs of two aphid picorna-like viruses. Arch. Virol. 109, 59-70. Williamson, C . , Von Wechmar. M. B. and Rybicki, E. P. (1989). Further characterisat ion of Rhopalosiphum padi virus of aphids and comparison of isolates from South Africa and Illinois. J. Invert. Pathol. 54, 85-96. Willison, J. H. M. and Cocking, E. C. (1972). Frozen fractured viruses: a study of virus structure using freeze etching. J. Microsc. 95, 397411. Wilson, M. E . . Mainprize, T . H., Friesen, P. D. and Miller, L. K. (1987). Location, transcription and sequence of a baculovirus gene encoding a small arginine-rich polypeptide. J. Virol. 61, 661-666.

ADVANCES IN INSECT VIROLOGY

73

Wood, H. A . , Hughes, P. R., Johnson, L. B . and Langridge, W . H. R. (1981). Increased virulence of Autographa californica nuclear polyhedrosis virus by mutagenesis. J. Invert. Pathol. 38, 236-241. Wrigley, N. G. (1969). A n electron microscopy study of the structure of Sericesthis iridescent virus. J . Gen. Virol. 5, 129-134. Xeros, N . (1954). A second virus disease of the leatherjacket, Tipula paludosa. Nature 174, 562-565. Yang, C . L., Stetler, D . A. and Weaver, R. F. (1991). Structural comparison of the Autographa californica nuclear polyhedrosis virus-induced RNA polymerase and the three nuclear RNA polymerases from the host, Spodoptera frugiperda. Virus Res. 20, 251-264. Yuen, L . , Dionne, J . , Arif, B . and Richardson, C. (1990). Identification and sequencing of the spheroidin gene of Choristoneura biennis entomopoxvirus. Virology 175, 427-433. Yuen, L., Noiseux, M. and Gomes, M. (1991) D N A sequence of the nucleoside triphosphate phosphohydrolase I (NPH I ) of the Choristoneura biennis entomopoxvirus. Virology 182, 40-06. Yule, B. G. and Lee, P. E. (1973). A cytological and immunological study of Tipula iridescent virus-infected Galleria mellonella larvae haemocytes. Virology 51, 409423. Zhong, W., Dasgupta, R . and Rueckert, R. (1992). Evidence that the packaging system for nodaviral RNA2 is a bulged stem-loop. Proc. Natl. Acad. Sci. USA 23. 11146-11150. Zummer, M. and Faulkner, P. (1979). Absence of protease in baculovirus polyhedra grown in vitro. J. Invert. Pathol. 33, 383-384.

This Page Intentionally Left Blank

Genetic Mechanisms of Early Neurogenesis in Drosophila rn elanogaster Jose A. Campos-Ortega Institut fur Entwicklirngshiologie. Uriiver.sitat zu Kiilri, 0-50931 Koln. Germany

1 Introduction 75 2 Cellular basis of neurogenesis 76 3 The segregation of neuroblasts 77 4 Cell interactions in the neuroectoderm 79 5 Neuralizing signals and intrinsic factors 80 6 Genetics of neurogenesis 80 7 The neurogenic genes are functionally interrelated 84 8 Physical interactions of Notch and Delta 85 9 The epidermal decision is controlled by the E(sPL)-C 87 10 The neural decision is controlled by the proneural genes 88 11 Interactions between neurogenic and proneural genes 89 12 Conclusions 90 Acknowledgements 95 References 96

1 Introduction

In insects, the cells of the central nervous system (CNS) are generated by the proliferation of progenitor cells called neuroblasts which develop from a special neurogenic region of the ectoderm. In Drosophila melanogasfer,the neuroectoderm consists of two different parts, the ventral neuroectoderm (VNE), from which the ventral cord and the suboesophageal ganglion will develop, and the procephalic neuroectoderm, from which the brain hemispheres emerge (Poulson, 1950; Hartenstein and Campos-Ortega, 1984; Technau and Campos-Ortega, 1985). Both regions give rise to neural progenitor cells; however, whereas the cells of the VNE have to decide between developing either as neuroblasts or as epidermoblasts (progenitor cells of the epidermis), there is no clear picture as to how the procephalic neuroectoderm is organized and how neuroblasts develop from this region (Hartenstein and Carnpos-Ortega, 1984; Technau and Campos-Ortega, ADVANCES IN INSECT PHYSIOLOGY VOL. 25 ISBN &1?424?25-7

CopvrrRhr 0f9Y4 Academic Press Lintired A / / righrs of repruducrion i n any form reserved

76

JOSE A. CAMPOS-ORTEGA

1985; Jiirgens et al. , 1986; Stiittem and Campos-Ortega, 1991). Accordingly, this review will deal with the VNE. A hypothesis has been recently formulated to account for the development of progenitor cells of sensory organs within the epidermis of Drosophila (Ghysen and Dambly-Chaudiere, 1989; see Jan and Jan, 1990; Ghysen et al., 1993, for reviews). Appropriately modified, this hypothesis can also be applied to the segregation of neuroblasts and epidermoblasts in the VNE. The hypothesis proposes a sequence of three steps to explain the segregation of epidermal and central neural progenitor cells. In the first step, all cells of the VNE acquire the capability to develop as neuroblasts, whereby contiguous groups of about four to five cells, so-called proneural clusters, are each enabled to give rise to a particular type of neuroblast. In the second step, one cell in each group is singled out by intervening neuralizing signals and segregates into the space between ectoderm and mesoderm to develop as a particular type of neuroblast. In the third step, the neuroblast sends inhibitory, epidermalizing signals to the surrounding cells preventing them from following a neural fate and permitting them to assume an epidermal fate. Therefore, three operations are included in this scheme: one confers upon the VNE cells the capability to produce neural or epidermal progenies, another permits the separation of the two classes of progenitor cells, and the third specifies particular types of neuroblasts and epidermoblasts. The first two of these operations are brought about by the participation of a large number of gene products, functionally interconnected to form a complex network. Major constituents of this network are two groups of regulatory proteins, which allow the development of the neural and epidermal cell fates, respectively, and a group of membraneassociated proteins which serve to pass the regulatory signals between neighbouring cells in the neuroectoderm and transduce them to the nuclei of the interacting cells. 2 Cellular basis of neurogenesis

In wild-type Drosophila melanogaster, the neuroectoderm becomes morphologically manifest during the initial phase of germ band elongation in stage 7 (stages of embryogenesis according to Campos-Ortega and Hartenstein, 1985). Within the segmented germ band, the ectodermal layer becomes subdivided into a lateral part, comprising small cylindrical cells, and a medial part with large cuboidal cells (Hartenstein and Campos-Ortega, 1984; see Fig. 1). The lateral part will differentiate during later stages into the dorsal epidermis, with its annexes, whereas the medial part is the VNE itself, from which the ventral cord and the ventral epidermis will develop (Technau and Campos-Ortega, 1985). Virtually all cells of the VNE, i.e. 100 rows of mediolaterally arranged cells on either side of the midline, with

NEUROGENESIS IN DROSOPHlLA

77

approximately nine cells in each row, enlarge to become conspicuously different from the cells of the dorsal epidermal anlage. After the neuroblasts have segregated from the VNE into the interior of the embryo, the cells remaining on the outside shrink. In striking contrast to the case in the fruit fly, only single cells of the W E within groups comprising several cells are reported to enlarge during early neurogenesis in grasshoppers; the enlarged cells are the prospective neuroblasts themselves, which will segregate from the remaining cells (Bate, 1976, 1982; Doe and Goodman, 1985a). 3 The segregation of neuroblasts

In Drosophila, segregation of neuroblasts lasts for approximately 3 h and is discontinuous, proceeding in discrete pulses which give rise to different subpopulations of neuroblasts. Originally, three pulses, which give rise to SI, SII and SIII neuroblasts, were distinguished (Hartenstein and CamposOrtega, 1984). However, Doe (1992), using molecular markers, has distinguished two additional pulses of segregation which give rise to two additional subpopulations of neuroblasts, SIV and SV. Single cells among the large cells of the VNE undergo conspicuous shape changes, leave the ectoderm and move internally to form the neural primordium. The cells immediately adjacent to each of the SI neuroblasts establish relationships with them by means of long basal processes, which transiently surround the segregated neuroblast forming a sort of sheath. The ensheathing processes are later retracted and the prospective epidermal progenitor cells diminish in size. Therefore, there is sufficient contact between the VNE cells both before and during lineage segregation to enable cellular interactions to take place. After their segregation, the neuroblasts form a monolayer between ectoderm and mesoderm (Fig. 1). Due to their sequential segregation and subsequent behaviour, various subpopulations of neuroblasts can be distinguished from each other on the basis of their size and location within the array (Hartenstein and Campos-Ortega, 1984; Hartenstein ef af., 1987; Doe et al., 1988; JimCnez and Campos-Ortega, 1990; Cui and Doe, 1992; Doe, 1992). Doe (1992) has reported on markers that permit reliable identification of individual neuroblasts, even when in late stages the pattern becomes more complex. Midway through stage 11, another group of neural progenitor cells develops in each segment from the mesectodermal cells, that comprises the unpaired median neuroblast (MNB) and a number of small, paired cells called midline precursor cells (MPs; Bate, 1976; Thomas et af., 1984; Klambt et af., 1991). All of these midline cells in each segmental group form an ovoid cluster which straddles the boundary between adjacent neuromeres. The first maps of neuroblasts in insects (Bate, 1976; Hartenstein and

JOSE A. CAMPOS-ORTEGA

FIG. 1 Illustration of the morphological modifications of the ectoderm accompanying neuroblast segregation. The dark shadowing in all drawings indicates neuroblasts and their progeny, as well as ectodermal cells with capabilities to produce neuroblasts. Epidermal cells are shown in lighter shadowing. (A) One row of neuroectodermal ( W E ) cells, from the midline to the dorsal epidermal anlage (DEA), is shown. All cells of the VNE enlarge considerably in stage 8 and become conspicuously different from the cells of the DEA. (B) Segregation of most of the SI neuroblasts (SI NB) takes place from median and lateral regions of the W E . Here, single cells undergo conspicuous shape changes to leave the outer layer and move

NEUROGENESIS IN DROSOPHILA

79

Campos-Ortega, 1984) were obtained by reconstructing serial sections. The advent of antibody staining techniques (Hartenstein et al., 1987; Doe et al., 1988; JimCnez and Campos-Ortega, 1990; Martin-Bermudo et al., 1991; Skeath and Carroll, 1992; Doe, 1992) and other markers, like lacZ expressing ‘enhancer trap’ lines (O’Kane and Gehring, 1987; Bier et al., 1989; Bellen et al., 1989; Campos-Ortega and Haenlin, 1992; Cui and Doe, 1992; Doe, 1992), has permitted a more accurate mapping of the embryonic neuroblasts. The most complete map of neuroblasts in the Drosophila embryo has been presented by Doe (1992), who describes an approximate number of 30 neuroblasts per hemisegment, including a glioblast that generates longitudinal glia. Klambt et al. (1991), in a careful analysis of the midline progenitor cell clusters using a number of ‘enhancer trap’ lines, describe eight cells per segmental midline cluster, i.e. one median neuroblast, three neuronal precursors, three glial precursors and one cell whose identity is not well established. 4 Cell interactions in the neuroectoderm

Two pieces of evidence indicate that the decision of the VNE cells to adopt the epidermal or the neural fate is mediated by cell-cell interactions. First, laser ablation experiments carried out in grasshoppers show that the cells remaining in the VNE after the neuroblasts have segregated are not firmly committed to their fate (Taghert el al., 1984; Doe and Goodman, 1985b). Under normal circumstances, these cells would develop as epidermoblasts; however, in the conditions of the experiment they may adopt a neural fate instead. These results provide the best evidence in support of the idea that

internally, where they become arranged in two longitudinal stripes. Notice that the VNE cells immediately adjacent to each of the SI NB establish intimate contact with the latter cells by means of ensheathing processes. Regulatory signals are assumed to be sent preceding the segregation. (C) SII neuroblasts segregate mainly from intermediate regions of the VNE, where the cells enlarge again and single cells separate from the germ layer. The processes which ensheathed the SI neuroblasts are retracted and the prospective epidermal progenitor cells diminish in size. Ectodermal cells are shown pale when they have lost the capability to produce neuroblasts and have entered the epidermal pathway. (D) Most SIII and subsequent neuroblasts (SIV and SV, according to Doe, 1992) segregate from the median cells and, again in this case, there is a conspicuous enlargement of the ectodermal cells that precedes the separation of the neuroblasts from the germ layer. (E) Once all the neuroblasts have segregated, the cells remaining on the outside shrink to form the epidermal sheath. Immediately after segregation, neuroblasts round up and begin to divide asymmetrically to produce ganglion mother cells (GMC) and neurones (or ganglion cells, GC). Since neuroblasts decrease in size after divisions, SI can be distinguished from SII and from SIII NB on the basis of their smaller size.

80

JOSE A. CAMPOS-ORTEGA

the prospective epidermoblasts are inhibited by the neuroblasts from adopting the neural fate (Doe and Goodman, 1985b). Second, results of cell transplantations in Drosophila suggest that regulatory signals in this species pass between the ectodermal cells and that these signals are involved in the cells’ commitment to one of the two developmental fates (Technau and Campos-Ortega, 1986a, 1987; Technau, 1986, 1987; Technau et al., 1988; Becker and Technau, 1990; Stiittem and Campos-Ortega, 1991). These experiments suggest that two types of signals, epidermalizing and neuralizing, operate during lineage segregation. The epidermalizing signals are required to inhibit the neural fate in the cells of the proneural cluster destined to become epidermoblasts; the neuralizing signals to implement the primary neural fate. 5 Neuralizing signals and intrinsic factors

Intrinsic properties of the neuroectodermal cells are of paramount importance in deciding whether any given cell can adopt a neural fate upon its transplantation in the neuroectoderm. For example, proctodeal cells (Stuttem and Campos-Ortega, 1991), endodermal (Technau and CamposOrtega, 1986b) and, probably, also mesodermal cells (Beer et al., 1987) do not develop neural fates upon transplantation into the VNE, but produce proctodeal, endodermal and mesodermal lineages. Two different factors appear thus to be involved in the neural decision of ectodermal cells. One is a particular environment that permits neurogenesis; the other factor is the competence of the transplanted cells to respond to whatever signals might be provided by the environment and permit neurogenesis; this competence is present in cells of the VNE, the DEA or the foregut anlage, but absent from cells of the proctodeal anlage. Therefore, intrinsic properties of the transplanted ectodermal cells as, for example, particular transcription factors, may confer upon them a bias toward adoption of the neural fate and facilitate a higher degree of autonomy in the choice. A recent study by Luer and Technau (1992) shows that DEA cells have intrinsic epidermogenic capabilities, which confer upon them a clear bias to develop as epidermoblasts in the absence of cell interactions. The authors have cultivated individual cells from well-defined regions of the ectoderm and observed what kind of progenies they form. They found a gradient of epidermogenic capabilities in the ectoderm, decreasing from dorsal to ventral, and another gradient of neurogenic capabilities in the opposite direction. 6 Genetics of neurogenesis

In Drosophila melanogaster, the correct separation of neuroblasts and epidermoblasts is controlled by two groups of genes (Table l), the so-called

TABLE 1

Genes involved in early neurogenesis of Drosophila melanoaoster

Genes

Loss-of-function phenotype

Proneural genes AS-C Neural hypoplasia vnd

da

Neural hypo p 1asi a Neural hypoplasia

Neurogenic genes N Neural hyperplasia

Gene product

Possible function

bHLH

Regulation of transcription

?

3

bHLH

Regulation of transcription

Transmembrane EGF-like repeats

Adhesion signal receptor?

Relevant references JimCnez and Campos-Ortega (1979, 1987, 1990), White (1980), Dambly-Chaudikre and Ghysen (1987), Villares and Cabrera (1987), Alonso and Cabrera (1988) White (1980), Jimenez and Campos-Ortega (1987, 1990) Caudy et al. (1988a), Caudy et al. (1988b)

Poulson (1937), Wharton et al. (1985), Kidd et al. (1986), Fehon et al. (1990). Heitzler and Simpson (1991) Lehmann et al. (1981). Vassin et al. (1987), Kopczynski et al. (1988). Haenlin et al. (1990), Fehon et al. (1990), Heitzler and Simpson (1991) Lehmann et al. (1983). Klambt et al. (1989). Knust et al. (1992). Schrons et al. (1992) Lehmann et al. (1981), Smoller et al. (1990)

DI

Neural hyperplasia

Transmembrane EGF-like repeats

Adhesion signal source?

E(sPL)-C

Neural hyperplasia Neural h yperplasia Neural hyperplasia

bHLH

Regulation of transcription

Nuclear protein

3

Homeobox bacterial repressors Transmembrane

Regulation of transcription?

Lehmann et al. (1981). Boulianne et al. (1991)

Channel? Transporter?

Lehmann et al. (1981), Rao et al. (1990)

3

Bourouis et al. (1989, 1990)

?

Knust et al. (1987b), Hartley et al. (1988), Delidakis et al. (1991), Schrons et al. (1992)

mum neu

big brain shaggy groucho

Neural h yperplasia Neural hyperplasia Neural hyperplasia

Serine-threonine kinase Nuclear protein

82

JOSE A. CAMPOS-ORTEGA

neurogenic genes (Poulson, 1937; Lehmann et al., 1981, 1983) on the one hand, and the so-called proneural genes (Ghysen and Dambly-Chaudiire, 1989, 1990; Romani et al., 1989), i.e. the various members of the achaete-scute complex (AS-C), the locus ventral nervous system condensation defective (vnd), and daughterless, and probably other as yet unidentified genes, on the other hand (Jimenez and Campos-Ortega, 1979, 1987, 1990; White, 1980; White et al., 1983; Caudy et al., 1988a; Brand and CamposOrtega, 1988). Neurogenic and proneural genes are currently referred to as though they form two different groups, and such a distinction is justified by the phenotypes of their mutations (to be discussed below). However, I would like to emphasize that the products of all of these genes are apparently involved in the same complex genetic network and together contribute to the process of separation of neural and epidermal cell progenitors. Since the functions of the AS-C, vnd and daughterless promote neural development, they have been generically called “proneural genes” (Ghysen and Dambly-Chaudiere, 1989, 1990; Romani et a f . , 1989). Poulson (1937) called Notch (the first neurogenic gene discovered) a “neurogenic” gene following the convention in Drosophila genetics of naming a gene according to the phenotype of the mutation that leads to its discovery. Accordingly, other genes which cause the same phenotype as Notch have also been called neurogenic. Loss of function of any of the neurogenic genes causes most ectodermal cells to develop as neuroblasts (Fig. 2) (Lehmann et al., 1981, 1983; Jiminez and Campos-Ortega, 1982). Neuralization of the ectoderm of neurogenic mutants follows the pattern of neuroblast segregation in the wild-type and proceeds in pulses (Campos-Ortega and Haenlin, 1992). In the mutants, all the VNE cells from which neuroblasts normally segregate at each pulse take on neural fate until, in mid-stage 11, all cells in the VNE have adopted neural fate. Regions from which sensory organs develop exhibit also a high proportion of neural cells (Hartenstein and Campos-Ortega, 1986; Ghysen and Dambly-Chaudiere, 1990; Goriely et al., 1991). Therefore, neurogenesis is initiated in these mutants by a much higher number of neuroblasts and sensory organ mother cells than in the wild-type and, consequently, the mutant embryos die with a highly hyperplasic central and peripheral nervous system and lacking ventral and cephalic epidermis. Thus, the wild-type functions of the neurogenic genes are formally required to suppress neural development of a large fraction of ectodermal cells and allow them to develop as epidermoblasts. The complete loss of the shaggy function leads to all the embryonic cells taking up neural fate; intermediate phenotypes are obtained when maternal or zygotic gene expression is affected (Bourouis et al., 1989; Sirnpson and Carteret, 1989). The molecular nature of the shaggy product, a putative serine-threonine kinase (Bourouis et al., 1990), together with the phenotype of its mutants, strongly suggests a role for this gene in the transmission of

NEUROGENESIS IN DROSOPHILA

a3

Neuroblast Epidermoblast

Neuroblast

Neuroblast

- Notch Delta

Epidermoblast Epidermoblast or cell death FIG.2 (A) Shows two interacting cells of the VNE in the wild-type; the white ovals represent the cell nuclei. The proneural and the neurogenic genes, including the genes of the E(sPL)-C,encode the proteins of a regulatory signal chain which allows the cells to develop either as neuroblasts or as epidermoblasts. The proneural genes are required to regulate the genetic activity of the neuroblasts, those of the E(sPL)-C to give neuroectodermal cells access to epidermal development. (B) Mutation of a neurogenic gene results in the development of all neuroectodermal cells as neuroblasts. This is probably due to derepression of the proneural genes in the cells which would have normally developed as epidermoblasts. (C) Mutation in the proneural genes results in either the development of additional neuroectodermal cells as epidermoblasts, at the expense of neuroblasts, or in cell death. This is probably due to inactivation of the proneural genes in the cells which would have normally developed as neuroblasts.

84

JOSE A. CAMPOS-ORTEGA

signals between neuroectodermal cells. However, there is as yet no further evidence to support this function. Embryos homozygous for loss-of-function mutations in the proneural genes exhibit a highly hypoplasic CNS and severe defects in the PNS (Fig. 2; Jimenez and Campos-Ortega, 1979, 1987, 1990; White, 1980; DamblyChaudiere and Ghysen, 1987; Ghysen and Dambly-Chaudiere, 1988; Caudy et al., 1988a). The origin of the neural hypoplasic defects of such mutants is complex. Embryos lacking the AS-C or vnd initiate neurogenesis with less than the normal complement of neuroblasts: 2&25% of all neuroblasts are missing (Jimenez and Campos-Ortega, 1990); mutants lacking both the AS-C and vnd lack up to 50% of all neuroblasts. In addition, during later stages, large numbers of cells degenerate within the neural primordium of all these mutants (JimCnez and Campos-Ortega, 1979, 1990; White, 1980; Brand and Campos-Ortega, 1988). However, the CNS of these mutants still contains a significant number of neurones, indicating that still other genes are necessary for neuroblast commitment. Since the populations of neuroblasts affected by mutations in the AS-C and vnd do not seem to overlap significantly, these genes, and probably other as yet unidentified ones, may each control the development of particular sets of neuroblasts. Brand and Campos-Ortega (1988) found that the epidermal sheath of neurogenic mutants is larger when they also lack the AS-C; this is suggestive evidence to support the notion that at least some of the cells which fail to develop as neuroblasts in the mutants develop as epidermoblasts instead. The participation of the gene daughterless in neuroblast commitment and segregation is not yet clear. The complement of neuroblasts in daughterlessmutants is initially normal; however, most of these cells die early in embryonic development and mutants show obvious hypoplasic defects in the CNS (Jimenez and Campos-Ortega, 1990; Brand and Campos-Ortega, submitted). However, daughterless- mutants completely lack the peripheral nervous system, due to death of the progenitor cells of the sensory organs (Caudy et al., 1988a; Brand and Campos-Ortega, 1988; Vassin et al., 1993). 7 The neurogenic genes are functionally interrelated

There is abundant evidence from various kinds of genetic analyses to support the assumption that the neurogenic loci are functionally interrelated (Campos-Ortega et a f . , 1984; Vassin et al., 1985; de la Concha et al., 1988; Shepard et af., 1989; Brand and Campos-Ortega, 1990); Godt et al., 1991; Xu et al)., 1990. Insofar as their participation in the segregation of neuroblasts from epidermoblasts is concerned, neurogenic loci were found to be functionally linked in a chain of epistatic relationships, in which the E(sPL)-C was found to be the last link; big brain was found to act independently of the others (Vassin el al., 1985; de la Concha et al., 1988).

NEUROGENESIS IN DROSOPHILA

85

Hence, the function of each of these genes appears to be dependent on that of another member of the group and, consequently, the function of the entire chain is perturbed if any of the links is missing. Recent data on the distribution of transcripts from neurogenic genes in neurogenic mutants indicate that some of the interrelationships between the neurogenic genes are likely to reflect transcriptional regulation (Godt et al., 1991).

8 Physical interactions of Notch and Delta The participation of Notch and Delta in cell communication processes is strongly suggested by the primary structure of the encoded proteins. The Notch protein, as deduced from sequences of genomic and cDNA clones (Wharton et al., 1985; Kidd et al., 1986), shows features typical of a transmembrane protein. Indeed, antibodies raised against different parts of the Notch protein confirm its location in the cell membrane (Kidd et al., 1989; Johansen et al., 1989). A striking feature of the extracellular domain is a tandem array of 36 cysteine-rich motifs each of about 40 amino acids with similarity to the epidermal growth factor (EGF) and other proteins of vertebrates and invertebrates. The Notch protein is, like the Notch RNA (Hartley et al., 1987), ubiquitously distributed during early developmental stages, although there are quantitative differences between cells suggesting a role for Notch in epidermogenesis (Johansen et al., 1989; IOdd et al., 1989; Fehon et a f . , 1991; see also Hoppe and Greenspan, 1986, 1990). Delta is also a transmembrane protein, with a hydrophobic signal sequence and a membrane-spanning domain (Vassin et al., 1987; Kopczynski et al., 1988). The extracellular domain contains nine EGF-like repeats, arranged in tandem like those in the Notch protein. The Delta protein includes 833 residues (Haenlin et al., 1990). Transcription of Delta is spatially regulated; however, during neuroblast segregation Delta RNA is apparently present in all VNE cells (Vassin et al., 1987; Kopczynski and Muskavitch, 1989; Haenlin et al., 1990). Recent data on the distribution of the Delta protein (Kooh et al., 1993) corroborate the RNA findings and show no evident asymmetric distribution within the neuroectoderm. The EGF-like repeats and other motifs in the extracellular domains of Delta and Notch may well represent essential parts of the cell communication pathway by mediating direct protein-protein interactions. Genetic mosaic analyses indicate that the products of Notch and Delta cannot diffuse (Dietrich and Campos-Ortega, 1984; Hoppe and Greenspan, 1986; de Celis et al., 1991; Heitzler and Simpson, 1991). Various pieces of experimental evidence support the view that the EGF-like repeats of Notch mediate protein-protein interactions. First, mutations at the Notch locus, i.e. split and several Abruptex alleles, each differ from the Notch wild-type protein by single amino acid exchanges in specific, distantly separated EGF-like

86

JOSE A. CAMPOS-ORTEGA

repeats, i.e. 14th with respect to split, and 24th, 25th, 27th and 29th with respect to Abruptex alleles (Hartley et af., 1987; Kelley et al., 1987). Two alleles of Delta have been recovered which suppress the phenotype of the split mutation but which do not modify the phenotype of Abruptex alleles (Brand and Campos-Ortega, 1990); that is to say, this is an allele-specific suppression of the split phenotype in both cases. These alleles have been sequenced and both are associated with single amino acid exchanges in the 4th and 9th EGF-like repeat respectively (Lieber et af., 1992). This is altogether strong evidence to support the contention of direct interactions between Notch and Delta. Two different pieces of biochemical evidence provide support for a direct association of Notch and Delta and show that, indeed, both proteins may mediate cell adhesion. Transient expression of both proteins under the control of a metallothionein promoter (Fehon et al., 1990; Rebay et al., 1991), or expression of clones comprising the coding regions of Notch or Delta under the control of a heat shock promoter after stable transfection (Lieber et af., 1992), confers upon Schneider cells adhesive properties which they normally lack. The adhesivity is cell concentration dependent and heterophilic, in that cells expressing Notch adhere only to Delta- but not to other Notch-expressing cells (Lieber et al., 1992). Schneider cells expressing either split or an Abruptex variant (AxR) are still capable of adhering to cells expressing one of the Delta proteins that suppress split (Deltasup’) with the same kinetics as cells expressing wild-type forms of Notch and Delta. However, the mutant proteins cannot compete with the corresponding wild-type proteins in cell aggregation assays, suggesting that their adhesivity is somehow impaired (Lieber et al., 1992). Notch-mediated cell adhesion has been found to depend on EGF-like repeats 11th and 12th (Rebay et af., 1991). Yet, the region of interaction between both proteins appears to be much larger. Suppression of split by Deltasup’ is not compatible with a mechanism based on cell adhesion in vivo, since no reversion of the diminished binding activity of split is observed after aggregation with cells expressing the suppressor protein. This result suggests that intracellular signalling by spfit relevant to compound eye development is mediated by specific parts of the protein, different from those required to mediate adhesion (Lieber et af., 1992). The experimental evidence thus indicates that Notch and Delta are directly associated at the membrane of the neuroectodermal cells. Data on the distribution of the proteins (Fehon et af., 1991; Kooh et af., 1993) confirm their co-localization in assumedly interacting cells. One possible mode of direct interaction between Notch and Delta at the cell membrane involves a receptor-ligand relationship. The available evidence points to Notch as a receptor and Delta as the source of regulatory signals required for epidermogenesis (Dietrich and Campos-Ortega, 1984; Hoppe and

NEUROGENESIS IN DROSOPHILA

a7

Greenspan, 1986, 1990; Technau and Campos-Ortega, 1987; de Celis et af., 1991; Heitzler and Simpson, 1991, 1993).

8 The epidermal decision is controlled by the E(sPL)-C The genetic analysis shows that the functions encoded by the E(sPL)-C are epistatic over those of the other neurogenic genes. This is compatible with the assumption that this locus is responsible for the epidermalizing function i.e. inhibition of neurogenesis, ascribed to the whole group of the neurogenic genes (Vassin et al., 1985; Knust et a f . , 1987b; de la Concha et al., 1988). Rather than being the function of a single gene, the epidermal decision of the VNE cells is controlled by contributions from seven partially redundant, ancestrally related genes, i.e. H L H - m a , H L H - m y , H L H - m P , H L H - d , H L H - m 5 , H L H - m 7 , and E(spf), which constitute the E(sPL)-C (Knust et al., 1987b, 1992; Klambt et af., 1989; Schrons et a f . , 1992). Causal relationships between a small deletion present in the m8 transcription unit of the E(spf)D mutant and the enhancement of split, were established by P element-mediated transformation using mutant and wild-type genes (Klambt etal., 1989; Tietze et a f . , 1992). Genetic analyses have demonstrated functional redundancy among the members of the E(sPL)-C (Ziemer et a f . , 1988; Schrons et a f . , 1992). This redundancy is apparently due to two factors. First, during SI and SII neuroblast segregation, transcripts from all genes of the complex but HLH-m3 exhibit virtually identical spatial distributions. The distribution of these RNAs is overlapping and, immediately before the segregation of SI and SII neuroblasts takes place, it matches remarkably well the regions of the VNE from which first the SI neuroblasts and later the SII neuroblasts Will segregate (Knust et af., 1987b, 1992). Second, sequencing of genomic and cDNA clones has shown a high degree of sequence similarity in the proteins encoded by H L H - m a , H L H - m y , H L H - m P , H L H - i d , HLH-m.5, HLH-m7 and E(sp1) (Klambt et al., 1989; Knust et al., 1992). All seven are members of the bHLH family of transcriptional regulators. The transcription unit m9-ml0, which is located in the immediate neighbourhood of the E(sPL)-C (Knust et af., 1987, 1992), corresponds to groucho (Preiss et a f . , 1988; Schrons et a f . , 1992), a mutation associated with various bristle defects (Knust ef al., 1987b). The putative groucho protein is nuclear (Delidakis et af., 1991) and contains a repeated motif which is similar to another motif in the p-subunit of transducin (Hartley et al., 1988), as well as in the product of a cell cycle gene (CDC4; Yochem and Byers, 1987) and in a component of the spliceosome (PRP4; Balroques and Abelson, 1989), both of yeast. Tietze et al. (1992) have shown that the grouch; phenotype itself is due to a lesion in m9-mlO and that an insertion

88

JOSE A. CAMPOS-ORTEGA

of a middle repetitive DNA element present in transcription unit m9/m10of the E(spl)D strain causes a hypomorphic, visible mutation in the groucho gene. groucho is in fact a neurogenic gene with a prominent maternal component of gene expression, which happens to be located in the neighbourhood of the E(sPL)-C (Schrons et al., 1992). 10 The neural decision is controlled by the proneural genes

The AS-C includes four genes, achaete, scute, lethal of scute and asense, the names being derived from the phenotypic effects of their mutations on bristle development and viability (Muller, 1935; Garcia-Bellido, 1979; Ghysen and Dambly-Chaudiere, 1988). The spatial distribution of achaete, scute and lethal of scute transcripts (Cabrera et al., 1987; Romani et al., 1987; Brand and Campos-Ortega, 1988; Cabrera, 1990; Martin-Bermudo et al., 1991; Skeath and Carroll, 1992; Ruiz Gomez and Ghysen, 1993) is in principle similar for all three and shows a high degree of correlation with the processes of neuroblast segregation and development of sensory organs, i.e. those processes in which, from the analysis of mutants, we know the functions of the genes to be required. During early neurogenesis, the three transcripts are expressed in partially overlapping clusters of cells within the VNE. One or two cells from among these clusters will delaminate as neuroblasts. After the segregation has occurred, RNA of these genes remains detectable in the neuroblasts for some time. This pattern of transcription is suggestive of a role for the AS-C genes in neuroblast commitment. Cabrera (1990) and Martin-Bermudo et al. (1991) have raised antibodies against lethal of scute, and Skeath and Carroll (1991, 1992) against achaete and scute, and they find a similar correlation between accumulation of protein and neuroblast segregation. Not all the neuroblasts express similar amounts of lethal of scute, suggesting that not all the neuroblasts require this gene product for their commitment (MartinBermudo et al., 1991; see also Jimenez and Campos-Ortega, 1990). Causal relationships between expression of proneural genes and commitment of VNE cells as neuroblasts are indicated by genetic studies (JimCnez and Campos-Ortega, 1987, 1990). As mentioned above, AS-C- mutants have 20-25% fewer neuroblasts than the wild-type; similar findings have been made with vnd mutants. Apparently, the AS-C and vnd control the commitment of non-overlapping populations of neuroblasts, since AS-Cand vnd- double mutants show additive effects and lack roughly 50% of all neuroblasts. In addition, increasing the number of copies of the AS-C genes and of vnd by using duplications of the region leads to development of additional neuroectodermal cells as neuroblasts at the expense of epidermoblasts. Therefore, the complementary phenotypes caused by decreasing and increasing the number of copies and, presumably, the amount

NEUROGENESIS IN DROSOPHILA

89

of product of the proneural genes, strongly support the hypothesis that these products are responsible for the development of neuroectodermal cells as neuroblasts. A similar role for these genes has been proposed with respect to the commitment of the progenitor cells of sensory organs. It has been shown that, contrary to the situation with the neuroblasts, the function of achaete, scute and asense is prominent in sensory organ development of the larva and imago, whereas lethal of scute is apparently dispensable for this process (Garcia-Bellido and Santamaria, 1978; Garcia-Bellido, 1979; Dambly-Chaudiere and Ghysen, 1987; Ghysen and Dambly-Chaudiere, 1988, 1989; Ghysen and O’Kane, 1989; Bodmer et al., 1989; Jan and Jan, 1990). Sequence analyses have shown that the proteins encoded by achaete, scute, lethal of scute and asense are similar to each other and contain the bHLH domain, involved in transcription activation (Villares and Cabrera, 1987; Alonso and Cabrera, 1988; Gonzalez et af., 1989). The daughterless locus has been found to encode a bHLH protein as well (Caudy et af., 1988b; Cronmiller ef al., 1989). Murre et al. (1989a,b) showed that proteins with the HLH motif are able to form homodimers and heterodimers and bind to specific DNA sequences. These findings strongly support the contention that all these proteins function in vivo as transcriptional regulators. A high degree of specificity and complexity in the regulatory functions of the corresponding genes may thus be achieved through the combination of different proteins to form heterodimcrs. Genetic interactions between some of the AS-C genes and daughterless during sensory organ development have been recently described (Dambly-Chaudiere et al., 1988), suggesting that these genes are involved in closely related functions (Jan and Jan, 1990). Thus, the finding (Murre et al., 1989b) that lethal of scute and daughterless may form DNA binding heterodimers corroborates the observations of and the inferences drawn from the genetic analysis. 11 Interactions between neurogenic and proneural genes

Evidence has been obtained that the AS-C genes and daughterless are functionally interconnected with the neurogenic genes within the same genetic network (Brand and Campos-Ortega, 1988, 1990; Cabrera, 1990; Skeath and Carroll, 1992; Hinz et a f . , 1994; Haenlin et al., 1994; Kramatschek and Campos-Ortega, 1994; Oellers et al., 1994. First, the severity of the neurogenic phenotype of double mutants for neurogenic genes and deletions of the AS-C or daughterfess is considerably reduced, compared with the phenotype of the same neurogenic mutation alone, i.e. without the concomitant presence of a AS-C- or daughterless mutation. AS-C- mutations were found to be epistatic over the neurogenic mutations, suggesting that the function of the AS-C genes follows that of the neurogenic genes in the functional chain (Brand and Campos-Ortega, 1988).

90

JOSE A. CAMPOS-ORTEGA

The distribution of products from t h e AS-C genes in neurogenic mutants suggests that at least some of the interactions between neurogenic and AS-C genes are likely to involve regulation of the transcription of these genes. Changes in the pattern of transcription of the genes lethal of scute and achuete have been observed in embryos carrying any of several neurogenic mutations (Brand and Campos-Ortega, 1988; Skeath and Carroll, 1992; Ruiz-Gomez and Ghysen. 1993). In these embryos, the early expression of lethal of scute and achaete is indistinguishable from the wild-type. However, already at the beginning of neuroblast segregation, RNA is found in larger clusters than in the wild-type probably due to the fact that neuroblasts do not segregate from the clusters. In the wild-type, a restriction of lethal of scute and uchaete transcription occurs from an initial group of several neuroectodermal cells to one, or a few neuroblasts, as they segregate from the epidermoblasts; in neurogenic mutants, this restriction fails to occur. Cabrera ( 1990) has recently presented similar evidence with respect to lethal o f s c u t e , and Skeath and Carroll (1992) with respect to achaete in that the number of neuroectodermal cells expressing these proteins is higher in neurogenic mutants than in the wild-type. All these observations indicate that the neurogenic genes define the normal expression of AS-C genes in that they suppress the transcription (or the accumulation) of RNA from at least lethal of scute and achaete in some of the VNE cells. Molecular evidence, both in v i m and in vivu, for interactions between proneural and neurogenic genes, derives from the analysis of the promoters of Delta (Haenlin et al., 1994) and of HLH-m5 and E(sp1) (Kramatschek and Campos-Ortega, 1994; Oellers et ul., 1994, as well as from the ectopic expression of lethal of scute (Hinz et al.. 1994). These studies have revealed that proneural proteins activate transcription of Delta, HLH-m5 and E(spl) by means of binding to multiple sites distributed throughout their promoters. The activation of E(sPL.)-C genes by proneural proteins within the VNE is a n early event. initiating the transcription of these genes in this particular region. Thc assumed function of this activation process will be discussed below. Activation o f transcription of Delta. which encodes the presumptive epidermalizing signal molecule, strongly suggests that it serves the function of increasing the efficacy of lateral inhibition in the prospective neuroblast during its singling out from the proneural cluster.

12 Conclusions

The emergence of neuroblasts and epidermoblasts from an undifferentiated anlage is the result of a rather complex set of operations. At the beginning of this review, I proposed a hypothesis in which these operations were arranged in three consecutive steps: acquisition of competence to assume the neural fate; selection of single cells from groups of a few cells to take on

NEUROGENESIS IN DROSOPHILA

91

the fate of particular neuroblasts; suppression of the primary neural fate by lateral inhibition of the remaining cells of the groups to enable their epidermal development. The competence to take on neural fate is likely to depend on the presence of products of the proneural genes in the neuroectodermal cells. Neural competence in the VNE exhibits a clear positional context, in that groups of four to five cells are enabled to give rise to particular types of neuroblasts depending on their position (Taghert et af., 1984; Doe and Goodman. 1985b; Doe et af., 1988; Skeath et af., 1992). These groups are comparable to the so-called proneural cell clusters in sensory organ development (Ghysen and Dambly-Chaudiere, 1989; Romani etaf., 1989; Simpson, 1990; Jan and Jan, 1990). However, the postulated proneural clusters of the VNE are probably in immediate contiguity to each other, such that eventually the entire VNE receives neural competence and can become neuralized in neurogenic mutants, whereas the clusters in the epidermis, from which progenitor cells of sensory organs will arise, may be widely spaced. In fact, both the phenotype of neurogenic mutations and the results of transplanting cells from the anterolateral ectoderm suggest that the neural competence extends outside of the VNE. Neural competence and specification of neuroblasts and epidermoblasts are probably two different operations which may require the functions of different genes (Skeath et af., 1992; see Rodriguez et af., 1990, for sensory organ specification). In the second step, one cell from each proneural cluster has to be singled out to become a neuroblast. The important element in this process seems to be the presence of a critical amount of proneural proteins, but it is unknown how this comes to be since all cells of the group are assumed to have initially comparable amounts. Oscillations in the content of proneural gene products, which would cause the predominance of proneural gene products in one cell and its entry into the neural pathway. may occur in a cellautonomous manner. Cubas and Modolell (1992) and Van Doren et a/. (1992) have presented evidence that within the epidermal proneural clusters, progenitor cells of the sensory organs develop in regions of minimal concentration of extramacrochaetae, a negative regulator of proneural genes (Ellis et al., 1990; Garreil and Modolell, 1990). Higher concentration of extramacrochaetae in a given cell would down-regulate proneural proteins and impede neural development in that cell. Unfortunately, there is no evidence for a participation of extramacrochaetae in neuroblast commitment. Here, the predominance of proneural gene products in one cell could be the consequence of information conveyed to the prospective neuroblast by way of its relations with its neighbours. With respect to bristle development, Heitzler and Simpson’s (1991, 1993) results suggest that the amount of Notch product present in a cell has an influence on its fate, less Notch than normal leading to neural, more Notch to epidermal development. Struhl er af. (1993) have recently obtained evidence that truncated forms of the Notch protein, which contain only the intracellular domain, are

92

JOSE A. CAMPOS-ORTEGA

constitutively active. These observations are compatible with a function of Notch as a ligand-activated receptor in neuroblast segregation. Finally, as far as the third step is concerned, the data show convincingly that cell interactions are somehow involved in development of the epidermoblasts, although they do not yet allow inferences about when the salient cellular interactions take place. Evidence suggests that the decision of one cell from each proneural group to become a neuroblast is followed by cell communication to impede neurogenesis in the remaining cells of the groups and permit in this way their epidermal commitment (Doe and Goodman, 198%; Hartenstein and Campos-Ortega, 1984). However, as with the neuroblasts, epidermogenesis may also be initiated autonomously and reinforced by lateral inhibition derived from the neuroblasts. In fact, observations of Luer and Technau (1902) on the in vitro behaviour of single cells from defined ectodermal regions indicate that epidermogenesis can in part be the result of a cell autonomous process. Cell communication could be required, for example, to suppress expression of proneural genes in the cells normally developing as epidermoblasts, for in neurogenic mutants the domains of expression of lethal of scute and achaete do not become restricted to single cells, as in the wild-type, but extend to all the cells of a proneural cluster. The results of transplanting E(sPL)-C- cells into the VNE of the wild-type were interpreted to mean that the E(sPL)-C encodes functions related to the reception of epidermalizing signals. The proteins encoded by HLH-mG, HLH-my. HLH-mP, HLH-m5, HLH-m7, and E(spl), with the basic DNA binding and the HLH dimerization motif (Klambt et al., 1980; Knust et al., 1992), are indeed compatible with a function at the side of the receptor. Therefore, the failure of E(sPL)-C- cells to develop epidermal fates following transplantation into the wild-type VNE was most probably due to the lack of proteins allowing epidermal development of the transplanted cells. Granted a regulation of the specific genetic activities of the epidermoblasts by the products of the E(sPL)-C, this regulation may occur by direct binding to DNA; indeed, the HLH-m5 and E(sp1) proteins are capable of binding to specific DNA sequences in vitro (Tietze et al., 1992; Oellers ef al., 1994). Or, alternatively, regulation may occur by heterodimer formation with other bHLH proteins, impeding their activity in this way. I t seems improbable that proteins encoded by the E(sPL)-C activate directly transcription of the ‘realizators’ (Garcia-Bellido, 1975) of the epidermal pathway, i.e. the genes whose products eventually make the epidermal cells, since the complete deletion of the E(sri>)-C still permits epidermal development in the dorsal-most embryonic regions (Knust et al, l087a). I would like to propose that the E(sPL)-Cregulates epidermogenesis indirectly via the proneural genes (Fig. 3). My assumption here is that epidermogenesis is a constitutive developmental pathway that is followed by any cell without intervening genetic regulation. In the cells destined to

NEUROGENESIS IN DROSOPHILA A

93

realisators, epilermis (-1

E(SFL) -c-

proneural

f- genes (-)

realisators, CNS

x

B

realisators epidermis

genes realisators, CNS 4 Neuroblasts

C

FIG. 3 A formal genetic model of the interrelationships between proneural and E(sPL)-C genes during neuro-epidermal lineage segregation. (A) Proneural and E(sPL)-C genes are thought to functionally inactivate each other; proneural genes activate the neural and repress the epidermal ‘realizator’ genes. If the balance between both groups of regulatory genes is perturbed, proneural (B) or E(sPL)-C genes (C) will predominate in the corresponding cell, leading to its development as a neuroblast or as an epidermoblast, respectively. ’

become neuroblasts, proneural gene products would suppress the epidermal ‘realizator’ genes; in the other cells, the proneural gene products would be suppressed by the products of the E(sPL)-C. I already mentioned how the neurogenic phenotype of double mutants AS-C- and, for example, E(sPL)C-, is reduced compared with that of the E(sPL)-C- mutants alone (Brand and Campos-Ortega, 1988). This observation suggests that the AS-C, as well as the other proneural genes, suppress the epidermal ‘realizator’ genes: the deletion of the AS-C eliminates this suppressive effect and permits the development of epidermis. I further assume that the proneural gene products activate in addition the neural ‘realizator’ genes and thus permit neural development of the neuroblasts. This assumption is also supported by the reduction of the neurogenic phenotype of AS-C- and E(sPL)-C- double mutant embryos. Activation and suppression here are meant operationally, without presupposing any particular molecular mechanism. Cell determination in the VNE can thus be envisaged as the result of a delicate balance

94

JOSE A. CAMPOS-ORTEGA

between the functional activity of the proteins encoded by the proneural genes and the E(sPL)-C. Both groups of proteins are assumed to be capable of reciprocally regulating each other, as suggested by the opposite effects, neural hyperplasia or hypoplasia, of duplications and deletions of the proneural and of the E(sPI.)-C genes, discussed in the previous sections. In fact, the complexity of both gene complexes and the structure and function of the bHLH proteins they encode, point to a great potential for interrelationships, of which very little is known. This model is based exclusively on genetics and there is as yet little molecular evidence to support it. The proposed regulation can be achieved at the transcriptional, at the post-transcriptional, or at both levels. There is both in vifro and in vivo evidence that the proneural proteins activate the transcription of the E(sPL)-Cgenes E(spl) and HLH-m5 (Kramatschek and Campos-Ortega, 1994; Hinz et id., 1994). This activation appears to be required to allow products of the E(sPL)-C genes to be present within the same VNE cells in which products of the proneural genes occur, in order for them to be able to compete with each other. In v i m studies using cell transfection assays (Oellers et ul.. 1994) have shown that addition of HLH-m5 and E(spl) to a specific DNA sequence in the E(sp1) promoter (N-box) reduces the transcriptional activation mediated by heterodimers between lethal of scute and daughterless. Therefore, the proteins encoded by the E(sPL)-Cmay act indirectly as negative regulators of transcription of proneural genes. The results of these interrelationships are likely to be either the predominance of the proneural proteins or their functional suppression in any particular VNE cell, leading to its commitment to one of the fates. In any cell in which proneural products predominate, one first step on the neural pathway would consist of the synthesis of the molecular machinery that establishes communication with ncighhouring cclls (Fig. 4). This initial communication WOUHperm2 reception of neurakzing sf&-zX? and fhus reinforce and stabilize the neural decision. A second step for the same cell on its way to neurogenesis would be the synthesis of the molecular apparatus for transmission of epidermalizing signals. for instance formation of functional Delta protein, in order to permit lateral inhibition to occur. Delta transcription can indeed be activated in the VNE by proneural proteins (Kunisch et ul., submitted). Initially, this activation takes place in all cells of the proneural cluster, but is likely to be dependent on the concentration of proneural proteins. Hence, this would lead to an increase of the amount of inhibitory signal in the cell that has initiated neurogenesis, to the consecutive inhibition of the surrounding cells of the cluster and the reinforcement of the neural decision in the neuroblast. Conversely, the behaviour of those cells of the competence group which follow the epidermal pathway is likely to reflect the functional suppression of the proneural proteins, and consequently the initiation of the constitutive

95

NEUROGENESIS IN DROSOPHILA Pioneural i3eiies act i ' J d t P t r a n s c r i p t i o n of E(SPL)-(. genes

111i i ~ u i r ? i ~ t (ni ~ ~ i ~ r

f ' r > m p p t i t i o nb e t w e e n E I S P L ) - C and p r o n e u r a l g p n e s

1(+' e p i d e r m a l is i n g

FIG. 4 Sequence of events assumed to occur during lineage segregation. The squares represent cells within a proneural cluster. Refer to text for further details.

developmental pathway, with the synthesis of receptor for epidermalizing signals in order to reinforce and stabilize the epidermal decision. Acknowledgements

I would like to thank the members of my laboratory for support, and Elisabeth Knust, Paul Hardy and Thomas Klein for constructive criticisms on the manuscript. The research reported here was supported by several grants of the Deutsche Forschungsgemeinschaft (DFG, SFB 243) and the Fonds der Chemischen Industrie.

96

JOSE A. CAMPOS-ORTEGA

References Alonso. M. C . and Cabrera, C.V. (1988). The uchaete-scute gene complex of Drosophilu melunogaster comprises four homologous genes. E M B O J. 7, 25852501. Balroques, J . and Abelson, J. N . (1989). PRP4: a protein of the yeast U4iU6 small nuclear ribonucleoprotein particle. Mol. Cell. B i d . 195, 233-245. Bate. C . M. (1976). Embryogenesis of an insect nervous system. I . A map of the thoracic and abdominal neuroblasts in Locitsta rnigrutoriu. J . Embryol. Exp. Morphol. 35. 107-123. Bate. C . M. (1982). Development of the neuroepithelium. Neurosci. Res. Bull. 20, 803-806. Becker. T. and Technau. G . M. (1990). Single cell transplantation reveals interspecific cell communication in Drosophila chimeras. Development 109. 821-832. Bccr. J . , Technau, G . M.and Campos-Ortega, J. A . (1987). Lineage analysis of transplanted individual cells in embryos of Drosophila melanogaster. IV. Commitment and proliferative capabilities of mesodermal cells. Roux’s Arch. Dev. Biol. 196, 222-230. Bellen. H. J.. O’Kane. C. J . , Wilson. C., Grossniklaus, U., Pearson, R . K. and Gehring. W. J. ( 1989). P-element-mediated enhancer detection: a versatile method to study development in Drosophilu. Genes Dev. 3 . 1288-1300. Bier. E . , VBssin, H . , Sepherd, S., Lee, K.. McCall, K., Barbel, S., Ackerman, L., Carretto. Uemura, T.. Grell, E., Jan. L. Y . and Jan. J. N. (1989). Searching for pattern and mutation in the Drrxnphilu genome with a P-lacZ vector. Genes Dev. 3 , 1273-1287. Bodmer. R., Carretto. R. and Jan, Y . N. (1989). Neurogenesis of the peripheral nervous system in Drosophila embryos: DNA replication patterns and cell lincages. Neuron 3 . 21-32. Boulianne. G . L.. de la Concha, A . . Campos-Ortega. J. A , . Jan. L. Y . and Jan Y . N . (1991). The Drosophila neurogenic gene neuralized encodes a novel protein which is expressed in precursors of larval and adult neurones. E M B O J. 10, 2975-2983. Bourouis. M., Heitzler. P.. El Messal, M. and Simpson, P. (1989). Mutant Drosophilu embryos in which all cells adopt a neural fate. Nature 341, 442444. Bourouis. M.. Moore. P., Ruel, L.. Grau. Y . . Heitzler, P. and Simpson. P. (1990). An early embryonic product of the gene shaggy encodes a serinehhreonine protein kinase related to the CDC281cdc2f subfamily. E M B O J. 9, 2877-2884. Brand. M. and Campos-Ortega, J. A . (1988) Two groups of interrelated genes regulate early neurogenesis in Drosophilu melanogaster. Roux’s Arch. Dev. Biol. 197. 457470. Brand. M. and Campos-Ortega, J. A . (1990). Second site modifiers of the split mutation of Notch define genes involved in neurogenesis in Drosophila melanogusrer. Roux’s Arch. Dev. Biol. 198, 275-285. Brand. M. and Campos-Ortega ( 1094). The role of daughterless during development of the embryonic and adult nervous system of Drosophila melanogaster. submitted. Cabrera. C. V. (1990). Lateral inhibition and cell fate during neurogenesis in Drosophilu: the interactions between scute. Notch and Delta. Development 109, 733-742. Cabrera. C . V.. Martinez-Arias. A . and Bate, M. (1087). The expression of three members of the uchaere-scute gene complex correlates with neuroblast segregation in Drosophilu. Cell 50. 425433.

NEUROGENESIS IN DROSOPHlLA

97

Campos-Ortega. J. A . and Haenlin, M. (1992). Regulatory signals and signal molecules in early neurogenesis of Drosophila melanogaster. Roux’s Arch. Dev. Biol. in press. Campos-Ortega, J. A . and Hartenstein, V. (1985). ”The Embryonic Development of Drosophila melanogaster.” Springer Verlag, Berlin. Campos-Ortega, J. A.. Lehmann. R . , Jimenez, F. and Dietrich, U. (1984). A genetic analysis of early neurogenesis in Drosophila. In “Organizing Principles of Neural Development” (Ed. S. C. Sharma). pp. 129-144. Plenum Press, New York. Caudy. M., Grell, E. H., Dambly-Chaudiere, C.. Ghysen, A , . Jan, L. Y. and Jan, Y. N. (1988a). The maternal sex determination gene daughterless has zygotic activity necessary for the formation of peripheral neurons in Drosophila. Genes Dev. 2. 843-852. Caudy, M.. Vissin. H., Brand. M., Tuma, R., Jan, L. Y., Jan, Y. N. (198%). daughterless, a gene essential for both neurogenesis and sex determination in Drosophila. has sequence similarities to m y and the achaete-scute complex. Cell 55, 1061-1067. Cronmiller, C.. Schedl, P. and Cline. T . W. (1989). Molecular characterization of daughterless, a Drosophila sex determination gene with multiple roles in development. Genes Dev. 2. 1666-1676. Cubas, P. and Modolell, J. (1992) The extramacrochaetae gene provides information for sensory organ patterning. E M B O J. 11. 3385-3393. Cui, X. and Doe, C. Q. (1992). ming is expressed in neuroblasts sublineages and regulates gene expression in the Drosophila central nervous system. Development 116, 943-952. Dambly-Chaudiere. C. and Ghysen. A . (1987). Independent subpatterns of sense organs require independent genes of the achaete-scute complex in Drosophila larvae. Genes Dev. 1, 297-306. Dambly-Chaudiere, C., Ghysen, A , , Jan, L. Y. and Jan, Y. N. (1988). The determination of sense organs in Drosophila: interactions of scute with daughterless. Roux’s Arch. Dev. Biol. 197, 419-423. de Celis, J., Mari-Beffa, M. and Garcia-Bellido, A. (1991) Cell autonomous role of Notch, an epidermal growth factor homolog. in sensory organ differentiation in Drosophila. Proc. Natl. Acad. Sci. USA 88, 632436. de la Concha. A . . Dietrich, U . . Weigel. D. and Campos-Ortega. J. A . (1988). Functional interactions of neurogenic genes of Drosophila melanogaster. Genetics 118, 499-508. Delidakis, C.. Preiss, A , , Hartley. D . and Artavanis-Tsakonas. S. (1991). Two genetically distinct functions reside within the Enhancer of split locus of Drosophila melanogaster. Genetics 129, 803-823. Dietrich. U . and Campos-Ortega, J. A . (1984). The expression of neurogenic loci in imaginal epidermal cells of Drosophila melanogaster. J. Neurogenef. 1, 315-332. Doe, C. Q. (1992). Molecular markers for identified neuroblasts and ganglion mother cells in the Drosophila central nervous system. Development 116. 855-863. Doe, C. Q. and Goodman, C. S. (1985a). Early events in insect neurogenesis. I . Development and segmental differences in the pattern of neuronal precursor cells. Dev. B i d . 111, 193-205. Doe. C. Q. and Goodman. C. S. (1985b). Early events in insect neurogenesis. 11. The role of cell interactions and cell lineages in the determination of neuronal precursor cells. Dev. Biol. 111, 20&219. Doe, C. Q.. Hiromi, Y., Gehring, W. J. and Goodman C . S. (1988). Expression and

98

JOSE A. CAMPOS-ORTEGA

function o f the segmentation gene firshi-turuzit during Drosophila neurogenesis. Sciencc 239, 17(k175. Ellis, H. M.. Spann. D. R. and Posakony. J. W. (1990) extramacrochaetae, a negative regulator of sensory organ development in Drosophilu, detines a new class o f helix-loop-helix proteins. Cell 61. 27-38. Fehon, R. G . . Kooh. P. J.. Rebay, I.. Regan. C. L., Xu. T., Muskavitch. M. A. T. and Artavanis-Tsakonas. S. (1990). Molecular interactions between the protein products of the neurogenic loci Notch and Delta. two EGF-homologous genes in Drosoplzilri. Cdl 62. 523-533. Fehon, R. G . . Johansen, K . . Rebay, I. and Artavanis-Tsakonas, S. (1991). Complex cellular and subcellular regulation of Notch expression during embryonic and imaginal development of Drosophila: implications for Notch function J . Cell Biol. 113, 657-669. Garcia-Bcllido, A . (1975). Genetic control of wing disc development in Drosophila. In “Cell Patterning” Ciba Foundation Symposium 29 (Ed. by S. Brenner), pp. 161-178. Elsevier. Excerpta Medica. North Holland. Assoc. Amsterdam, Oxford, New York. Garcia-Bellido, A . (1979). Genetic analysis o f the acliric>te-.mite system of Drosophilu melanoguster. Genetics 91. 491-520. Garcia-Bellido, A . and Santamaria. P. ( 1978). Developmental analysis of the nchaetc.-scure system of Drosophilu rnelanogaster. Genetics 88. 469-486. Garrell. J. and Modolell, J . ( 1990). The Drosophilu extranzacrocliuetae locus, an antagonist of proneural genes that, like these genes, encodes a helix-loop-helix protein. Crll61, 39-48. Ghysen. A . and Damhly-Chaudiere. C . (1988). From DNA t o form: the achaetescute complex. Genes Dev. 2. 495-501. Ghysen, A. and Dambly-Chaudiere. C . ( 1989). Genesis of the Drosophila peripheral nervous system. Trend in Genetics 5. 251-255. Ghysen. A . and Dambly-Chaudiere, C. (1990). Early events in the development of Drosophila peripheral nervous system. J . Plivsiol. (Paris) 84, 1 1-20. Ghysen, A . and O’Kane, C. (1989). Detection of enhancer-like elements in the genome of Drosophilu. Develoiiment 105, 35-52. Ghysen. A . . Dambly-Chaudiere. C.. Jan. L. Y. and Jan. Y.-N (1993). Cell interactions and gene interactions in peripheral neurogenesis. Genes Dev. 7, 723-733. Godt, D.. Schrons, H . . Guth, S. and Camps-Ortega, J. A . (1991). The distribution o f transcripts of neurogenic genes in neurogenic mutants of Drosophila melanogaster. J . Neurogeriet. 7. 241-252. Gonzilez. F., Romani, S., Cubas. P.. Modolell. J. and Campuzano. S. (1989). Molecular analysis of userise. a member of the uchuete-scure complex of Drosophila rnelanogristrr. and its novel role in optic lobe development. E M B O J. 8. 3553-3562. Goriely. A . , Dumont. N.. Dambly-Chaudiere. C. and Ghysen. A . (1991). The determination of sense organs in Drosophilri. effect of the neurogenic mutations in the embryo. Drvelopmenr 113, 1395-1404. Haenlin. M., Kramatschek. B. and Campos-Ortega, J. A. (1990). The pattern of transcription of the neurogenic gene Delta of Drosophilu nielanogaster. Developrnent 110, 905-914. Haenlin, M., Kunisch, M., Kramatschek. B. and Campos-Ortega. J. A . (1994). Genomic regions regulating expression of the Drosophila neurogenic gene Delta. Meclz. Dev. (in press). Hartenstein. V. and Campos-Ortega, J . A . (1984). Early neurogenesis in wildtype Drosophila melunogrister. Roux’s Arch. Dev. B i d . 193, 308-325.

NEUROGENESIS IN DROSOPHILA

99

Hartenstein, V. and Campos-Ortega. J. A . (1986). The peripheral nervous system o f mutants of early neurogenesis in Drosophila melanopster. ROILX'S Arch. Dev. Biol. 195. 210-221. Hartenstein. V., Rudloff, E. and Campos-Ortega, J. A . (1987). The pattern o f proliferation of the neuroblasts in the wild-type embryo of Drosophila melunogasrer. Roux's Arch. Dev. Biol. 196. 473485. Hartley, D. A.. Xu. T. and Artavanis-Tsakonas, S. (1987). The embryonic expression of the Notch locus o f Drosophifa melanogaster and the implications o f point mutations in the extracellular EGF-like domain of the predicted protein. EMBO J . 6. 3407-3417. Hartley. D. A . , Preiss. A . and Artavanis-Tsakonas. S. (1988). A deduced gene product from the Drosophilu neurogenic locus Enhancer of split shows homology to mammalian G-protein p subunit. Cell 55. 785-795. Heitzler, P. and Simpson, P. (1991). The choice of cell fate in the epidermis of Drosophila. Cell 64. 1083-1092. Heitzler, P. and Simpson. P. (1993). Altered epidermal growth factor-like sequences provide evidence for a role o f Notch as a receptor in cell fate decisions. Development 117. 1 1 13-1 123. Hinz, U.. Giebel. B. and Campos-Ortega. J. A . (1994). The basic-helix-loop-helix domain of the Drosophilu lethal of scute protein is sufficient for proneural function and activates neurogenic genes. Cell 76, 77-87. Hoppe, P. E . and Greenspan, R. J. (1986). Local function of the Notch gene for embryonic ectodermal choice in Drosophila. Cell 46, 773-783. Hoppe, P. E. and Grecnspan. R . J . (1990). The Notch locus of Drosophila is required in epidermal cells for epidermal development. Development 109. 875-885. Jan, Y. N. and Jan. L. Y. (1990). Genes required for specifying cell fates in Drosophila embryonic sensory nervous system. Trends Neurosci. 13, 493498. Jimenez. F. and Campos-Ortega. J . A . (1979). A region of the Drosophilu genome necessary for CNS development. Nature 282. 310k312. Jimenez, F. and Campos-Ortega. J . A . (1982) Maternal effects of zygotic mutants affecting early neurogenesis in Drosophila. Roux's Arch. Dev. Biol. 191. 191-201. Jimenez. F. and Campos-Ortega, J . A . (1987). Genes in subdivision 1B of the Drosophila melanoguster X-chromosome and their influence on neural development. 1. Neurogenet. 4, 179-200. Jimenez. F. and Campos-Ortcga. J . A . (1990). Defective neuroblast commitment in mutants of the achaete-scute complex and adjacent genes of Drosophila melanogaster. Neuron 5. 81-89. Johansen. K. M . . Fehon. R. G. and Artavanis-Tsakonas, S. (1989) The Notch gene product is a glycoprotein expressed o n the cell surface of both epidermal and neuronal precursor cells during Drosophila development. J . Cell. Biol. 109. 2427-2440, Jiirgens, G.. Lehmann. R.. Schardin. M. and Niisslein-Volhard, C. (1986). Segmental organisation of the head in the embryo of Drosophila melanogaster. A blastoderm fate map of the cuticle structures of the larval head. Roux's Arch. Dev. Biol. 195. 359-377. Kelley. M. R., Kidd, S., Deutsch. W. A. and Young, M . W. (1987). Mutations altering the structure of epidermal growth factor-like coding sequences at the Drosophilu Notch locus. Cel! 51, 539-548. Kidd. S.. Kelley, M. R. and Young. M. W . (1986). Sequence of the Notch locus of Drosophilu mdonugaster: Relationship o f the encoded protein to mammalian clotting and growth factors. Mol. Cell. Biol. 6. 3094-3108. Kidd. S.. Baylies, M. K., Gasic. G . P. and Young, M. W. (1989) Structure and

100

JOSE A. CAMPOS-ORTEGA

distribution of the Notch protein in developing Drosophila. Genes Dev. 3, 11 13-1 129. Klambt, C., Knust, E., Tietze, K. and Campos-Ortega, J. A . (1989). Closely related transcripts encoded by the neurogenic gene complex Enhancer of split of Drosophila melanogaster. EMBO J . 8, 203-210. Klambt, C., Jacobs, J. R. and Goodman, C. S. (1991). The midline of the Drosophila central nervous system. A model for the genetic analysis of cell fate, cell migration, and growth cone guidance. Cell 64, 801-815. Knust, E . , Bremer, K. A., Vassin, H . , Ziemer, A , . TepaB, U. and Campos-Ortega, J. A . (1987a). The Enhancer of split locus and neurogenesis in Drosophila melanogaster. Dev. Biol. 122, 262-273. Knust, E.. Tietze, K. and Campos-Ortega, J. A . (1987b). Molecular analysis of the neurogenic locus Enhancer of split of Drosophila melanogaster. EMBO J . 6, 4113-4123. Knust, E . , Schrons, H., Grawe, F. and Campos-Ortega, J. A . (1992) Seven genes of the Enhancer of split complex of Drosophila melanogaster encode helix-loop-helix proteins. Genetics 132, 505-518. Kooh, P. J . , Fehon, R. and Muskavitch, M. A . T . (1993). Implications of dynamic patterns of Delta and Notch expression for cellular interactions during a Drosophila development. Development 117, 493-507. Kopczynski, C. C. and Muskavitch, M. A. T. (1989) Complex spatio-temporal accumulation of alternative transcripts from the neurogenic gene Delta during Drosophila embryogenesis. Development 107, 623436. Kopczynski. C. C., Alton, A . K., Fechtel, K.. Kooh, P. J. and Muskavitch, M. A. T. (1988). Delta, a Drosophila neurogenic gene, is transcriptionally complex and encodes a protein related to blood coagulation factors and epidermal growth factor of vertebrates. Genes Dev. 2, 1723-1735. Kramatschek, B. and Campos-Ortega, J . A . (1994) Neuroectodermal transcription of the Drosophila neurogenic genes E(sp1) and HLH-m5 is regulated by proneural and neurogenic trans-activators. Development 120, 815-826. Kunisch, M., Haenlin, M. and Campos-Ortega, J. A . (1994) Lateral inhibition mediated by the Drosophila neurogenic gene Delta is activated by proteins encoded by proneural genes. (submitted) Lehmann, R., Dietrich, U., Jimenez, F. and Campos-Ortega, J. A . (1981). Mutations of early neurogenesis in Drosophila. Roux’s Arch. Dev. Biol. 190, 226229. Lehmann, R., Jimenez. F.. Dietrich. U . and Campos-Ortega, J . A . (1983). On the phenotype and development of mutants of early neurogenesis in Drosophila melanogaster. Roux’s Arch. Dev. Biol. 192, 62-74. Lieber, T., Alcamo, E.. Hassel, B . , Krane, J. F., Campos-Ortega. J . A . and Young, M. W. (1992). Single amino acid substitutions in EGF-like elements of the Notch and Delta proteins modify Drosophila development and depress cell adhesion in vitro. Neuron 9, 847-859. Luer, K. and Technau, G. M. (1992). Primary culture of single ectodermal precursors of Drosophila reveals a dorsoventral prepattern of intrinsic neurogenic and epidermogenic capabilities at the early gastrula stage. Development 116, 377-385. Martin-Bermudo, M. D . , Martinez, C . and Jimenez. F. (1991). Distribution and function of the lethal of scute gene product during early neurogenesis in Drosophila. Development 113, 445454. Muller. H. J. (1935). The origination of chromatin deficiencies as minute deletions subject to insertion elsewhere. Genetica. 17, 237-252.

NEUROGENESIS IN DROSOPHILA

101

Murre, C., Schonleber McCaw, P. and Baltimore, D. (1989a). The amphipathic helix-loop-helix: a new DNA-binding and dimerization motif in immunoglobulin enhancer binding, daughterless, MyoD and myc proteins. Cell 56, 777-783. Murre, C., Sehonleber McCaw, P . , Vaessin, H., Caudy, M., Jan, L. Y., Jan, Y. N . , Cabrera, C. V., Buskin, J. N., Hauschka. S. D., Lassar, A. B., Weintraub, H. and Baltimore, D. (1989b). Interactions between heterologous helix-loop-helix proteins generate complexes that bind specifically to a common DNA sequence. Cell 58, 537-544. Oellers, N . , Dehio, M. and Knust, E. (1994). Basic-helix-loop-helix proteins of the Enhancer of split complex of Drosophila negatively interfere with transcriptional activation by proneural proteins. Mol. Gen. Genet. (in press). O’Kane, C. and Gehring, W. J. (1987) Detection in situ of genomic regulatory elements in Drosophila. Proc. Natl. Acad. Sci. USA. 84, 9123-9127. Poulson, D. F. (1937). Chromosomal deficiencies and embryonic development of Drosophila melunogaster. Proc. Natl. Acad. Sci. USA 23. 133-137. Poulson, D. F. (1950). Histogenesis, organogenesis and differentiation in the embryo of Drosophilu melunogaster. In “Biology of Drosophila” (Ed. M. Demerec), pp. 268-274. Preiss, A., Hartley, D. A. and Artavanis-Tsakonas, S. (1988). The molecular genetics of Enhancer of split, a gene required for embyonic neural development in Drosophilu. EMBO J . 7,3917-3928. Rao, Y . , Jan, L. Y. and Jan, Y. N. (1990). Similarity of the product of the Drosophilu neurogenic gene big brain to transmembrane channel proteins. Nature 345, 163-167. Rebay, I . , Fleming, R. J . , Fehon, R. G., Cherbas, L., Cherbas, P. and ArtavanisTsakonas, S. (1991). Specific repeats of Notch mediate interactions with Delta and Serrate: implications for Notch as a multifunctional receptor. Cell 67. 687-699. Rodriguez, I., Hernindez, R., Modolell, J . and Ruiz-Gomez, M. (1990). Competence to develop sensory organs is temporally and spatially regulated in Drosophila epidermal primordia. EMBO J . 9, 3583-3592. Romani, S.. Campuzano. S. and Modolell, J . (1987). The uchaete-scute complex is expressed in neurogenic regions of Drosophila embryos. EMBO J . 6, 2085-2092. Romani, S . , Campuzano. S . , Macagno, E. R. and Modolell, J. (1989). Expression o f uchaete and scute genes in Drosophila imaginal discs and their function in sensory organ development. Genes Dev. 3, 997-1007. Ruiz-Gomez, M. and Ghysen, A. (1993). The expression and role of a proneural gene, achaete, in the development of the larval nervous system of Drosophila. E M B O J . 12, 1121-1 130. Schrons, H.. Knust. E. and Campos-Ortega. J. A. (1992) The Enhancer of split complex and adjacent genes in the 96F region of Drosophilu melunogaster are required for segregation of neural and epidermal progenitor cells. Genetics 132, 481-503. Shepard, S. B., Broverman. S. A. and Muskavitch, M. A. T. (1989). A tripartite interaction among alleles of Notch, Delta and Enhancer of split during imaginal development of Drosophila melanogaster. Genetics 122, 429-438. Simpson, P. (1990). Lateral inhibition and the development of the sensory bristles of the adult peripheral nervous system of Drosophila. Development 109, 509-519. Simpson, P. and Carteret, C. (1989). A study of shaggy reveals spatial domains of expression of achaete-scute alleles on the thorax of Drosophila. Development 106. 57-66. Skeath. J . B. and Carroll, S. B. (1991). Regulation of achuete-scute gene expression

JOSE A. CAMPOS-ORTEGA

102

and sensory organ pattern formation in the Drosophilri wing. Get1t.s Dev. 5 , 9 x 4 9 9 5.

Skeath, J . B. and Carroll. S. B. (1992). Regulation of proneural gene expression and cell fate during neuroblast segregation in the Drosophila embryo. Developmenr 114. 930-946. Skeath. J . B., Panganiban. G.. Selcgue, J . and Carroll. S. B. (1992). Gene regulation in two dimensions: the proneural rickric.re and sc'iite genes are controlled by combinations (if axis-patterning genes through a common intergenic control region. Genes D ~ L6, , . 260&2619. Smoller. D.. Friedel. C.. Schmid. A., Bettler. D.. Lam. L. and Yedvobnich, B. ( 1990). The Drosopfzila neurogenic gene musterrnind encodes a nuclear protein unusually rich in amino acid homopolymers. Genes & Dew/. 4, 1688-1 700. Struhl. G . , Fitzgerald. K. and Greenwald, I. (1993). Intrinsic activity of the Lin-iz and Notch intracellular domains in vivo. Cell 74. 331-345. Stiitten. I. and Camps-Ortega, J . A. (1901).Cell commitment and cell interactions in the ectoderm of Drosophilri melrrnoguster. Development. Suppl. 2, 3946. 7'aghert. P. H.. Doe. C. 0 . and Goodman. C. S. (1984).Cell determination and reguhtion during development of neuroblasts and neurones in grasshopper embryos. Nritiire 307, 163-165. Technau. G. M . (1986). Lineage analysis of transplanted individual cells in embryos of Drosophilci tnelririogasrcr. I. The method. Roiix's Arch. Dev. Biol. 19.5, 389-398. Technau, G . M . (1087).A single cell approach to problems of cell lineage and commitment during embryogenesis of Drosoplzila melunogaster. Development 100, 1-12. Technau. G . M . and Campos-Ortega. J . A. (1985). Fate mapping in wildtype Drosophilrr melutiogrisfer, I1 Injections of horseradish peroxidase in cells of the early gastrula stage. Roiix's Arch. D P V .Riol. 194, 196212. Technau. G. M . and Campos-Ortega, J . A. (IYX6a).Lineage analysis of transplanted individual cells in embryos of Drosophilu melanogusrer. 11. Commitment and proliferative capabilities of neural and epidermal cell progenitors. Roux's Arch. Dev. Biol. 195. 445454. Technau. G. M. and Campos-Ortega. J . A. (1986b).Lineage analysis of transplanted individual cclls in embryos of Drosophilci melrinogcister. 111. Commitment and proliferative capabilities of pole cells and midgut progenitors. Roux's Arch. Dev.

Biol. 19.5. 389498. Technau, G . M . and Camps-Ortega. J . A . (1987). Cell autonomy of expression of neurogenic genes of Drosophilu rtieluriogrister. Proc. Nritl. A u I ~ .Sci. U S A , 84, 45003504. Technau. G . M.. Becker, T. and Canipos-Ortega, J . A . (1988). Reversible commitment of neural and epidermal progenitor cells during embryogenesis of Drosophilu melritioguster. Roiix's Arch. Dev. B i d . , 197. 413-418. Thomas. J . B . . Bastiani, M . J . , Bate, M. and Goodman. C . (1984). From grasshopper to Drosophilu: ii common plan for neuronal development. Nature 310, 203-207. Tietze, K., Oellers, N . and Knust. E. (1992) Enhancer of .split'). ;I dominant mutation of Drosophila and its use in the study o f functional domains of a helix-loop-helix protein. Proc. Narl. A c d . Sci. U S A 89. 61524156. Van Doren, M . , Powell, P. A., Pasternak, D . , Singson, A. and Posakony, J . W . ( 1992). Spatial regulation of proneural gene activity: auto and cross-activation of rdiciere is antagonized by extr(It7iucrocharrie. Genes Dev. 6, 2592-2605. Viissin, H.. Vielmetter. J . and Campos-Ortcgz. J . A . (1985). Genetic interactions in early ne u roge ncs is o f Drosophilti tnelrinogri.srer. J . Neii rogcn . 2. 29 1-308.

NEUROGENESIS IN DROSOPHlLA

103

Vassin, H., Bremer. K. A.. Knust. E. and Campos-Ortega. J. A. (1087). The neurogenic locus Deltrr of Ilrosophilu nirlritiogmster is expressed in neurogenic territories and encodes ii putative transmembrane protein with EGF-like repeats. EMBO J. 6. 343 1-3440. Vassin. H.. Brand. M.. Jan. L. and Jan. Y. N . (1094). daughterless is essential for neuronal precursor differentiation but not for initiation of neuronal precursor formation in Drosophila embryos. Development (in press). Villares, R. and Cabrera. C. V . (1087). The achuere-scirfe gene complex of Drosophila melrrriogastcr: conservcd domains in a subset of genes required for neurogenesis and their homology to tnyc. Cdl 50. 415424. Wharton. K. A . , Johansen. K . M . , Xu. T. and Artavanis-Tsakonas. S. (lY85). Nucleotide sequence from the neurogenic locus Notch implies a gene product that shares homology with proteins containing EGF-like repeats. Cell 43, 567-581. White, K. ( 1980). Defective neural development i n Drosophila melanogusfer embryos deficient for the tip of the X-chromosome. Dev. B i d . 80. 322-344. White. K . , DeCelles. N . L. and Enlow. T. C. (1083). Genetic and developmental analysis of the locus t w r l in Drosophilri niel(rnogaster. Genetics 104. 433488. Xu, T., Rebay. 1.. Fleming. R. J.. Scottgale, T. N . . Artavanis-Tsakonas. S. (1090). The Notch locus and the genetic circuitry involved in early Drosophila neurogenesis. Genes Dev. 4. 4641175. Yochem, J. and Byers. B . (1087). Structural comparison of the yeast cell dimension cycle gene CDC4 and a related pseudogene J. Mol. B i d . 195. 233-234. Ziemer. A.. Tietze, K . , Knust, E . und Campos-Ortega, J. A. (1988). Genetic analysis o f Enhancer qf’ .split. a locus involved in neurogenesis in Drosophila melanogaster. Genetics 119, 63-74.

This Page Intentionally Left Blank

Molecular Biology of the Honeybee Robin F.A. Moritz lnstitut fur Biologie, Technische Universitat Berlin, Franklinstr 28/29, 10587 Berlin, Germany

1 Introduction 105 2 Genes and sequences

107 2.1 Nuclear genes 107 3 Mitochondrial genome 114 3.1 Mitochondrial genes 116 3.2 Non-coding sequences and length variation 117 4 Gene activity in embryonic development 124 5 Population variability 125 5.1 Nuclear DNA markers 125 5.2 Mitochondrial DNA markers 129 6 Molecular evolution and biogeography 130 6.1 Molecular phylogeny of bees 130 6.2 Genetic variability among honeybee species 131 6.3 Molecular variability within species 133 7 Outlook 140 Acknowledgements 141 References 144

1 Introduction

The biological rules that govern the honeybee colony have fascinated scientists since Aristotle. The bustling activities of the thousands of workers are seemingly at random and yet in the end a unified and apparently coordinated system is achieved. The honeybee riddle of chaos on the one hand and yet pattern on the other was to occupy generations of scientists after Aristotle. It took a long time to grab some pieces of the great puzzle and understand the basics of communication in honeybee colonies. It was the great pioneers in the field of bee research who unravelled the mystery of identifying the dance language of the worker bee (von Frisch, 1965) and of queen control through pheromones (Butler, 1973). In the second half of this ADVANCES IN INSECI'PHYSIOLOGY VOL. 25 ISBN lL1?4l?1??5-7

106

R. F. A. MORITZ

century bee research was dominated by studies focusing o n behavioural ecology and communication under the influence of Karl von Frisch's and Martin Lindauer's epochal work on honeybees. However. the focus on behavioural ecology was not always so strong. and in fact honeybees also entered the arena of modern research as a genetic model system. Mendel conducted experiments with bees but unfortunately he failed to achieve controlled matings in netted cages and could not produce any F2 offspring which he required to confirm his theories based on the pea experiments (Johannson, 1980). Real breakthroughs were achieved by Dzierzon (1845) who found that drones are parthenogenetically produced. Petrunkewitsch (1903) and Nachtsheim (1912, 1914) used honeybees in their studies to elucidate the mode of fertilization and chromosome duplication. Honeybees were among the prime genetic model systems used until Castle and Morgan introduced Drosoplzila to the field of genetics in the early 20th century. Yet after the wide acceptance of fruit flies as the genetic test organism, honeybee genetics became a very specialized field and even the introduction of artificial insemination by Laidlaw (1944) and Mackensen (1947) could not stop the decline of honeybee genetics. The honeybee system was difficult to handle, had a slow generation cycle and an annoyingly large number of chromosomes ( n = 16) which made linkage studies tedious. Controlled matings could only be achieved through artificial insemination, and the generation cycle was too long to generate a rapid series of high-quality publications, an important basis for successful research and the recruiting of funds. Maintenance costs were high and other organisms offered better, swifter, and less expensive ways to solve pending questions. One of the last strongholds of honeybee genetics was in behavioural genetics, but after the work of Rothenbuhler (1964) on the genetics of hygienic behaviour of worker honeybees, a depressing silence also broke out in international scientific journals in this field. Clearly. the molecular revolution of genetics took place leaving the honeybees aside. The molecular genetics of Drosophila melanogaster boomed while bee geneticists felt more attracted to applied breeding research. Only recently, after a surprisingly long abstinence of several decades, are geneticists slowly re-entering the scene in honeybee research, trying to catch up with modern techniques and addressing new and (more important) unique questions that can only be solved by using honeybees as a test system. Although honeybees as an animal model for basic genetics have their pitfalls, they are far from being useless in basic genetical research. In fact, they have several unique features that make them more attractive than any other genetical test system available in higher organisms. Particularly important are the following traits for genetical research: 1. Male haploidy allows the study of gene expression in haplotypes. This is important for both selection experiments and gene mapping studies.

MOLECULAR BIOLOGY OF THE HONEYBEE

107

2. The rich behavioural repertoire and the social organization makes it a primc system in behavioural genetics. 3. The slow embryonic development offers plenty of opportunities to study gene regulation and expression during early development. These are just a few characteristics which make the honeybee profitable for basic genetical research. Since honeybees are of significant economical and ecological importance, there is obviously also a compelling need to understand their population and breeding genetics. Here also molecular techniques are aiding us in understanding processes in natural and artificial selection and the undcrlying genetic mechanisms. Today the stage is set for rapid progress in molecular honeybee genetics, since we can capitalize on the very detailed knowledgc obtained from Drosophila. Cloned DNA probes are available which enable us to rapidly map the honeybee genome; sequenced genes allow us a better understanding of gene regulation and evolution. In addition, mitochondria1 DNA is used in population genetic studies revealing the dynamics of natural selection in feral honeybee populations. Thus, molecular genetics of honeybees can both incrcase our understanding of basic genetic mechanisms and improve our knowledge of honeybee specific genetic problems.

2 Genes and sequences

2.1

NUCLEAR GENES

2.1.1 Elongation factor I The search for genes in honeybees has been limited and only a few genes have been isolated and sequenced so far. The most productive way of isolating genes has been through hybridization of honeybee DNA to known Drosophila genes. Walldorf and Hovemann (1990) studied a DNA coding for one of the three proteins ( a , p , y ) of the cytoplasmic elongation factor 1 (EF-1). EF-la catalyses the transport of the aminoacyl-tRNA to the 80s ribosome. In Drosophila melanogaster two independent genes (F1 and F2) code for EF-la (Hovemann et al., 1988). Walldorf and Hovemann (1990) found the EF-la gene of honeybees to be closely related to the corresponding coding region in Drosophila melanogaster. They cloned two fragments of 1.0kb and 1.1 kb. respectively, which revealed as much as 77% sequence homology to the EF-la F1 Drosophila reading frame. The same degree of homology was found for another elongation factor gene (EF-lcu, F2). Through this high sequence conservation Walldorf and Hovemann (1990) were able to locate the translation start and stop sequence in the gene (nucleotides 365 and 2121, respectively) and they also found two introns

108

R. F. A. MORlT o p e n reading f r a m e

intron

FIG. 1 Putative structure of the E F - l a gene in Apis melliferu. The base pair numbering is adopted from Walldorf and Hovernann (1990).

similar to the Drosophila EF-la F2gene (Fig. 1). Walldorf and Hovemann (1990) suggested that the Apis EF-la gene evolved from a common ancestral =-type gene rather than the F1 gene which is free of introns. The isolated EF-la sequence of Apis codes for 461 amino acids with a calculated mass of 50.5 kDa. It is unclear whether Apis has more than one EF-la gene. In a southern hybridization of the Drosophila probe on an EcoRI digest of Apis mellifera DNA, Walldorf and Hovemann (1990) found a weak second signal at about 9.4 kb and they could not exclude the possibility of a second EF-la gene in honeybees. The high homologies between Drosophila and Apis are not necessarily surprising since the elongation factor is known to be encoded in an extremely conservative gene region with only little variance among different taxa. Only two more variable amino acid sites appear, at positions 186 and 315, respectively, by adding the Apis sequence to the already known sequence variability among eight different species (Walldorf and Hovemann, 1990). 2.1.2 Segmentation genes Fleig et a/. (1988) studied the homologies of several homeobox genes of Drosophila rnelanogaster and Apis rnellifera. They constructed a genomic library of Apis mellifera and identified a homeobox DNA sequence in a cloned 500 bp Cla-Sull fragment of Apis rnellifera showing 82% homology to the Drosophila gene Dfd, and coding for the identical amino acid sequence. They termed the honeybee gene H42 and found additional homologies that extended beyond the 5' and 3' ends of the box. They also argued that the position of the intron may be at an identical site in both honeybees and fruit flies. In further work, Walldorf et a/. (1989) compared other homeobox genes of Apis rnellifera and Drosophila rnelanogusfer in more detail. Drosophila DNA probes containing Antennupedia ( A n f p ) ,fushi tarazu ( f t z ) , sex combs

109

MOLECULAR BIOLOGY OF THE HONEYBEE

TABLE 1 Sequenced homeobox genes in Apis mellifera

Drosophila probe abdominal-A Antennapedia Deformed fushi tarazu engrailed" invected"

muscle segment homeobox Sex combs reduced W-13

Honeybee clone

% amino acid similarity in the homeobox

E30 E6O E30 E60 H17

96.7 98.3 100 not detected 96.7 91.7 81.7 83.3 96.7 (with 5.4 kb intron)

H55 H40

98.3 54.2

H15 H90 H42 -

~~

T h e sequence similarity between eti and inv is smaller between than within species. It is therefore not possible to assign E30 and E60 to either gene.

reduced (Scr) and Deformed ( D f d ) hybridized to EcoRl digests of total genomic DNA of Apis rneffifera (Table 1). Although frz did yield hybridization signals in genomic honeybee DNA (Fleig et al., 1988), the authors were unable to detect a homologous conserved sequence in the honeybee. Screening their library with f t z , the only clones that were reisolated were those that had already been isolated with the Antp probe. Walldorf et al. (1989) concluded that, given the honeybee has an ftz homologous gene, it must have considerably diverged from the Drosophila gene. Alternatively, they suggest that the gene is completely lacking and other loci perform the ftz function. The genes of the engrailed class (en = engrailed and inv = invected) appeared to be conservative. Two clones, ,560 and E30, were highly homologous, yet it was not possible to assign the two Drosophila genes to the two honeybee clones because sequence divergence between the two genes within the species was less than between the species. The homology of the homeobox in clone H40 showed only 55.2% similarity to the Drosophila W-13 gene. No homeobox has been found in Drosophila with strong homologies to H40. Sommer et al. (1992) studied segmentation genes with the Cys2-His2 zinc-finger DNA binding motif. They found evolutionary conserved patterns between Drosophila and Apis for hunchback, Kriippel, and snail (Fig. 2). Particularly, Kriippel showed high similarities between both species with a 94.6% amino acid homology, and the authors suggest that it might be duplicated in the honeybee. It appeared that those amino acids which are believed to have a direct contact with the bases in the DNA are completely conserved. The escargot sequence from Drosophila had a higher homology

110

R. F. A. MORlTi

hunchback

Kiiippel

snail

FIG. 2 Sequence alignment of finger fragments obtained from the genes of Drosophilu (top) and Apis (bottom). Identical amino acids are denoted with a dash, a deletion with a dot. The stars indicate those amino acids that are believed to be in direct contact with the bases in the DNA. For the snail gene (c) also the escargot sequence is plotted to reveal the sequence similarity to the Apis sequence.

to thc Apis fragment (91.4%) than had the paralog snail sequence (75.6%) which was initially used to isolate the Apis DNA (Sommer et ul., 1992).

2.1.3 Genes coding for honeybee venom compounds There are also few studies focusing on honeybee specific genes. These mainly include DNA regions coding for enzymes and peptides of bee venom (Table 2 ) which have been analysed because their composition is known from various detailed biochemical studies (Habermann and Jentsch, 1967; Shipolini er al., 1971; Bachmeyer et al., 1972; Habermann, 1972; Gauldie et ul., 1976, 1978; Suchanek et al., 1978; Kreil et al., 1980). Melittin is the main lytic peptide of the honeybee venom and is found in both queens and workers. Like most peptides it is initially synthesized as a larger precursor which is then proteolytically cleaved to the final product (Kreil, 1990). Typical of the melittin precursor is the pro-region which is composed of an amino acid sequence in which every other position is either alanine or

MOLECULAR BIOLOGY OF THE HONEYBEE

111

TABLE 2 Composition ot t h e worker honeybee venom in 54 dry weight Enzymes Phospholipase A? Hyaluronidase Acid Phosphatase

l(klS% 2 (7,

Peptides Mclittin A pa mi n MCD-peptidc Secapin T e rt i a pi n

45-605? 2-3 %' 2 %' 1%

<1%

< 1%

proline. This region is removed by a stepwise cleavage of dipeptides through dipeptidylaminopeptidase (DPAP; Kreil et al., 1980). The end product is thus produced by removing 11 dipeptides of the sequence X-Ala or X-Pro, with X being alanine, glutamine or asparagine, by DPAP activity. Kreil et al. (1980) found that the end product, melittin, is only synthesized after the precursor has entered the venom sac. Vlasak et al. (1983) isolated mRNA from venom glands of queen bees and were able to synthesize a cDNA coding for prepromelittin; which is a transient precursor of melittin (Suchanek et ul., 1978). A clone with a 374 bp insert proved to contain an open reading frame which corresponded exactly to the prepromelittin amino acid sequence as determined by biochemical peptide analysis (Fig. 3 ) . Vlasak and Kreil (1984) isolated and sequenced a cDNA that coded for preprosecapin, which is a precursor for secapin, a minor compound in the venom glands of workers with an as-yet unknown function (Gauldie et al.,

ATG AAA 'ITC TTA GTC AAC GTT GCC CIT GTT TIT ATG GTC GTG TAC ATT TCT TAC ATC Met -Lys- Phe - Leu - Val -Asn - Val - M a -Leu - Val - Phe -Met - Val - Val - Tyr -1le -Ser-Tyr- IleTAT GCG GCC CCT GAA CCG GCA CCA GAG CCA GAG GCG GAG GCA GAC GCG GAG GCA Tyr - A l a - Ala -P ro - G l u - P r o - A l a - P r o - G l u - P r o - G l u - Ala - G b - A l a - Asp- Ala-Glu - A h

0 GAT CCG GAA GCG GGA ATT GGA GC4 GTT CTG AAG GTA ?TA ACC ACA GGA TTG CCC Asp - Pro - Glu - Ala - Gly - Ile - Glv - Ala - Val - Leu - Lvs - Val -Leu - Thr - Thr - Glv -Leu - PrQ

GCC CTC ATA AGT TGG A'IT AAA CGT AAG AGG CAA CAG GGT TAG.. . . Ala - Leu - Ile - Ser - Tro - Ile - Lvs - Are - Lvs - ATP - Gln - G h - Gly

***

FIG. 3 DNA and amino acid sequence of prepromelittin and melittin (underlined). The putative peptide bond which is hydrolysed by the signal peptidase is marked by an arrow.

R. F. A. MORITZ

112

ATG AAG AAC TAT TCA AAA AAT GCA ACA CAC 'ITA ATT ACG GTT (7IT CTA I T C AGC TlT Met - Lys - Asn - Tyr - Ser - Lys - Asn- Ala- Thr - His - Leu - Ile -Thr - Val - Leu - Leu - Phe - Ser - Phe,

GTT G'IT ATA CIT I T A ATT ATT CCA TCA AAA TGT GAA GCC G'IT AGC AAT GAT AGG Val - Val - Ile - Leu - Leu - Ile - Ile - Pro- Ser - Lys- Cys- Glu -Ala - Val - Ser - Asn -Asp - Arg-

0 CAA CCA 'ITG GAA GCA CGA TCT GCT GAT I T A GTC CCG GAA CCA AGA TAC ATT A l T Gln-Pro - L e u - G l u - A l a - Arg-Ser-Ala-Asp-Leu-Val -Pro - G l u - P r o - Arg-Tvr -1le - Ile GAT G'IT CCT CCT AGA TGT CCT CCA GGT TCT AAA TTC A'IT AAG AAC AGA TGT AGA ASR- Val - Pro - Pro - Ara - Cvs - Pro - Pro - Glv - Ser - Lvs - Phe - Ile - Lvs - Asn - A r p - Cvs - Are GTC ATA GTG CCT TAA Val - Ile - Val -Pro- **rt

.. . .

FIG. 4 DNA and amino acid sequence of preprosecapin and secapin (underlined). The putative end of the signal protein is marked by an arrow.

1976). Vlasak et al. (1986) argue that there is caste-specific variation for this compound, because the venom of queen honeybees has a higher secapin content than worker venom. The sequenced secapin precursor (Fig. 4) corresponds almost exactly to the secapin amino acid sequence as derived from biochemical analysis (Gauldie et al., 1978). A gene coding for the enzyme phospholipase A2 has been sequenced by Kuchler et al. (1989). Phospholipase A2 is the principal allergen and a major compound (up to 15% of the dry mass) of worker venom. Surprisingly this enzyme is caste specific and almost lacking in queen bee venom (Marz et al., 1981). Using polyclonal antibodies, Kuchler et al. (1989) prepared a cDNA coding for phospholipase A2 from honeybee venom glands. The isolated clone had a 540 bp insert with a deduced amino acid sequence that showed substantial homologies to phospholipase A2 from lizards and bovine pancreas. The similarities are particularly obvious in a three-dimensional analysis of the molecule structure. The critical residues that are essential for the catalytic activity are conserved in both the bovine and the honeybee venom enzyme. Using Northern blot technology, Marz et al. (1981) were able to confirm the apparent lack of an mRNA coding for phospholipase A2 in queens: they only found an mRNA of about 850 bp in worker bee venom RNA extracts but not in queen RNA. In a Southern blot with total honeybee DNA, the cloned cDNA probe hybridized to a single 1600 bp fragment and the authors concluded that only a single, rather small gene codes for phospholipase in the honeybee genome. Finally a DNA region coding for the enzyme hyaluronidase, which has been described as a 'spreading factor' (Habermann, 1972) to facilitate the rapid diffusion of the other venom compounds, has been sequenced by Gmachl and Kreil (1993).

MOLECULAR BIOLOGY OF THE HONEYBEE

113

2.1.4 In situ hybridization Although several genes are now known and have been sequenced from the honeybee genome, we are still lacking any information on a genetical map. The production of a classical linkage map of the honeybee has proven difficult in the past. Only very few linkage groups are known (Tucker, 1986; Del Lama et al., 1985, 1993), and we have no information on which chromosomes these linkage groups are located. This is mainly due to the large number of chromosomes but also to the difficulties in identifying each of the 16 chromosomes with classical G or C banding techniques (Hoshiba and Kusangi, 1978; Hoshiba and Okada, 1986). Chromosome identification is primarily done by size, but since the size differences are small, it is usualiy only possible to identify the largest chromosome, numbered 1. A first step towards a better characterization of the chromosomal set can be achieved through in situ hybridization. Beye and Moritz (1993) recently adopted the fluorescence in situ hybridization technique (FISH) to the honeybee system. A DNA probe of Drosophila containing the repeat coding for the rRNA subunits 28S, 18S, 5.8s and 2s (Tautz et af., 1988), including the intergenic sequences, was used for hybridization to honeybee chromosomes prepared from larval drone testes. The tandem repeat gene occurs at about 250 copies on the X chromosome and 200 copies on the Y chromosome of Drosophila. Similar tandem repeat gene structures have been found in many other organisms (Spear, 1974). In the honeybee this probe hybridized at the telomeric position of two chromosomes (Fig. 5 ) . In subsequent studies (Beye and Moritz, 1994) probes with repetitive DNA cloned from honeybee

FIG. 5 In situ hybridization of an rDNA D. melunoguster probe to a haploid set of metaphase chromosomes of A. rnellifera (courtesy of M. Beye).

114

R. F. A. MORITZ

FIG. 6 In situ hybridization of a repetitive DNA probe of A . meilifera to a haploid set of metaphase chromosomes isolated from drone testes. The probe is centromere specific and the two chromatids are clcarly visible (Beye and Moritz. 1994).

genomic DNA revealed chromosome-specific patterns and could be used to unambiguously identify the chromosomal set. Another probe proved to be ccntromere specific for 14 chromosomes (Fig. 6).

3 Mitochondria1 genome Thc mitochondria1 genome of the honeybee is much better known than its nuclear counterpart. With a length of 16300-17000 base pairs it has been shown to contain all those genes we know from Drosophila yakuba, currently the only other insect for which the complete sequence of mitochondria1 (mt) DNA is known (Clary and Wolstenholme, 1985). The study of honeybee mtDNA genes started with the work of Vlasak et al. (1987), who sequenced the region coding for the large ribosomal RNA. Crozier et al. (1989) sequenced a region of 2950 bp in which they recognized genes coding for four tRNAs (tryptophan, Icucine, aspartate and lysine) and two genes coding for two cytochrome oxidasc subunits (CO-1. CO-11). Cornuet and Garnery (1991) identitied another tRNA coding for tyrosine at

115

MOLECULAR BIOLOGY OF THE HONEYBEE

FIG. 7 Map and gene order o f the circular mitochondrial genome of A . rnel/iferu (modified and redrawn from Crozier and Crozier, 1992) and D . yakubu (modified and redrawn from CIary and Wolstenholme, 1085). The dotted areas represent non-coding A +T-rich regions. The Bcll sites are given for A . rnellifera. The asterisks indicate differences in the codons of DNA sequences coding for tRNAs (hatched areas).

the beginning of this 2950 bp region. from position 56 to 124, being transcribed, however, in the opposite direction. Furthermore, they detected another tRNA-like sequence, with a GGG anticodon between the tRNA"y and tRNAtrPgenes. Meanwhile the complete sequence of mtDNA of Apis meflijera ligustica has been analysed by Crozier and Crozier (1992) yielding precise information on the genes located on the honeybee mtDNA. Much of the following section is based on their results (the sequence can be accessed through Genbank #L 06178). A map of the Apis mellifera mitochondrial genome is shown in Fig. 7 and compared with that of Drosophila yakuba, for which the mtDNA has also been sequenced in full (Clary and Wolstenholme, 1985). The similarities are striking with an almost identical gene order of the 37genes in both species. The high A + T content is typical for both mtDNAs. For example the CO-I region is composed of 75% and 80% A T in D. yakuba and A. rnellifera, respectively (Crozier et a[., 1989). A total of 84.9% of the entire mitochondrial genome is either A or T. In spite of these strong similarities there are some significant differences that need special consideration. For example guanine is the rarest nucleotide in A . mellifera (5.5%) whereas in D. yakuba cytosine is the least frequent. In its low guanine content A. melliferu mtDNA resembles that of the echinoderm Paracentrotus lividus (Crozier and Crozier, 1992). Also the total size of A. mellifera mtDNA can be substantially longer than that of D. yakuba, as will be seen later in a discussion of the non-coding regions in the honeybee mtDNA.

+

116

3.1

R.

F. A. MORlT2

MITOCHONDRIAL GENES

Although the genes encoding proteins, the control region and the rRNAs are in the same order as in D.yakuba, 11 of the 22 tRNA coding regions are in different positions. These position shifts are all located between the A+T-rich region and the ATPase8 gene (upper left quadrant in Fig. 7). The overlap of the ND2 and tRNACy" genes does not occur in D. yakuba, whereas the other overlap between ATPase8 and ATPase6 is found in both species. Crozier and Crozier (1992) suggest that the CO-I and CO-I1 genes do not overlap although no complete T A A stop codon is found. Citing Ojala et al. (1981), they argue that mRNAs from mtDNA can be polyadenylated at the 3' end, in this way completing the otherwise incomplete termination codon. In fact this phenomenon seems to be very common: Wolstenholme (1992) reports such incomplete termination codons as a general feature of those animal systems for which the complete mtDNA sequence is available. An exception is the cnidarian Metridiurn senile which Wolstenholme (1992) interprets as an indication for an early origin of the cleavage-polyadenylation mechanism. Initiation codons either code for methionine (three ATG and three ATA) or isoleucine (one ATC and six ATT) and no anomalous codons, like ATAA in D. yakuba (Clary and Wolstenholme, 1985), were found. All but the two stop codons for CO-I and CO-I1 (a single T) are TAA Surprisingly the A+T-rich region seems to lack the typical structure for the initiation of replication as found in vertebrates. Another A+T-rich duplicate region is found between the tRNA'"" and the CO-I1 region (Cornuet et af., 1991). However, it is as yet unclear whether this region can have replicate activity (see below), because some species and races of Apis seem to lack this duplication. The tRNA genes have striking similarities to the sequence of D. yakuba (Clary and Wolstenholme, 1985). The anticodons are identical but for the tRNALayS (TIT instead of CTT) and tRNAS" (TCT instead of GCT). Crozier and Crozier (1992) inferred the secondary structure of all tRNAs, and found a high A + T content (87.1%) which is higher than for the entire mitochondria1 genome (84.9%). Furthermore, they found various cases of base pair mismatch which has been shown to be a common phenomenon also in other insects (Clary et af., 1982, 1984; Clary and Wolstenholme, 1983a,b, 1984; DeBruijn, 1983; HsuChen and Dubin, 1984; Uhlenbusch et af., 1987). Four tRNAs coding for asparagine, leucine, threonine, and valine are characterized by an unusually long TUrCG stem with six nucleotide pairs (Fig. 8). Moreover, the TUrCG loop of the tRNATh' is surprisingly long (10 nucleotides). Typical, however, is the lack of the DHU arm in the tRNASer which is very likely to be universal (Wolstenholme, 1992). The genes coding for the large and small ribosomal unit and the 13 protein coding genes are listed in Table 3. Again we find strong similarities

717

MOLECULAR BIOLOGY OF THE HONEYBEE

TABLE 3 The protein coding genes and the large and short rRNA coding regions of the Apis rnellifrra mitochondria1 genome. Nucleotide composition, sequence similarity to Drcxophila yakuba, and number of nucleotides Gene ATPase8 ATPase6 Cytochrome oxidase I Cytochrome oxidase I1 Cytochrome oxidase I11 Cytochrome b NADH dehydrogenase 1 NADH dehydrogenase 2 NADH dehydrogenase 3 NADH dehydrogenase 4 NADH dehydrogenase 4L NADH dehydrogenase 5 NADH dehydrogenase 6 Small rRNA Large rRNA

3' 6 A + T

% similarity to D. yakuba

Number of nucleotides

90 85 84 81 84 69 86 87 87 94 86 87 87 81 85

64 62 74 68 66 66 64 51 63 64 57 61 53 68 73"

156 678 1560 675 777 1149 915 999 35 1 1341 26 1 1662 501 786 1371

Data from Crozier and Crozier (1992). 'The similarity is slightly higher than given by Crozier and Crozier (1992) after improving the alignment of the D.yrkuba and A . meNiferu sequence.

to the D. yakuba genome. Similarities between D . yakuba and A . mellifera range from 51% to 75%. The G+C/A+T ratio is much higher in D. yakuba (0.43; Jukes and Bhushan, 1986) than in A . mellifera (0.18; Crozier and Crozier, 1992). This strong A + T bias in the honeybee mtDNA genome causes some differences in amino acid usage compared with Drosophila. For example alanine in Drusophila is replaced by serine in the honeybee at 43 sites (conserved at 42 sites). The most conserved amino acid appears to be arginine with 95% of the honeybee sites being conserved in Drosophila and 63% of the Drusophila sites being conserved in the honeybee.

3.2

NON-CODING S E Q U E N C ~ SA N D LENGTH VARIATION

Smith and Brown (1990) were the first to report on length variation and detected six locations with variable base pair numbers. Two sites showed small variability (20 bp) whereas the other four sites varied by about 100 bp. The small size variability is located within the C O - K O - I 1 region and can be easily detected o n Bcll digests. Cornuet and Garnery (1991) provided a plausible explanation for this length variation. Exactly 20 bp upstream of the Bcll site e (see Fig. 7) the following motif can be found: ..TGACCA ... With

A A

A A A A A A T

T T T T T T A

T A T A A A T T G A T A A C A T A

T T A G T T T G

A

T T A A A T A A T T T A A

A A T C A

T

C T

T A A

A A

T

G-T T A T A A T T A T A T A T T A T A A A A T T A A T A T A C G T T A A T T T A C T T Q T G C A A T T T A T A T A A T A T A T T A T G T A G

C

t RNA ASP

t RNA Leu Ja

7

D

A A A A T T T

A A T G C C G

A A A A A T T T T G T A A A C A T A A

T

T A AA

A T TT

A T T A T A T A A T A A T T T A T

T

T

C

A A

T A

~

~

A G

A A T A T A A T A T T A T A T A T G T

T TA

A A T T G T A A ? A

T T

T T T C T

A T T T T A A A G A T A T T T T T T A A A A A A T T A A G A A A G T G

A T A C

t RNAThr FIG. 8 Cloverleaf structures of four tRNAs with 6 bp T q CG stems. The T q C G loop of the tRNAThr(bottom left) is exceptionally long.

t RNAVal

I 0

z rn

<

m rn rn

120

R. F. A. MORlTZ

just a single transition at the third position of the GAC codon (coding for leucine) to G A T (also coding for leucine) to ..TGATCA ... a new Bcfl site is obtained. Combined with the loss of the Bcll site e, the gain of this new site could plausibly explain the two size variations without many presumptions. Other scenarios, however, are also possible and as long as the actual size variants have not been sequenced, the underlying mechanism will remain in the realm of speculation. The A . rnelfifera mitochondrial genome is 300-800 bp longer (depending on size variants) than that of D. yakuba, although the control region has been found to be 151 bp shorter. The larger total size is mainly due to non-coding intergenic sequences. As many as 618 non-coding nucleotides were found in the complete sequence data for the shortest Apis mitotype found in Apis mellifera ligustica (Crozier and Crozier, 1992) whereas only 183 were counted in D.yakuba (Clary and Wolstenholme, 1985). Particularly, the intergenic sequences (up to 831 bp; Cornuet and Garnery, 1991) between the tRNA"" and the CO-I1 genes are almost completely lacking in D.yakuba ( 5 bp only). Large mtDNA size variability is usually found near the control region of the mitochondrial genome (Moritz and Brown, 1987). According to Cornuet and Garnery (1991) two of the larger size variations found in honeybees are also located in the A+T-rich region which may function as the control region. Smith and Brown (1990) suggested that a variable number of tandem repeats between 80 and 100 bp causes this size variation. Yet again sequence data are lacking to support this hypothesis. The best studied size variability is between the (20-11 region and the tRNAl'" gene of the mitochondrial genome, and this has been found to be polymorphic in both European (Cornuet et a f . , 1991) and African subspecies (Garnery et al., 1992; Meusel and Moritz, 1992; Moritz et al., 1994). Cornuet et ul. (1991) analysed this region in detail and found repetitive sequences in the intergenic region between the tRNA"" and the CO-11 genes. They could identify two types of repeats which they named P and Q. Q is an obligatory sequence of 194-196 bp length found in all bees so far tested (Fig. 9). P with a length of 54 bp is facultative and is also present in a form Po (Cornuet and Garnery, 1991) with a 15 bp deletion which seems to be typical for the African subspecies of A . mellifera (Garnery et a / . , 1992) (Fig. 9). P is completely composed of A + T , whereas in the Q sequence 7.3% G + C is found. The size variation observed in several subspecies of honeybees is composed of various combinations of these two repeats. Whereas the Q repeat is variable within species, the P sequence has been found to be typical for various subspecies (Garnery et al., 1992). Table 4 lists all those types found to date. The repeat Q is composed of three subunits, Q,, Qz, Q3, which have conspicuous similarities with adjacent DNA regions. Cornuet and Garnery (1991) suggest that Q I might have diverged by duplication from the 3' end of

COI

I tRNAleuI

P

I

Q1 Q2 Q3

I COII 0 0

41-I

I

m

I

0

6 < W

rn rn

P TTAATAAATTAATAT~TAT~T---------------TATATTTATTAAATTTAATTTATTAAA

******************* ******

...........................

TTAATAAATTAATAT~T~TAT~TATATTTATT~~TTAATT~ATTAAA Q3

FIG. 9 Gene order in the CO-I-CO-I1 region of the mitochondria1 D N A of A. meffifera.The intergenic region between tRNAL" and CO-I1 consists of a fragment P and one or various repeats of fragment Q which is composed of three subunits Q1, Qz, and Q3. The sequences of the intergenic fragments are shown below. P has a high similarity to Q3, and Q, closely matches the 3' end of CO-I. The similarity of Qz to tRNAL" is shown in Fig. 10.

R. F. A. MORITZ

122

TABLE 4 Apis rnellifera subspecies and mitotypes Type of length variability

Subspecies

Reference

Q

carnica caucasica ligustica lamarc kii

Cornuet et al. (1991) Crozier and Crozier (1992) Crozier et al. (1989) Garnery et al. (1992) Smith (1991)

rnellifera iberica

Cornuet et al. (1991) Garnery et al. (1992)

scutellata capensis rnonticola unicolor

Cornuet et al. (1991) Hall and Smith (1991) Moritz and Meusel (1992) Smith (1991) Moritz et al. (1993)

scutellata

Moritz et al. (1993)

~~

X is a yet unknown sequence following the P,, fragment.

the CO-I gene, whereas Q2 bears strong similarities to the tRNA"" 3' end. In particular, the putative secondary structure of Q2 reveals striking similarities to the adjacent gene (Fig. 10). The aminoacyl arm, anticodon stem, and the T q C G stem are almost identical (16 out of 17 nucleotides). The three loops, however, differ substantially. The DHU loop of Q2 is twice as large in the corresponding tRNA'"" sequence, whereas the T K G loop lacks four nucleotides. Cornuet et af. (1991) suggest that the large DHU loop may impede the transcription of a functional tRNA. Q3 is very similar to Po and thus offers no such simple explanation for its evolutionary origin. Cornuet et af. (1991) argue that the entire region may function as an additional origin of replication finding support by the high A + T content (92.2%), and the secondary hairpin and cloverleaf structure. Such an additional origin of replication is surprising, and it is clearly lacking in D. yakuba, the closest species relative to honeybees for which complete sequence data are available. In fact a second origin of replication is uncommon in the animal systems studied so far (Harrison, 1989). Crozier and Crozier (1992) expressed doubts as to whether this highly variable region could function as an additional origin of replication. In attine ants the same region has been found to be extremely variable and sometimes completely lacking (Wetterer, cited in Crozier and Crozier, 1992). Yet, although this sequence may be lacking in ants, the Q sequence is present in one copy at least in all honeybees tested so far, thus leaving the issue still open.

T T T A A T A T

A A G

G A C G

A A A A T T A T

T T T A A T T A

T A

A A T T T T C A T A A A G T T

A

T

A

A T T T

T T A

G C A T A T C A T G T A A

A A A T A T T A T T A

T

A A A A T T A A T

41 -I I

A A T T T C T T A A A G A T A T T A T C A C T T A A

T T G A A T T T T A

t RNA Leu FIG. 10 Inferred cloverleaf structure of tRNALe"and the intergenic fragment Q2. Note the sequence identity in the aminoacyl arm, the anticodon stem and the W C G stem. Sixteen out of 17 nucleotide pairs are identical.

42

rn

I

0 Z

rn

-c

m

rn

rn

124

R. F. A. MORlTZ

4 Gene activity in embryonic development

The segmentation of the embryo is one of the best understood steps in insect embryogenesis. In particular the findings in Drosophila rnelanogaster have given an important insight into gene regulation in early embryonic development (Niisslein-Vollhard and Wieschaus, 1980; Nusslein-Vollhard er ul., 1982). The pattern of gene activity corresponds to the external rnetameric subdivision in the larval body (Martinez-Arias and Lawrence, 1985; Akam, 1987; Lawrence et al., 1987 among others). The genetics of segmentation has also been studied in other insects, like the flour beetle, Tribolium custuneurn (Beeman et al., 1989), and the grasshopper (Patel et al., 1989).In honeybees Fleig et al. (1988) were the first to study gene expression of segmentation genes. The honeybee offers various attractive features as a model system to study embryogenesis. The potential banding and segmentation patterns can be more easily interpreted since the head of the honeybee embryo is not involuted as in Drosophilu. Furthermore, the gastrula of the honeybee is straight and not folded like in Drosophila. Gastrulation starts about 33 h after oviposition (Fleig and Sander, 1986). The first metameric units become visible at this very early developmental stage as slight transverse grooves in the gnathal and thoracic region. The most anterior groove is slightly angled and separates the antennae anlagen from the intercalary segments. By using an antibody staining technique, Fleig (1990) showed that the expression of the engrailed gene product directly correlates with the groove pattern in this early developmental stage. An alternating pattern of intense and weak stained stripes was found along the gastrulating embryo. The intensely stained mandibular, labial, and mesothoracic stripes were about two to three cell rows wide, whereas the faintly stained stripes of the maxillary and the prothoracic region consisted of a single irregular row of cells. As gastrulation continues, the engrailed stripes are seen along the whole body, appearing and vanishing one after the other in an antero-posterior sequence. The alternating pattern is replaced by a regular striped pattern of three cell rows wide coinciding with the rnetarneric groove pattern. The labial stripe, however, remains broader whereas the procephalic stripe is very weak and has only a few stained cells. After mid-gastrulation all nuclei in the pre-serosa (and later in the serosa) show engrailed activity, which has not been described for any other insect but the honeybee. During the germ band stage, and in prehatching larvae, engrailed activity persists. Activity is particularly strong in the four cell row just anterior to the bottom of each metameric groove. The general pattern of engrailed expression in Apis rnellifera is similar to that found during embryonic development of other arthropods (Kornberg el al., 1985: DiNardo et a!., 1985; DiNardo and O’Farrell, 1989; Patel et ul.,

MOLECULAR BIOLOGY OF THE HONEYBEE

125

1989; Karr et af., 1989; Foe, 1989). The development of the stripe pattern and its topography on the embryo is somewhere between those found for Drosophila and Schistocerca. The honeybee has proved to be a most useful test system because of its size, its availability, and the straight structure (lack of head involution and no abdominal folding) of the embryo. In the future these features might even allow a direct analysis of the actual intercellular control processes that regulate the early metameric groove pattern in insects in general.

5 Population variability

Classically, genetic variability in populations has been analysed with isozyme polymorphisms. Although this technique has proved very powerful for the study of many insect populations, isozyme analysis in honeybees has often posed a problem. Honeybee population geneticists suffered for a long time from a notorious lack of suitable isozymes which revealed only a few polymorphisms (Sylvester, 1986). This was often attributed to male haploidy (Pamilo et af., 1978). However, Gan ef af. (1991) found 10 enzyme systems to be polymorphic in honeybees with an average degree of heterozygosity of 0.137. Thus it is also possible that the choice of gel running conditions was not always optimal in previous studies. The dynamics of gene frequency changes in honeybee populations can offer more ways of analysis than solitary insects can. Most interesting is the dual system of individual and colonial reproduction which results in potentially conflicting evolutionary strategies between workers, the sexual reproductives, and the colony as a whole. To examine the various individual and colonial fitness parameters one needs suitable genetic markers. This has proved to be a problem in the past when using allozyme markers. The progress in molecular analytical techniques, however, has allowed the direct utilization of DNA variability for population screening. Both nuclear and mitochondria1 DNA studies have been repeatedly conducted in the past few years. The analysis of exclusively queen-inherited markers, in combination with markers passed on by both sexes, can be studied by combining both nuclear and mitochondria1 DNA analysis. Such an approach is extremely helpful to dissect the selective importance of individual and colonial reproduction in honeybee populations. 5.1

NUCLEAR DNA MARKERS

Various nuclear DNA markers have been used to document genetic variability at the population and the colony level.

126

R. F. A. MORlTZ

5.1.1 Variability at the population level Usually nuclear DNA variability is detected through restriction fragment length polymorphisms (RFLP) which are visualized by hybridization to a labelled DNA probe. Classically the DNA probe is isolated from a gene bank. Hall (1986, 1988) was the first to use random fragment probes of honeybee DNA cloned with the plasmid pBR322 in Escherichia coli. Hall (1990) found that 19 of his isolated probes were useful for discrimination between Africanized and European honeybees. He explicitly stressed the usefulness of three probes (P130, P138, P170) in this context. However, although Hall (1990) was able to reveal DNA variability among various subspecies, the discriminating power of the probes was limited. The frequency for the ‘typical’ European allele was either 0 or 1 for the probes P130 and P170, and probe P138 was only found at a frequency of 0.06 in Apis mellifera mellifera. This high degree of variability for a diagnostic allele was obtained from a small sample size of two to four colonies per European race. Even given that the only goal was to find discriminative alleles between the Africanized honeybee and US-American ‘European’ honeybees (a racial mix of honeybees imported from Europe), the discriminatory power was unsatisfactory. In honeybee populations sampled in Arizona and Florida (20 and 15 colonies, respectively), ‘European’ gene frequencies ranged from 0.70 to 0.83, indicating substantial polymorphisms in each of the samples. In statistical terms this allows for a one-sided- test only. The presence of a ‘European’ allele may qualify a honeybee as ‘European’; the lack of this allele, however, does not qualify a honeybee as ‘African’. Clearly a polymorphic locus has only limited power as a diagnostic test system to discriminate between two subspecies and further screening seems to be desirable to isolate DNA probes with better discrimination ability. Hall was well aware of this problem, and I fully agree with him that ‘future research to establish the nuclear DNA RFLP specificity among pure European races (including A . m. carnica and A . m. iberica) should prove valuable’ (Hall, 1990, p. 618). However, also in follow-up studies (Hall, 1992; Hall and McMichael, 1992), no unambiguous diagnostic and racespecific probes were found, although there were significant frequency differences in the different subspecies for various alleles. For example, allele 2A2-B was found at frequencies between 0.83 and 0.91 in African but only at 0.25 in west European bees. Nevertheless, such variability is inappropriately high for the use of these probes as diagnostic tools for the identification of unknown honeybees. Multivariate analysis based on allelic variability at several loci is likely to solve the problem but has not been utilized for molecular data so far. Genomic DNA probes, as found in Hall’s studies, are very powerful tools for studying the dynamics of population genetics, and they are very helpful for understanding the problem of the Africanized bee problem in the

MOLECULAR BIOLOGY OF THE HONEYBEE

127

Americas, a subject to which I will return in detail later in this review. Currently the use of genomic DNA markers for the classification of racial types is premature. Successful and correct discrimination between European and African honeybees based on a few alleles only is very unlikely in the light of the vast number of highly diverse subspecies on each continent (Ruttner, 1988). Another approach to reveal intrapopulation variability was presented by Estoup et al. (1993). They isolated a series of microsatellites from a genomic library of the honeybee. One microsatellite which was composed of 13 CT tandem repeats associated with 9 G G T repeats proved to be highly variable in a population of A . mellifera. Using the flanking primers of the microsatellite, they found a strong length polymorphism in the PCR products yielding a total of 13 allelic forms. The same primers yielded considerably shorter PCR products in other honeybee species like A . cerana, A . florea and A . dorsata, which were not screened for intrapopulation polymorphism. 5.1.2 Variability within the colony Intracolonial genetic variability is high in honeybee colonies. This is due to the high degree of polyandry of the queen, with up to 20 drones mating a queen (Adams et al., 1977). As a consequence, up to 20 worker subfamilies or patrilines may coexist in the same colony. Since all workers are offspring from the same mother queen, workers with the same father are related by r = 0.75 (super-sisters), and those with different fathers are related by r = 0.25 (half-sisters). This intracolonial genetic variability is of particular interest for evolutionary and sociobiological aspects of honeybee biology. On the one hand the low genetic relatedness between patrilines might cause conflict in the colony, if workers recognize super-sisters and behave nepotistically. On the other hand genetic diversity in the colony may be advantageous (Moritz, 1989). For example genetic task specialization has been shown to be an important factor for division of labour in the colony (Page and Robinson, 1991) and genotypic intracolonial variability has been suggested to be an important fitness parameter for groups of bees and colonies (for recent reviews of these aspects of honeybee biology see Breed and Page, 1989; Moritz and Southwick, 1992). Until recently only isozyme or phenotypical mutant markers were used in studies to reveal patriline identity. This had many drawbacks because of the limited number of patrilines that can be tested with allozymes and the use of artificial insemination. Honeybee DNA probes have proved to be very useful in a study focusing on intracolonial variability. Various techniques have been tested and have shown the great potential of molecular tools to increase our understanding of the function of the honeybee colony as a whole. In a recent study by

128

R. F. A. MORITZ

Oldroyd et al. (1994), a DNA probe from a genomic library of the honeybee proved useful to identify different patrilines in a colony of A . florea, the Asian dwarf honeybee. They could identify as many as eight different RFLP patterns suggesting a mating with about eight drones. This is a surprisingly high number in the light of observations by Koeniger et al. (1989) who claimed that only a very limited number (less than five) of A . frorea drones mate with the queen, based on semen counts of queens returning from their mating flight. They found no semen in the oviducts and concluded that the A . frorea drones deposit their semen directly into the spermatheca with the complexly structured endophallus. The DNA data of Oldroyd et al. (1994) do not support this view, and the estimate of eight matings may be conservative since rare patrilines may not have been detected and some drones might have had the same RFLP type. Oldroyd et a f . (1994) could also confirm those observations made by Page and Robinson (1991) in A. meffifera, who found a variety of genetically determined specialists in the colony. Another approach using molecular tools for the identification of patrilines in a colony of honeybees has been DNA fingerprinting. Moritz et al. (1991) could discriminate between super- and half-sisters in a colony by using multilocus fingerprinting with the synthetic (GATA)4 oligonucleotide. Using the same oligonucleotide, Haberl and Moritz (1994) found 12 different subfamilies in a colony of A . rnellifera. A similar result was obtained by Blanchetot (1991) who used M13 as a probe for DNA fingerprinting. He was able to identify 11 patrilines in a single colony. Both figures are conservative in that patrilines may not have been identified if two drones had the same genotype and if the frequency of a patriline in the colony was very low. The genetic effective number of males was lower than the number of subfamilies found, because they were not equally represented. Twenty-five per cent of the workers belong to the most frequent subfamily, whereas rare subfamilies had frequencies of less than 2%. Using the estimator for the effective male number of Laidlaw and Page (1984), N , is between 6 and 7. Both empirical studies deviate substantially from the previous widely accepted estimate of an average of 17 effective matings per queen (Adams et a f . , 1977), which may be either an overestimate or a specific trait of the sampled Africanized honeybee population in South America. DNA amplification using random amplified polymorphic DNA technique (RAPD) has also been used to identify paternity in the honeybee colony. Hunt and Page (1993) and Fondrk et a f . (1993) used a 10 bp random primer with 50% GC content to screen a colony. They could clearly reidentify the four patrilines in a queen artificially inseminated with four drones. Some interpretational problems arose from non-paternal bands that resulted from chimaeric amplification products, known as PCR recombination. Nevertheless, Fondrk et al. (1993) could unambiguously classify each individual worker on the basis of 20 markers only.

MOLECULAR BIOLOGY OF THE HONEYBEE

5.2

129

MITOCHONDRIAL DNA MARKERS

Mitochondria1 DNA has been found to be inherited exclusively maternally in most animal systems (Moritz and Brown, 1987). In several cases, however, this rule might have been broken since heteroplasmic individuals have been found in various organisms. Brown el a f . (1992) found 42% heteroplasmy in sturgeon, and Hoeh et a f . (1991) observed 57% heteroplasmic individuals in Mytifus mussels. Kondo et a f . (1990) showed that in three out of 331 lineages paternal mtDNA replaced the original mtDNA in artificial crosses in Drosophifa. Gyllenstein et a f . (1991) reported on paternal leakage in mice. Looking at the special fertilization mechanism of the egg, honeybees seem to be another candidate for a significant paternal mtDNA inheritance. Apis meffifera has a polyspermic mode of fertilization, with many sperms entering the egg, including the mitochondria-rich tail (Blochmann, 1889; Petrunkewitsch 1901). Although many sperms have been found to enter the egg, only one of them fuses with the egg nucleus. The others remain as accessory sperms in the egg and usually disintegrate rapidly after the first cell divisions (Nachtsheim, 1912, 1914). Only rarely do they show mitotic activity yielding gynandromorph individuals (reviewed by Woyke and Hillesheim, 1990). Before an mtDNA marker can be used as an authentic maternal marker proportion paternal mtDNA

30%

I

20%

10%

0% 0

12

24

36

48 60 72 development (hrs)

84

96

108

120

FIG. 11 Decrease of paternal rntDNA after fertilization of the egg. When the larvae hatch after 96 h only traces of paternal rntDNA can be detected.

R. F. A. MORITZ

130

in honeybees, it seemed necessary to test whether any paternal leakage of mtDNA could contribute to mtDNA variability in the population. Meusel and Moritz (1993) pursued this in a study using hybrids of two races utilizing length polymorphism in the CO-I1 and tRNA"" region (see Section 2.3). They inseminated an A . meflifera capensis queen (0.4 kb insert) with A . meflifera carnica drones that did not have this insert and tested the progeny at various larval stages. Initially they found up to 23% paternal mtDNA in freshly laid eggs, but with ongoing development paternal mtDNA rapidly decreased (Fig. 11). In late larval stages only maternal mtDNA was found, indicating that in spite of the large initial paternal contribution only maternal mtDNA is genetically effective. The loss of paternal mitochondria has been reported for a variety of organisms. For example Anderson (1968) observed the degradation of paternal mitochondria in the sea urchin, Paracentrotus lividus, and even in isogamous fungi, in which mtDNA recombination can occur, the rapid disappearance of one parental type has been observed (Meland et al., 1991). In honeybees mtDNA seems to provide a safe tool to study maternal gene flow in populations. Several markers are available and particularly the size variants listed in Table 4 have been used in the past to characterize populations and racial types of the honeybee. 6 Molecular evolution and biogeography

6.1

MOLECULAR PHYLOGENY OF BEES

Besides studies at the population level molecular techniques are also helpful in phylogenetic studies. Similarities between different taxa can be evaluated by comparing RFLPs, restriction site maps or sequence data of specific DNA regions. The now classical techniques of using highly conserved ribosomal DNA have also been applied to honeybees, but with only limited success. Sheppard and McPheron (1991) report on a paucity of 18s rDNA variation in the genus Apis, not allowing for any phylogenetic analysis. Nevertheless, they did find sequence variability at the tribal level of the Apidae which could be used to produce a phylogenetic tree. Indeed, the need for a molecular approach to the phylogeny of the Apidae seems highly promising, because the classical phylogenies based on morphological or behavioural traits vary substantially among a variety of authors (see Fig. 12) The results from the 18s and 28s ribosomal DNA (Sheppard and McPheron, 1991) are based on seven informative sites and clearly more information is needed to get a more precise picture of the possible phylogenetic trees. Using the branch and bound routine of the PAUP 3.0 software package (Swofford, 1990) they found a parsimony tree as shown in Fig. 12 which differed from a similar tree produced by Cameron (1991; also using branch

131

MOLECULAR BIOLOGY OF THE HONEYBEE Euglossin i

Born bini Meliponini

Euglossini

Born bini Meliponini Apini

Melipon in i

Euglossini

<

Meliponini

$

!!‘;::1

Born bini Apini

Euglossini Apjni

Meliponini Meliponini

Born bini

FIG. 12 Alternative phylogenetic trees of the four tribes of the Apidae. (A) Michener (1944, 1990); (B) Michener (1974); (C) Winston and Michener (1977); (D) Plant and Paulus (1987); (E) Sheppard and McPheron (1991); (F) Cameron (1991). The phylogenetic trees E and F are based on sequence data from mtDNA (rRNA).

and bound of the PAUP 3.0 package) based on mitochondria1 DNA variability (Fig. 12). Although in both studies bumblebees and stingless bees are more similar than the Apini and Euglossini, the topologies of the two most parsimonious trees are different, indicating the need for complementary data in support of one or the other derived phylogenetic tree.

6.2

GENETIC VARIABILITY AMONG HONEYBEE SPECIES

Genetic variability within the genus can also be revealed using mtDNA variability. These studies are of particular interest because of the recent rediscovery of various ‘forgotten’ species of Apis. Maa (1953) recognized as many as 24 species divided into three genera: Micrapis (dwarf honeybees), Megapis (giant honeybees) and Apis (cave-breeding honeybees). Subsequent papers, however, ignored or criticized Maa’s system as unwarranted by the data set. For decades the genus Apis was composed of only four species (Ruttner, 1988), until the work on Asian honeybees was reintensified. Today Apis andreniformis (Wu and Kuang, 1987; Wongsiri et al., 1990) and Apis koschevnikovi (Koeniger et al., 1988; Tingek et al., 1988; Rinderer et al., 1989; Ruttner et al., 1989) are well established species, renewing the interest in the system originally developed by Maa. In spite of this interest, the molecular contribution towards a new systematics of the

132

R. F. A. MORITZ

Duplication Po

9 A. mellifera

Elongation

co

I1

A. c e r a n a

Regression A. d o r s a t a

co

I

co

I1

A. florea

Regression

tRNA leu

modification of leu

FIG. 13 Putative evolutionary patterns of the mtDNA CO-I-CO-I1 intergenic region in Apis (redrawn from Cornuet et al., 1991).

genus Apis has been scarce. Smith (1991) found that mtDNA of A . koschevnikovi and A. melliferu were both highly divergent from A . cerana samples, with more than 10% sequence divergence between each of the species. Garnery et ul. (1991) screened mtDNA of A. melliferu, A. cerana, A. floreu and A . dorsutu. They sequenced the intergenic region between the tRNA"" region and the CO-I1 gene, and found sequence divergence ranging from 7% to 11%. Based on the sequence data, Cornuet and Garnery (1991) present a most parsimonious phylogenetic tree, supporting the early divergence of A . floreu, and the close relationship of the two cave-breeding species A. melliferu and A . cerunu, which is in agreement with morphometrical, behavioural (Alexander, 1991) and allozyme data (Sheppard and Berlocher, 1989). Cornuet and Garnery (1991) developed a putative evolutionary pathway of the mtDNA region sequenced by Garnery et al. (1991), explaining the evolutionary changes through DNA duplication, elongation, and regression (Fig. 13). Cornuet and Garnery (1991) extended the sequence data 185 bp into the region of the large ribosomal unit yielding a total of 61 informative sites. Based on both a neighbour joining (Saitou and Nei, 1987) and a parsimony analysis (PAUP 3.0 software package) they obtained phylogenetic trees with Bombus lucorum as an outgroup, confirming the topology of the tree presented in Fig. 13. Cornuet and Garnery (1991) suggested that both cave-breeding species A. melliferu and A . ceruna diverged about 5.9 million years from their common ancestor. Willis et ul.

MOLECULAR BIOLOGY OF THE HONEYBEE

133

(1992) also presented a phylogenetic tree on the basis of mtDNA variability. They analysed the CO-I1 sequence of six different species, including A . koschevnikovi and A . andreniformis, using the CO-I1 sequence of the wasp (Excrisres roborator) as an outgroup. The most parsimonious tree derived in this study has a different topology. A. dorsara is considered as the most ancestral group which is in contradiction to Garnery’s results (Garnery er al., 1991; Cornuet and Garnery, 1991). The most surprising result is the placement of the cave-breeding A . koschevnikovi together with the two free breeding dwarf honeybees. Willis et a f . (1992) admit that this result is rather puzzling since the cave-breeding species are considered to be very similar, behaviourally , morphologically and ecologically (Smith, 1991). Also the low bootstrap value for the fork between the closely related A . meflifera and A . cerana (45.1%) may be an indication that additional data are required to obtain plausible phylogenetic trees. One would not necessarily need to share Willis er af.’s (1992) view, that because the oldest fossil bee is of a dorsara type (Culliney, 1983) therefore this supports A . dorsara as the most ancestral type. Also Ruttner (1988) found in a morphometrical study that the fossil Apis armbrusleri (10-12 m.y. b.p.) has a wing venation pattern very similar to that of A . dorsara. Although such evidence shows that dorsaru-type bees did exist in the lower Miocene, this does not mean that florea type bees did not exist. It could well be that an olderflorea-like fossil exists but has not been found or not been dated. The controversial results make it very clear that more molecular data are required to present a concise picture. Particularly in the light of the high A T content and the rapid evolution of honeybee mtDNA (Crozier er al., 1989), it seems wise to include other genes in a phylogenetic study as well and not only rely on a sequence analysis of a single mitochondria1 gene. 6.3

MOLECULAR VARIABILITY WITHIN SPECIES

The majority of mtDNA research has been used to study genetic variability within species. Mitochondria1 DNA has also been successfully used in a variety of studies characterizing subspecies and the biogeography of honeybee populations.

6.3.1 Apis cerana Smith (1991) constructed a distance tree based on RFLPs of various A . cerana populations (Fig. 14). Based on this information she suggested an evolutionary scenario in which the subspecies were connected during the late Pleistocene. At this time the sea level was about 120m lower and the Sunda Shelf, including Borneo, Sumatra and Java, was connected to the mainland explaining the close relationship (0-1.9% sequence divergence) among the populations found in Borneo, Malaysia, Japan, Thailand and

R. F. A. MORlTZ

134

M a 1a y si a Ma1ays i a Borneo

-

India

A . florea

6

4

Thailand

2

0

% sequence divergence

FIG. 14 UPGMA distance tree of Apis ceruna populations (redrawn from Smith, 1991).

India. Populations that were isolated on islands in that time period (Luzon, Andaman Islands) showed substantially higher sequence divergence from the other populations (2.9-5.6% divergence). Both other cave-breeding honeybees ( A . koschevnikovi and A . rnellifera) showed a sequence divergence of more than l o % , which is much higher than any intraspecific distance among subspecies.

6.3.2 Apis dorsata The systematics of the giant honeybees is controversial (as is systematics for most organisms). Maa (1953) recognized four different species which he grouped in a separate genus Megapis. Ruttner (1988) worked with only one species and gave the different types subspecies status. Sakagami et al. (1980) accepted Apis luboriosa of the Himalayan highlands as an individual species of giant honeybees in addition to Apis dorsata. Alexander (1991) and Otis (1991a) discussed the ambiguity of the dorsata group, forming either a genus-like group of four species or a single species with four subspecies. Smith (1991) favoured Ruttner’s (1988) nomenclature for her studies on mtDNA variability among the giant honeybees. Nevertheless, the sequence divergence she found between the different haplotypes was much larger than for any other honeybee species. She estimated as much as 12.23% sequence divergence between A . dorsata binghami and A . dorsata dorsata from the Asian mainland. This is a substantial difference for bee populations living in

MOLECULAR BIOLOGY OF THE HONEYBEE

135

India India India India Borneo Ma lays i a Thailand Pakistan

I. d. dorsata

Malaysia Andaman Is.

A . d. b i n g h a m i Sulawesi

1 10

8 6 4 2 % s e q u e n c e divergence

0

FIG. 15 UPGMA distance tree of the Apis dorsata group (redrawn from Smith, 1991).

the same geographic region. Apis mellifera from northern Europe has an estimated sequence divergence of 3% to A . m. capensis from South Africa (Smith, 1991). Although the large variability in the dorsata group does not in itself explain whether we are dealing with true species or subspecies, it clearly indicates that the giant honeybees are a much more diverse group than the cave breeding honeybees (Fig. 15).

6.3.3

Dwarf honeybees

Currently two species of dwarf honeybees are recognized, Apis florea and Apis andreniformis. Apis florea is found from the Persian gulf to Thailand whereas A. andreniformis is found in tropical Asia east of this distribution area. Smith (1991) screened samples from five different origins to determine sequence divergence (Fig. 16). Interestingly the distance between the two species is less than half of that of the giant honeybees! This certainly gives no support to the four subspecies classification of the dorsata group as presented by Ruttner (1988). This is particularly so because the dwarf bees are much poorer dispersers than are the giant honeybees. On the other hand, the data are still of very preliminary character, based on a scarce and limited sample set.

R. F. A. MORITZ

136

Borneo Malaysia Japan Thailand India Luzon

I

Andaman Is.

1 5

4 3 2 1 % sequence divergence

0

FIG. 16 UPGMA distance tree of the dwarf honeybees (redrawn from Smith, 1991).

6.3.4 Apis melliferu

The western honeybee, A. melliferu, is certainly the best studied honeybee species concurring with its mtDNA variability. The most striking variability is the length variation in the CO-I-CO-I1 region which is composed of various numbers of tandem repeats as discussed above (Garnery et ul., 1992). There are three distinct haplotypes comprising a fragment P (lineage M), the African subspecies with a fragment Po which is 15 bp shorter (lineage A) and a lineage C lacking the P fragment altogether. Based on these data and additional restriction site and sequence data, Garnery et ul. (1992) developed a new phylogeographic tree for the species A. melliferu. Ruttner (1988) concluded from his biogeographic studies a Y-shaped branching in northern Africa. Two branches dispersed westward north and south of the Mediterranean, and one spread southward into central Africa. In this model west European A. melliferu melliferu were at the end of the branch along the North African coast, bridging to Europe at Gibraltar. East European A. m. curnicu were morphologically very different from A. m. melliferu and thought to be the end of the branch spreading at the south European Mediterranean coast, both subspecies divided by the Alps. Garnery et ul. (1992) constructed a triple-branched phylogeographic tree with the origin in Iran. They found that lineage A is more similar to C than A to M. Furthermore, C is less similar to M than to A (Fig. 17). This clearly

137

MOLECULAR BIOLOGY OF THE HONEYBEE

3

I

FIG. 17 Biogeography of the lineages of Apis mellifera around the Mediterranean (redrawn from Garnery et al., 1992). Lineage M is the most northern branch leading to the A . mellifera subspecies of western Europe. Lineage C represents the races of the Balkan and the northern Mediterranean and lineage A comprises the African subspecies.

does not support Ruttner’s (1988) models which predict a higher similarity between the West European and African races. Garnery et al. (1992) plausibly explain the racial admixture in the Spanish population through secondary hybridization with African populations. Since various mitotypes have been suggested to be race typical (Cornuet et al., 1991; Cornuet and Garnery, 1991; Garnery et al., 1992; Meixner et al., 1993; Moritz et al., 1986; Sheppard et al., 1991a, b; Smith, 1988, 1991;

138

R. F. A. MORITZ

Smith and Brown, 1990; Smith et al., 1989), they have been used to verify the racial origin of feral honeybee populations. For example Oldroyd et al. (1992) studied a honeybee population on Kangaroo island in southern Australia using restriction enzyme analysis on mtDNA in combination with isozyme variability. Interestingly they found that the isozyme data support the theory that no hybridization occurred after the introduction of Italian A. melfifera ligustica and the protection of this population by the Ligurian Bee Act of 1886. Yet surprisingly, mitochondria1 types proved to be those thought to be typical for A . mellifera mellifera (Sheppard et al, 1991a). Oldroyd et al. (1992) concluded that the degree of mtDNA variability found in honeybee populations may have been underestimated. The high variability of mitotypes is particularly apparent in the length variation between the CO-I and CO-I1 region (discussed above in detail). The variability for the number of repeats of the Q fragment is equal within and between races in southern Africa (Moritz et al., 1994). In a single race, A. mellifera mellifera, three different size types were found (Garnery et al., 1992) offering no obvious diagnostic power to discriminate among populations or subspecies.

6.3.5 The Africanized bee problem The problem of honeybee dispersal has become quite an important one during the past decades as the Africanized honeybee problem has developed into a serious issue of international scope. Among the various narratives about the outbreak of the problem, the following was favoured in a recent monograph on the problem (Spivak et a f . , 1991). Forty-seven or 48 queens (from a lot of 170 queens) from the Transvaal highveld (South Africa) and one queen from Tabora (Tanzania) were successfully introduced into colonies at Piracicaba (SP) in Brazil in 1956 to increase the productivity of the European honeybee stock which was believed to be poorly adapted to the tropical conditions of South America. In early 1957 queen excluders were removed from the flight entrances and 26 of these colonies swarmed. This is believed to be the beginning of the Africanized bee problem (Kerr, 1957). In addition to these accidentally escaped swarms it has been claimed that African honeybee queens have been systematically reared and distributed to beekeepers (Spivak el al., 1991). This clearly would be a much more plausible basis for the rapid spread of the African type honeybees. However, all of these claims are based on a questionable interpretation of data, unconfirmed personal communications and anecdotal material. The introduced stock, which was considered undesirable due to its strongly expressed stinging behaviour, spread rapidly throughout the continent, virtually eliminating the occurrence of European stock. Its current distribution ranges from Texas in the north to Argentina in the south. Because of its negative effects on the public, the ‘Africanized bee problem’

MOLECULAR BIOLOGY OF THE HONEYBEE

139

is one of the better studied cases in population biology (for a comprehensive review, see Spivak et al., 1991). We know that the Africanized bees have a swarming rate about 13 times higher than European honeybees (Otis, 1982, 1991b). Furthermore, Africanized drones have been found to penetrate into European colonies and simply by their presence reduce the production of European drones (Rinderer et al., 1985). The phenotypical result of the Africanization process is quite clear. Feral honeybees in the Americas behave like true African bees and they phenotypically resemble the African rather than the European honeybee (Daly, 1991). In spite of the seemingly clear ecological data there is considerable controversy about the actual spread mechanisms of Africanized honeybees. Particularly, two testable hypotheses have been discussed to explain the process of Africanization. One model assumes the production of ‘hybrid’ populations and the spread of a phenotypically well adapted genotype through natural selection (Kerr and Bueno, 1970; Michener, 1975; Rinderer et al., 1985; Rinderer, 1986). The other model assumes the spread of the introduced queens as a pure African gene pool without hybridization, for example through hybrid inviability or pre-mating isolation (Taylor, 1985; Fletcher, 1991). Thus the problem is whether the Africanized bees spread exclusively through swarming or whether hybridization is also a significant factor in this process. In support of both hypotheses, a variety of studies focusing on mtDNA have been conducted, but unfortunately this has not yet solved the problem and still left space for contention. One set of studies on mtDNA (Hall and Muralidharan, 1989; Hall and Smith, 1991; Smith et al., 1989) supports the view that the African honeybees spread as pure maternal lineages. Theoretically mtDNA is a prime choice marker because it is inherited maternally (Meusel and Moritz, 1993). However, as soon as hybridization occurs between the two populations, which seems plausible in light of the increased fitness of the Africanized drones, European mitotypes are to be expected in the hybrid population. The higher the fitness of the Africanized drones, the higher the frequency of the European mitotypes. Moritz and Meusel (1992) modelled the spread of a neutral mitochondria1 marker which is initially linked to a fit nuclear genome. They could show that a determination of the cyto-nuclear disequilibrium in the hybrid zone would allow for an analysis of the significance of swarming and hybridization. Although a pure maternal spread through swarming might theoretically be possible in principle, Moritz and Meusel (1992) showed in a population genetic model that the spread of pure maternal lineages requires some extreme assumptions that may not be very likely in nature. Only strong assortative mating (as suggested by Taylor, 1985), a cyto-nuclear incompatibility (as suggested by Hall and Muralidharan, 1989) or functional hybrid sterility (as claimed by Fletcher, 1991) can explain the persistence of pure African mitotypes in a honeybee population with racial admixture. Although such mechanisms are possible

140

R. F. A. MORITZ

and have been found in other organisms, there is as yet no such evidence for honeybee populations. Harrison and Hall (1993) found indications for a negative heterosis concerning the metabolic capacity of hybrid honeybees. The metabolic capacity (determined as watts/kg body weight) during agitated flight of a caged worker was about 15% less in the hybrid bees compared with pure African bees. It remains unclear however whether these differences in a rather artificial bioassay have any implications for the reproductive capacity of colonies. Rinderer et af. (1991) and Sheppard ef al. (1991a,b) found substantial hybridization in Africanized populations in Mexico, Brazil and Argentina, which is in line with the theoretical expectations as suggested by Moritz and Meusel (1992). Lob0 et af. (1989) also found racial admixture in feral honeybee populations in South America. These studies seem to contradict the view that the few imported African queens spread as pure maternal lineages in America. Most significant are those results in which a European mitotype is found with an African phenotype or nuclear genotype. These cases are in line with a view that hybrids do contribute to the gene pool and there are no biological barriers to hybridization. The issue may initially look academic but it is of substantial practical impact. It directly affects concepts for controlling the Africanized honeybee. If there is no hybridization between Africanized and European honeybees, any effort to interfere in the Africanization process through mating control and breeding would be void. In the light of the ongoing controversy it would certainly be premature to omit breeding control as a means to manage the Africanized honeybee. As long as contradicting empirical data are obtained all potentially possible control options have to be used. A shortcoming of all empirical studies on mtDNA variability is the lack of repeated sampling of the same transect in the hybrid zone between Africanized and European honeybees. It is impossible to determine a dynamic population process by taking a single ‘snap shot’ of gene and allele frequencies. It is only possible to determine nuclear and mitochondria1 gene flows through repeated sampling over several generations. Only then shall we be able to understand the genetic basis of the Africanized bee problem. 7 Outlook

Much of the cited work in this review reveals the major shortcomings of recent molecular research in honeybees. The field is trying hard to catch up in methodology with current standards in molecular research but it is in the early stages. Information is however rapidly accumulating. With the fast progress in molecular methodology and the increasing ease of performance of the various protocols, it is not difficult to predict that the boom of research in molecular honeybee research is yet to come. A typical example

MOLECULAR BIOLOGY OF THE HONEYBEE

141

may be the analysis of the mitochondria1 genome of the honeybee, where within a few years the data accumulated from some patchy restriction fragment information (Moritz et af., 1986) to the complete sequence (Crozier and Crozier, 1992). The development of highly variable nuclear DNA markers will allow a further analysis of the intracolonial demography and the genetic basis for division of labour and colony organization. The analysis of structure and sequence of important genes will follow soon. Several groups are currently focusing on the analysis of the sex locus which indeed is one of the major interesting genes in Hymenoptera. The mapping of the genome will form the basis for such a study. The consequent use of the advantages offered by honeybees over other test systems will ensure a rapidly increasing body of most exciting studies, re-establishing its position as a most rewarding study organism in both physiological and genetical research. Acknowledgements

I wish to thank M. Beye, P. Kryger and H. G. Hall for their most helpful comments. I am grateful to J. M. Cornuet, R. H. Crozier, G. J . Hunt, B. Oldroyd, R. E. Page, T. E. Rinderer, and M. Winston, who provided preprints or submitted manuscripts which were most helpful to update this review.

References Adams, J.. Rothman, E. D., Kerr, W. E. and Paulino, Z . L. (1977). Estimation of sex alleles and queen matings from diploid male frequencies in a population of Apis mellifera. Genetics 86, 583-596. Akam, M. (1987). The molecular basis for metameric pattern in the Drosophila embryo. Development 101, 1-22. Alexander, B. (1991). A cladistic analysis of the genus Apis. In "Diversity in the Genus Apis" (Ed. D. R . Smith), pp. 1-28. Westview Press, Boulder, CO. Anderson, W. A. (1968). Structure and fate of paternal mitochondrion during early embryogenesis of Paracentrotus lividus. J . Ultrastruct. Res. 24, 31 1-321. Bachmayer, H., Kreil, G. and Suchanek, G. (1972). Synthesis of the promelittin and melittin in the venom gland of queen and worker bee patterns observed during maturation. J . Insect Physiol. 18, 1515-1521. Beeman, R . W., Stuart, J. J., Haas. M. S. and Denell, R . E. (1989). Genetic analysis of the homeotic gene complex (HOM-C) in the beetle Tribolium castaneum. Dev. Biol. 133, 196209. Beye, M . and Moritz R. F. A . (1993a). In situ hybridization of rDNA on chromosomes of the honeybee, Apis mellifera L. Experientia 49, 337-338. Beye, M. and Moritz R. F. A. (1994). Characterization of Apis melliferu L. chromosomes using fluorescence in situ hybridization. J . Hered. (in press).

142

R. F. A. MORITZ

Blanchetot, A. (1991). Genetic relatedness in honeybees as established by DNA fingerprinting. J . Hered. 82, 391-396. Blochmann, F. (1889). Uber die Zahl der Richtungskorper bei befruchteten und unbefruchteten Bieneneiern. Morphol. Jb. 15, 85-96. Breed, M. D. and Page, R . E. (Eds) (1989). “The Genetics of Social Evolution”. Westview Press, Boulder, CO. Brown, J. R., Beckenbach, A. T. and Smith, M. J. (1992). Mitochondrial DNA length variation and heteroplasmy in populations of white sturgeon (Acipenser transmonfanus). Genetics 132, 221-228. Butler, C. G. (1973). The queen and “the spirit of the hive”. Proc. R . SOC. Entomof., Lond. A . 48, 59-65. Cameron, S. A. (1991). A new tribal phylogeny of the Apidae inferred from mitochondria1 DNA sequences. In “Diversity in the Genus Apis” (Ed. D. R. Smith), pp. 72-87. Westview Press, Boulder, CO. Clary, D. 0. and Wolstenholme, D. R. (1983a). Genes for cytochrome c oxidase subunit I, URF2, and three tRNAs in Drosophila mitochondrial DNA. Nucleic Acids Res. 11, 685943872. Clary, D. 0. and Wolstenholrne, D. R. (1983b). Nucleotide sequence of a segment of Drosophila mitochondrial DNA that contains the genes for cytochrome c oxidase subunits I1 and 111 and ATPase subunit 6. Nucleic Acids Res. 11, 421 1-4277. Clary, D. 0. and Wolstenholme, D. R. (1984). A cluster of six tRNA genes in Drosophila mitochondrial DNA that includes a gene for an unusual tRNASer. Nucleic Acids Res. 12, 2367-2379. Clary, D. 0. and Wolstenholme, D. R. (1985). The mitochondria1 DNA molecule of Drosophila yakuba: nucleotide sequence, gene organization, and genetic code. J . Mof. Evol. 22, 116125. Clary, D. O., Goddard, J. M., Martin, S. C., Fauron, C.M.-R. and Wolstenholme, D. R. (1982). Drosophila mitochondrial DNA: a novel gene order. Nucleic Acih Res. 10, 6619-6637. Clary, D. O., Wahleithner, J. A. and Wolstenholme, D. R. (1984). Sequence and arrangement of the genes for cytochrome b, URF1, URF4, URFS, URF6 and five tRNAs in Drosophila mitochondrial DNA. Nucleic Acids Res. 12, 3747-3762. Cornuet, J . M. and Garnery, L. (1991). Mitochondria1 DNA variability in honeybees and its phylogenetic implications. Apidologie 22, 627-642. Cornuet, J. M., Garnery, L. and Solignac, M. (1991). Putative origin and function of the intergenic region between COI and COII of Apis mellifera L. mitochondrial DNA. Genetics 128, 393403. Crozier, R. H. and Crozier, Y. C. (1992). The mitochondrial genome of the honeybee Apis mellifera: complete sequence and genome organization. Genetics 133, 97-117. Crozier, R. H., Crozier, Y . C. and Mackinlay A. G. (1989). The CO-I and CO-I1 region of honeybee mitochondrial DNA: Evidence for variation in insect mitochondrial evolutionary rates. Mol. Biof. Evol. 6 , 399-41 1. Culliney, T. W. (1983). Origin and evolutionary history of the honeybees Apis. Bee World 64, 29-37. Daly, H. V. (1991). Systematics and identification of Africanized honey bees. In “The ‘African’ Honey Bee” (Eds M . Spivak, D. J . C. Fletcher and M . D. Breed), pp. 1344. Westview Press, Boulder, CO. DeBruijn, M. H. L. (1983). Drosophila melanogaster mitochondrial DNA, a novel organization and genetic code. Nature 304, 234-241. Del Lama, M. A., Soares, A. E. E. and Mestriner, M. A. (1985). Linkage studies in

MOLECULAR BIOLOGY OF THE HONEYBEE

143

Apis mellifera using biochemical and morphological markers. J . Hered. 76, 427-430. Del Lama, M. A , , Mascarenhas, R . and Soares, A. E . E. (1993). Further linkage studies with biochemical markers in Apis mellifera. J . Hered. 84, 129-130. DiNardo, S. and O’Farrell, P. H . (1989). Establishment and refinement of segmental pattern in the Drosophila embryo: distributions of the fushi tarazu, engrailed, Sex combs reduced, Antennapedia, and Ultrabithorax proteins. Genes Dev. 2, 350360. DiNardo, S., Kuner, J. N., Theis. J . and O’Farrell, P. H. (1985). Development of embryonic pattern in Drosophila melanogaster as revealed by accumulation of the nuclear engrailed protein. Cell 43, 59-69. Dzierzon, J . (1845). Gutachten iiber die von Hrn. Direktor Stohr im ersten und zweiten Kapitel des General-Gutachtens aufgestellten Fragen. Eichstadter Bienenzeitung 1, 109-113, 119-121. Estoup, A . , Solignac, M., Harry, M. and Cornuet, J. M. (1993). Characterization of (GT), and (CT), microsatellites in two insect species: Apis mellifera and Bombus terrestris. Nucleic Acids Res. 21, 1427-1431. Fleig, R. (1990). Engrailed expression and body segmentation in the honeybee Apis mellifera. Roux Arch. Dev. Biol. 198, 467-473. Fleig, R. and Sander, K. (1986). The embryogenesis of the honeybee Apis mellifera L. (Hymenoptera): a SEM study. Int. J . Insect. Morphol. Embryol. 15, 449462. Fleig, R., Walldorf, U., Gehring, W. J. and Sander, K. (1988). In situ localization of the transcripts of a homeobox gene in the honeybee Apis mellifera L. (Hymenoptera). Roux Arch. Dev. Biol. 197, 269-274. Fletcher, D . J. C. (1991). Interdependence of genetics and ecology in a solution to the African bee problem. In “The ‘African’ Honey Bee” (Eds M. Spivak, D. J. C. Fletcher and M. D. Breed), pp. 77-79. Westview Press, Boulder, CO. Foe, V. E. (1989). Mitotic domains reveal early commitments of cells in Drosophila embryos. Development 107, 1-22. Fondrk, M. K., Page, R . E . and Hunt, G . J. (1993). Paternity analysis of worker honeybees using random amplified polymorphic D N A (RAPD). Naturwissenschaften 80, 226231. Gan, Y. Y . , Otis. G . , Mardan, M. and Tan, S. G. (1991). Allozyme diversity in Asian Apis. In “Diversity in the Genus Apis” (Ed. D . R . Smith), pp. 117-130, Westview Press, Boulder, C O . Garnery, L., Vautrin, D . , Cornuet, J. M. and Solignac, M. (1991). Phylogenetic relationships in the genus Apis inferred from mitochondrial D N A sequence data. Apidologie 22, 87-92. Garnery, L., Cornuet, J . M. and Solignac, M. (1992). Evolutionary history of the honey bee Apis mellifera inferred from mitochondrial D N A analysis. Mol. Ecol. I , 145-154. Gauldie, J.. Hanson, J . M.. Rumjanek. F. D., Shipolini, R. A. and Vernon, C. A. (1976). The peptide components of bee venom. Eur. J . Biochem. 61, 369-376. Gauldie, J . , Hanson, J . M., Shipolini, R . A . and Vernon, C. A . (1978). The structure of some peptides from bee venom. Eur. J . Biochem. 83, 405-410. Gmachl, M. and Kreil, G . (1993). Bee venom hyaluronidase is homologous to a membrane protein of mammalian sperm. Proc. Natl. Acad. Sci. USA 90, 3569-3573. Gyllenstein, U., Wharton, D . , Josefsson. A . and Wilson, A . C . (1991). Paternal inheritance of mitochondria1 D N A in mice. Nature 352, 255-257. Haberl, M. and Moritz, R . F. A . (1994). Determination of intracolonial relationship in a honeybee colony (Apis mellifera L.) with D N A fingerprinting. Insectes Soc., in press.

144

R. F. MORITZ

Habermann, E. (1972). Bee and wasp venoms. Science 177, 314-318. Habermann, E. and Jentsch, J . (1967). Sequenzanalyse des Melittins aus den tryptischen und peptischen Spaltstucken. Hoppe Seylers Z . Physiol. Chem. 348, 37-50. Hall, H. G. (1986). DNA differences found between Africanized and European honeybees. Proc. Natl. Acad. Sci. USA 83. 4874-4877. Hall, H. G. (1988). Characterization of the African honey-bee genotype by DNA restriction fragments. I n "Africanized Honey Bees and Bee Mites" (Eds G. R. Needham, R. E. Page, M. Delfinado-Baker and C. E. Bowman), pp. 294-302, Ellis Horwood, Chichester. Hall, H. G. (1990). Parental analysis of introgressive hybridization between African and European honeybees using nuclear DNA RFLPs. Genetics 125, 611-621. Hall, H. G. (1992). Further characterization of nuclear DNA RFLP markers that distinguish African and European honeybees. Arch. Insect. Biochem. Physiol. 19, 163-175. Hall, H. G. and McMichael, M. A. (1992). European honey bee (Apis mellifera L.) (Hymenoptera: Apidae) colonies kept at high elevations in Costa Rica tested for African DNA markers. Bee Science 2, 25-32. Hall, H. G. and Muralidharan, K. (1989). Evidence from mitochondrial DNA that African honeybees spread as continuous maternal lineages. Nature 339, 21 1-213. Hall, H. G. and Smith, D. R. (lY91). Distinguishing African and European honeybees matrilines using amplified mitochondrial DNA. Proc. Natl. Acad. Sci. USA 88, 45484552. Harrison, J . F. and Hall, H. G. (1993). African-European honeybee hybrids have low nonintermediate metabolic capacities. Nature 363, 258-260. Harrison, R. G. (1989). Animal mitochondrial DNA as a genetic marker in population and evolutionary biology. Trends Ecol. Evol. 4, 6 1 1. Hoeh, W. R., Blakley, K . H. and Brown W. M . (1991). Heteroplasmy suggests limited biparental inheritance of Mytilus mitochondrial DNA. Science 251, 1488- 1490. Hoshiba, H. and Kusangi, I . (1978). Karyological study of honeybee. J . Apic. Res. 17, 105-109. Hoshiba, H. and Okada, I . (1986). G-banding analysis of male chromosomes in Apk cerana and Apis mellifera Iigusrica. Apidologie 17, 101-106. Hovemann, B.. Richter, S . , Walldorf. U. and Cziepluch, C. (1988). Two genes encode related cyotoplasmic elongation factors la (EF-la) in Drosophila melanogaster with continuous and stage specific expression. Nucleic Acids Res. 16, 3 175-3 194. HsuChen, C.-C. and Dubin, D. T. (1084). A cluster of four transfer RNA genes in mosquito mitochondrial DNA. Biochem. Int. 8. 385-391. Hunt, G. J . and Page, R. E. (1993). Patterns of inheritance with RAPD molecular markers reveal novel types of polymorphism in the honey bee. Theor. Appl. Genet. 85, 15-20. Johansson, I. (1980). Meilensteine der Genetik. Paul Parey, Hamburg, Berlin. Jukes, T. H. and Bhushan, V. (1986). Silent nucleotide substitutions and G+C content of some mitochondrial and bacterial genes. J . Mol. Evol. 24, 39-44. Karr, T. L., Weir, M. P., Ali, Z . and Kornberg, T. (1989). Patterns of engrailed expression in early Drosophila embryos. Development 105, 605-612. Kerr, W. E. (1957). Introduqao de abelhas africanas no Brasil. Brasil Apicola 3, 21 1-213. Kerr, W. E. and Bueno D. (1970). Natural crossing between Apis rnellifera adansonii and Apis mellifera ligustica. Evolution 24, 145-148.

MOLECULAR BIOLOGY OF THE HONEYBEE

145

Koeniger, N., Koeniger, G.. Tingek, S . , Mardan, M. and Rinderer, T. E. (1988). Reproductive isolation by different time of drone flight between Apis cerana (Fabricius 1793) and Apis koschevnikovi (Buttel-Reepen 1906). Apidofogie 19, 103-106. Koeniger, N.. Koeniger, G. and Wongsiri S. (1989). Mating and sperm transfer in Apis porea. Apidologie 20, 413418. Kondo, R.. Satta. Y . , Matsuura. E. T., Ishiwa, H., Takahata, N. and Chigusa, S. I. (1990). Incomplete maternal transmission of mitochondrial D N A in Drosophila. Genetics 126, 657463. Kornberg, T., Siden I . , O’Farrell, P. H. and Simon, M. (1985). The engrailed locus of Drosophila: in situ localization of transcripts reveals compartment-specific expression. Cell 40, 45-53. Kreil, G. (1990). Processing of procursors by dipeptidylaminopeptidase: a case of molecular ticketing. Trends Biochem. Sci. 15, 23-25. Kreil, G., Haiml. L. and Suchanek. G. (1980). Stepwise cleavage of the pro part of Promelittin by dipeptidylpeptidase IV. Evidence for a new type of precursorproduct conversion. Eur. J . Biochem. 111, 49-58. Kuchler, K.. Gmachl, M., Sippl, M. J. and Kreil, G . (1989). Analysis of the cDNA for phospholipase A? from honeybee venom glands. The deduced amino acid sequence reveals homology to the corresponding vertebrate enzymes. Eur. J . Biochem. 184, 249-254. Laidlaw, H . H. (1944). Artificial insemination of the queen bee (Apis melfifera L . ) : Morphological basis and results. J . Morphof. 74, 429-465. Laidlaw, H . H . and Page, R . E. (1984). Polyandry in honey bees (Apis mellifera L.): sperm utilization and intracolonial relationships. Genetics 108, 985-997. Lawrence, P. A , , Johnston, P., Macdonald, P. and Struhl, G. (1987). Borders of parasegments in Drosophila embryo are delimited by fushi tarazu and even skipped genes. Nature 328, 44W42. Lobo, J. A . , Del Lama, M. A . and Mestriner, M. A . (1989). Population differentiation and racial admixture in the Africanized honey bee (Apis mellifera L.). Evolution 43, 794-802. Maa, T. C. (1953). An inquiry into the systematics of the tribus Apidini or honeybees (Hym.). Treuhia 21, 525-640. Mackensen, 0. (1947). Effect of carbon dioxide on initial oviposition of artificially inseminated and virgin queens. J . Econ. Entomol. 40, 344-349. Martinez-Arias, A . and Lawrence, P. A . (1985). Parasegments and compartments in the Drosophila embryo. Nature 313, 639442. Marz, R., Mollay, C . and Kreil, G. (1981). Queen bee venom contains much less phospholipase than worker bee venom. Insect Biochem. 11, 685490. Meixner, M. D . , Sheppard, W. S. and Poklukar, J. (1993). Asymmetrical distribution of a mitochondrial D N A polymorphism between 2 introgressing honey bee subspecies. Apidologie 24. 147-153. Meland, S., Johansen. S . , Johansen, T . , Haugli, K . and Haugli, F. (1991). Rapid disappearance of one parental mitochondria1 type after isogamous mating in the myxomycete Physarum polycephalum. Curr. Genet. 19, 55-60. Meusel. M. S. and Moritz. R . F. A . (1992). Mitochondria1 D N A length variation in the cytochrome oxidase region of‘ honeybee (Apis mellifera L.). 2pidologie 23, 147-1 50. Meusel, M. S. and Moritz, R. F. A. (1993). Transfer of paternal mitochondrial DNA during fertilization of honeybee (Apis mellifera L.) eggs. Curr. Genet. 24, 539-543. Michener, C . D . (1944). Comparative external morphology, phylogeny, and classification of bees (Hymenoptera). BUN. A m . Mus. Nut. Hist. 82, 151-326.

146

R. F. A. MORll

Michener, C. D. (1974). “The Social Behavior of the Bees”. Harvard Universil Press, Cambridge, MA. Michener, C. D. (1975). The Brazilian bee problem. Annu. Rev. Entornol. 2( 399-416. Michener, C. D. (1990). Classification of the Apidae. Univ. Kans. Sci. Bull. 54 75-1 64. Moritz, C. F. and Brown W. M. (1987). Evolution of animal mitochondria1 DNA relevance for population biology and systematics. Ann. Rev. Ecol. Syst. 18 269-292. Moritz, R. F. A. (1989). Colony level and within colony level selection in honeybees A two allele population model for Apis rnellifera capensis. Behav. Ecol. Sociobioi. 25, 437-444. Moritz, R. F. A. and Meusel, M. S. (1992). Mitochondrial DNA polymorphism in Africanized honeybees Apis rnellifera L.: theoretical model and empirical evidence. i . Evol. Biol. 5, 71-81. Moritz, R. F. A. and Southwick, E. E. (1992). “Bees as Superorganisms. An Evolutionary Reality”. Springer Verlag, Berlin. Moritz, R. F. A , , Hawkins, C. F., Crozier, R. H. and Mackinley, A. G . (1986). A mitochondrial DNA polymorphism in honeybees (Apis rnellifera L.). Experientia 42, 322-324. Moritz, R. F. A., Cornuet. J. M., Kryger, P. and Hepburn, H. R. (1994). Mitochondrial DNA size variability in African honeybees (Apis rnellifera). Apidologie 25, 169-178. Moritz, R. F. A , , Cornuet, J . M., Kryger, P. and Hepburn, H. R. (1994). Mitochondrial DNA size variability in African honeybees (Apis rnellifera). Apidologie 25. 169-178. Nachtsheim, H. (1912). Parthenogenese, Eireifung und Geschlechtsbestimmung bei der Honigbiene. Sitzungsber. Ges. Morph. Physiol. Munchen 28, 22-29. Nachtsheim, H. (19 14). Cytologische Studien uber die Geschlechtsbestimmung bei der Honigbiene (Apis rnelllfica). Arch. Zellforsch. 11, 169-241. Nusslein-Vollhard, C. and Wieschaus, E. (1980). Mutations affecting segment number and polarity in Drosophila. Nature 287, 795-801. Nusslein-Vollhard. C., Wieschaus, E. and Jurgens, G. (1982). Segmentierung bei Drosophila - eine genetische Analyse. Verh. Dtsch. Zool. Ges. 1982, 91-104. Oldroyd, B. P., Sheppard, W. S. and Stelzer, J . A. (1992). Genetic characterization of the bees of Kangaroo Island, South Australia. J . Apic. Res. 31, 141-148. Oldroyd, B. P., Sylvester, H. A , , Wongsiri, S . , Kleinpeter, S. J. and Rinderer, T-.E. (1994). Task specialization in a wild bee, Apis florea, revealed by RFLP banding. Behav. Ecol. Sociobiol. 34, 25-30. Ojala, D., Montoya, J. and Attardi, G. (1981). tRNA punctuation model of RNA processing in human mitochondria. Nature 290, 470-474. Otis, G. W. (1982). Population biology of the Africanized honey bee. In “Social Insects in the Tropics” (Ed. P. Jaisson). pp. 209-219. Universite Paris-Nord, Paris. Otis, G. W. (1991a). A review of the diversity of species within Apis. In “Diversity in the Genus Apis” (Ed. D. R. Smith), pp. 29-49. Westview Press, Boulder, CO. Otis, G . W. (1991b), Population biology of the Africanized honey bee. In “The ‘African’ Honey Bee” (Eds M. Spivak, D. J. C. Fletcher and M. D. Breed), pp. 213-234, Westview Press, Boulder, CO. Page, R. E. and Robinson, G. E. (1991). The genetics of division of labour in honey bee colonies A d v . insect. Physiol. 23, 117-171. Pamilo, P., Varvio-Aho, S. and Pekkarinen, A. (1978). Low enzyme gene variability in Hymenoptera as a consequence of haplodiploidy. Hereditas 88, 93-99.

MOLECULAR BIOLOGY OF THE HONEYBEE

147

Patel, N. H., Kornberg, T. B. and Goodman, C. S. (1989). Expression of engrailed during segmentation in grasshopper and crayfish. Development 107, 201-212. Petrunkewitsch, A. (1901). Die Richtungskorper und ihr Schicksal im befruchteten und unbefruchteten Bienenei. Zool. Jahrb. 14, 573-608. Petrunkewitsch, A. (1903). Das Schicksal der Richtungskorper im Drohnenei. Ein Beitrag zur Kenntnis der naturlichen Parthenogenese. Zool. Jahrb. Abt. Anat. Ontog. Tiere 17, 481-516. Plant, J. D. and Paulus, H. F. (1987). Comparative morphology of the postmentum of bees (Hymenoptera: Apidea) with special remarks on the evolutions of the lorum. Z . zool. Syst. Evolutionsforsch. 25, 81-103. Rinderer, T. E. (1986). Africanized bees: The Africanization process and potential range in the United States. Bull. Entomol. SOC.Am. 32, 22-227. Rinderer, T. E., Hellmich, R. L., Danka, R. G. and Collins, A. M. (1985). Male reproductive parasitism. A factor in the Africanization of honeybee populations. Science 228, 1119-1121. Rinderer, T. E., Koeniger, N., Tingek, S . , Mardan, M. and Koeniger, G. (1989). A morphological comparison of the cavity dwelling honeybees of Borneo Apis koschevnikovi (Buttel-Reepen, 1906) and A . cerana (Fabricius, 1793). Apidologie 20, 40541 1. Rinderer, T. E., Stelzer, J. A., Oldroyd, B. P., Buco, S. M. and Rubink, W. L. (1991). Hybridization between European and Africanized honeybees in the neotropical Yucatan peninsula. Science 253, 309-311. Rothenbuhler, W. C. (1964). Behaviour genetics of nest cleaning in honeybees. Anim. Behav. 12, 578-583. Ruttner, F. (1988). “Biogeography and Taxonomy of Honeybees.” Springer Verlag, Berlin, Heidelberg, New York. Ruttner, F., Kauhausen, D. and Koeniger, N. (1989). Position of the red honey bee, Apis koschevnikovi (Buttel-Reepen 1906) within the genus Apis. Apidologie 20, 395404. Saitou, N. and Nei, M. (1987). The neighbour-joining method: a new method for reconstructing phylogenetic trees. Mol. Biol. Evol. 4, 406425. Sakagami, S. F., Matsumura, T. and Ito, K. (1980). Apis laboriosa in Himalaya. The little known world’s largest honeybee (Hymenoptera, Apidae). Insecta Matsumurana New Series 19, 47-77. Sheppard, W. S. and Berlocher, S. H. (1989). Allozyme variation and differentiation among four Apis species. Apidologie 20, 419431. Sheppard, W. S. and McPheron, B. A. (1991). Ribosomal DNA diversity in Apidae. In “Diversity in the Genus Apis” (Ed. D. R. Smith), pp. 89-102. Westview Press, Boulder, CO. Sheppard, W. S., Rinderer, T. E., Mazzoli, J. A., Stelzer, J. A. and Shimanuki H. (1991a). Gene flow between African- and European-derived honey bee populations in Argentina. Nature 349, 782-784. Sheppard, W. S., Soares, A. E. E., De-Jong, D. and Shimanuki H. (1991b). Hybrid status of honeybee populations near the origin of Africanization in Brazil. Apidologie 22, 643-652. Shipolini, R. A., Callewaert, G. L., Cottrell, R. C . and Vernon, C. A. (1971). The primary sequence of phospholipase-A from bee venom. FEBS Lett. 17, 39-40. Smith, D. R. (1988). Mitochondrial DNA polymorphisms in five Old World subspecies of honey bees and in New World hybrids. In “Africanized Honey Bees and Bee Mites” (Eds G. R. Needham, R. E. Page, M. Delfinado-Baker, C. E. Bowman), pp. 303-312. Ellis Horwood, Chichester. Smith, D. R. (1991). Mitochondrial DNA and honeybee biogeography. In “Diversity in the Genus Apis” (Ed. D. R. Smith), pp. 131-176. Westview, Boulder, CO.

148

R.

F. A. MORITZ

Smith, D. R. and Brown, W. (1990). Restriction endonuclease cleavage site and length polymorphisms in Apis mellifera mellifera and A . m. carnica (Hymenoptera Apidae). Ann. Entomol. SOC. Am. 83, 81-88. Smith, D. R., Taylor, 0. R. and Brown, W. M. (1989). Neotropical Africanized honey bees have African mitochondrial DNA. Nature 339, 213-215. Sommer, R. J., Retzlaff, M., Goerlich, K., Sander, K. and Tautz, D. (1992). Evolutionary conservation pattern of zinc-finger domains of Drosophila segmentation genes. Proc. Natl. Acad. Sci. USA 89, 10782-10786. Spear, B. B. (1974). The genes for ribosomal RNA in diploid and polytene chromosomes of Drosophila melanogaster. Chromosoma 48, 159-179. Spivak, M., Fletcher, D. J. C. and Breed, M. D. (Eds) (1991). “The ‘African’ Honey Bee”. Westview Press, Boulder, CO. Suchanek, G., Kreil, G. and Hermodson, M. A. (1978). Amino acid sequence of honeybee prepromelittin synthesized in vitro. Proc. Natl. Acad. Sci. USA 75, 701-704. Swofford, D. L. (1990). PAUP: Phylogenetic Analysis Using Parsimony. Version 3.0. Illinois Natural History Survey, Champaign, IL. Sylvester, H. A . (1986). Biochemical genetics. In “Bee Genetics and Breeding” (Ed. T. E. Rinderer), pp. 177-203, Academic Press, Orlando, FL. Tautz, D., Hancock, J. M., Webb, D. A., Tautz, C. and Dover, G. A . (1988). Complete sequence of the rRNA genes of Drosophila melanogaster. Mol. Biol. Evol. 5, 366-376. Taylor, 0. R. (1985). African bees: potential impact in the United States. Bull. Entomol. SOC.Am. 31, 15-24. Tingek, S. M., Mardan, M., Rinderer, T. E., Koeniger, N. and Koeniger, G. (1988). Rediscovery of Apis vechti (Maa 1953): the Saban honey bee. Apidologie 19, 97-1 02. Tucker, K. W. (1986). Visible mutants. In “Bee Genetics and Breeding” (Ed. T. E. Rinderer), pp. 57-90, Academic Press, Orlando, FL. Uhlenbusch, I . , Rippe. R. M. and Gelissen (1987). Mitochondria1 tRNA genes in Locusta migratoria. Naturwissenschaften 74, 142-143. Vlasak, I . , Burgschwaiger, S. and Kreil, G. (1987). Nucleotide sequence of the large ribosomal RNA of honeybee mitochondria. Nucleic Acids Res. 15, 2388. Vlasak, R. and Kreil, G. (1984). Nucleotide sequence of cloned cDNAs coding for preprosecapin, a major product of queen-bee venom glands. Eur. J . Biochern. 145, 279-282. Vlasak, R., Unger-Ullmann, C., Kreil, G. and Frischauf A-M. (1983). Nucleotide sequence of cloned cDNA coding for honeybee prepromelittin. Eur. J . Biochern. 135, 123-126. Vlasak, R., Malec, I . and Kreil, G. (1986). cDNA encoding precursors of the bee venom peptides melittin and secapin. In “Molecular Cloning of Hormone Genes” (Ed. J. F. Habener), pp. 405412. Humana Press, Clifton, NJ. von Frisch, K. (1965). “Tanzsprache und Orientierung der Bienen.” Springer Verlag, Berlin, Heidelberg, New York. Walldorf, U . and Hovemann, B. T. (1990). Apis mellifera cytoplasmic elongation factor la (EF-la) is closely related to Drosophila melanogaster EF-la. FEBS Lett. 267, 245-249. Walldorf, U., Fleig, R. and Gehring, W. J . (1989). Comparison of homeoboxcontaining genes of the honeybee and Drosophila. Proc. Nad. Acad. Sci. USA 86, 9971-9975. Willis, L. G., Winston. M. L. and Honda, B. M. (1992). Phylogenetic relationships in the honeybee (genus Apis) as determined by the sequence of the cytochrome oxidase region of mitochondrial DNA. Mol. Phylogenet. Evol. 1, 169-178.

MOLECULAR BIOLOGY OF THE HONEYBEE

149

Winston, M. L. and Michener, C. D. (1977). Dual origin of highly social behavior among bees. Proc. Natl. Acad. Sci. USA 74, 1135-1137. Wolstenholme, D. R. (1992). Animal mitochondria1 DNA: structure and evolution. Int. Rev. Cytof. 141, 173-216. Wongsiri, S . , Limbipichai, K., Tangkanasing, P., Mardan, M., Rinderer, T. E., Sylvester, H. A., Koeniger, G. and Otis, G. (1990). Evidence of reproductive isolation confirms that Apis andreniforrnis (Smith 1858) is a separate species from sympatric Apis florea (Fabricius 1787). Apidofogie 21, 47-52. Woyke, J. and Hillesheim, E. (1990). Genetic aspects of instrumental insemination. In “The Instrumental Insemination of the Queen Bee” (Ed. R. F. A. Moritz), pp. 105-123. Apimondia, Bucharest. Wu, Y. and Kuang, B. (1987). Two species of small honeybee-a study of the genus Micrapis. Bee World 68, 153-155.

This Page Intentionally Left Blank

Information Processing in the Insect Ocellar System: Comparative Approaches t o the Evolution of Visual Processing and Neural Circuitsa Makoto Mizunami laboratory of Neuro-Cybernetics, Research Institute for Electronic Science, Hokkaido University, Sapporo 060,Japan a

This chapter is dedicated to Prof. Hideki Tateda, Shimonoseki City University, Japan.

1 Introduction 152 2 Distribution and structure of insect ocelli 154 2.1 Phylogeny and distribution 154 2.2 Structure 157 2.3 Ontogenetic development 161 2.4 Summary 162 3 Functional properties of ocellar neurones 162 3.1 Spatial properties 162 3.2 Absolute sensitivity 163 3.3 Detection of absolute intensity levels 164 3.4 Speed of signal transmission 164 3.5 Spectral sensitivity 165 3.6 Polarization sensitivity 168 3.7 Summary 168 4 Behavioural roles of ocelli 169 4.1 Ocelli as a stirnulatory organ 169 4.2 Contribution to phototactic orientation 169 4.3 Visual course control in flight 170 4.4 Orientation toward edges 171 4.5 Light intensity perception for the control of diurnal activity 173 4.6 Detection of polarized light 177 4.7 Control of neuroendocrinic secretion 177 4.8 Summary 179 5 Neural organization of ocellar pathways 179 5.1 Synaptic organization of ocellar plexus 180 5.2 Morphology of second-order ocellar neurones 181 5.3 Synaptic organization of ocellar tract neuropil 189 5.4 Synaptic organization of posterior slope neuropil 193 5.5 Morphology of third-order ocellar neurones 195 5.6 Diversity and evolution of neural circuits of ocellar systems 199 5.7 Summary 203 ADVANCES IN INSECT PHYSIOLOGY VOL. 25 ISBN &12-024225-7

Copyright 01994 Academic Press Limited of reproduction in any form reserved

A / / rights

152

MAKOTO MIZUNP

6 Molecular basis of the ocellar system 204 6.1 Visual pigment and its molecular evolution 204 6.2 Neurotransmitters and neurohormones 206 6.3 Structural mutants in Drosophila 210 7 Information processing in the ocellar system 21 1 7.1 Information processing in the ocellar plexus 211 7.2 Information processing in second-order neurones 218 7.3 Modulatory roles of efferent neurones 228 7.4 Signal processing between second- and third-order neurones 231 7.5 Multimodal integration in third-order and higher ocellar neurones 236 7.6 Principles of visual processing in an ocellar system 239 8 Comparative approaches to the evolution of visual systems 244 8.1 Some higher visual functions are explained by modifying neural circuits fa simpler visual functions 245 8.2 Evolution of neural circuits by modifying simpler circuits 247 Acknowledgements 252 References 252

1 Introduction

Most adult insects possess two or three dorsal ocelli in addition to a pair oi compound eyes. The dorsal ocelli differ in their ontogenetical origin from both larval lateral ocelli of holometamorphic insects and from compound eyes. The compound eyes are sophisticated visual organs responsible for functions that require good spatial resolution, such as the perception of movement, fixation of objects and pattern recognition. In contrast, the dorsal ocelli are simple photoreceptive organs with a very poor spatial resolution. Why do insects need simple photoreceptors, such as ocelli, even though they have sophisticated compound eyes? Do these simple eyes have advantages over coexisting sophisticated eyes in some visual behaviours? Although these questions have remained unresolved for a long time, some of the answers have emerged from recent extensive anatomical, physiological and behavioural studies. Most earlier behavioural works concluded that the ocelli were rarely responsible for initiating a motor response on their own, but appeared to participate in the modulation of phototactic reactions mediated by compound eyes. Thus, the ocelli have been labelled as photo-stimulatory organs, whose functions are restricted to influencing the activity of the central nervous system (Wolsky, 1930, 1933). An important breakthrough toward elucidating the direct behavioural roles of ocelli came from the work of Wilson (1978a) who examined anatomical and physiological properties of locust ocelli. He proposed that the ocelli are most suited to detect the contrast between the earth and the sky, and thus in detecting instability in flight. This hypothesis, originally proposed by Hesse (1908), was supported by subsequent behavioural studies (Stange and Howard, 1979; Stange, 1981; Taylor, 1981a,b), and it is now established that one of

P L A T E 1 External structures of ocelli in ;I n u m b e r of representative insects. Gcnerally three ocelli arc grouped in ;I triangle o n t o p of the head. hut there is ;I wide variation in their n u m b e r and porition\ Front;il (a) a n d front-lateral ( h ) view of the head of :I dragonlly. Swriptwfiin kiiffckrli (Lihellulidac) ( x X ) . T h e lateral and median ocelli cover different fields of view in the horizontal pl;ine. ( c ) A dorsal view of the head o f a dragonfly. Lexk7.s .spotisu (Lcstidac). The three ocelli cover almost 360" fields of view i i i thc horizntal plane ( X 11). ( d ) A frontal view of the head of ii locust. I.ocii.srii inigriiroriu ( X Y ) . ( e ) T h e head o f a cwkronch. Peripluitrtii ufmwcufiii. Frontnl view ( X Y ) . ( f ) T h e head of a stinkbug. t ' i w u f o r i i i i j u p ~ t i i i c i i .l)or\al view ( X I S ) . (9) T h e head of a m o t h , Nhcfiiidii sp. Front-lateral view ( x I Y ) . ( h ) Dorsal view of the head of ;I winged male ant. Forniiciificxii jiiptiicii. Dorsal view ( X 2 4 ) . T h e occlli a r c ldrgcr in \ize i n the winged forms than in w i n g l e s soldiers in this species. a s a r e In many other ants.

PLATE 2 External structures of ocelli in a number of representative insects. Frontal (a) and front-lotcriil ( b ) view of the head of a bee. Apis niellifercr ( X 12). (c) A frontal view of the head o f ii wasp. Vrspirla sp ( X 16). (d) The ocelli of a bee. Ichnrurnonidn sp.. is large relative to the size of the head ( ~ 2 3 )(e. . f ) Dorsal view of the head of a female ( c ) and a male (f) of a hoverfly. Eri.smlis cercwlis (Syrphidae) ( X 13). Front-lateral (g) and dorsal ( h ) view of the head o f a male hihionid Hy. Bihioiiicicr sp. T h e ocelli are located in an elevated position at the vertex o f the head. The three ocelli cover almost 360" field of view in the horizontal plane (X22).

THE DIVERSITY OF INSECT OCELLAR SYSTEMS

153

the major roles of insect ocelli is to contribute to flight steering, at least in dragonflies and locusts. The neural mechanisms subserving the ocellar contribution to flight steering have now been established (Simmons, 1980; Reichert et al., 1985; Griss and Rowell, 1986; Rowell and Reichert, 1986). Further behavioural studies made on a number of insect species suggest that the behavioural roles of ocelli are multiple and differ in different insects. During the 1970s and 1980s, the neural organization of ocellar systems was examined in a number of insects by cobalt-filling of the ocellar nerve and using electron microscopy. These studies revealed an unexpected complexity in the neural wiring of ocellar systems, thereby demonstrating that the ocelli are far from rudimentary organs and play important roles in the survival of insects. These studies also showed that the neural organization of ocellar systems differs among different insects, thereby supporting the diversity in ocellar functions. The ocellar system contains some of the largest neurones in the insect brain, and the accessibility of these interneurones for microelectrode penetration makes the ocellar system a suitable model system to reveal basic

principles of information processing in visual systems. Following the pioneering work of Chappell and Dowling (1972), who studied the neural basis of visual processing in the ocellar retina, a number of researchers have analysed the electrophysiological activities of ocellar neurones. Recently, detailed analyses of information processing have been made using white noise and sinusoidally modulated light in the cockroach ocellar system. These studies successfully revealed some of the basic principles of neural processing, a part of which may be applicable to more advanced visual systems, including insect compound eyes and vertebrate retinas. The aim of this article is twofold. First, it attempts to synthesize the large volume of present knowledge on the anatomy, physiology, and behavioural roles of insect ocellar systems to provide a search image for this still enigmatic visual system. In this respect, the aim of this review is to update the reviews of Goodman (1970, 1981). The second aim of this article is to theorize on the diversities of ocellar systems. Surprisingly, little attention has been paid to the fact that huge diversities exist in the anatomy, physiology and behavioural roles among insect ocelli which, I believe, prevents a full understanding of the functions of insect ocelli. Here I summarize the diversities of ocellar systems in their structure, neural organization, physiology and behavioural roles and attempt to interpret them in terms of phylogenetic pathways and adaptations to the environment. This attempt is also based on my belief that insect ocellar systems will prove to be a rich field from which to study the evolution of neural circuits. The evolution of neural systems is one of the focal points of neurobiology, but the lack of adequate model systems has hindered the extension of this important research field. Earlier studies of insect ocelli have been compre-

MAKOTO MlZUNAMl

154

hensively reviewed by Goodman (1970, 1975, 1981), and some aspects of insect ocelli have been concisely reviewed by several authors (Wilson, 1978a; Laughlin, 1981; Wehner, 1981; Mobbs, 1985; Toh and Tateda, 1991). 2

2.1

Distribution and structure of insect ocelli PHYLOGENY A N D DISTRIBUTION

Possession of two types of visual organs, i.e. compound eyes (lateral faceted eyes) and ocelli (median single-lens eyes) are typical characteristics of the major arthropod groups. Typical compound eyes and ocelli are seen in Silurian trilobites (Paulus, 1979). Paulus (1979) proposed that the three ocelli of the pterygote insects, i.e. two lateral and one median ocelli, are homologous to the two pairs of ocelli of Limulus, i.e. two median eyes and two endoparietal eyes, the four neuplius eyes in Crustacea, and the four frontal ocelli of apterygote insects, Collembola. According to this hypothesis, the ancestors of insects possessed four ocelli, and two of them were fused to form the median ocelli of pterygote insects. Indeed, Caesar (1913), Mobbs (1979) and Eaton (1983a) noted that the median ocelli of some pterygote insects are formed by fusion of two ocellar primordia during ontogenetic development. Ocelli are seen in some orders of apterygote insects and in most, but not all, orders of pterygote insects (Fig. 1). According to Kalmus (1945), all Palaeoptera possess ocelli whereas in many orders of Neoptera, both ocellate and anocellate species occur (Fig. l ) , thereby suggesting that ocelli have frequently been lost at various stages of evolution, after the possession of ocelli has been established as a typical characteristic in an early stage of evolution of Pterygota. Three ocelli are the usual number in pterygote insects, but sometimes only two ocelli are present, as for example, in cockroaches, Peripfaneta (Plate le), stinkbug, Puntaroma (Plate If) and moths (Plate lg). In Coleoptera, the possession of dorsal ocelli is considered to be plesiomorphic and the number of ocelli is never three (Crowson, 1981). Recently, an individual of Lesteva sp. (Omaliinae, Oxytelidae) which possesses three ocelli was discovered (Naomi, 1987). As this species usually possesses a pair of ocelli on the vertex, this phenomenon is probably regarded as an example of atavism, suggesting that the primitive adult Coleoptera might have three ocelli like other insects. Ocelli are located upon the frons (e.g. Plate la,b,d and e) or the vertex (e.g. Plate If, and Plate 2c,e,f), in a position that normally affords them horizontal or dorsal fields of view. The three ocelli cover very wide fields of view. In some insects, such as the dragonfly, Lestidae (Plate lc) and the fly, Bibionidue (Plate 2g,h), the three ocelli cover nearly 360" of the horizontal

THE DIVERSITY OF INSECT OCELLAR SYSTEMS APTERYGOTA

155

PTERYGOTA 'PALAEOPTERA

NEOPTERA

I POLYNEOPTERA ' PARANEOPTERA

HOLOMETABOLA

I

I

P s A n He T h St Co Ne Hy Tr Le Mc Ap Di

Pr T y I

II -_ --- - ---T---J

FIG. 1 The distribution of wings and ocelli in the insect orders. (B) All species no species possess wings possess wings (upper column) or ocelli (lower column); (0) (upper) or ocelli (lower); ( 0 )some species possess wings (upper) or ocelli (lower) but others do not. CI, Collernbola; Pr, Protura; Ty, Thysanura; Ep, Ephemeroptera; Od, Odonata; PI, Plecoptera; Em, Embioptera; Or, Orthoptera; De, Dermaptera; Is, Isoptera; Ps, Psocoptera; An, Anoplura; He, Herniptera; Th, Thysanoptera; St, Strepsiptera; Co, Coleoptera; Ne. Neuroptera; Hy, Hymenoptera; Tr, Trichoptera; Le, Lepidoptera; Mc, Mecoptera; Ap, Aphaniptera; Di, Diptera. Based on Kalmus (1945). The dendrograrn is based on Kristensen (1981).

plane. In most other insects, however, the ocellar field of view is restricted behind and beneath the insect. The distribution of ocelli, both within a class and within particular orders, has been examined in the search for clues to ocellar function by several authors. Kalmus (1945) found that the correlative development of ocelli and wings exists in respect of insect orders (Fig. 1). The dendrogram in Fig. 1 is based on Kristensen (1981). Of the four orders where exclusively winged species occur, two (Ephemeroptera, Odonata) also have ocelli, whereas the species of the other two orders (Neuroptera, Trichoptera) may or may not have them. Of the five entirely wingless orders, three (Pronura, Anoplura, Aphaniptera) have only anocellate species and two (Thysanura, Collembola) have ocellate and anocellate species. No order is completely ocellate and wingless o r anocellate and winged, nor is there an order in which all species are ocellate but some wingless. In 11 orders winged and wingless as well as ocellate and anocellate species occur.

156

MAKOTO MlZUNAMl

Kalmus (1945) also reported that a correlative development of ocelli and wings exists within particular insect orders, such as Strepsiptera and Thysanoptera, where both ocellate and anocellate as well as winged and wingless species exist. An example of a strong correlation between ocelli and wings is seen in the castes of Austrian termites (Hill, 1942). Five species of Culotermes have not lost but only reduced the ocelli in the wingless soldiers, 121 species have lost them in this caste but retained them in the winged sexual forms, and nine species have lost them altogether. Not a single species is recorded where the wingless soldiers have ocelli (or compound eyes) while the winged adults have none; most species have ocellate adults and anocellate soldiers. Thus, Kalmus (1945) concluded that if it is assumed that termites are descendants of winged and ocellate casteless ancestors, it would appear that among the Austrian species none has lost the ocelli in the winged adult before it has lost them in the wingless soldier. Kalmus (1945) suggested that one of the selective advantages of ocelli to flying insects might be in orientating the flying insect with its back to the sky, a hypothesis which is confirmed at least in some insects (see Section 4.3). In holometamorphic insects, wings and ocelli develop during the pupal stage and are wholly absent in the larvae, but in hemimetabolous insects both structures develop gradually in the last instar larvae. Thus, the absence or reduction of one or both of them can be explained as a larval feature and the reduced form can then be described as neotenic (Kalmus, 1945). Kerfoot (1967a,b) found a correlation between ocellar size and the light intensity during foraging activities in bees. The occurrence of enlarged ocelli in nocturnal foraging bees (Michener, 1944) has led to a quantitative study of the relation between ocellar size and foraging activity in a number of species of bees which shows a negative correlation between ocellar size and the light intensity threshold for foraging. An example of large ocelli in Hymenoptera is shown in Plate 2d. Ocelli also increase with head size, but this relationship is not as good a predictor of ocellar size as is the light intensity threshold. No nocturnal foraging bees have small head size, and it is suggested that this may be the result of selection for larger ocelli for bees that forage at very low light intensities. The roles of ocelli in low light intensity perception is discussed in Sections 3.3 and 4.5. Some Lepidoptera possess a pair of internal ocelli (Eaton, 1971), reduced ocelli located within the head capsule. Electroretinogram (ERG) recordings from internal ocelli suggest that they are functional (Eaton, 1971; Pappas and Eaton, 1977). It appears that the internal ocelli monitor light spread into the head capsule. Some moths, such as the sphingid moth, Munduca sextu, possess pairs of external and internal ocelli, thus possessing a total of four ocelli (Dickens and Eaton, 1973).

157

THE DIVERSITY OF INSECT OCELLAR SYSTEMS

A

CT

ON

F

(TO BR&lN)

FIG. 2 The structure of the insect ocellus. (A) A schematic drawing of the longitudinal section through two lateral ocelli of a fleshfly. CL, corneal lens; CT, cuticle; C G , corneagenous cell; RH, rhabdom; RC, retinula cell; PC, pigment cells surrounding outside of the rhabdomal layer; BM, basement membrane; SON, distal process of the second-order neurone; ON, ocellar nerve. From Toh et af. (1971). (B, C) A schematic comparison of the optical system of the ocellus (B) and the compound eye (C). Dragonfly ocellus and its underfocused optical system is shown in (B). The dashed and dotted line marks the optical axis of the ocellus. Light entering the ocellus from any particular direction is distributed over an extended part of ocellar retina rather than focused on a single retinal point, while in the compound eye images are formed in an array of photoreceptors (C). Cc, corneagen cells; F, focal point; L, lens; N , ocellar nerve; Np, ocellar neuropil; Rt, rhabdomal layer. (B) is from Wehner (1987); (C) is from Land and Fernald (1992).

2.2

STRUCTURE

Dorsal ocelli possess high-aperture dioptrics which exhibit wide visual fields (Cornwell, 1955). The focal lengths of the ocellar lens have been estimated for about 20 species (review by Goodman, 1981) and in all cases the image plane appears to lie well behind the retinal layer as illustrated in Fig. 2B. In the fly, Culliphora erythrocephulu, for example, the microvillar zone of the lateral ocellus begins 7-19 pm behind the inner surface of the lens and has a thickness of about 8 p m , whereas the focal plane is located at 58f21 pm behind the lens (Schuppe and Hengstenberg, 1993). The ocellar lens of Cufliphoru also has a second focal plane at 96_+29p m behind the lens. Due to underfocusing, an object entering the ocellar visual field results in a change in the light intensity impinging on photoreceptor layers rather than

158

MAKOTO MlZUNAMl

the formation of an image. The output of a large number of photoreceptor cells converges upon a small number of large second-order neurones, called L-neurones. Such a system is most suited for the detection of small changes of light intensity integrated over its wide visual field (Wilson, 1978a). However, the possibility of coarse image formation by ocelli cannot be fully ruled out, at least in flies (Schuppe and Hengstenberg, 1993). Beneath the lens, the ocellar receptor cells (Fig. 2A,B) are not arranged in an ordered array as in the compound eye where its rhabdomeres are presented as a mosaic of light guides for axial light (Fig. 2C). In the ocellus the rhabdomeres of the irregularly packed cells form a network of baffles (Wilson, 1978a), presumably maximizing their ability to trap light spread diffusely through the retina. A variety of rhabdomal patterns are seen among different insects. In the ocelli of honeybee, Apis mellifera (Toh and Kuwabara, 1974) and Apis meflijica carnica (Kral et al., 1985), and wasp, Paravespufa vulgaris (Kral, 1979), a rectangular rhabdom is formed from apposed borders of two retinula cells. In dragonflies (Dowling and Chappell, 1972), three cells normally contribute to form a Y-shaped rhabdom. In the cockroach, Periplaneta (Weber and Renner, 1976), the rhabdoms are composed of two to five retinula cells, and individual cell groups can be encountered at several levels throughout the receptor layer. In the fruit fly Drosophifa melanogaster (Stark et al., 1989), receptor cells have open rhabdomeres. In the flesh fly Boettcherisca peregria (Toh and Kuwabara, 1975), bibionid fly Dilophus febrilis, (Wunderer et al., 1988), and moth Trichoplusiu ni (Dow and Eaton, 1976) the microvilli originating from neighbouring receptors cells are interdigitated to form a continuous network of rhabdom. Some insects like cockroaches, crickets and dragonflies have a tapetal layer which serves to improve light absorption by reflection back through the receptor cells. An extreme specialization of ocelli is seen in nocturnal insects, especially in the cockroach, Peripluneta americana. The cockroach has two ocelli with a lens diameter of 0.7 mm, perhaps the largest among insect ocelli and with an extremely large receptive field. The corneal lens is slightly concave or flat which is not useful as a dioptric apparatus but serves as a window for photon entry. The number of photoreceptors contained in the ocellar retina is about 10000 (Weber and Renner, 1976), the largest among known ocelli, thus the volume of the rhabdom is very large. There are no pigmented cells to restrict the entry of photons into the rhabdom. Beneath the rhabdom layer there is a developed tapetal layer which increases optical length (Weber and Renner, 1976; Toh and Sagara, 1984). All of these structures are designed to improve photon capture, sacrificing spatial resolution, apparently an adaptation to a low light habitat. Interestingly, similar specialization, flat corneal lens, advanced tapetal, and a large volume of rhabdom, are seen in the compound eyes of the deep sea amphipod crustaceans Eurythenes and

THE DIVERSITY OF INSECT OCELLAR SYSTEMS

159

Ampefisca, which seem to have given up spatial resolution to attain improved sensitivity (Hallberg et a l . , 1980; Nilsson, 1989), thus providing an example of convergent evolution. Some ocelli exhibit structural changes upon light adaptation, to control the amount of light entering into the rhabdom layer. In the ocelli of the locust Schisfocerca gregaria (Goodman, 1970) and the moth Trichoplusia ni (Dow and Eaton, 1976), the rhabdom complex is surrounded by endoplasmic cisternae in the dark-adapted state while the cisternae are much reduced in extent and displaced radially within the cell in the light-adapted state. Formation of a palisade presumably assists the retention of light within the rhabdom. A number of ocelli, including those of grasshoppers, dragonflies and bees, possess a movable pigment sheath between the lens and the receptor cell layer. Many grasshoppers, including Austraciris gutfulosa and Valanga nigricornis, possess a radial, pigmented iris that stops down to a fixed aperture upon light adaptation (Wilson, 1975, 1978a). When movement is complete, a round pupil is formed in the median'ocellus, while lateral ocelli generally have an elliptical pupil with the long axis vertical. In Austraciris the maximum reduction in intensity of the pupil is 0.7 log units; thus the pupil is not very efficient (Wilson, 1978a). Dragonfly ocelli have movable pigment which covers the ventral part of the ocelli upon light adaptation. In the dark-adapted ocelli of the dragonfly Anax junius, pigment is concentrated in a store along the median ventral ridge in the median ocellus and in the posterior ventral corner of the lateral ocellus (Stavenga et al., 1979). Upon intensive illumination the pigment disperses, resulting in a reduction of the tapetal reflection in the ventral part of the ocellus. Stavenga et al. (1979) have suggested that the reason that the pupil mechanism is only active in the ventral part of the ocellus may be linked with the use of the ocelli in flight stabilization systems where, it is suggested, the ocelli monitor body movement relative to the horizon (Wilson, 1978a). Dragonflies usually fly with their body axis horizontal. The field of view of the ocellus is centred near the horizon, and direct sunlight is incident from above. Stavenga et af. (1979) pointed out that this light is unwanted, since a distinct vertex in Anax functions as a visor shading the ocellus from above and pupillary pigment screens the ventral part of the ocelli (Fig. 3B). They also noted that in those dragonflies such as Anax and Aeschna, which spend their daily routine patrolling above ponds, a large frons has been developed (Fig. 3B) which may function to prevent reflections from the water surface entering the ocellus so that no confusions occur for such systems as flight stabilization. Goodman (1981) pointed out that it is possible that a similar function is served by the filiform hairs of worker bees, i.e. that they screen the retina from unwanted reflections from outside the visual field (Plate 2a,b).

MAKOTO MlZUNAMl

160

A

lEOD v e r t i c d

T

C

1oc

* I I

a,*-,

/I

I

C

I

lob

lo8

'01

lor axial lull moon assuming UV+Gr

equivalent PAQ tor axial tull moon I assumlng only I

lo p

,

I

,

10"

IO'O

Peck A x i a l Quanta q ~ r n - ~ s - l

D

B

I

10 s

L

/ ,,

I

,

'

-

'

Dark

-3

-2

0

Dark

FIG. 3 (A) Angular extent of the three ocellar fields of Locusta expressed relative to longitudinal body axis. Field of median ocellus approximately symmetrical about the longitudinal axis and smaller than eccentric lateral fields. Vertical axis corresponds to animal's sagittal plane and horizontal axis lies in frontal plane. From Wilson (1978a). (B) A diagram demonstrating the set of visors with which a median ocellus of the dragonfly Anax is equipped. The dorsal vertex of the ocellus occludes the main area of the skies. The protruded frons shields the ground parts and the reflections from the water surface above which this dragonfly often flies. Direct sunlight impinging upon the ventral part of the ocellus is prevented from reaching the photoreceptor layer by the pupillary pigment. From Stavenga et al. (1981). (C) Averaged sensitivity of ocellar L-neurones and compound eye cells of the locust expressed in terms of peak axial irradiance in quanta. PAQ+ are shown on the 50% V,,, line and are as follows: (0)PAQSo for 8 L-cells with most sensitive and least sensitive cells shown (0).Averaged measured PAQ5" for 37 locust compound eye retinula cells shown (*) together with the estimated PAQSo for locust lamina neurones (0) thought to be 2 log units more sensitive than retinula cells. Continuous responseiintensity functions for the transient (0)and plateau (A)measured after 1 s averaged from 21 incompletely dark adapted cells and fitted to the PAQ5,1 for

THE DIVERSITY

2.3

OF INSECT OCELLAR SYSTEMS

161

ONTOGENETIC DEVELOPMENT

In hemimetabolous insects whose larvae have compound eyes and ocelli, dorsal ocelli start their development in the embryonic period (Mobbs, 1976, 1979). In holometabolous insects whose larvae have lateral ocelli (stemmata) but not dorsal ocelli, development of the adult dorsal ocelli begins in the larval period or in the prepupae (Caesar, 1913; Eaton, 1983a; Pabst and Kral, 1989). In both hemimetabolous and holometabolous insects, the three ocelli of the adult are derived from four embryonic rudiments of epidermal origin (in Hymenoptera: Caesar, 1913; in Orthoptera: Mobbs, 1979; in Lepidoptera: Eaton 1983a). Two of these rudiments each form a lateral ocellus, and two fuse in the midline to form a single median ocellus. In the moth, the ocellus which arises from epidermal cells is detectable on the day of pupation (Day 1 pupae; Eaton, 1983a,b). The ocellar corneagen cells and retinula cells have differentiated by Day 2. By Day 4 the retinula cell axons have entered the protocerebrum and large ocellar interneurones have become visible. The ocellus and interneurones assume the appearance of those in the adult by Day 5 , two days before adult emergence. Mobbs (1979) observed that the retinular cell axons of the locust, Schistocerca gregaria, grow across the space between the rudiment and the brain, and later interneurone axons grow out along the pathway pioneered by retinula cell axons and form a peripheral synaptic region. Similar observations have been made in the ocellus of the cockroach, Periplanetu arnericunu (Toh and Yokohari, 1988). To test if axonal growth of interneurones involves an interaction between receptor axons and interneurones, Eaton and Sprint (1985) produced unilaterally ocellate adult moths by cauterizing one of the pairs of ocellar primordia in fifth instar larvae. They found that two interneurones from the ablated ocellus grow into the synaptic region of the remaining ocellus, presumably making functional connections with ocellar receptor cell axons, findings which support the view that the axonal growth of the interneurones is guided by receptor cell axons. Pabst and Kral (1986, 1989) suggested a role for gap and septate junctions in the differentiation of photoreceptor cells in honeybee ocelli. They found that gap junctions and septate junctions between differentiating photoreceptor cells occur only as long as the rhabdoms are beginning to form. Their disappearance after differentiation indicates that they may play a part in cell determination. In Drosophilu, the three visual organs, i.e. compound eyes, transient response. Dotted vertical lines: number of peak axial quanta equivalent to full moon on optical axis of ocellus on assumption that the cell under consideration has only a UV peak or has a 100% UV peak and a 50% green peak. From Wilson (1978a). (D) Responses of a cockroach ocellar L-neurone to prolonged illuminations. The light intensities are indicated as loglo attenuation (0 log units = 20 p W cm2). From Mizunami and Tateda (1988a).

MAKOTO MlZUNAMl

162

larval lateral ocelli and adult dorsal ocelli, contain a photoreceptor cell-specific membrane protein, chaoptin, which is recognized by monoclonal antibody MAb24B10 (Zipursky et al., 1984). This protein is required for normal morphogenesis of the adult rhabdomeres (Zipursky et al.. 1984). It will be of interest to see if there are also proteins which specifically guide the formation of ocellar systems, or whether all visual systems share the same proteins when their structures are formed during ontogenetic development. 2.4

SUMMARY

Dorsal ocelli are simple photoreceptive organs coexisting with compound eyes, sophisticated organs for spatial vision. The ocellar lens is underfocusing, allowing for effective collection of light from a large visual field at the cost of spatial resolution. The ocelli often have structures to improve photon capture, such as the tapetal layer and endoplasmic cisternae. The ocelli are distributed among the major orders of apterygote insects. A positive correlation between the possession of ocelli and wings is reported among and within insect orders, and a negative correlation between the size of ocelli and the light intensity threshold of foraging is noted in bees. Both of these factors hint at functional roles for ocelli. 3 Functional properties of ocellar neurones

What kind of visual signals are encoded in ocellar neurones? What kind of visual signals are the ocelli designed to detect? Wilson (1978a) has demonstrated that a practical way to answer these questions is to compare functional properties of ocellar neurones to those of compound eye neurones. In this section, spatial, temporal, and spectral properties, as well as the absolute and polarization sensitivity of ocellar interneurones, are summarized, and compared with those of neurones in compound eyes. I will focus on large second-order neurones, called L-neurones, which form the principal pathway to transmit ocellar signals to the brain and are comparable with lamina monopolar cells, a major class of second-order neuranes of the compound eyes. 3.1

SPATIAL PROPERTIES

The most careful measurements of the receptive fields of ocellar L-neurones have been carried out in the locust by Wilson (1978a). The extent of the receptive fields of locust L-neurones is at least 130" (Fig. 3A), reflecting the high aperture of the ocellar lens and the high degree of convergence of ocellar photoreceptors onto L-neurones. Due to the large receptive field, the ocelli ignore the movement of small objects and specifically respond to

THE DIVERSITY OF INSECT OCELLAR SYSTEMS

163

the movement of large objects. This feature, referred to as a low-pass spatial filtering property, is in sharp contrast with the high-pass spatial filtering nature of the compound eyes where photoreceptors and subsequent columnar elements, including lamina monopolar cells, are tuned to narrow visual fields of typically 1-3" (compare Figs 2B and C). In the locust (Plate Id), as well as in the dragonfly (Plate la,b,c) and the bibionid fly (Plate 2g,h), three ocelli are directed horizontally relative to the longitudinal body axis in the normal flight posture and act as independent monitors of light intensity within three large visual cones oriented toward the horizontal, as can be seen from the receptive field of their L-neurones (Fig. 3A). These ocelli are most suited for the detection of shadowing by movement of the body relative to the horizon (Wilson, 1978a). In the stinkbug Pentatoma japonica (Plate If), hoverfly Eristalis cerealis (Plate 2e, f), and blowfly Calliphora erythro cephala (Schuppe and Hengstenberg, 1993), the ocelli are arranged on the top of the head and view mainly the dorsal hemisphere of the surroundings. 3.2

ABSOLUTE SENSITIVITY

Light intensity is defined as the average number of photons per unit time, and the fewer the photons, the larger is the uncertainty about the true average. Detection and encoding of light intensity are the basis of any visual functions. Thus the effectiveness in catching a photon and converting it into a voltage signal, which is termed as sensitivity or absolute sensitivity, is an essential limiting factor for visual systems. In any arbitrary visual system, increased sensitivity can be used to see at lower intensities, to see faster events, or to detect objects of less contrast. There is evidence to suggest that high sensitivity is one of the essential features of insect ocelli and that the attained high sensitivity is used for all three purposes mentioned above (see later sections). Measurements of the absolute sensitivity of locust L-neurones suggest that the sensitivity of ocelli are at least several times as high as that in compound eyes. Wilson (1978a) concluded in locusts that after several hours of dark adaptation, ocellar L-neurones are five times more sensitive to a point source than lamina monopolar cells, a major class of second-order neurones of compound eyes (Fig. 3C). When the large number of ocellar receptor cells are considered, this indicates that ocellar photoreceptors perform poorly in capturing axial photons compared with those in the compound eyes. One of the consequences of the large receptive field of L-neurones is that they have an extremely high sensitivity to extended visual sources. Wilson (1978a) concluded that ocellar L-neurones of locusts are 5000 times more sensitive to extended sources than the lamina monopolar cells in the compound eye, although pooling of signals from monopolar cells could improve the sensitivity of higher visual neurones in compound eyes.

MAKOTO MlZUNAMl

164

Goodman (1968) reported in locusts that in unidentified neurones of the thoracic ganglia which respond to both ocellar and compound eye stimuli, the sensitivity of ocelli to dimming is at least one order of magnitude greater than that of compound eyes. Some behavioural studies have also concluded that the ocelli are more sensitive than the compound eyes. Schricker (1965), for example, examined phototactic runs where bees could choose between two lights, and noted that a significantly greater number of them chose the brighter light. Below 1 lux, only a 2-fold difference of intensity was needed for this to occur. Bees with one ocellus occluded needed a 4-fold difference, bees with two ocelli occluded needed a 6-fold difference and bees with all the ocelli occluded needed an 8-fold difference before a significantly greater number of the bees chose the brighter of the two lights. Other behavioural evidence to show the high absolute sensitivity of ocelli is discussed in Section 4.5. 3.3

DETECTION OF ABSOLUTE INTENSITY LEVELS

A major class of second-order neurones of both the ocelli and the compound eyes, L-neurones and large monopolar cells (LMCs) , principally code changes in intensity, rather than the absolute intensity (Laughlin and Hardie, 1978; Mizunami et a f . , 1986; see Section 7.1). Signals about absolute intensity levels are not retained in LMCs, since steady illumination of the compound eye does not induce a steady membrane potential change in LMCs (Laughlin el al., 1987). In contrast, ocellar L-neurones retain signals about absolute intensity levels. L-neurones of the cockroach, for example, exhibit a hyperpolarization to a prolonged illumination with a transient peak which slowly recovers and reaches a steady level within 30-40s (Fig. 3D; Mizunami and Tateda, 1988a). This steady potential level is maintained as long as the steady illumination is continued for up to 10min. The level of the hyperpolarization changes depending on the intensity of the steady illumination, over at least 4 log ranges of intensity change (Mizunami and Tateda, 1988a). Coding of absolute intensity levels by ocellar neurones has also been noted in locust L-neurones (Simmons, 1993), third-order neurones of the cockroach (M. Mizunami, unpublished) and unidentified units in the ocellar nerve of the fly Cafliphora (Metschl, 1963). 3.4

SPEED OF SIGNAL TRANSMISSION

One important advantage of ocelli over compound eyes is the higher speed of signal transmission. In bees, the latency of ocellar and compound eye pathways were measured in descending multimodal neurones which receive inputs from both the compound eyes and ocelli (Guy et a f . , 1979). In situations where ocellar and compound eye pathways were simultaneously

THE DIVERSITY OF INSECT OCELLAR SYSTEMS

165

stimulated, the descending cells received ocellar input before the arrival of the delayed compound eye input. The latency of ocellar pathways was 9 ms, while that of the compound eye pathway was 25-35 ms (Fig. 4A). Do such small differences in speed have any functional significance? Taylor (1981a) measured the delay of head motion response of the tethered locust, Schisrocerca gregariu, after the motion of a diffuse horizon. Both the ocelli and compound eyes are involved in this response (Section 4.3). The delay is 45.4k4ms when all eyes are intact. The delay was almost unchanged when compound eyes were ablated (47.4+ 12 ms) but significantly increased when the ocelli were cauterized (103.4f7 ms), thereby demonstrating that a high speed of signal transmission by the ocelli contributes to the shortening of the latency of the visual response. Why do ocellar inputs reach the thoracic motor systems earlier than compound eye inputs? One of the reasons is that the number of interneurones intervening between photoreceptors and thoracic motor centres are much smaller in the ocellar pathways. In the bee, for example, signals from ocellar photoreceptors reach the thoracic ganglia via a single set of interneurones called LD-neurones, whereas in the compound eyes four interneurones are serially intervened at the shortest pathway from photoreceptors to thoracic motor centres (Guy et a f . , 1979). Second, L-neurones of the ocelli are among the largest in the insect nervous system which allows for the faster transmission of signals. Although compound eye systems have some large neurones in the third optic neuropil, lobula plate, their neurones are generally small especially in medulla, i.e. the second optic neuropil. In addition, the simplicity of signal processing in the ocellar pathway should facilitate faster signalling than the compound eye pathways where complicated and probably time-consuming signal integrations occur in their complicated neural circuits.

3.5

SPECTRAL SENSITIVITY

The ocellar L-neurones are characterized by a broad spectral tuning curve with sensitivity to both UV and visible light (Fig. 4B). Ocellar spectral sensitivity curves have a marked UV peak with an additional peak in either the blue (in Calliphora: Kirschfeld and Lutz, 1977; Drosophila: Hu er al., 1978) or green (dragonflies: Chappell and DeVoe, 1975; locusts: Wilson, 1978a; moths: Manduca sexra, Pappas and Eaton, 1977; Trichoplusia ni, Eaton, 1976; mantids: Sontag, 1971; honeybees: Goldsmith and Ruck, 1958; bumblebees: Meyer-Rochow, 1980). Wilson (1978a) argued that the enhanced UV sensitivity of locust ocelli facilitate the detection of the contrast between the sky and the land. The combined UV-blue or UV-green sensitivity in the ocelli appears to be an adaptation for increased sensitivity. UV-green receptors of dragonfly

166

MAKOTO MlZUNAMl

ocelli show another adaptation to attain maximum sensitivity, referred to as the reverse Purkinje shift (Chappell and DeVoe, 1975; Mobbs et al., 1981). At low light intensity these cells are very sensitive to green light, and at high intensity they are maximally sensitive to UV, thus matching the natural changes in environmental light. Although most ocelli have sensitivity to both UV and visible light, a few insects living in extreme light conditions have ocelli with only UV or green sensitivity. Mote and Wehner (1980) found that the ocelli of desert ants, which live under bright and UV-rich sunlight, have only a UV peak and are not sensitive at wavelengths longer than 445nm. The ocelli of some nocturnal insects which rarely encounter UV-rich sunlight, such as field crickets, Gryllus firrnus (Lall and Trouth, 1989) and cockroaches, Periplaneta arnericana (Goldsmith and Ruck, 1958; Mizunami, in preparation) have only single receptor systems maximally sensitive to green and not sensitive to UV. These nicely fit the idea t h a t visual pigments have been evolved in accordance with environmental light conditions. Studies in Calliphora and Drosophila suggested that the combined UV-visible sensitivity was attributable to only one spectral mechanism and, therefore, to only one cell type (Kirschfeld and Lutz, 1977; Hu et al., 1978). Kirschfeld et al. (1988) found that the pigment system in the ocelli of Musca and Calliphora closely resembles the pigment system in a type of photoreceptor of the compound eye, called R7y, in that there is a sensitizing pigment absorbing in the UV, combined with a visual pigment with an absorption maximum at 425 nm. Similar studies in the honeybee, dragonfly

FIG. 4 (A) Timing relationships between ocellar and compound eye visual signals in the bee. An L-neurone LOCz of the lateral ocellus and a descending neurone are drawn to scale and have been superimposed on an outline of the brain. Silver stains demonstrate that the two cells shown are in intimate contact with one another. Neural layers of the compound eye are outlined. The times are latency values for a response to a ‘light on’ stimulus measured for slow potentials in the lamina and ocellar pathway, and for spike potentials in the lobula. LA, lamina; LOB, lobula, M, medulla; R , compound eye receptor cells; VNC, ventral nerve cord. Arrows indicate direction of information flow. From Guy et al. (1979). (B) Relative spectral sensitivities of dark-adapted, UV-green cells of dragonflies Anax junius (top curve) and Aeschna tuberculifera (bottom two curves). Vertical axis: log units of relative spectral sensitivity for 4 mV criterion responses; curves have been displaced vertically an arbitrary amount for clarity. Horizontal axis: wavelengths of 100 ms test flashes, in nm. The log ratios of UV sensitivity to green sensitivity in the three curves are shown; a negative ratio indicates that the cell was more sensitive in the green. From Chappell and DeVoe (1975). (C) Spectral sensitivities of the UV, blue and green ‘single pigment’ cells found in the ventral part of the compound eye of the dragonfly Hemicordulia tau. The dotted curves are the theoretical absorbances of rhodopsins with absorption maxima at 360, 440 and 510 nm. From Laughlin (1976).

167

THE DIVERSITY OF INSECT OCELLAR SYSTEMS

t

C 100

.->r :._ . 50 n

c

0

-0.2

n

x

0.2 a

0.8

,

0 35 0 wovelength

- nm

,

4 50 550 wovelength , n m

\, 650

MAKOTO MlZUNAMl

168

and moth, however, showed that separate spectral mechanisms were needed to explain their UV-green ocellar response (Goldsmith and Ruck, 19.58; Chappell and DeVoe, 197.5; Eaton, 1976; Pappas and Eaton, 1977; Yamazaki and Yamashita, 1991). Intracellular studies on dragonfly photoreceptors established that UV-green sensitivity resulted from two pigments, but it was not possible to state whether these were located in one cell or in different cells coupled electrically (Chappell and DeVoe, 1975). In the compound eye, there are typically three types of colour receptors, UV, blue and green receptors, which are characterized by relatively narrow spectral tuning curves (Fig. 4C). The colour signals are processed by higher order neurones which receive excitatory inputs from one type of colour receptor and inhibitory inputs from another type, thus encoding colour contrast (Kien and Menzel, 1977). It appears that different strategies have been adopted in the ocelli and compound eyes to process signals from multiple pigment systems. In the compound eye, signals from different colour receptors are subtracted to enhance colour contrast, but in the ocelli, signals from different pigment systems are summed to enhance sensitivity.

3.6

POLARIZATION SENSITIVITY

Polarization sensitivity has been rarely examined for insect ocelli, but there is a report which shows that ocellar photoreceptors of the desert ant Catagfyphis are sensitive to the plane of polarized light. Mote and Wehner (1980) found that all ocellar receptors of the desert ant demonstrated polarization sensitivity ratios of about four. This is comparable with the polarization sensitivity of the compound eyes which ranges from 1.5 to 6. The ocellar photoreceptors are UV sensitive and thus are similar to UV photoreceptors of the compound eyes in terms of spectral and polarization sensitivities. Further behavioural studies show that Cutaglyphis can utilize the ocelli to detect the celestial polarization pattern (Fent and Wehner, 1985) as will be discussed in Section 4.6. The structural and neural basis of polarized light detection in Cutaglyphis remains to be clarified. 3.7

SUMMARY

The high absolute sensitivity and high speed of signal transmission are possibly the most advantageous features of insect ocelli over the compound eyes, which are attained by abandoning spatial resolution. Ocellar Lneurones are characterized by high absolute sensitivity, a broad spectral tuning curve with a peak at UV, and a large receptive field. These characteristics are most suited to detect contrast between the sky and land for stability control in flight, when the ocelli are directed horizontally in a normal flight posture (Wilson, 1978a). The high absolute sensitivity of ocelli is also suited to other visual behaviours in low light conditions.

THE DIVERSITY OF INSECT OCELLAR SYSTEMS

169

4 Behavioural roles of ocelli

4.1

OCELLI AS A STIMULATORY ORGAN

There is a theory which suggests that some sensory organs have, in addition to their specific sensory function, an essential function in maintaining the central excitatory state by sending sustained stimuli to the central nervous system and so raising the non-specific reactivity of the organism (Wolsky, 1930). This interesting but unproved (Bullock and Horridge, 1965) function is termed a general (or unspecific) stimulatory function as opposed to a specific sensory function. Wolsky (1930, 1933) noted the loss of tonus and of speed of walking in several insects when the ocelli were occluded and claimed that the ocelli are examples of stimulatory organs. A number of subsequent behavioural studies have also claimed that the ocelli have a general stimulatory role. None of these studies, however, proved that ocelli do function in this manner, as has been pointed out by Goodman (1970). A question which remains to be solved, concerning the concept of a general stimulatory role, is whether or not insects possess a system analogous to reticular activating systems in vertebrates, as pointed out by Bullock and Horridge (1965). In vertebrates, the reticular formation of the brain stem plays a role in producing arousal and sleep states of the animal and thus producing a circadian activity rhythm. The activity of the reticular formation is regulated by specific sensory pathways, called the reticular activating system, thus sensory inputs to the system result in changes in states of arousal (Shepherd, 1988). If insects possess a system comparable to the reticular activating system, and ocellar and other sensory inputs regulate the state of arousal by affecting the activating system, the once abandoned concept of a general stimulatory function may be revived in a refined fashion. This awaits future examination.

4.2

CONTRIBUTIONTO PHOTOTACTICORIENTATION

Most earlier behavioural studies of ocelli have been carried out with regard to their contribution to phototaxis. It has been concluded that the ocelli alone are rarely responsible for producing phototactic orientation, but rather that the ocelli appear to contribute to positive phototactic orientation mediated by the compound eyes, since ocellar occulution results in a decrease in the accuracy of such orientation (Cornwell, 1955; Cassier, 1965; Jander and Barry, 1968; Schricker, 1965; Meyer, 1978; reviewed by Goodman, 1970). In flies, locusts and crickets, the ocelli complement the compound eyes only under low levels of ambient illumination (Cornwell, 1955; Jander and Barry, 1968). In the fly Drosophilu, the contribution of ocelli and three types of photoreceptors, R1-6, R7 and R8, of compound eyes to positive phototaxis have been studied by utilizing receptor-deficient

MAKOTO MlZUNAMl

170

mutants (Fischbach and Reichert, 1978; Hu and Stark, 1977, 1980; Miller et al., 1981). All three photoreceptor types in the compound eyes evoke positive phototaxis. The sensitivity of the input from R1-6 receptors is higher than that from R7 and R8 receptors. In bright light, R7 dominates positive phototaxis and the presence of ocellar receptors facilitates R7 input (Hu and Stark, 1980). In dim light, R1-6 dominates positive phototaxis. Interestingly, strains with R1-6 do not exhibit positive phototaxis in dim light without ocelli, demonstrating strong facilitatory effects of the ocelli. Earlier studies of phototaxis were too limited to fully understand the behavioural roles of ocelli. First, phototaxis represents only the simplest form of a rich variety of visual responses insects exhibit in a natural environment. Furthermore, some earlier experiments on phototaxis were made in an extreme light condition which insects rarely encounter in a natural environment. Thus, it is not very surprising that these earlier studies did not contribute much to the clarification of the direct behavioural role of ocelli. Most of our knowledge on the behavioural roles of ocelli are based on recent experiments carried out in more natural behavioural situations and under natural light conditions.

4.3

VISUAL COURSE CONTROL I N FLIGHT

Many insects have a tendency to turn their back towards the centre of brightness during flight and walking, to keep their course straight (reviewed by Wehner, 1981). This reaction is referred to as the dorsal light response. A way to attain a dorsal light response is to monitor position and movement of the horizon relative to the body. Hesse (1908) argued that the ocelli are most suited to detect movement of the horizon for stability control in flight, and Wilson (1978a) reformulated Hesse’s argument based on new evidence. Wilson (1978a) argued that locust ocelli have a large receptive field directed horizontally, providing the animal with heavily blurred neural images of the skyline, where unwanted information about structural details is eliminated. Ocellar sensitivity to UV facilitates horizon detection since the contrast between bright sky and dark ground is highest in UV. The high speed of signal detection and transmission in the ocellar system is ideal for rapid course control. Pitch and roll deviation of the flight course are independently detectable by the combination of signals from three ocelli. A roll (turning around the long axis of the body) will cause no change in signal from the median ocellus but will tend to cause a decrease of illumination in one lateral ocellus and an increase in the other lateral ocellus. Detection of pitch could be achieved through measurement of the output of the median ocellus . This hypothesis has received support from behavioural studies. In dragonflies, Stange and Howard (1979) and Stange (1981) observed that

THE DIVERSITY OF INSECT OCELLAR SYSTEMS

171

stimulation of ocelli can evoke a steering response. The stimulation of the median ocellus evoked a head movement around a pitch axis, and the stimulation of the lateral ocellus evoked head movements around the roll axis. In locusts, Goodman (1965) noted the ocellar contribution to stability control in flight and Taylor (1981a,b) further examined the contribution of compound eyes and ocelli to flight steering (Fig. 5 ) . Locusts were tethered inside a simulated horizon display (Fig. 5A,B). Rotation of the horizon elicited following motions of the animal’s head and rudder-like movements of the body during flight. Head and steering body motions were still elicited after either the compound eyes, or the ocelli, were surgically ablated. Head motion after the cautery of the compound eyes (Fig. 5C,E,F) suggested that the ocelli may function synergistically with the compound eyes to (a) minimize the delays of visual responses and (b) augment visual responses at a dim light condition and when n o sharp horizontal border is present. Taylor (1981a) noted that hoverflies (Eristalis renax, Syrphidae) and damselflies (Argia vivida) also followed horizon rotators with head motion after their compound eyes had been ablated. Kastberger (1990) suggested that the honeybee ocelli help to control the flight course. Although it is well established that one of the principal functional roles of ocelli is to detect the horizon for flight stabilization in some insects, ocelli of some other insects seem not to participate in the stability control in flight. In blowfly, Cafliphora, simulated roll of an artificial horizon evokes little ocelli-mediated head movements (Hengstenberg, 1984, 1993) and likewise, changes of the brightness in the visual field of the median and one lateral ocellus elicit only a weak steering response (Schuppe and Hengstenberg, 1993). The ocelli of blowfly are directed dorsally, thus they may not be suited for horizon detection. Tomioka and Yamaguchi (1980) also concluded that ocelli play little role in posture control during flight in the night-flying cricket, Gryllus bimaculafus. It is difficult to suppose that well-developed and specialized ocelli of the cockroach Peripfanera (see Section 2.2) have been evolved to attain better flight steering for this weak flier. Therefore, functional roles of ocelli differ among insects: to contribute to flight steering is a principal function in some insects, but not in others.

4.4

ORIENTATION TOWARD EDGES

Walking houseflies can use ocelli for orientation. Wehrhahn (1984) tested the orientation of walking houseflies with blinded compound eyes, and found that by using only their ocelli, flies orient toward edges and relatively small bright objects situated in the frontal equatorial part of the visual field (Fig. 6). This finding may indicate that the fly ocelli resolve coarse visual images, a possibility which needs to be pursued by future examination.

MAKOTO MlZUNAMl

172

pickups for head position transducer

Head Dosi tion

Head position

Horizon position

E

-

50"

Head position

L 7 J l2.0Js - L n l

-

Head oosition

LLeft-side illuminated - - -

- - - - -

Kiglit

0.5 s

FIG. 5 (A) The artificial horizon apparatus used to evoke rapid rolling motions in a locust. A locust was suspended inside the apparatus with its long axis coaxial to the horizon tubes. The blackened hemisphere was rotated by an axle connected to a chart recorder pen motor. The small spheres are capacitative pickups used to sense the position of the animal's head. A 40 kHz electrical signal was applied to a fine wire within the glass wand shown attached to the locust's head. Comparison of the signals induced in wires leading from the two spheres allowed determination of head position about the roll axis. (B) Apparatus used to study pitch axis motions of the horizon. Similar to (A), except that the animal is mounted transversely in the illuminated tube. (C-F) Responses to visual stimulation of the ocelli. Except for (D), all locusts have compound eyes disconnected by section of the optic lobes leaving the ocelli intact. Unless noted otherwise, downward rotation signifies clockwise rotation with respect to the animal. 1000 lux. (C) Head rolling response to horizon rotation. After three cycles of horizon oscillation, the display was rotated 180" counterclockwise (i.e. dark above, illuminated below). (D) A similar experiment to that shown in (C), except with all eyes intact. After five cycles of motion, the horizon was inverted. (E) Head-rolling response of a non-flying locust to horizon rotation (no wind applied). (F) Head rolling response of a flying locust to alternate illumination of the two sides of the animal's head by a pair of mechanically switched light sources subtending 5". Illuminance of the animal's head was approximately 75 lux. From Taylor (1981a).

THE DIVERSITY OF INSECT OCELLAR SYSTEMS

173

FIG. 6 (a) Frequency distribution of the orientation of freely walking flies whose compound eyes only were blinded (open circles) or whose compound eyes and ocelli were both blinded (filled triangles). The panorama consisted of a black half and a white half generating two vertical edges. The flies with only ocelli show a marked preference to walk towards the edges. Average values from 10 flies. Bars denote standard deviation of the mean. (b) Frequency distribution of the orientation of freely walking flies with blinded compound eyes. The pattern consists of a white square (22.5" side length) and opposite to it a white rectangle on a black background as indicated in the top of the figure. The flies prefer the white square. From Wehrhahn (1984).

4.5

LIGHT INTENSITY PERCEPTION FOR THE CONTROL OF DIURNAL ACTIVITY

Initiation and cessation of diurnal activities of insects often depend on the light intensity levels (Schricker, 1965; Dreisig, 1980). Under natural environmental conditions in which the light-dark cycle consists of dawn and dusk ramps, it would be advantageous for insects to be sensitive to minute changes in illumination. Thus, the ocelli are suited to control the diurnal activity whose absolute sensitivity is at least several times as high as that of compound eyes (Section 3 . 2 ) . Indeed, signals from ocelli are utilized to determine the threshold light intensity for the diurnal activities in bees (Schricker, 1965; Gould, 1975) and moths (Eaton et al., 1983; Sprint and Eaton, 1987). Foraging flight activity in bees is chiefly governed by two factors: the weather and the light intensity (Schricker, 1965). The first and last of the daily flights is dependent upon the intensity level. If the ocelli are occluded, foraging bees behave normally in most respects (but see Renner and Heinzeller, 1979). Occlusion of the ocelli, however, does interfere with the timing of the first and last foraging flights. Bees with one, two or three

B

-

1001

*NOCELL

-e- CONTRROI

--r TIME

TIME

'w O0I

0 TIME

TIME

FIG. 7 Percentage of male moths, Trichopfusiu ni, flying during the sunset period in the phase-advance (PA) experiments. (A) Day before PA. (B) First day of PA. (C) Second day of PA. (D) First day of return to original time of sunset. Anocell, anocellate moths; L. light intensity. The range of light intensity was between 100 and <1 lux. From Sprint and Eaton (1987).

THE DIVERSITY OF INSECT OCELLAR SYSTEMS

175

ocelli occluded start to collect food later in the morning and cease collecting earlier in the evening than normal workers. The light intensity required for the first and last collecting flight is increased by a factor of 2 if one ocellus is covered, 3.3 if two ocelli are covered and 4.5 if all the ocelli are covered. Eaton et al. (1983) studied the role of ocelli in adjusting the time of flight initiation of cabbage looper moths, Trichoplusia ni, at dusk under simulated sunset conditions. Ocellar occlusion delayed flight initiation on the first day following treatment but was less effective on the subsequent days. Ocellus ablation delayed flight initiation (Fig. 7). Control and sham-operated males can adjust to changes in the time of sunset better than anocellate male moths (Sprint and Eaton, 1987), thereby indicating that moths make use of input from the ocelli in determining the threshold light intensity for flight. Sprint and Eaton (1987) also noted that control and sham-operated males flew significantly more than anocellate males and thus concluded that the ocelli regulate also the intensity of flight activity. Wunderer and Kramer (1989) studied the control of the mating system of the arctiid moth Creatofos transiens, in which both males and females produce pheromone, using continuous measurements of oxygen consumption. Luring activity started some time after the ambient light fell below a certain threshold and also when daylight cycles were shortened considerably. Occlusion of the dorsal ocelli induced a significant delay in the onset of activity, suggesting that the ocelli detect threshold intensity for initiating diurnal activity. The ocelli cannot mediate light stimuli for an entrainment of the circadian clock (Nishitsutsuji-Uwo and Pittendrigh, 1968; Roberts, 1965; Rivault, 1983). In crickets, however, the ocelli play an indirect role in circadian rhythmicity by controlling the sensitivity of the compound eyes (Rence et al., 1988). The free-running period of the circadian stridulation rhythm of male crickets depends on the light intensity in constant light (LL) conditions. With the increase in LL intensity, the free-running period increases. In male field crickets, Tefeogryllus commodus (Fig. 8), and house crickets, Acheta domesticus, the severance of the three ocellar nerves significantly decreases the circadian stridulation period in LL which is indicative of a reduced perception of the available light intensity. The shortening of the LL free-running period was around 30 min, which can be translated to a 10-fold decrease in the actual light level used for setting the period of the free-running behaviour. These authors further discovered that the size of the ERGS of the compound eyes was reduced by 20% with ocellar occlusion (Fig. 8), suggesting that the ocelli control the sensitivity of the compound eyes at the level recorded in an ERG. Cobalt fills revealed that neurones travel from the lateral ocelli out into the optic lobes of the compound eyes. Rence et a f . (1988) thus concluded that ocelli play an indirect role in circadian rhythmicity, augmenting the sensitivity of compound eyes to better perceive photic entrainment signals.

MAKOTO MlZUNAMl

176

. . .

.. . -

.

-101

C

B

Shutter

w

Ocelll Occluded Ocelli Restored

I sec Ocelli Occluded

Ocelli Restored

FIG. 8 (A) Sample double-plotted record of free-running stridulatory activity in a male cricket, Tekogryllus commodus, before and after surgical severance of the three ocellar nerves. Surgery occurred on Day 47 at the arrow and the period of the rhythm shortened from a pre-surgical T = 24.0 h to a final stable T = 23.15 h post-surgically. Environmental conditions throughout were LL (0.00025lux) and 2S_+0.S°C. (B. C) The effect of ocellar occlusion on the ERG recorded in the compound eye of a male cricket T. commodus. (B) Raw record of ERG responses

THE DIVERSITY OF INSECT OCELLAR SYSTEMS

4.6

177

DETECTION OF POLARIZED LlGHT

Some bees and ants utilize celestial polarized patterns for compass orientation and navigation. Detection of the plane of polarized light is mainly done by specialized ommatidia in the dorsal rim areas of the compound eyes (Wehner and Strasser, 1985). In some insects, however, perception of polarized light is not restricted to the compound eyes. Western bumblebees fly straight while homing by polarized light, but zigzag when they use landmarks. Wellington (1974) used the pattern of flight to determine the roles of the ocelli and parts of the compound eyes in homing. The presence of celestial polarized light and uncovered ocelli can prolong foraging at twilight, when landmarks are no longer visible, suggesting that the bumblebee can use its ocelli to steer by polarized light from the sky when the ground is too dimly lit for homing by landmarks. In a walking insect, the desert ant Cataglyphis, the ocelli can read compass information from the blue sky (Fent and Wehner, 1985). When the ant’s compound eyes are occluded and both sun and landmarks are obscured, the ocelli, using the pattern of polarized light in the sky as a compass cue, help in guiding the ant back home (Fig. 9).

4.7

CONTROL OF NEUROENDOCRINIC SECRETION

Some authors propose that ocelli play roles in the control of neurosecretion. The amount of neurosecretion in the pars intercerebralis (PI) and in the corpora cardiaca (CC) of the honeybee changes during the course of the day (Heinzeller, 1976). The amount of secretion in the PI and CC is minimal at midday, and is negatively correlated with the density of nuclei in the corpora allata (CA). When normal animals are kept in a daylight cage, the content of secretion in the PI and CC does not decrease during the morning as it does in free-flying foragers. If bees with black varnished ocelli are placed in the cage, the amount of secretion increases very much, and some bees show an extremely high density of nuclei in the CA. Based on these results, Heinzeller (1976) proposed that ocelli play roles in humoral regulation of motor activity. Brousse-Gaury (1968, 1975) observed in the cockroach that some interneurones of the ocellar tract enter the tracts of neurosecretory axons in the evoked by the shutter alone (no light stimulus) and a 0.1 lux light stimulus in an animal with all three ocelli opaqued and then uncovered. Stimulus presentation is indicated by the black bar at the top. (C) Histogram of the mean percentage (+ SD) change in the amplitude of the compound eye E R G response from intact crickets (n = 16) t o those with all three ocelli occluded and then restored by peeling off the opaquing agent. There is a statistically significant reduction in the E R G response when the ocelli are occluded. From Rence et al. (1988).

MAKOTO MlZUNAMl

178

A

t

B

t

C

t

D

.r

FIG. 9 Return directions of desert ants Cutaglyphis that were allowed to view the natural blue sky (with the sun obscured) or individual E vectors. The ants' bearings are given for distances of 2 , 4. and 6 m from the start (+). Black arrow (in A , B, and D) represents the homeward direction. (A) Controls (0)with neither compound eyes nor ocelli occluded and ocelli occluded animals (COMP animals; 0)viewing the sky but prevented from seeing the sun. The two series d o not differ in mean direction and deviation. (B) The compound eyes occluded animals (OC animals) tested as in (A). Two series (0 and 0) that differ in the azimuth position of the sun relative to the homeward direction. The mean directions of the ants differ significantly from the sun's azimuth, thus excluding the possibility that the ants moved merely toward the brightest part of the sky. (C) The OC animals viewing a spot of artificially polarized light; E vector orientation, ,y = -73" (0)or -60"(W)I. The black arrow indicates the

THE DIVERSIP/ OF INSECT OCELLAR SYSTEMS

179

brain. Cooter (1975) also reported one fibre of the ocellar nerve of the cockroach reaches to a region where the nervus corporis cardiaci I1 has its ongin. Subsequent cobalt-filling studies of cockroach ocellar neurones, however, failed to confirm these observations (Koontz and Edwards, 1984; Mizunami and Tateda, 1986; M. Mizunami, personal observations). The neural pathways from the ocelli to the neurosecretory cells still remain to be established. 4.8

SUMMARY

Behavioural roles of insect ocelli can be classified into two categories according to their relationship to those of the principal visual organs, the compound eyes. First, the ocelli directly participate in some visual behaviours to complement the compound eyes. Second, the ocelli modulate behaviours mediated by the compound eyes. Due to the superiority of ocelli over compound eyes in sensitivity and speed, participation of ocelli in complementing and modulating compound eye functions improves visual performance of insects. The ocelli appear to play different behavioural roles in different insects. In dragonflies and locusts, the principal role of ocelli is to detect movement of the horizon for stability control in flight. In some other insects, however, a number of different behavioural roles are implicated for ocelli including low light intensity perception. If we emphasize that dragonflies (Palaeoptera) and locusts (Polyneoptera) appeared in an earlier stage of evolution of Pterygota, it can be hypothesized that the possession of three ocelli has been established to perform flight steering, and in the later stage of the evolution of Pterygote insects, other functions have been acquired by the ocelli due to their specific advantage in sensitivity and speed, thus generating the present functional diversity. Understanding the behavioural roles of insect ocelli has just begun, and I believe a consideration of evolutionary history may help to further clarify the roles of ocelli.

5 Neural organization of ocellar pathways In the last 20 years many studies have been made on the neural organization of insect ocellar systems using cobalt filling of the ocellar nerve, intracellular

homeward direction that the ant should choose when relying exclusively on E vector information. White arrows are hypothetical homeward directions to be selected by the ants if they took the spot of polarized light for the sun. (D) Blind animals tested as in (A) and (B). Within the time interval of 5 min only 5 out of 19 blind animals moved as far as 6 m from the start. From Fent and Wehner (1985).

MAKOTO MlZUNAMl

180

staining of individual ocellar neurones and electron microscopy. These studies have revealed both common features and fundamental differences in insect ocellar systems. In this section, I will begin with a summary of our knowledge of the neural organization of ocellar systems. Then I will discuss the possible functional and evolutionary implications of some of the differences in the basic neural organizations of different insects.

5.1

SYNAPTIC ORGANIZATION OF OCELLAK PLEXUS

Electron microscopic studies have revealed that synaptic interactions within the ocellar plexus are complicated (Fig. 10). Typically, the major elements of the ocellar plexus are photoreceptors, and large and small second-order neurones. In addition, possible efferent neurones have been noted in honeybees (Toh and Kuwabara, 1974), fleshflies (Toh and Kuwabara, 1975) and cockroaches (Toh and Sagara, 1984). In all ocelli so far studied, the synapses most frequently seen in the ocellar plexus are those made by the many photoreceptors onto a small number of L-neurones (in moths, Fig. 10A; fleshflies, Fig. l0B). The synapses often exhibit diadic configurations, where a photoreceptor is presynaptic to an L-neurone and another photoreceptor. There are notable differences in synaptic organization of the ocellar plexus among different insects. Numerous synaptic contacts between photoreceptor axons have been observed in dragonflies (Dowling and Chappell, 1972), locusts (Goodman et a / . , 1979), cockroaches (Cooter, 1975; Toh and Sagara, 1984) and Bibionid flies (Wunderer et a / . , 1988). In the moth, Trichoplusia ni, there are numerous electrotonic junctions between photoreceptors (Fig. 10A; Dow and Eaton 1976). In cockroaches, however, synapses between photoreceptor axons are infrequent (Toh and Sagara, 1984). It is not known if such infrequent synapses have a functional significance or are a casual by-product of developmental promiscuity and thus functionally less important. In fleshflies (Toh et a / . , 1971; Toh and Kuwabara, 1975), honeybees (Toh and Kuwabara, 1974) and wasps (Kral, 1979), n o reciprocal synapses are found between photoreceptors. Eaton and Pappas (1977) proposed that the synapses between photoreceptors may act to generate synchronized activity in the receptor cell axons by reciprocal excitation. This hypothesis awaits future electrophysiological and pharmacological examination. In dragonflies (Dowling and Chappell, 1972), locusts (Goodman et al., 1979), the moth, Trichoplusiu ni, (Eaton and Pappas, 1977), wasps (Kral, 1979) and Drosophilu (Stark er a / . , 1989), L-neurones make feedback synapses onto photoreceptors. In cockroaches, however, feedback synapses onto photoreceptors are infrequent (Toh and Sagara, 1984). In honeybees (Toh and Kuwabara, 1974) and fleshflies (Fig. 10B; Toh et a / . , 1971; Toh and Kuwabara, 1975), L-neurones appear not to make feedback synapses

THE DIVERSITY OF INSECT OCELLAR SYSTEMS

181

B

FIG. 10 (A) Diagram of synaptic conriections between four branches of receptor cell axons (stippled circles) and three dendrites of second-order neurone (open circles) of the moth Trichoplusia ni. Data derived from 25 serial sections of synaptic region of one ocellus. Arrows indicate chemical synapses, bars electronic synapses. From Eaton and Pappas (1977). (B) Synaptic connections in fleshfly ocellus schematically drawn in one plane between two retinula axons (Rl) and (R), one thick ocellar nerve fibre (T) and seven thin ocellar nerve fibres ( t l 4 and t). The figure represents 26 serial sections. From Toh and Kuwabara (1975).

onto the receptors. In fleshflies (Fig. 11B) and also in honeybees and locusts, photoreceptors are presynaptic to small fibres. These small fibres are considered to be second-order neurones in locusts (Ammermiiller and Weiler, 1985) or efferent neurones in honeybees (Toh and Kuwabara, 1974) and fleshflies (Toh et al., 1971; Toh and Kuwabara, 1975). The functional roles of feedback synapses onto photoreceptors have not been established (see Section 7.1). In summary, we have not yet fully understood the functional roles of each synaptic pathway in the ocellar plexus, thus it is premature to discuss the possible functional significance of the differences in neural wiring of the ocellar retina among different insects.

5.2

MORPHOLOGY OF SECOND-ORDER OCELLAR NEURONES

5.2.1 General features of second-order neurones The ocellar nerve of insects typically contains a small number of largediameter axons (L-neurones) and a larger number of smaller diameter fibres (S-neurones). The L-neurones represent some of the largest neurones in the insect brain, the diameters of which range from 8 p m to in excess of 20 pm. The L-neurones receive extensive synapses from photoreceptors and form a principal pathway to transmit ocellar signals to the brain (Fig. 11A,B). The

MAKOJO MlZUNAMl

182

B

FIG. 11 (A) Morphology of a second-order ocellar neurone (L-neurone) and a third-order neurone (PS-I neurone) of the cockroach, Periplaneta arnericana, viewed dorsally. The neurones were drawn from cobalt-filled, silver-intensified preparations. OC, ocellus; ON, ocellar nerve; O T , ocellar tract; PC, protocerebrum; PS, posterior slope. From Mizunami (1990a). (B) Morphology of locust L-neurone, revealed in intensified, cobalt-stained preparations. The protocerebrum drawn from the dorsal aspect, showing an L1-3 neurone in the left ocellar tract and an L4-5 neurone in the right tract. Locust L1-3 neurones project into the ocellar tract, as d o cockroach second-order neurones, while locust L4-5 neurones terminate in the posterior slope, as do some cockroach third-order neurones. From Littlewood and Simmons (1992).

S-neurones have a range of diameters but usually do not exceed about 5 pm. The identities of S-neurones have not been fully understood. In most insects a large number of the S-neurones are probably second-order neurones, although some others are possibly third order or efferent.

5.2.2

Morphology of L-neurones

Morphologies of ocellar L-neurones have been examined in a number of insects by cobalt filling of the ocellar nerve and by intracellular staining of ocellar neurones. These studies have shown that L-neurones can be classified into three types, based on their input regions (Goodman, 1976b, 1981): ‘L’ types which receive synapses from the photoreceptors of a lateral ocellus (e.g. Fig. 11B), ‘M’ types which receive synapses from photoreceptors of the median ocellus and ‘ML’ types which receive synapses from photoreceptors of both a median and a lateral ocellus (Goodman, 1981).

THE DIVERSITY OF INSECT OCELLAR SYSTEMS

183

L (large)-neurones can also be classified into three types from their output regions (Mizunami, in preparation; see also Goodman, 1981): (1) LOT neurones which project into the ocellar tract (e.g. Fig. 11A and L1-3 of 11B); (2) Lps neurones which project into the posterior slope (e.g. L;15 of Fig. 11B and Fig. 12A,B); and (3) LD neurones which run on into the ventral nerve cord and terminate in the thoracic ganglia (e.g. Fig. 12C). LD neurones form a short pathway connecting the ocellar plexus and thoracic flight motor systems (Goodman, 1981; Goodman et al., 1987). LOT neurones typically possess characteristic arborizations in an area posterior to the protocerebral bridge, called the anterior ocellar focus (Goodman, 1981). These arborizations are prominent in locusts (Fig. 11B: Goodman, 1976b) and crickets (Koontz and Edwards, 1984) but less prominent in cockroaches (Fig. 11A; Mizunami et al., 1982). Two distinct output areas of LOT neurones, i.e. the ocellar tract and anterior ocellar foci, are functionally distinct, as will be discussed later (Section 5.3.2). In cockroaches, all L-neurones are the Lor type (Mizunami et al., 1982). Locusts (Goodman, 1976b) and crickets (Koontz and Edwards, 1984) possess LOT and Lps neurones. In dragonflies, Chappell et al. (1978) reported that all L-neurones are the Lps type. However, Chappell et al. (1978) also noted characteristic arborizations in the anterior optic foci typical of LOT neurones of other insects, suggestive of the presence of second-order neurones projecting in the ocellar tract, although they may not be L (large)-neurones but rather medium-sized or small neurones. Honeybees (Pan and Goodman, 1977; Goodman et al., 1987), wasps (Kral, 1979) and flies (Nassel and Hagberg, 1985) possess Lps and LD neurones. In moths, all L-neurones are reported to be the Lps type (Pappas and Eaton, 1977). In summary: (1) LOT neurones have been found only in hemimetabolous insects; (2) all insects, except for cockroaches, possess Lps neurones; and ( 3 ) LD neurones have been observed only in holometabolous insects. For more information about the number and morphology of L-neurones of each insect species, see Goodman (1981). Possible functional differences among different morphological classes of L-neurones are discussed in Section 5.3. 5.2.3 Genetic determination of L-neurones Variations in the morphology and number of L-neurones have been noted in all ocellar systems in which L-neurones have been examined by cobalt staining. Goodman (1974, 1977) has examined the forms of variability and found that duplications and deletions of identified L-neurones can occur with a high degree of genetic control. In duplications, an identified cluster of neurones contains an extra neurone and in deletions, one cell of a cluster is missing. Duplications and deletions are strongly influenced by parentage and must be subject to a degree of genetic control. Among two laboratory

184

MAKOTO MlZUNAMl

FIG. 12 (A-C) Summary diagram of median ocellar L-neurone projections in the brain of the worker honeybee. (A) L-neurones of the median ocellus projecting into the posterior slope. (B) L-neurones of the lateral ocellus projecting into the posterior slope. (C) L-neurones of median and lateral ocelli projecting into the thoracic ganglia. Bar = l00pm. LO, left lateral ocellus; MO, median ocellus; RO, right lateral ocellus; CX, calyx of the mushroom body; OF, oesophageal foramen; SG, suboesophageal ganglion; PB, protocerebral bridge; PTG, prothoracic ganglion; MMTG, meso/metathoracic ganglion. From Pan (1981); appeared in Mobbs (1985).

THE DIVERSITY OF INSECT OCELLAR SYSTEMS

185

populations, each initiated from a small number of individuals, one population had a tendency for duplications within one cell cluster (ML1-2) whereas locusts of the other population had a tendency for duplications within a different cell cluster (Ll-3). Goodman (1977) investigated the influence of genotype upon the identity and frequency of duplications by examining isogenic clones of locusts, produced by parthenogenetic breeding. The variability between clones is larger than that between the isogenic members of any one clone, but the variability still exists within clones. Genetic control of neural duplication is demonstrated by the finding that the number of locusts with duplications within one cell cluster (Ll-3) ranged from 52% in one clone to 0% in another clone, In single locusts, duplications can occur in a cluster of neurones on one side of the brain without affecting the contralateral homologues. Thus the genetic factors express themselves against a background of other factors during neural development. As has been discussed by Goodman (1977) and Laughlin (198l), duplications and deletions could have played an important role in the evolution of ocellar L-neurones. Ohno (1970) suggested that gene duplications promote evolutionary change of proteins, by introducing redundancy into the genotype. Similarly, neurone duplications could make the nervous system more tolerant of structural change. When a duplication is made, one cell can sustain normal function while its partner can generate novel connections, thus allowing novel functional connections to be generated without necessarily eliminating existing functions. Goodman (1977) compared the anatomy of L-neurones among locusts, crickets and cockroaches, and proposed homologies among them. In Fig. 13, the possible homologies of L-neurones among three insects are illustrated, where original suggestions by Goodman (1977) are revised based on the results of subsequent studies in crickets (Koontz and Edwards, 1984) and cockroaches (Mizunami et af., 1982; Mizunami and Tateda, 1986). One of the cells in a three-cell cluster of L1-3 in the locust and cricket is duplicated to form a four cell cluster, L1-4, in the cockroach. One of the cells in a L4-5 cell cluster in the locust is reduced in size in the cricket and both of them are deleted in the cockroach. M and ML neurones of locusts and crickets are deleted in the cockroach which has completely lost the median ocellus. One of the M2 cells in the cricket is smaller in size than that in the locust. In conclusion, simple duplication, deletion and modification are sufficient to explain differences in the morphologies and numbers of L-neurones among three Polyneopteran insects (see Fig. l ) , locusts, crickets and cockroaches. Extension of such a comparison to other Polyneopteran insects, such as termites and praying mantis, will reveal how the morphology of single identified neurones has changed during the course of evolution. It is also interesting to compare the morphology of L-neurones of Polyneopteran insects with that of other Pterygote insects.

186

MAKOTO MIZUNAM

I*I I : I

0

3

Locust

e e

ML1-2 L4-5

0

M2/MS

e 0

ML1-2

Cockroach

e 0

LL LS

0

O 0 M1 1 . 0 0 . L1-3

C

Cricket

0 .

0 . L1-3

I

0-0

0-0

L1-4

FIG. 13 Schematic diagram of the number and position of somata of the large ocellar interneurones in the brain of (A) the locust, Schistocerca nitens; (B) the cricket, Acheta domesticus; and ( C ) the cockroach, Periplaneta americana. The locust has three ocelli (two lateral and one median) and three ocellar nerves, each containing the axons of seven large interneurones. (0)Cells with axons in the left or right lateral ocellar nerve; (0)cells with axons in the median ocellar nerve; (0)cells with axons in both a lateral and the median ocellar nerve. The cricket has one large and one small cell body where the locust has a two-cell cluster, L4-5, and two large cell bodies (M2iM2) where the locust has one large and one small cell bodies (M2/MS). The cockroach does not have an median ocellus, thus lacks ML and MUM2 neurones. Cockroach has four cell bodies where the locust and cricket have a three-cell cluster, L1-3. Based on reports by Goodman (1977), Mizunami et al. (1982), Koontz and Edwards (1984), and Mizunami (submitted). From M. Mizunami (unpublished),

5.2.4 Small neurones in the ocellar nerve The total number of small-diameter neurones within the three ocellar nerves typically exceeds 30. Goodman and Williams (1976) distinguished eight types of small ocellar interneurones amongst the 61 S-neurones identified in Acridid grasshoppers. The projection areas include the posterior slope, the protocerebral bridge, the ventral bridge, the optic ganglia, and the antenno-glomerular tracts (Fig. 14A). The projection areas of S-neurones have also been examined in bees (Pan and Goodman, 1977), crickets (Koontz and Edwards, 1984), cockroaches (Mizunami, in preparation), moths (Eaton and Pappas, 1977) and blowflies (Fig. 14B; Nassel and Hagberg, 1985). In these studies, a more or less similar pattern of projections has been observed, although specificity in each species is also evident (see Fig. 14A and B; Goodman, 1981). Unfortunately, it is difficult to reliably identify a large number of very small neurones, thus our present knowledge is not as yet sufficient to discuss the generality and specificity of projection patterns of S-neurones among insects. For a more detailed

THE DIVERSITY OF INSECT OCELLAR SYSTEMS

187

FIG. 14 (A) Schematic diagram of the brain of the locust (posteriorly viewed, dorsal is up) showing on the left side some of the tracts and neuropil regions of the brain and on the right side eight specific areas of neuropil containing arborizations of the small ocellar interneurones (Types I-VIII). LOT, MOT, lateral and median ocellar tracts; PL, protocerebral lobe; PI, pars intercerebralis; C , S, calyx and part of stalk of mushroom body; A G , antenna1 glomeruli; AGT, antenno-glomerular tract (only partly shown); OL, optic lobe; L, lobula (shown along with part of a posterior optic tract); PB, protocerebral bridge; CB. central body; EB, ellipsoid body; VB, ventral bridge; C-OC, circum-oesophageal connective; calibration bar, 200 p m . From Goodman and Williams (1976). (B) Tracing of cobalt-filled ocellar interneurones from whole-mount preparation of the fly Calliphora. This tracing shows the total set of large- and medium-sized neurones (except the left lobula fibre OMLo). The large-diameter fibres (OL) end in the posterior slope (Post 9). The mediumdiameter fibres project to the medulla (Me), the lobula (Lo), the protocerebrum (Pc) and via the suboesophageal ganglia (Sog) to thoracic ganglia. They are indicated by OMMe, OMLo, OMPc and OMDN, respectively. Arrows indicate one thin-diameter fibre running to the medulla. Ost, ocellar stalk; Oes, oesophagus. The neurones are seen in frontal (transverse) views; dorsal is up. From Nassel and Hagberg (1985).

188

MAKOTO MlZUNAMl

discussion of projection patterns of S-neurones in some insects, see Goodman (1981). Both the direction of information flow in S-neurones and the order of S-neurones are not known with certainty in most insects, but it appears that many are second-order neurones receiving synapses from photoreceptors in the ocellar plexus (Goodman, 1981). Other neurones are possibly thirdorder neurones and some others are possibly efferent neurones. Thus, it appears that the signals from ocellar photoreceptors are transmitted to the brain via two separate channels, i.e. a small number of large second-order neurones and a large number of small second-order neurones in insect ocelli (Goodman, 1976a,b; Goodman and Williams, 1976; Chappell et al., 1978; Goodman, 1981). An exception to this rule has been recently found in the cockroach, where the ocellar nerves d o not contain small second-order neurones (Mizunami, in preparation). Injections of cobalt into the ocellar nerve and the ocellar tract revealed the morphology of 25 interneurones, none of which, except for four L-neurones, extended dendritic branches into the ocellar plexus where the photoreceptors make output synapses. Thus, in the cockroach, ocellar L-neurones form a signal afferent channel to transmit photoreceptor signals to the brain. The multiplicity in the termination areas of ocellar S-neurones (Fig. 14) hints at diversity in the functional roles of the ocelli. The terminal areas of the ocellar S-neurones can be classified into: 1. sensory centres (visual, olfactory and mechanical centres) which include the medulla and lobula neuropils of the optic lobe, antenna1 lobe and dorsal deutocerebrum and tritocerebrum; 2. higher association centres including the mushroom body and central complex; 3. premotor centres from which descending brain neurones originate, including the posterior slopes and lateral protocerebrum (Fig. 14A). Possible functional roles of the S-neurones, therefore, include: 1. to modulate the activity of visual, olfactory and mechanosensory systems; 2. to influence the activity of higher brain centres; 3. to modulate the activity of multimodal descending neurones. S-neurones projecting into the optic lobes are particularly interesting since they may serve as a neuronal basis for the modulatory actions of the ocelli over the compound eyes widely observed among insects (e.g. Hu and Stark, 1980; Rence et al., 1988). Future analysis of S-neurones may clarify some ocellar functions that are, so far, only poorly understood, including the possible participation in regulating the state of arousal. Furthermore, a careful comparison of S-neurones among insects will point out diversities in ocellar functions among insects.

THE DIVERSITY OF INSECT OCELLAR SYSTEMS

5.3

189

SYNAPTIC ORGANIZATION OF OCELLAR TRACT NEUROPIL

The ocellar tract is one of the major output areas of L-neurones in locusts (Guy et al., 1979), crickets (Koontz and Edwards, 1984) and cockroaches (Toh and Hara, 1984). In these insects, L-neurones are not only presynaptic in the ocellar tract but also postsynaptic to many small profiles, thus the tract is a zone in which complex integration of ocellar information takes place. In honeybees (Guy et al., 1979) and wasps (Kral, 1979), however, the ocellar tract seems to be a mere neural tract rather than an integration centre. L-neurones of honeybees and wasps rarely make output synapses until they enter the posterior slope. It appears that there are fundamental differences in neural organizations of ocellar systems among holometabolous and hemimetabolous insects, as will be discussed in Section 5.6. 5.3.1 The ocellar tract of cockroaches In the cockroach (Toh and Sagara, 1984) and the locust (Goodman et al., 1979), the ocellar nerve is a synaptic area that continues to the ocellar tract neuropil where L-neurones make output synapses onto third-order neurones (see Fig. 16). In the cockroach, the ocellar tract and ocellar nerve usually contain four thick fibres of L-neurones, 10-30 retinular axons, a few efferent processes, a few dense cored-vesicle-containingprocesses, a distal extension of the thick third-order processes, and fibres extending to the ventral nerve cord (Toh et al., 1983; Toh and Sagara, 1984; Toh and Hara, 1984). Studies of ontogenetic development in cockroach ocellar systems suggest that the role of the long retinular axons is to guide the axonal growth of interneurones from the brain to the ocellus, and that their participation in visual processing is doubtful (Toh and Yokohari, 1988). The L-neurones are both pre- and postsynaptic to surrounding thin processes throughout the ocellar nerve and ocellar tract. Thus, the ocellar nerve and ocellar tract are not merely the area where L-neurones make output synapses onto dendrites of third-order neurones but also an area where local interactions occur among many types of neurones.

5.3.2

The ocellar tract and anterior ocellar focus of locusts

Electron microscopic studies have revealed that synaptic interactions in the lateral ocellar tract of the locust include: (1) interaction between secondorder neurones; (2) interaction between the second-order neurones and dendrites of descending third-order ocellar neurones running to the ventral nerve cord; ( 3 ) entry into the tract of intrinsic brain interneurones from several different sources, some of them possibly efferents (Goodman et al., 1977; Guy et al., 1977; Goodman et al., 1979).

MAKOTO MlZUNAMl

190

C

ocellus

PS

FIG. 15 (A) Drawings of an L-neurone and a DNI neurone in locust brains viewed dorsally. A neurone of class L1-3 is drawn on the left. O n the right, the major branches of a DNI neurone, a descending third-order ocellar neurone, are drawn. The single arrow points to the projection into the median ocellar tract; the double arrow points to the fan-shaped branch which extends through the tritocerebrum. From Simmons and Littlewood (1989). (B) Diagrammatic representation of the synaptic relationships of the lateral L-neurone LOC2 of the bee, Apis melliferu. This diagram combines a camera lucida drawing of the cell with electron microscopic

THE DIVERSITY OF INSECT OCELLAR SYSTEMS

191

Simmons and Littlewood (1989) studied the ultrastructure and distribution of synapses between L-neurones and a descending third-order neurone, DNI, in the ocellar tract (Fig. 15A). Among three classes of L-neurones of the lateral ocellar tract, two classes, L1-3 (Fig. 11B) and ML1-2, make synapses with the DNI neurone (see Fig. 37A). No output synapses were found from L4-5 (Fig. 11B) in the ocellar tract or in the anterior optic focus, which are entirely wrapped by glial cells for most of the length of the lateral ocellar nerve. Electron microscopic examination of neurones filled with hexamminecobalitic ions (Littlewood and Simmons, 1988) showed that a single L1 neurone makes as many as 9000-10000 anatomical synapses with a DNI neurone. Littlewood and Simmons (1992) further studied the ultrastructure and distribution of discrete anatomical synapses which constitute two distinct types of output connections made by ocellar L-neurones, L1-3 (Fig. 11B). Outputs to neurones L4-5 are excitatory and transmitted tonically, whereas reciprocal connections among the three L1-3 are inhibitory and incapable of transmission for longer than a few milliseconds (Simmons, 1982a, 1985). Transmission in the former synapses occurs by way of graded potentials, while that in the latter synapses requires for action potentials. The tonically transmitting synapses are located in the lateral ocellar tract and are made from the axons of L1-3 to short branches of L4-5. Each excitatory connection is composed of a few hundred discrete anatomical synapses, each characterized by a bar-shaped presynaptic density. Associated with tonic synapses are abundant invaginations of the presynaptic membrane. Synapses of the reciprocal, inhibitory, phasic connections occur in arbors of L1-3 located at the anterior ocellar focus. Each phasic connection is composed of a few tens of discrete anatomical synapses. Compared with the tonic, excitatory connections, there are fewer vesicles and less invagination of the presynaptic membrane associated with each synapse. Interestingly, in the cockroach where there are few synaptic connections among L-neurones (Toh and Hara, 1984; Mizunami, 1990a), the arborization of L-neurones in the anterior ocellar focus is sparse (Mizunami et al., 1982). The results of Littlewood and Simmons (1992) have indicated that, although the axons of lateral L-neurones lie parallel to, and in close

evidence from several preparations. Solid arrows: input synapses from receptor cells in the ocellar retina. Open arrows with tails: input synapses from S-neurones. Open arrows: output synapses in the distal axonal region and arborization in the posterior slope. PB, protocerebral bridge; OS, oesophagus; LO, lateral ocellus; T, trachea. From Guy et al. (1979). (C) Reconstruction of two parallel large second-order ocellar neurones of the wasp, P. gerrnanica, based on photographs of Procion yellow-filled ocellar nerves. Sites of lateral synapses are marked by curved arrows. Direction of information flow is indicated by arrows. PS, posterior slope. From Kral (1982).

MAKOTO MlZUNAMl

192

Cockroach

m L

Bee

Locust

L1-3

L4,5

L

LD

FIG. 16 Schematic comparison of synaptic organization of ocellar L-neurones of three representative insects, cockroaches, locusts and bees. Based on the research of Toh and Hara (1984) and Toh and Sagara (1984) for the cockroach, Periplaneta americana; Simmons and Littlewood (1989) and Littlewood and Simmons (1992) for the locust, Schistocerca gregaria; and Guy et al. (1979) for the bee, Apis mellifera. ON, ocellar nerve; AOF, anterior ocellar foci; VNS, ventral nerve cord. The major input areas (stippled) and output areas (filled) are indicated. Note that synaptic organization of a class of locust L-neurones (Ll-3) is similar to that of cockroach L-neurones, while synaptic organization of the other class of locust L-neurones (L4,5) is similar to that of a class of bee L-neurones. LD-neuronesof the bee appear to receive synaptic inputs in the posterior slope (Goodman et al., 1990).

proximity between the ocellus and protocerebrum, the synapses which L1-3 make with L4-5 are restricted to a short length of the ocellar tract. The synapses which the L1-3 neurones make among themselves are located only in their arbors in the anterior ocellar focus. In addition, the outputs of L4-5 are restricted to the posterior slope, whereas those of L1-3 to third-order neurones are distributed throughout the ocellar tract. In short, the three major output areas of the L-neurones, i.e. ocellar tract, anterior ocellar foci and posterior slope, are functionally separated: (1) ocellar tracts are the major output area of L1-3 and ML neurones where signal transmissions occur by way of graded potentials; (2) the anterior optic foci are the areas where interactions between L1-3 neurones occur, by way of action potentials; and ( 3 ) the posterior slope is the output area for L4-5 neurones (see Fig. 16).

THE DIVERSITY OF INSECT OCELLAR SYSTEMS

193

The individual anatomical synapses which make up each of the two types of physiological connections, i.e. tonic excitatory connections and phasic inhibitory connections, have shown quantitative, rather then qualitative, differences in ultrastructure (Littlewood and Simmons, 1992). Details of the ultrastructure are similar including the shape and size of individual presynaptic densities, and the diadic output configuration (one output to two receptive profiles). The synaptic vesicles associated with the presynaptic densities are of the same diameter at both types of synapses. These observations suggest that the different postsynaptic effects of two types of synapses of L-neurones may not be due to the difference in the types of transmitters released, but due to the differences in the amount of transmitters and the actions of the transmitters upon the postsynaptic membrane.

5.3.3

The ocellar tract of bees and wasps

The synaptic organization of the ocellar tract of holometabolous insects has been studied in honeybees (Guy et al., 1979) and wasps (Kral, 1979). In both species, the ocellar tract is a simple fibre tract, rather than an integration neuropil as in the case of locusts, crickets and cockroaches. In the honeybee (Guy et al., 1979) and wasps, Paravespula vulgaris and Paravespula germanica (Kral, 1983), the axons of L-neurones are separated from one another by glial cells in the ocellar tract until they enter the posterior slopes of the protocerebrum where the glial cells disappear so that the axons come very close to each other and form junctions. L-neurones of honeybees (Guy et al., 1979) and wasps (Kral, 1979) rarely make output synapses until they enter the posterior slope (Fig. 15B,C). Interestingly, the morphology and synaptic organization of the L4-5 neurones of locusts are analogous to the L-neurones of bees whose major output area is the posterior slope (Fig. 16). In contrast, the L1-3 neurones of locust ocelli are analogous to the L-neurones of cockroaches whose output area is the ocellar tract and the anterior ocellar focus (Fig. 16).

5.4

SYNAPTIC ORGANIZATION OF POSTERIOR SLOPE NEUROPIL

The posterior slope neuropil is one of the major terminal areas of L-neurones. In locusts and bees, large lobula fibres similar in form and disposition to the movement-sensitive fibres of flies enter the posterior slope (Goodman, 1981). In addition, several commissural tracts containing large and small fibres cross from lobula to lobula and from medulla to medulla through the region of the L-neurone arborizations in locust and bee, and collaterals from these tracts can be seen to mingle with the dendrites of the L-neurones (Guy et al., 1979; Goodman, 1981). In holometabolous insects, such as bees and flies, this area includes the perioesophageal neuropilar area which receives extensive L-neurone terminal arborizations. In flies, descend-

MAKOTO MlZUNAMl

194

A

B

100um

I

1

100”rn

FIG. 17 (A,B) Convergence of ocellar and compound eye inputs to descending neurones of the fly Calliphora. (A) Arrangements of ocellar interneurone terminals at DNOVS 1. At Idr ocellar interneurones converge with lobula plate endings. Specializations invest the Idr and the palmate dendrites of its m3 field. Other endings, segregated from VS terminals, invest the adr, especially its d l field. The lower drawing shows two similar endings, one from each side, converging to the right DNOVS 1 and a bilateral terminal investing both left and right DNOVS 1. Terminals figured here represent four of several morphological variants amongst the twelve large interneurones. (B) Forms of VS neurones (filled profiles) superimposed by DNOVS 1 (open profile). Collaterals (open arrows) extend ventrally into the deutocerebrum. VS4,5: the dendrites are limited to the v-stratum (v) and represent a vertical strip of front-lateral retina. Their branched terminals at DNOVS 1 give rise

THE DIVERSITY OF INSECT OCELLAR SYSTEMS

195

ing neurones originating from the posterior slope receive inputs directly from sensory systems and indirectly via higher centres and peptidergic interconnections from the protocerebrum (Strausfeld, 1976; Strausfeld et af., 1984). Direct inputs include primary mechanosensory afferents, interneurones from the olfactory lobes and ocellar L-neurones, and higher order visual neurones that interact with retinotopic inputs from compound eyes. Each cluster receives a unique combination of inputs which are shared by the dendritic trees of its constituent descending neurones. Strausfeld and Bassemir (1985) have further reported that, in flies, a cluster of three Y-shaped descending neurones receives ocellar interneurones and vertical cell terminals (Fig. 17). Some of these descending neurones extend through the pro-, meso- and metathoracic ganglia, branching ipsilaterally within their tract and into the inner margin of the leg motor neuropil of each ganglion, and one other terminates as a stubby ending in the dorsal prothoracic ganglion onto the main dendritic trunks of neck muscle motor neurones. These descending neurones appear to participate in saccadic head movement and in the stabilization of flight course (Strausfeld and Bassemir, 1985). In the posterior slope, three distinct termination areas of the L-neurones are observed within flies (Nassel and Hagberg, 1985) and two distinct termination areas are found in dragonflies (Chappell et al., 1978). One consequence of the direct projection of L-neurones from the ocellar retina to the posterior slope is that ocellar signals can reach descending neurones more rapidly than those in the compound eye pathway (see Section 3.4). In the cockroach, no L-neurones extend into the posterior slope but instead, some ocellar third-order neurones extend terminal arborizations into this area (Mizunami and Tateda, 1986). Recently, Mizunami (in preparation) has observed several classes of higher ocellar neurones originating from the posterior slope of the cockroach including: (1) multimodal descending neurones; (2) neurones projecting into the mushroom bodies, protocerebral bridge, central complex, or lamina neuropils of the optic lobe; and (3) neurones connecting bilateral posterior slopes. 5.5

MORPHOLOGY OF THIRD-ORDER OCELLAR NEURONES

Because ocellar L-neurones, except for LD, terminate in either the ocellar tract or the posterior slope of the brain (Section 5.2.2), third-order ocellar to form one to three longer processes, which penetrate ventrally. VS6-9: the h-dendrites ( h ) represent the upper one-third of the retina and the v-dendrites, a band spanning the lower two-thirds of the retinotopic mosaic. VS6 and 7 have tuberous collaterals lateral to DNOVS 1 and short branches at the DNOVS cluster. VS8 and 9 terminals are comparatively simple. A, anterior; L, lateral; ca, calyx of mushroom body; oe, oesophagus. From Strausfeld and Bassemir (1985).

196

MAKOTO MlZUNAMl

neurones of the brain can be classified into two types: (1) neurones originating from the ocellar tract neuropil; and (2) neurones originating from the posterior slope. The former neurones have been studied in detail in the cockroach, where all L-neurones terminate in the ocellar tract. The latter neurones have been examined in the honeybee, where all L-neurones, except for LD, terminate in the posterior slope. Target neurones of LD in the thoracic ganglia have not been reported. In the cockroach, Mizunami and Tateda (1986) described nine types of ocellar interneurones with arborizations in the ocellar tract. Subsequent simultaneous intracellular recordings from L-neurones and these neurones have shown that these neurones receive monosynaptic excitatory synapses from L-neurones (Mizunami and Tateda, 1988a). Further intracellular and extracellular staining of ocellar tract neurones has revealed that the ocellar tract contains at least 25 neurones (Mizunami, in preparation), with 15 neurones shown to be third-order neurones (numbered 5-10, 13-15, 17-20, 23-24 in Fig. 18; Mizunami and Tateda, 1986; Mizunami, in preparation). The other 10 neurones are: (1) four second-order neurones (numbered 1-4; Mizunami et af., 1982); (2) two possible efferent neurones (numbered 11, 12; Ohyama and Toh, 1990a); (3) three third-order or efferent neurones (numbered 21, 22, 25); and (4) one neurone still to be characterized (numbered 16). Third-order ocellar neurones project into a number of discrete neuropil areas of the brain. The flow of ocellar and other sensory information in the cockroach brain is schematically illustrated in Fig. 19, which shows that the projection areas of third-order ocellar neurones include: (1) visual, olfactory and mechanosensory centres; (2) the mushroom body (a higher association centre; Mizunami et af., 1993); (3) premotor centres, including the posterior slope, from which descending brain neurones originate; and (4) the thoracic motor systems. Interestingly,

FIG. 18 Summary diagrams of the neural organization of the ocellar system of the cockroach, Periplunetu urnericanu. Morphologies of 2.5 ocellar interneurones identified in the ocellar tract neuropils (OTs). viewed posterodorsally, are drawn. Ocellar photoreceptors synapse onto four second-order neurones (numbered 1 4 ) . The large second-order neurones (L-neurones) exit the ocellus (OC) and project into the ocellar tract neuropil (OT) of the protocerebrum (PC), through the ocellar nerve (ON). In the ocellar tract neuropil, L-neurones make synapses onto at least 15 third-order neurones (numbered 5-10, 13-15, 17-20, 23-24) which project into a variety of neuropil areas of the brain including posterior slope (PS; premotor centre where descending brain neurones originate), medulla (ME) and lobula (LO) of the optic lobe (visual centres where signals of the compound eyes are processed), tritocerebrum (TC. mechanosensory centre), antenna1 lobe (AL, olfactory centre), calyx (CA) of mushroom body (higher associative centre). There are, in addition, at least two efferent neurones (numbered 11 and 12; Ohyama and Toh, 1990a). Whether the remaining three neurones (numbered 16, 21, 22, 2.5) are third-order or efferent remains to be determined. From Mizunami (submitted).

2

MAKOTO MIZUNAM

198

the projection areas of ocellar third-order neurones of the cockroach are more o r less similar t o those of S-neurones of other insects (see Fig. 14), most of which appear t o be second-order neurones. Therefore, ocellar signals a r e transmitted to more o r less similar target neuropil areas in all insects despite the difference in the order of neurones which transmit ocellar signals to these areas. In the bee, Milde (1988) has described neurones which respond to both ocellar and compound eye inputs in the posterior protocerebrum, an area equivalent to the posterior slope of other insects. Because their dendritic arborizations overlap the terminal branches of L-neurones, they most probably receive direct inputs from L-neurones (see Fig. 36a,b). Interestingly, morphologies of some possible fourth-order ocellar neurones in the posterior slope of the cockroach (Mizunami, in preparation) are very similar to those of possible third-order ocellar neurones in the posterior slope of the bee.

-I

I1

CENTRAL COMPLEX

I

SENSORY AFFARENTS

MOTOR OUTPUTS

SECOND-OROER NEURONS OCELLAR PLEXUS

t

PH$gERFt:PTORS

FIG. 19 Schematic representation of ocellar pathways in the cockroach brain. The insect brain consist of: (1) sensory centres, i.e. mechanosensory, visual, and olfactory centres; ( 2 ) higher associative centres including mushroom body and central complex; and (3) premotor centres from which descending brain neurones originate. Ocellar signals are first processed in the ocellar plexus, and then further processed in the ocellar tract neuropil. Third-order neurones originating from the ocellar tract neuropil transmit ocellar signals to a variety of target neuropil areas including sensory centres, higher centres, premotor centres and thoracic motor centres. From Mizunami (in preparation).

THE DIVERSITY OF INSECT OCELLAR SYSTEMS

5.6

199

DIVERSITY AND EVOLUTION OF N E U R A L CIRCUITS I N OCELLAR SYSTEMS

By comparing neural organizations of ocellar systems among different insects, Mizunami (in preparation) has proposed that insect ocellar systems can be classified into three types (Fig. 20): 1. The ‘cockroach-type’ ocellar system where signals from photoreceptors, at first, converge onto a small number of second-order neurones and then diverge to a large number of third-order neurones which project into a number of neuropil areas of the brain (Mizunami and Tateda, 1986; Mizunami, in preparation). 2. The ‘bee-type’ ocellar system where photoreceptor signals are passed to target neuropils by a large number of large and small second-order neurones (Guy et al., 1979; Milde, 1986). 3. The ‘locust type’ ocellar system where ocellar signals are transmitted to target neuropils by both second- and third-order neurones (Goodman, 1976b; Goodman and Williams, 1976; Pan and Goodman, 1977; Reichert et af., 1985). The ocellar system of bees can be termed a bisynaptic type since only two synapses intervene when photoreceptor signals are transmitted to target neuropil areas, while those of cockroaches can be termed a trisynaptic type. The ocellar system of locusts contain both bi- and trisynaptic pathways (Fig. 20). Mizunami (in preparation) has proposed a possible functional difference among the three types of ocellar systems, based on the discussion of functional properties of insect ocelli. There is evidence to suggest that insect ocelli supplement the compound eyes in speed and sensitivity (Section 3). The higher speed of signal transmission in the ocelli compared with the compound eyes is due, at least in part, to the smaller number of neurones intervened when signals of ocellar photoreceptors are transmitted to target areas including motor centres (Wilson, 1978a; Goodman, 1981; Reichert et al., 1985). The higher sensitivity of ocelli mainly reflects the higher ratio of convergence between photoreceptors and second-order neurones (Wilson, 1978a; Goodman, 1981). It appears that there are different compromises between sensitivity and speed in the ocellar systems of different insects. The smaller number of neurones intervened in the bee ocelli, compared with those in the cockroach ocelli, possibly allows for a faster transmission of signals. In the cockroach ocelli, signals from 10 000 photoreceptors first converge onto a very low number (only four) of neurones before they are transmitted to target brain areas. This is possibly a design to enhance sensitivity by attaining a high ratio of convergence from photoreceptors to second-order neurones; Weber and Renner, 1976; Mizunami et al., 1982), although at a possible cost of speed. Indeed, the intracellularly measured absolute sensitivity of cockroach L-neurones (Mizunami and Tateda, 1986;

200

MAKOTO MlZUNAMl Cockroach

-I

I

Locust

Bee

t

Photoreceptors Ocellar Plexus Second-order Neurons

Ocellar tract Neuropil Third-order neurons

Target neuropils

FIG. 20 Comparison of neural organizations of ocellar systems of three representative insects, cockroaches, bees and locusts. In the cockroach, Peripfaneta americana, ocellar photoreceptors converge onto four large second-order neurones (L-neurones) at the ocellar plexus. L-neurones synapse onto at least 15 third-order neurones in the ocellar tract neuropil. Third-order neurones transmit ocellar signals to a variety of neuropil areas, including the posterior slope, thoracic motor centres, optic lobe, and tritocerebrum. In the bee, Apis melfiferu, a large number of large (L) and small (S) second-order neurones originate from the ocellar plexus. Dendritic arborizations of L-neurones (thicker lines) appear to cover wider areas of the ocellar plexus than S-neurones (thinner). Second-order neurones project into a variety of neuropil areas: Some L-neurones project into the posterior slope while others project into thoracic motor centres; S-neurones project into various neuropil areas of the brain, including the optic lobe and tritocerebrum (Goodman, 1981; Milde, 1986). In the locust, Schisrocerca gregaria, some L-neurones project into the ocellar tract where they synapse onto third-order neurones, while other L-neurones project into the posterior slope (Simmons and Littlewood, 1989). Some third-order neurones originate from the ocellar tract and project into thoracic motor centres (Simmons, 1980). A number of S-neurones originate from the ocellar plexus and project into various neuropil areas of the brain (Goodman, 1981). including the optic lobe and tritocerebrum. In the locust, thus, ocellar signals are transmitted to a variety of neuropil areas by both second- and third-order neurones. From Mizunami (submitted).

M. Mizunami, unpublished), is one or two log units higher than that reported for bee L-neurones (Baader, 1989), while the response time course is much slower in the cockroach L-neurones. I thus propose that the bisynaptic ocellar circuits of the bee are a high-speed type and the trisynaptic ocellar circuits of the cockroach are a high-sensitivity type. The ocellar circuits of the locust, which contain both bi- and trisynaptic pathways, are characterized as an intermediate type, where both the speed and sensitivity are possibly important.

THE DIVERSITY OF INSECT OCELLAR SYSTEMS

201

Large second-order neurones (L-neurones) of bisynaptic pathways appear to receive synapses from a larger number of photoreceptors than other smaller second-order neurones (S-neurones) , since dendritic arborizations of the former appear to cover wider areas of the ocellar plexus than the latter (Goodman, 1981; Milde, 1986; see Fig. 20). Thus, L-neurones in the bisynaptic pathways may form a channel to transmit ‘sensitive’ signals to a few specific target neuropils, i.e. the posterior slope in the case of locusts (Goodman, 1976) and the posterior slope and thoracic motor centres in the case of bees (Goodman, 1981), while all other target neuropils receive ‘less sensitive’ signals from a large number of S-neurones. This differs from the trisynaptic pathways where ‘sensitive’ signals of L-neurones are transmitted to all target neuropils via third-order neurones. Among the eight families of the six orders of insects in which neural organization of ocellar systems have been reported in some detail, including dragonflies (Aeschnu and Anux: Chappell et u f . , 1978; Patterson and Chappell, 1980), crickets (Achetu: Koontz and Edwards, 1984), wasps (Puruvespulu: Kral, 1982, 1983), moths (Trichopfusiu: Pappas and Eaton, 1977; Eaton and Pappas, 1977) and flies (Muscu and Culliphoru: Strausfeld, 1976; Nassel and Hagberg, 1985), the neural circuits of the ocellar systems of all holometabolous insects examined (bees, wasps, moths and flies) can be classified as bisynaptic (fast) types. Among the four hemimetabolous insects so far studied (dragonflies, locusts, crickets and cockroaches), ocellar circuits of cockroaches can be classified as a trisynaptic (sensitive) type, and those of locusts and crickets as an intermediate type having both bi- and trisynaptic pathways. The ocellar circuits of dragonflies were classified as an intermediate type, although the evidence to show that they have trisynaptic pathways is not perfect. The phyletic distribution of different types of ocellar circuits is shown in Fig. 21. It is well established that the compound eyes and ocelli of insects living in different visual environments often exhibit different properties. Autrum (1950) measured ERGs of the compound eyes of a variety of insects, and classified the compound eyes into two functionally different categories: fast eyes and slow eyes. Fast eyes were found in diurnal, swiftly flying insects such as honeybees (Apis), wasps (Vespu), fleshflies (Culliphoru), and dragonflies (Aeschnu). They were characterized by high flicker fusion frequency (the ability to resolve high rates of flicker) and low photic sensitivity. Slow eyes, found in nocturnal insects such as cockroaches (Peripfuneru) and crickets ( Tuchycines) were characterized by low flicker fusion frequency and high sensitivity. Ruck (1958) compared the ERGs of ocelli of three diurnal, swiftly flying insects, the dragonfly (Puchydipfux), honeybee (Apis),and fleshfly (Phormiu) and one nocturnal, weak flyer, the cockroach (Periplunetu). He observed that: (1) the fleshfly and honeybee ocelli had a high flicker fusion frequency and a low absolute sensitivity, i.e. their ocelli were characterized as a fast and less sensitive type; (2) cockroach

MAKOTO MlZUNAMl

202

Palaeoptera

Neoptera I

-1

Pol yneoptera

Holornetabola

I

Dragonflies

IM

I

I 1

Locusts Crickets

IM

Cockroaches

SE

Bees wasps

Flies

Moths

F

F

F

FIG. 21 The phyletic distribution of different types of ocellar circuits. Among the eight families from six order of insects, which are dragonflies (Aeschna tuberriculiferu and Anax junius; Odonata) (Chappell et al., 1978; Patterson and Chappell, 1980), locust (Schistocerca gregaria, Orthoptera) (Goodman, 1976; Goodman and Williams, 1976; Pan and Goodman, 1977; Reichert et al. 1985; Simmons and Littlewood, 1989; Littlewood and Simmons, 1992), crickets (Acheta domesticus, Orthoptera) (Koontz and Edwards, 1984), cockroaches (Periplaneta americana, Dictyoptera, Blattaria), bees (Apis mellifera, Hymenoptera) (reviewed by Goodman, 1981; Milde, 1986), wasps (Paravespula vulgaris and P. germanica, Hymenoptera) (Kral, 1982, 1983), moths (Trichoplusia ni, Lepidoptera) (Pappas and Eaton, 1977; Eaton and Pappas, 1977). and flies (Musca dornestica and Calliphora erythrocephala. Diptera) (Strausfeld, 1976; Nassel and Hagberg, 1985), ocellar circuits of bees, wasps, moths and flies can be classified as a bisynaptic or a fast (F) type, ocellar circuits of cockroaches as a trisynaptic or a sensitive type (SE), and ocellar circuits of dragonflies, locusts and crickets as an intermediate type (IM) having both bisynaptic and trisynaptic pathways. The dendrogram is based on the research of Kristensen (1981). From Mizunami (submitted).

ocelli were a slow and sensitive type; and (3) dragonfly ocelli were a fast and sensitive type. The observation fit the hypothesis of Mizunami that proposes that fleshfly and honeybee ocelli are concerned more with speed than sensitivity, cockroach ocelli are concerned more with sensitivity than speed, and the dragonfly ocelli are concerned with both speed and sensitivity. This agreement is remarkable because ERGS reflect, at least in part, properties of photoreceptors, while the classification of ocellar circuits was based solely on the properties of neural wiring underneath the photoreceptors. Further

THE DIVERSITY OF INSECT OCELLAR SYSTEMS

203

careful studies are, of course, necessary to clarify the validity of the functional classification of neural circuits of ocellar systems discussed here. Does the present phyletic distribution of different types of ocellar systems reflect functional adaptation or evolutionary history? A notable observation related to this question is that moths, nocturnal holometabolous insects, do not possess trisynaptic, i.e. presumably sensitive pathways, although there is evidence to show that one of the major roles of their ocelli is to perceive low light intensity for the control of diurnal activity (Eaton et al., 1983; Sprint and Eaton, 1987; Wunderer and Kramer, 1989). This apparent mismatch between the type of ocellar circuits and the functional need can be explained if the evolutionary change in the basic wiring of ocellar circuits, which requires for a fundamental change in the developmental program, is less likely to occur than that in other characters of the ocellar system and thus, the change in the life style, which occurred very frequently, resulted in the change in characters such as the size of the ocellar lens and the number of photoreceptors but not the patterns of the neural wiring. The fact that a diurnal (a locust, S . gregaria) and a nocturnal Orthoptera (a cricket, A . domesticus) share the same type of ocellar circuits is in agreement with the hypothesis that the change in ocellar circuits occurred only rarely and consequently, the present phyletic distribution of different types of ocellar circuits reflects not only functional adaptation but also evolutionary history. Consider the most parsimonious scheme where the present phyletic distribution of different types of ocellar circuits (Fig. 21) are explained by evolutionary changes in the types of ocellar circuits of only twice. In the scheme: (1) the ocellar circuits of common ancestors of eight pterygote insects (dragonflies, locusts, crickets, cockroaches, honeybees, wasps, moths and flies) were the intermediate type containing both bi- and trisynaptic pathways: (2) the trisynaptic (sensitive-type) circuits of cockroaches evolved by a deletion of bisynaptic pathways of the intermediate-type circuits of their ancestors; and (3) the common ancestors of four holometabolous insects (flies, bees, wasps, and moths) attained bisynaptic (fast-type) circuits by a deletion of trisynaptic pathways. This hypothesis awaits future evaluation in comparative studies of ocellar systems of a large number of species within a class and within particular orders. 5.7

SUMMARY

Electron microscopic studies have shown that synaptic organizations of ocellar plexus are complicated and differ among different insects. The functional significance of the complexity and diversity needs to be clarified. The morphologies of L-neurones can be compared among different individuals of conspecific insects and among different species, providing a model to study evolutionary changes of single identified neurones. The morphologies and synaptic organizations of second- and third-order ocellar neurones differ among insects. By comparing neural organizations of insect

MAKOTO MlZUNAMl

204

ocellar systems, I have proposed that insect ocellar circuits can be classified into three functional types: (1) ocellar circuits of ‘bisynaptic’ type which are concerned more with speed than sensitivity; (2) ocellar circuits of ‘trisynaptic’ type which are concerned more with sensitivity than speed; and (3) ocellar circuits having both bi- and trisynaptic pathways which are concerned with both speed and sensitivity. It appears that the present phyletic distribution of different types of ocellar circuits reflects not only functional adaptation but also evolutionary history. I conclude that, because of their great accessibility to detailed anatomical and electrophysiological investigations, insect ocellar systems provide a rich field from which to study the evolution of complex neural systems at the level of synapses, neurones and neural circuits. 6 Molecular basis of the ocellar system

In this section, I will summarize present knowledge on the substances which play important roles in the ocellar system, such as visual pigments and neurotransmitters. Although we know little about such molecules at present, we have at least four reasons to study such molecules in the future. First, and most importantly, we can extend our understanding of ocellar systems by knowing their molecular basis. Second, the identification of substances used in ocellar systems facilitates pharmacological analysis of ocellar systems. For example, if we know the transmitter of a particular ocellar pathway we can pharmacologically examine the functional roles of that pathway. Similarly, by knowing the neuromodulators of ocellar systems, long-term modulation of ocellar functions can be examined. Third, by comparing molecules participating in the ocellar systems and those in other visual systems, e.g. compound eyes, we can discuss the generality and specificity of visual systems. Fourth, a comparison of the amino acid sequences of proteins often provides information about the evolution of visual systems. An example is discussed in Section 6.1. In the final part of Section 6, I will discuss some structural mutants of Drosophih, which will provide useful material to examine the molecular and developmental mechanisms in the formation of ocellar structures. 6.1

VISUAL PIGMENT A N D ITS MOLECULAR EVOLUTION

Visual pigment contained in the photoreceptor cells of both vertebrates and invertebrates consists of a protein, opsin, covalently bound to a chromophoric group, the 11-cis retinal. Absorbed light quanta cause the 11-cis chromophore of the visual pigment to isomerize to the all-trans form. This in turn leads to conformational changes of the protein. These changes initiate a

THE DIVERSITY OF INSECT OCELLAR SYSTEMS

205

complex sequence of biochemical and biophysical events that eventually lead to the excitation of the photoreceptor cell (Schwemer, 1989). Different spectral types of photoreceptors in insects contain different opsins. Four opsin genes have been identified in Drosophila: Rhl is expressed in R1-6 of the compound eye which contain green-sensitive pigment. Rh3 and Rh4 are expressed in different subsets of R7 which contain UV-sensitive pigment (Zuker et a f . , 1987). It has been postulated that Rh2 may be expressed in R8 (Cowman et af., 1986), but subsequent studies have failed to confirm this (Feiler et al., 1988; Pollock and Benzer, 1988). Lateral ocelli of Drosophifa larva express R h l , Rh3 and Rh4 but not Rh2 (Mismer and Rubin, 1987; Pollock and Benzer, 1988). Recent studies have shown that Rh2 is predominantly expressed in the ocellar photoreceptors (Pollock and Benzer, 1988; Feiler et al., 1988; Mismer et a f . , 1988), which are sensitive to both blue and UV. The blue sensitivity of ocellar photoreceptors is due to Rh2 opsin and the UV sensitivity is due to sensitizing pigment (Kirschfeld et al., 1988). The structure of the Rh2 opsin proposed by Cowman et al. (1986) is shown in Fig. 22A. Rh2 opsin contains all of the structural features typical of opsins, seven hydrophobic domains separated by hydrophilic sequences, a putative retinal-binding site in the seventh transmembrane domain (Lys 326), a glycosylation site in the extracytoplasmic face (Asn 27), and a series of potential phosphorylation sites in the C-terminal region (Ser and Thr residues). Zuker et a f . (1987) found that when Rh2 genes are expressed in R1-6 cells, the cells are capable of responding to light. This indicates that there must be common features in the transduction process in the ocellar and Rl-6 photoreceptors. A phylogenic tree of visual pigments (Fig. 22B), constructed on the basis of amino acid identity (Zuker et a f . , 1987), indicates that: (1) P-adrenergic receptors and opsins of both humans and flies are derived from a common ancestral protein; (2) different types of opsins diverged in the compound eyes before ocellar opsins (Rh2 opsins) were separated from R l - 6 opsins (Rhl opsin). This second point indicates that multiple colour receptors were differentiated in the compound eye before ocellar photoreceptors were separated from compound eye photoreceptors, i.e. ancestors of insects attained colour vision before attaining ocelli. Interestingly, Okano et a f . (1992) analysed the molecular evolution of visual pigments of vertebrates and concluded that pigments for scotopic vision (rod pigments) evolved out of pigments for colour vision (cone pigments). Thus, if we regard the ocelli as receptors for low light vision, visual pigments of vertebrates and insects share the same evolutionary history in the advancement of visual functions. Further analysis of molecular evolution of visual pigments will also help to solve a long-lasting controversy about the evolution of arthropod visual systems, i.e. the monophylic o r polyphylic origin of the compound eyes and ocelli (Paulus, 1979; Nilsson, 1989; Land and Fernald, 1992).

206

6.2

MAKOTO MlZUNAMl NEUROTRANSMITTERS AND NEUROHORMONES

Immunocytochemical and pharmacological evidence suggests that histamine is the major transmitter of photoreceptors of both the ocelli and the compound eyes of insects. In flies, Calliphora erythrocephala and Musca dornestica, Nassel et al. (1988) have found histamine-like immunoreactivity in some types of photoreceptors, R1-6 and R8, of the compound eye and from ocellar photoreceptors. The other photoreceptor type of the fly compound eye, R7, is immunoreactive to GABA (Datum et al., 1986). Histamine-like immunoreactivity is found also in the photoreceptors of locust ocelli (Schlemermeyer et a f . , 1989) and in the intracranial ocelli of the moth, Manduca sexta (Homberg and Hildebrand, 1991). Pharmacological experiments suggest that histamine is the major neurotransmitter of photoreceptors of fly compound eyes (Hardie, 1987, 1988). Simmons and Hardie (1988) have noted that when histamine is injected by iontophoresis into the locust ocellar plexus, L-neurones hyperpolarize in a dose-dependent manner (Fig. 23A-C). Hyperpolarizing potentials in L-neurones evoked by histamine had the same reversal potential as hyperpolarizing potentials evoked by ocellar illumination (Fig. 23D). Curare blocked the responses of the L-neurones to histamine and to light. Schlemermeyer et al. (1989) have observed an uptake of histamine by photoreceptors and glial cells, and depolarization-induced, calcium-dependent release of histamine from photoreceptors of locust ocelli. In the cockroach, Lin et al. (1990) have noted that hyperpolarization of L-neurones induced by iontophoretic application of histamine persisted after the blockage of synaptic transmission by Co2+, thereby suggesting that histamine receptors exist on the L-neurones. Histamine receptors of L-neurones differ pharmacologically from those so

Fig. 22 (A) Proposed structure of the Rh2 opsin of Drosophila. Amino acid residues are indicated by single-letter codes. Black solid circles indicate differences between the Drosophila ninaE opsin and the Rh2 opsin sequences. The positions at which the ninaE and Rh2 opsins are identical are indicated by open circles; stippled circles indicate conservative changes (M:V:I:L, A:G, F:Y, S:T:Q:N, K:R, and E:D). Squares indicate the positions of the charged (E,D and K,R,H) amino acids. Arrows indicate the positions of the three introns in the Rh2 opsin gene. From Cowman et al. (1986). (B) Phylogeny of visual pigments and the P-adrenergic receptor. This phylogenetic tree relating the different animal visual pigments and the p-adrenergic receptor was constructed on the basis of the principle of minimal mutation distances (parsimony principle). Branch lengths (number of mutational events) are indicated in the scale below the tree. Dm-Rhl stands for the Drosophila ninaE gene, Dm-Rh2, Dm-Rh3 and Dm-Rh4 refer to the Drosophila Rh2, Rh3, and Rh4 opsin genes. The hamster P-adrenergic receptor (Dixon et al., 1986) is referred to as ha-p-AR. h-Green, h-Red, and h-Blue refer to the human colour opsins, and h-Rhod of human rhodopsin (Nathans and Hogness, 1984; Nathans et al., 1986). From Zuker et al. (1987).

207

THE DIVERSITY OF INSECT OCELLAR SYSTEMS

A

B

ha-8 -RR Om-Rhl Om-Rh2 Om-RhJ Om-Rh4

,- h-Green h-Blue

I 280

I

I

I

I

I

I

1

240

200

160

120

80

40

0

Ruerage number o f amino acid rubrIlIuIions

MAKOTO MlZUNAMl

208

D

6

A

& lOnA

................

L-

20m"L

-,

5 nA

. .--

....................

-

L

I _ . -

A.B. 4 s

C

C.D.8s

FIG. 23 (A-D) Responses of an L-neurone of locust lateral ocellus to ionophoretically applied histamine. ( A ) The ocellus was repeatedly illuminated with identical pulses of light (monitor on bottom trace), and a 20 n A pulse of histamine was delivered (middle trace). The histamine hyperpolarized the L-neurone (top trace), and reduced the amplitude of its response to light. (B) Two L-neurones were recorded simultaneously. Histamine hyperpolarized one (second trace), but not the other (top trace). (C) A series of responses of an L-neurone to pulses of histamine of increasing size. The amplitude of the current used to eject histamine is given beneath each record. The record on the right is the response to a longer, 1 0 n A pulse of histamine. (D) Potential changes and currents in an L-neurone evoked by pulses of light and by a 10 nA pulse of histamine. The upper recording is of potential changes in the neurone, and in the lower recording, the potential has been clamped (upper

THE DIVERSITY OF INSECT OCELLAR SYSTEMS

209

far reported in vertebrates (Hardie, 1988; Lin et al., 1990). For example, they are sensitive to some antagonists of acetylcholine receptors, such as curare (Hardie, 1987, 1988; Lin ef al., 1990, Fig. 23F), and thus are most probably a novel class of receptor. In cockroaches and locusts, L-neurones receive depolarizing inputs mediated by GABA (y-aminobutyric acid), in addition to hyperpolarizing inputs mediated by histamine. In cockroaches, Lin et al. (1990) have found that GABAergic inputs to L-neurones derive from efferent neurones. Spike discharge of efferent neurones induced a depolarization in L-neurones in cockroaches (Ohyama and Toh, 1986). Depolarization of L-neurones caused by efferent spikes and that by iontophoretical application of GABA was depressed by perfusion of picrotoxin (Fig. 23E). Depolarization of Lneurones evoked by iontophoretic application of GABA was maintained after the blockade of synaptic transmission by Co2+,suggesting that GABA receptors exist on the L-neurones. In locusts, Taylor (1981~)observed that some L-neurones of the medial ocellus depolarized when a lateral ocellus was illuminated. The depolarizing response was abolished with picrotoxin, thus it is probably mediated by GABA. In the ocellar plexus of locusts, about 10 GABA-immunoreactive fibres form a dense network of fine arborizations (Ammermiiller and Weiler, 1985). It is not known if these neurones are small second-order neurones or efferent. Some GABA immunoreactive small fibres are also seen in the ocellar plexus of honeybees (Schafer and Bicker, 1986) and cockroaches ( Y . Toh, personal communication). An interesting feature of GABA-mediated synaptic transmission to L-neurones is that it causes a depolarization. In most GABA-mediated synapses, activation of GABA receptors results in an increase in the permeability to chloride ions whose equilibrium potential is negative to, or near the resting membrane potential. Thus, it is an exceptional observation that GABA induces a depolarization in L-neurones. Furthermore, there is evidence to show that the histamine-induced hyperpolarization of large second-order neurones (lamina monopolar cells) of the fly compound eye is mediated by chloride ions (Hardie, 1989). Thus, it is most likely that the histamine-induced hyperpolarization of ocellar L-neurones is mediated by chloride whereas its GABA-induced depolarization is mediated by an ion other than chloride, possibly sodium.

trace) to reveal transmembrane currents (second trace). From Simmons and Hardie (1988). (E and F) Effects of picrotoxin (E) and curare (F) upon responses of cockroach L-neurones to intermittent flashes and an air puff to cerci (horizontal bars). The upper traces are responses in normal saline. The lower traces are responses to perfusion of 0.1 mM picrotoxin for 40 min (E), and to perfusion of 0.05 mM curare for 15 min (F). Calibrations: 10 s and 5 mV. From Lin et al. (1990).

MAKOTO MlZUNAMl

210

In the first optic ganglion (lamina) of the compound eye of the fly, the C2 centrifugal neurone is immunoreactive to GABA (Datum et al., 1986). The C2 neurone synapses onto two types of lamina monopolar cells, L1 and L2 (Strausfeld and Campos-Ortega, 1977). Upon iontophoretic application of GABA, lamina monopolar cells of the fly exhibit a depolarization which is associated with an inhibition of the light-induced hyperpolarizing response (Hardie, 1987). Thus, GABA appears to mediate a modulation of a major class of second-order neurones by efferent neurones in both compound eyes and ocelli. Kral (1980) has observed a high acetylcholinesterase (AChE) activity in the receptor axons and in the surrounding glial cells of the honeybee ocelli. However, it is unlikely that acetylcholine (ACh) acts as a major transmitter of ocellar photoreceptor, since iontophoretically applied ACh does not mimic well the action of the transmitter from photoreceptors onto Lneurones (Simmons and Hardie, 1988; Lin et al., 1990). The role of ACh in insect ocellar retina remains to be determined. Other substances reported to be involved in the ocellar system include taurine, glutamate and FMRF-amide. Schafer et al. (1988) have observed taurine-like immunoreactivity from all photoreceptor cells of both the compound eyes and the ocelli of the honeybee. Bicker (1991) has also found taurine-like immunoreactivity in ocellar and compound eye photoreceptors of Drosophila and Locusta. He postulated that taurine may act as a neurohormone, playing a role in maintaining the structure and function of photoreceptors. Bicker et al. (1988) have observed glutamate-like immunoreactivity from large second-order neurones of both compound eyes and ocelli. The possibility that the transmitter of large second-order neurones of insects is glutamate needs to be examined in pharmacological experiments. Schiirmann and Erber (1990) have observed that some unidentified elements in the ocellar neuropils of the honeybee, Apis melliferu, exhibit FMRF-amide-like immunoreactivity. FMRF-amide may act as a neuromodulator.

6.3

STRUCTURAL MUTANTS IN DROSOPHILA

Stark et al. (1989) have investigated the ocellar structures of several Drosophila mutants. The none mutant has unusual compound eye and ocellar corneas. The compound eye is devoid of differentiated photoreceptors but some axons from undifferentiated cells form synapses. No receptors were found in the ocelli of none. The oc mutant has no ocelli, although sometimes an ocellar cornea like that of none is seen: the compound eye is normal. The rdo mutant is also specific to ocelli with smaller ocelli having half the wild-type allotment of receptor cells; the number of giant afferents is unaffected. The ocellar structures of mutants best known for their compound eye defects have also been examined (Stark et a f . , 1989). The

THE DIVERSITY OF INSECT OCELLAR SYSTEMS

21 1

norpA mutant loses its ocellar rhabdomeres with age but has normal feedforward and feedback synapses. This normal synaptology prevails despite the electrophysiological defects in norpA ocelli reported earlier (Labhart, 1977; Hu et al., 1978). The rdgAES” mutant has poorly formed ocellar receptors which show some degeneration with age but synapses survive. Drosophila mutants will provide a useful means by which to examine the molecular mechanisms used to form the neural organization of ocellar systems. 7 information processing in the oceilar system

In this section I will discuss how the photic signals encoded in the ocellar receptors are processed as they are passed along the course of ocellar pathways. I will focus on how the ocellar systems process signals concerning intensity changes around a mean illuminance. There have been some earlier studies on the responses of ocellar neurones using a flash of light in darkness. Because photic inputs which visual systems receive in a natural environment are intensity changes around a mean illuminance, visual systems, including insect ocelli, should be designed to process signals concerning intensity changes. Here, I will summarize our present knowledge of visual processing in insect ocellar systems and, in the final section (Section 7.6), I will propose some principles concerning information processing in an ocellar system. Some of the discussion in this section is based on comparative approaches where signal processing in insect ocellar systems is compared to other visual systems, including insect compound eyes and vertebrate retinas. I will discuss: (1) specific features of insect ocellar systems which should reflect specific functional roles of ocelli; and (2) common features among different visual systems which should reflect general principles of visual systems. 7.1

INFORMATION PROCESSING I N THE OCELLAR PLEXUS

7.1.1 Enhancement of response to intensity changes Ocellar photoreceptors exhibit a depolarizing response to illumination of the ocellus, whereas large second-order neurones, L-neurones, exhibit a hyperpolarizing response (Fig. 24A,B; Chappell and Dowling, 1972; Klingman and Chappell, 1978; Simmons, 1982b). The hyperpolarizing response of the L-neurone is due to an increase in membrane permeability to ions whose equilibrium potential is about 40 mV negative to the membrane potential in the dark (in locusts; Wilson, 1978b). Chappell and Dowling (1972) have compared the responses of photoreceptors and L-neurones of the dragonfly to a flash of light given in the

212

MAKOTO MlZUNAMl

- 2

35

-

30

-

25

-

20

-

15

-

10

-

R e c e p t o r Response

1 B

J Log I.

E >

lronsient

*-/

,*A

o/L 2

I

sustained

5 -

-1

r

-3

2-*

- - a d

0 -

-20

I

-

I

I*-*\

Postsynaptic Response

--10 15

\a

-

v-4 -5

L

-I

-5

-

0 -

nsI

en

sustained

/*/0-0<

:/@ I

-7

-6

O-0-0

, -5

I

I

-4

-3

LOG

I

I

-2

-I

I

0

I

FIG. 24 Responses of a photoreceptor and a large second-order neurone of the median ocellus of the dragonfly Aeschna. Intracellularly recorded responses of a photoreceptor (A) and a large second-order cell (B) to illumination of the median ocellus with pulses of white light of different intensities. Log I, loglo relative intensity of illumination. Scales: 200 p m ; 20 mV. (C) Intensity-response relationships for receptor and postsynaptic units. The amplitude of the peak of the transient wave and the sustained component of the receptor response and the postsynaptic response are plotted as function of intensity. The sustained component of the receptor response was measured 3 s after the start of illumination. (A) and (B) are from Patterson and Chappell (1980); (C) is from Chappell and Dowling (1972).

THE DIVERSITY OF INSECT OCELLAR SYSTEMS

213

dark, and have noted two important differences in the response properties between photoreceptors and L-neurones. First, the sensitivity of L-neurones is larger than that of photoreceptors by 1-2 log units (Fig. 24A,B). This is explained by a high ratio of convergence of photoreceptors onto L-neurones (Dowling and Chappell, 1972). Second, the waveform of the response of L-neurones is much more transient (Fig. 24A,B), and the sustained component of the response of L-neurones is much less prominent (Fig. 24C) than that of photoreceptors. The functional significance of the enhancement of transience was clarified by further examination of responses of photoreceptors and second-order neurones in a light-adapted condition (Chappell and Dowling, 1972). The photic inputs that visual systems receive naturally is a modulation of light intensity around a mean illuminance. The mean illuminance changes slowly but covers a large range in the course of one day. The depth of fluctuation around the mean is moderate and remains constant. The photoreceptor response, the amplitude of which is roughly a logarithmic function of the absolute light intensity for a wide intensity range, can be divided into two components, a steady mean potential and a time varying component, which signal mean illuminance levels and intensity changes, respectively. In the responses of L-neurones, however, the steady component is completely eliminated (in bees: Baader, 1989) or highly compressed (in dragonflies: Chappell and Dowling, 1972; Simmons, 1982b; in cockroaches: Mizunami et al., 1986), which means that L-neurones signal relative intensity changes and not absolute intensity levels. In short, signals concerning absolute intensity are converted into those concerning relative intensity change. The enhancement of the response to intensity changes by removing the steady response component between photoreceptors and second-order neurones has been observed in most visual systems so far studied, including barnacle ocelli (Stuart and Oertel, 1978; Hayashi et al., 1985), Lirnufus lateral eyes (Purple and Dodge, 1965), insect compound eyes (Laughlin and Hardie, 1978) and vertebrate retinas (Werblin, 1972; Naka et a f . , 1979; Normann and Perlman, 1979). This common principle reflects the fact that visual systems are designed to see objects in the light, rather than the light itself. The mechanisms that enhance the response to relative intensity changes in the ocellar plexus have been studied by several authors. Dowling and Chappell (1972) observed feedback synapses from L-neurones to photoreceptors in the dragonfly ocellar plexus, and proposed that they play roles in the enhancement of response transience. Klingman and Chappell (1978) have examined this hypothesis by studying the effects of various drugs on the response of L-neurones in dragonflies. Curare blocked the response completely, while both picrotoxin and bicuculline eliminated the offovershoot of the L-neurone. They suggested that the receptor synapses are inhibitory (sign-inverting) and curare-sensitive, whereas there is an excitatory (sign-conserving) GABAergic feedback synapse from L-neurones which

214

MAKOTO MlZUNAMl

is facilitatory to receptor transmitter release. Stone and Chappell (1981) have suggested that the hyperpolarizing oscillation at the off-set of illumination in the dragonfly photoreceptors reflects GABAergic feedback synapses onto photoreceptor terminals. Subsequent studies, however, have failed to confirm the feedback hypothesis for the enhancement of response transience. First, Simmons (1982a) made simultaneous intracellular recordings from a photoreceptor and an L-neurone from the dragonfly ocellar retina. No response was evoked in the photoreceptors when depolarizing or hyperpolarizing currents were injected into L-neurones, although L-neurones produced responses when current was injected into photoreceptors. Second, no immunoreactivity to GABA was observed from L-neurones at least in locusts (Ammermiiller and Weiler, 1985) and bees (Schafer and Bicker, 1986). Unfortunately, GABA-immunocytocheniical studies have not been made on dragonfly L-neurones. Third, the response of bee L-neurones is very phasic (Guy ef al., 1979), although no feedback synapses have been observed between L-neurones and receptors in the bee (Toh and Kuwabara, 1975). There should be a mechanism to enhance the transience other than the feedback from L-neurones, at least in the bee. Some authors, thus, have proposed alternative mechanisms for the enhancement of response transience. Ammermiiller and Weiler (1985) proposed a possible contribution of GABAergic small second-order neurones in a feedback system of the locust. Simmons (1982a) proposed that intrinsic membrane properties of the photoreceptor terminals and L-neurones, and excitatory synapses made by small second-order neurones on L-neurones are included in the mechanisms to enhance the response transience of dragonfly L-neurones. In summary, feedback synapses from L-neurones to photoreceptors, feedback from small second-order neurones, and active membrane properties of photoreceptors and L-neurones have been proposed, and partially demonstrated experimentally, to contribute to enhance the response transience in the ocellar retina, but none of these are conclusive. The enhancement of response transience between photoreceptors and second-order neurones appears to be a general principle of visual systems as has been discussed, and its neural mechanism has been examined in a number of visual systems. Mechanisms suggested for the enhancement of response transience include: feedback synapses in vertebrate retinas (Baylor ef al., 1971); voltage-dependent calcium and potassium currents at photoreceptor t e r m i d s in barnacle ocellus (Stockbridge and Ross, 1984) and fly compound eye (Weckstrom et al., 1992); inhibition mediated by extracellular field potentials in fly compound eye (Shaw, 1979); and feedback inhibition from self-same eccentric cells in Limulus (Purple and Dodge, 1965). Considering the diversities in the neural mechanisms for the enhancement of transience, it would not be surprising if different mechanisms have been adopted to enhance response transience in the ocelli of

THE DIVERSIP/ OF INSECT OCELLAR SYSTEMS

215

different insects. There is even a possibility that multiple mechanisms coexist to enhance the response transience in a given ocellar retina.

7.1.2 Dynamics and sensitivity of light-adapted L-neurones Mizunami et al. (1986) have measured the dynamics and sensitivity of light-adapted responses of cockroach L-neurones using white noisemodulated light with various mean illuminances. The kernels, obtained by cross-correlating the white noise input against the resulting response, provided a measure of incremental sensitivity as well as of response dynamics (Fig. 25A). A brief step of light given in the dark produced step-like responses in L-neurones. With continued white noise stimulation, the membrane potential reaches a steady level within 30 s (Fig. 25A). At this dynamic steady state, the actual responses of L-neurones to white noise stimulus can be predicted by a linear model (Fig. 25B) with mean square errors of about 11%, indicating that the response is practically linear, i.e. the magnitude of the response is proportional to the depth of modulation. The linear nature of the response to intensity changes has been reported from peripheral visual neurones of a variety of visual systems including photoreceptors of Limulus compound eye (Fuortes and Hodgkin, 1964), vertebrate retina (Baylor and Hodgkin, 1974; Naka et al., 1987) and insect compound eye (Pinter, 1972), as well as the second-order neurones of vertebrate retina (Naka et al., 1979; Tranchina et al., 1983; Chappell et al., 1985), Limulus compound eye (Knight et al., 1970) and insect compound eye (French and Jarvilehto, 1978). Linear coding of photic signals thus appears to be a general principle of peripheral visual systems. In Fig. 26A, first-order kernels which represent the linear component of the response, obtained at four log ranges of mean luminance levels, have been plotted on a contrast sensitivity scale. The waveforms are almost identical, with constant peak response times of about 50ms, while the amplitudes differ only by 30%. The results show that: (1) the response is an exact Weber-Fechner function, that is, contrast sensitivity is independent of the mean luminance; and (2) the response dynamics are independent of the mean luminance. For comparison, an example of kernels from a horizontal cell of turtle retina obtained under comparable conditions is shown in Fig. 26B, where: (1) the amplitude of kernels differs for different levels of mean illuminance; and (2) the peak response times shorten from 100 to 50 ms and the waveforms become more biphasic (differentiating) with an increase in mean luminance. These studies show that: (1) the cockroach L-neurones are ideal contrast detectors since their response amplitude is exactly proportional to the contrast of intensity change for at least four log ranges of mean illuminance levels; and (2) signal processing in the cockroach ocellus differs from that in other visual systems, including Limulus lateral eyes (Fuortes and Hodgkin, 1964), insect compound eyes (Pinter, 1972; Dubs, 1981), and

216

MAKOTO MIZUNAM

vertebrate retinas (Baylor and Hodgkin, 1974; Naka et al., 1979), and the human visual system (Kelly, 1971), in which the system’s dynamics changes depending on the levels of mean illuminance. Sinusoidally modulated light has also been used to study frequency-response characteristics of Lneurones (Mizunami and Tateda, 1988b; Mizunami, 1990a). The response of L-neurones has a bandpass property with an optimal frequency of 1-3 Hz, the lower cut-off frequency (-3 dB) of 0.05-0.1 Hz and the higher cut-off frequency of 12-15 Hz, indicating that this neurone can respond to slowly occurring events. Interestingly, there are marked differences in incremental responses of light-adapted L-neurones among different insects. Contrast sensitivity of bee L-neurones (Baader, 1989) is much lower than that of the cockroach (Mizunami et al., 1986). However, some L-neurones of the bee exhibit a higher cut-off frequency of about 30 Hz (Baader, 1989), and thus can respond to stimuli of much higher frequency than can cockroach Lneurones. These findings are consistent with the hypothesis that the ocellar system of the bee is more concerned with speed than sensitivity, while that of the cockroach is more concerned with sensitivity (see Section 5.6). In addition, the waveform of the response of bee L-neurones to sinusoidally modulated light highly deviates from a sinusoid (Baader, 1989), indicating that the response contains a high degree of non-linearity. The bee L-neurones are not suited to faithfully monitor the stimulus contrast, and are perhaps more related to the initiation of direct behavioural reactions. Furthermore, Simmons (1993) has recently measured the frequency response characteristics of locust L-neurones using sinusoidally modulated light, and has noted that the optimal frequency changes from 2 to 10 Hz with an increase in mean illuminance levels. Since such changes have not been observed in the responses of cockroach L-neurones (Mizunami et al., 1986), ocellar systems of locusts and cockroaches must adopt different strategies to adjust their response dynamics to environmental light intensity levels. Such differences imply that there are differences in behavioural roles of ocelli between locusts and cockroaches. FIG. 25 (A) Responses from a cockroach ocellar L-neurone evoked either by steps of light given in the dark or by white noise-modulated light. The relationship between I, and Vp or V,, is the cell’s D C (static) sensitivity and the relationship between I ( t ) and V ( t ) is the incremental sensitivity. Spike potentials are seen at the offset of step stimulation as well as during white noise stimulation. (B) Time records of part of a white noise stimulus and the resulting response of a cockroach L-neurone (continuous line). Superimposed o n the response trace is the linear model (broken line). Probability distribution function (PDF) for the light stimulus and the recorded response are also shown. The light P D F is also superimposed on the response PDF. (C) Power spectra of the light stimulus, response (continuous line), and linear model (broken line). The mean illuminance of the stimulus is 20 p W cm-*. From Mizunami et af. (1986).

THE DIVERSITY OF INSECT OCELLAR SYSTEMS

I

L

0

B

217

8

4 TIME

30 sec

RECORD

PDF

C 0

Y

U W

3 0

a

I

I !

1I I

I

0.1

I

1 10 FREQU E NCY ( H z )

I

10 0

218

MAKOTO MlZUNAMl

I

I

0

a

0.1

I

0.2

I

0.3

m

t Z 3 A

w Z

tx

w

Y

vo

-24 I

0

I

0.1 TIME

I

1

0.2

0.3

(S)

FIG. 26 (A) First-order kernels from a cockroach ocellar L-neurone, plotted on a contrast sensitivity scale, obtained at five mean illuminance levels. The first-order kernels were calculated by cross-correlating the white noise light stimuli with the recorded responses. Kernel;s are labelled 0 through -4 to indicate the log density of the filters interposed. Note that the amplitudes of the kernels did not differ by more than 30% and the peak response times were constant at 5 0 m s for all kernels, although the mean levels covered a range of 1:lOOOO. Stimuli dimmer than -4 log units did not produce any reliable results. (B) shows first-order kernels from a turtle retinal horizontal cell, plotted as in (A). The peak response times, waveforms, and amplitudes differed for different levels of mean illuminance. Kernel units are in mV .p W - ’ . cm2. s. The larger incremental sensitivity for ocellar kernels was due to the dimmer mean illuminance (20 p W cm-’ at 0 log) of the white noise stimulus than in the turtle experiment (50 pW cm-’ at 0 log). From Mizunami et al. (1986).

7.2

INFORMATION PROCESSING I N SECOND-ORDER NEURONES

7.2.1 Passive and active membrane properties of locust L-neurones Measurements of membrane properties of L-neurones show that the graded synaptic potentials generated in the ocellar plexus propagate passively along its axons toward their terminal areas. Wilson (1978b) has studied the electrical properties of L-neurones of locusts by penetrating a single L-neurone with two intracellular microelectrodes. Hyperpolarizing signals produced in the ocellar plexus exhibit decrement along L-neurone axons,

THE DIVERSITY OF INSECT OCELLAR SYSTEMS

219

indicating passive propagation. Ammermuller (1986) has further measured cable properties of locust ocellar L-neurones, and concluded that the passive cable properties guarantee, to a reasonable degree, the transmission of hyperpolarizing potentials from the ocelli to the brain, and should also allow graded interactions between different ocelli. In locusts (Wilson, 1978b; Simmons, 1982b) and cockroaches (Mizunami et al., 1982, 1987), L-neurones generate a few spikes at the off-set of ocellar illumination or at the termination of hyperpolarizing current injected into the neurone. The amplitude of the spike depends both upon the amplitude and the duration of the preceding hyperpolarization. Some L-neurones of bees generate tonic spike discharges whose frequency decreases when the ocellus is illuminated (Milde, 1984). In locusts (Wilson, 1978b) and cockroaches (Mizunami et ul., 1987), the spike initiates in the ocellar tract and passively propagates along its axon toward the ocellus, i.e. the spike is non-conducting . Ammermiiller and Zettler (1986) have carried out voltage-clamp analyses of active membrane properties of locust L-neurones. Under voltage-clamp two fast currents were observed. A fast inward current reflecting the action potentials in the current clamp experiments had a threshold of -65 mV. The steady-state inactivation curve shows that the channels of this current were completely inactivated at the L-neurone’s normal resting potential. A fast outward current, which partially shunted the fast inward current, could be blocked with 4-aminopyridine. The voltage characteristics were similar to those of the transient potassium current in Drosophilu flight muscle (Salkoff and Wyman, 1983). Additional slowly activating time- and voltagedependent currents were observed under voltage-clamp. The results demonstrate that locust L-neurones are capable of generating action potentials of small amplitude. The voltage dependence shows that the ion channels of these currents are mostly in an inactivated state at the normal resting potential of the L-neurones. Inactivation can be removed by shifting the membrane to a more negative level; thus the low resting potential of L-neurones explains the ‘non-spiking’ state of an L-neurone. Mizunami et al. (1987) have examined the ionic basis of action potentials of L-neurones of cockroaches. The action potential was blocked by replacing saline Ca2+ with Mg2+ but maintained when Ba2+ was substituted. A block was produced by 2 mM Cd2+ or 20 mM Co2+. The peak amplitude of the action potential increased by 26 mV for a 10-fold increase in external Ca2+ concentration, at concentrations below 1.8 mM. The action potential M tetrodotoxin. These was not affected by sodium-free saline or by 3 X findings suggest that calcium ions are the major carrier for the inward current of the action potential. The spike of the L-neurone is generated in the ocellar tract where a large number of presynaptic terminals are distributed. Possibly the spikes are generated by calcium channels of presynaptic terminals.

220

MAKOTO MlZUNAMl

On account of the study by Mizunami et al. (1987), much remains to be clarified about the ionic basis of action potentials of L-neurones. In the bee, spikes of LD neurones are actively conducted for a long distance along its axon from the brain to the thoracic ganglia. It is most probable that such conducting action potentials are produced by sodium channels, not by calcium channels. If LD neurones of the bee generate sodium spikes, it is probable that other types of L-neurones of the bee also have voltagedependent sodium channels. Furthermore, in their voltage-clamp study of locust L-neurones, Ammermuller and Zettler (1986) considered that their action potential may be sodium dependent, although they did not provide direct evidence to show this. Further pharmacological and voltage-clamp analyses are necessary to clarify if the non-propagating spikes of locust L-neurones are carried by calcium or sodium ions. Recently, Simmons and Rind (1993) have noted unusual physiological features of L-neurones of the blowfly, Calliphora erythrocephala. First, the hyperpolarizing response to light increases in amplitude when the neurone is hyperpolarized by injecting extrinsic current, in contrast to L-neurones of other insects where membrane hyperpolarization causes a decrease in amplitude of hyperpolarizing light response. Second, the L-neurones generate rhythmic oscillations in membrane potential to hyperpolarizing current. Membrane mechanisms underlying such unusual properties need to be clarified. It appears that the membrane properties of insect L-neurones are much more diverse than has been previously considered. 7.2.2 Spiking and non-spiking L-neurones of bees Milde (1981, 1984, 1986, 1987) found that the response of L-neurones of the bee to light flashes given in the dark ranged from graded responses with fast on and off transients to inhibition of spontaneous discharge (Fig. 27A). Since the same morphological type of L-neurones responded differently in different animals, response types could not be correlated with morphology.

FIG. 27 (A) Stimulation of honeybee ocellar L-neurones with wide-field white light flashes showing responses ranging from pure graded, mixed to pure spiking patterns. Numbers indicate arbitrary division of responses into four response classes, 1, graded responses only; 2 , graded responses with one or a few additional spikes; 3, graded responses and spontaneous discharge of action potentials; 4,spiking responses only. Vertical bar = 20 mV. (B) Response spectrum of a descending LD-neurone (Locth 1) of a honeybee to multimodal stimulation (n = 8). Stimuli were: light, removal of a Styrofoam substrate from the bee’s tarsi, airpuffs onto the head and abdomen, a drop of sugarwater applied to the antennae and the proboscis and an airstream containing odour. Closed circles indicate response, open circles no response, blank not tested. Bar represents 1 0 0 p m . LOC, lateral ocellus; MOC, median ocellus; CX, calyx of mushroom body; T, trachea; OS, oesophagus. From Milde (1986).

THE DIVERSITY OF INSECT OCELLAR SYSTEMS

22 1

A

,

2 -

13 300 rns

300 rns

B

3s

222

MAKOTO MlZUNAMl

However, purely graded responses were not found in LD-neurones which run directly from the ocellar retina to the thoracic ganglia, and Lps-neurones of lateral ocellus which project into the posterior slope never exhibited an inhibition of the tonic spike discharge without a graded component. Presumably non-spiking L-neurones are refractory due to maintained depolarization (Milde, 1981). Milde (1984) proposed that a given L-neurone may use either spiking or non-spiking signal transmission depending on different behavioural situations. Milde and Homberg (1984) further showed that L-neurones with mixed responses of graded and action potentials were unimodal, i.e. they reacted exclusively to stationary illumination of the ocelli, as do non-spiking L-neurones. In contrast, spiking L-neurones that lacked a graded response component could also respond to stimuli of other sensory modalities. The bee possesses five pairs of LD neurones, which respond to ocellar illumination with very phasic spike discharges at light ‘on’ and ‘off‘ (Pan and Goodman, 1977; Milde, 1984). One of these neurones, Locthl, responds to moving patterns, compound eye illumination, airstream, mechanical and gustatory stimulation (Fig. 27B). Goodman er al. (1987, 1990) found that the other LD neurones, Locth2 and Locth3 of Apis rneflifera and their homologues in the wasp, Paravespula vulgaris, are directionally sensitive to wide field motion over the frontal and lateral part of the compound eye. The vertical upward movement produced excitation in these units, whereas downward movement produced inhibition. These LD neurones of the bee are implicated in the stabilization of flight course (Goodman ez af., 1987, 1990), as are descending third-order ocellar neurones of the locust (see Section 7.5.2; Simmons, 1980; Reichert et al., 1985). 7.2.3 Dynamics of spiking responses of cockroach L-neurones Are the signals coded in the spike responses of L-neurones the same as those coded in the graded responses? This question was examined by Mizunami and Tateda (1988a) who studied the relationship between the slow potential and spikes of L-neurones of the cockroach, using a sinusoidally modulated light with various mean illuminances. A solitary spike was generated at the depolarizing phase of the modulation response (Fig. 28A). The relationship between the peak-to-peak amplitude of the slow potential response and the rate of spike generation was sigmoidal (Fig. 28B). In Fig. 28C, the spike rate has been plotted against the amplitude of the slow potential response at five different frequencies. The extrapolated regression lines for each frequency cross the vertical axis at almost the same point, suggesting that the non-linear threshold is frequency independent. On the other hand, the slope of the lines changes with frequency, which indicates that the spike initiation process contains a dynamic linearity. Subsequent studies have shown that the spike threshold at optimal fre-

THE DIVERSITY OF INSECT OCELLAR SYSTEMS

'

vn

223

v v v

L -10

mV

0

SLOW

3

30

S

P O T E N T I A L (mV)

FIG. 28 (A) Responses of a cockroach ocellar L-neurone evoked either by a step light stimulus given in the dark o r by a sinusoidally modulated light stimulus. The L-neurone responded to the sinusoidal stimulation with a sinusoidal voltage modulation, V M , around a mean voltage, V,. A spontaneous voltage fluctuation (voltage noise), V,, was superimposed on the modulation response. Spikes were seen at the offset of step stimulation and at the peak of the voltage modulation. The mean illuminance of the stimulus, f,, was 20 pW cm-2; the modulation frequency, f, was 2Hz; the depth of modulation of the stimulus, was 60%. (B) The rate of spike generation plotted against the peak-to-peak amplitude of the slow potential response of a cockroach L-neurone, obtained at a frequency of 1 Hz. The form of the curve was sigmoidal, with the linear part covering the range of spike rate of -10-90%. (C) The spike rate plotted against the amplitude of slow response of a cockroach L-neurone, obtained at five different frequencies. The extrapolated dashed lines are the regression lines for each frequency. These lines cross the vertical axis at almost the same point. The stimulus had a mean illuminance of 2 p W cm-*. From Mizunami and Tateda (1988a).

fo,

224

MAKOTO MlZUNAMl

3J ,

0.2

I

0.5

1

FREQUENCY

B

5

2

10

(Hz)

I h

> E 5v

n 1

-

0 I

Ln

w L-

cc

I

* -2

OK 0.2

0.5 1 2 F R E Q U E N C Y (Hz)

5

,

10

FIG. 29 (A) 50% threshold of spike response, defined as the peak-to-peak amplitude of the potential modulation at a spike rate of 50%, plotted against the modulation frequency. The plots are from the responses of a cockroach L-neurone to sinusoidal lights with a mean illuminance of 0.005 (-3.6 log), 0.2(-2 log), and 20 pW cm-2 (0 log). (B) Effects of noise current and steady current injection on the 50% threshold of a cockroach L-neurone. A noise current having a peak-to-peak amplitude of -4nA or a steady depolarizing current of 4 n A was injected during sinusoidal light stimulation. The light stimulus had a mean illuminance of 20pW cm-2. The estimated potential change produced by the injection of current 4 nA was -3 mV (the estimated input resistance at that mean illuminance was -0.8 M a ) . The 50% threshold at a low-frequency range decreased when the noise current was injected. The steady depolarizing current had little effect on the 50% threshold. From Mizunami and Tateda (1988a).

THE DIVERSITY OF INSECT OCELLAR SYSTEMS

225

quency (0.5-5 Hz) remains unchanged over a mean illuminance range of 3.6 log units, whereas the spike threshold at frequencies of t0.5Hz was lower at a dimmer mean illuminance (Fig. 29A). At a dimmer mean illuminance, the L-neurones exhibited a larger voltage noise, and their mean membrane potential levels were most positive (Fig. 3D). Steady or noise current injection during sinusoidal light stimulation showed that: (1) the decrease in the spike threshold at a dimmer mean illuminance was due to the increase in the noise variance: the noise had fascilitatory effects on the spike initiation; and (2) the change in the mean potential level had little effect on the spike threshold (Fig. 29B). In summary, the spike response of L-neurones is represented by a simple cascade model (Fig. 30A). Light stimulus is passed through a bandpass linear filter and produces a slow potential response in L-neurones. The slow potential is passed through a cascade of a linear filter followed by a non-linear filter and produces a spike discharge in L-neurones. The linear filter is bandpass containing both a differentiating and an integrative nature. The non-linear filter is a static threshold with a sigmoidal probabilistic input/output relationship. It is concluded that fundamental signal modifications occur during the spike initiation in the cockroach L-neurones, a finding that differs from the spike initiation process in other visual systems, including Limulus compound eye (Knight, 1972a,b) and catfish retina (Sakuranaga er al., 1987), in which it is presumed that little signal modification occurs at the analog-to-digital conversion process. It is interesting to see if the slow potential or the spike signals of L-neurones, or both, are encoded in the response of third-order neurones. Figures 30B and C show typical examples of responses of third-order ocellar neurones to sinusoidally modulated light. A type of third-order neurone, OL-1 neurone, had spontaneous spike activity and exhibited a modulation of the spike frequency around a mean (Fig. 30B). The pattern of the response was similar to the slow potential response of L-neurones. The other type, the D-1 neurone, had no spontaneous spike activity and exhibited single spikes at the decremental phase of light modulation (Fig. 30C). The pattern of the response was similar to that of the spike response of L-neurones. Thus, it is concluded that both graded and spike signals of L-neurones are encoded in the spike responses of third-order neurones (Mizunami and Tateda, 1988a; Mizunami, 1989). The graded response of L-neurones appears to continuously monitor the contrast of intensity changes, whereas the spike response possibly signals an urgent event which requires rapid behavioural reactions. 7.2.4 Synaptic interactions among locust L-neurones Taylor (1981~)found that two anatomical types of L-neurones (Ml, ML) of the locust responded with hyperpolarization to the illumination of the median ocellus, and with depolarization to the illumination of the lateral

MAKOTO MlZUNAMl

226

A LIGHT FILTER

FILTER

FILTER

B

C

FIG. 30 (A) A model for the spike response of cockroach ocellar L-neurones. Light signals are passed through a bandpass linear filter and produce a slow potential response in L-neurones. The slow potential contains a noise, which reflects that contained in the synaptic potential from photoreceptors. The slow potential is further passed through a lineahon-linear cascade and produces a spike discharge. The linear filter is a bandpass, and the non-linear filter is a static threshold. (B, C) Typical responses of two types of third-order ocellar neurones of the cockroach to sinusoidal light stimulation. One type of third-order neurone, the OL-I neurone (type 1, neurone projecting into the optic lobe), showed sinusoidal modulation of spike frequency (B), whereas the other type, the D-I neurone (type 1 neurone descending to the thoracic ganglia), generated solitary spikes at the decremental phase of light modulation (C). The lower traces indicate the stimulus light, monitored by a photodiode. T h e stimulus had a modulation frequency of 1 H z and a modulation depth of 50%. The mean illuminance was 2 p W cm-2. From Mizunami and Tateda (1988a).

ocellus. The depolarizing responses were abolished with picrotoxin, indicating that the depolarizing response is mediated by GABAergic neurones, possibly small second-order neurones. Simmons (1982b, 1986) demonstrated that L-neurones of locust ocelli make both excitatory and inhibitory connections with each other (Fig. 31). In the lateral ocellus, all three members of one anatomical class of L-neurones (L1-3) make reciprocal inhibitory connections with each other

THE DIVERSIP/ OF lNSECT OCELLAR SYSTEMS

227

1

inhibitory

A

e x t itatory

A

n

.

.

.

FIG. 31 Summary of the interactions among lateral (Ll-5) and among median (ML and M1) L-neurones of the locust, Schistocerca. In order to simplify the diagram, boxes have been drawn around neurones which make similar and parallel interactions; for example, L1, L2 and L3 each excite L4 and L5. From Simmons (1986).

(Fig. 31). Some of these neurones are presynaptic at excitatory connections with another class, L4-5. The median ocellus is connected with each lateral ocellus by a pair of neurones, called ML neurones. A bilateral pair of M1 neurones also connect the median ocellus with the brain. ML neurones make a reciprocal inhibitory interaction with the other three ML neurones. The ML neurones also make excitatory interactions with the left and right M1 neurones. Simmons (1982b, 1985, 1986) examined the properties of excitatory and inhibitory connections among L-neurones. Small, graded depolarizations and hyperpolarizations are transmitted at the excitatory connections but, at the inhibitory connections, a spike in the presynaptic neurone is required for transmission. At the excitatory connections, the resting potential of the presynaptic neurone normally lies depolarized from the threshold for transmission, so that both small hyperpolarizations and depolarizations effect changes in the postsynaptic potential. Spikes in the presynaptic neurone usually ensure that the postsynaptic neurone also generates spikes. At the inhibitory connections, the postsynaptic potential decrements within 10-20 ms. Because of this, rapidly rising presynaptic potentials, such as spikes, are required for transmission. The most likely reason why the duration of the inhibitory postsynaptic potentials (IPSPs) is limited is that calcium channels in the presynaptic terminal inactivate within 7 ms of first opening (Simmons, 1985). Simmons (1982b) suggested that the excitatory connections may sharpen responsiveness to decreases in illumination, and the inhibitory connections may enhance the detection of rapid movements of large objects, such as the visual horizon.

MAKOTO MlZUNAMl

228

Simmons (1990) measured changes in properties of locust L-neurones which depended on temperature. Such changes include: (1) a decrease in input resistance, which typically halves in value as temperature increases from 15°C to 35°C; (2) an increase in the size of rebound spikes; (3) a decrease in the latency to response to light; and (4)a decrease in the latency of transmission at both excitatory and inhibitory synapses between Lneurones. The temperature of the L-neurones in a flying locust will normally be between 25°C and 35°C or perhaps warmer, and it is concluded that there would be little functional changes in the operation of L-neurones for changes in temperature within this range, apart from a slight decrease in synaptic latencies and an increased phasic character in responses to light. 7.2.5 Responses of small second-order neurones Although it is generally agreed that signals of ocellar photoreceptors are transmitted to the brain by a large number of small neurones (S-neurones) and a small number of L-neurones, not much is known about response properties of S-neurones. In the fly, a small second-order ocellar neurone has been anatomically identified (Hengstenberg and Hengstenberg, 1981). The neurone runs from the lateral ocellus to the contralateral mesothoracic neuromere. The neurone was silent in the dark, tonically active in the bright, and responded vigorously to flickering light. Intracellular recordings from unidentified light inhibitory and light excitatory spiking neurones have been made in the ocellar plexus of the dragonfly (Chappell and Dowling, 1972), ocellar nerve of the locust (Wilson, 1978a; Ammermiiller and Weiler, 1985) and ocellar tract of the bee (Guy et al., 1979). These recordings are assumed to be from small second-order neurones. In Trichoplusia ni, only light excitatory second-order neurones have been recorded (Eaton, 1976). In the median ocellar nerve of the locust, Simmons (1986) observed some interactions between L-neurones and small neurones which produce trains of spikes. L-neurones make excitatory interactions with S-neurones which are inhibited by light, and inhibitory interactions with S-neurones which are excited by light. Further studies are necessary to clarify the order and the direction of signal transmission of these S-neurones. 7.3

MODULATORY ROLES OF EFFERENT NEURONES

There is evidence to show that the activities of ocellar neurones are modulated by inputs from other sensory organs, especially by ascending activities from the thoracic ganglia. In the dragonfly, Kondo (1978) recorded afferent and efferent impulses from the distal and proximal cut-end of the lateral ocellar nerve, respectively. The large efferent fibre responded to illumination of the compound eyes and also to movement of the wings. The large afferent fibre was also activated during wing beat, and the evoked

THE DIVERSITY OF INSECT OCELLAR SYSTEMS

229

discharge was inhibited by ocellar illumination. It is interesting to examine the functional roles of the modulatory effects of efferent systems in a tethered flight condition, since it has been shown that dragonfly ocelli contribute to flight steering (Stange and Howard, 1979; Stange, 1981). Possible efferent neurones have been also observed in the lateral ocellar nerves of the locust (Rotzler, 1989). A variety of sensory inputs excite the ocellar nerve units, including illumination of the compound eyes, active and passive movement of the wings, wind stimuli to the thorax, and sound. Most ocellar interneurones are influenced transsynaptically by electrical stimulation of the cervical connectives: L-neurones are depolarized and their response to a rectangular light pulse is changed in amplitude. The descending third-order ocellar neurones (DNs) are also influenced by stimulation of the contralateral connective, perhaps via efferents to ocellar receptors or to L-neurones. In the cockroach, Ohyama and Toh (1986, 1990a) identified possible efferent neurones, referred to as small multimodal (SM) neurones. The SM-neurone is activated by the illumination of compound eyes, movement of antennae, air puffs to cerci, vibrations to legs and wing beats (Fig. 32A-G). Cercal stimulation triggered most effectively a train of spikes in the SM-neurone: some of the seven giant axons in the ventral nerve cord were involved in this pathway. Cercal stimulation also evoked depolarization in the L-neurones, and it is suggested that SM-neurones intervene to produce a cercal response of L-neurones (Fig. 32H). The evoked depolarization in the L-neurone modifies the response of the L-neurones to the ocellar illumination. The cell body of the SM-neurone is located at the ventral crotch and its arbors cover the posterior protocerebrum and deutocerebrum, possibly the input areas, and the ocellar tract and the ocellar nerve, possibly the output areas. Neurones with a very similar morphology have been reported in locusts (Goodman and Williams, 1976) and crickets (Koontz and Edwards, 1984). Efferent control of the spectral sensitivity of ocellar photoreceptors has been demonstrated by Yamazaki and Yamashita (1991) in the cucumber looper moth, Anadewidia peponis, by recording ERGS. The peak sensitivities were observed at 340nm in the ultraviolet and at 520-540nm in the green. Selective spectral adaptation revealed the existence of at least two receptor types in the ocellar retina. The ratio of green to ultraviolet sensitivities for an ocellus whose ocellar nerve was cut was higher than that for an intact ocellus (Fig. 33). It is suggested that efferent signals control the spectral sensitivity of ocellar photoreceptors in the moth ocelli, rather than affecting the activity of L-neurones. The control of the sensitivity of photoreceptors by efferent inputs has been demonstrated in orb-weaving spiders (Yamashita and Tateda, 1981), Limulus (Barlow et al., 1977, 1987) and scorpions (Fleissner and Schliwa, 1977; Fleissner and Fleissner, 1978). In these species, the efferent inputs control the circadian sensitivity changes.

A

B

C

[

G

I

H

L*N SM .N wind to c e r c i

5mV (upper trace) 100pv (lower trace) 500ms

FIG. 32 (A-G) Responses of multimodel ocellar units to various stimuli recorded in mid-regions of the ocellar nerves of the cockroach by suction electrodes (A-E), and by a glass microelectrode (F, G ) . Stimuli are as follows: (A) and (F) illumination of compound eyes (1 lux); (B) an air puff to an antenna (3 m s-'); (C) a mechanical stimulus to legs by vibration of floor (amplitude 100 p m , 200 Hz); (D) and (G) an air puff to cerci ( 3 m s - I ) ; (E) rise and fall of a wing by external manipulation, frequency and amplitude not being exactly controlled. Stimulus duration is indicated by a bar under each recording. Recordings B and C are from the same nerve. Recordings F and G are intracellular recordings from the same SM-neurone. Time scale, 500 ms; calibration of amplitude, 250 pV for B-D, 100 pV for A and E, 30 mV for G and F. From Ohyama and Toh (1986). (H) Simultaneous recordings of L-neurone potentials (L-N; upper trace) and SM-neurone spikes (SM,N; lower trace) from a cockroach. From Ohyama and Toh (1990a).

23 1

THE DIVERSITY OF INSECT OCELLAR SYSTEMS

2.0>

.-

I

> ._ c .ln C

aJ

ln

$1.0-

.-

c

0 -

B

0 LL 300

400 500 Wavelength, nm

600

FIG. 33 Average spectral sensitivity curve of the dark-adapted ocelli of five ON-denervated (open circles) and 10 intact (closed circles) noctuid moths. The data for the intact moths are the same as shown in Fig. 1 . P
In the moth ocelli, however, circadian changes in the response amplitude have not been observed (S. Yamashita, personal communication).

7.4

SIGNAL PROCESSING BETWEEN SECOND- AND THIRD-ORDER NEURONES

Simmons (1981) described the operation of synapses which L-neurones make with a pair of large descending third-order ocellar neurones, DNI, in the locust. Both L-neurones and DNI neurones hyperpolarize when their ocellus is illuminated. The hyperpolarizing responses of DNI neurones to a step of light given in the dark saturate at a light intensity which is well within the range of intensities to which the L-neurones respond. L-neurones and DNI neurones produce sharply rising regenerative responses when a bright light is switched off, and DNI neurones spike at less intense changes in illumination than L-neurones do. L-neurones make excitatory (signconserving) chemical synapses with the DNI neurones. In steady daylight illumination, L-neurones continually release transmitters onto the DNI neurones. The hyperpolarizing responses of DNI neurones to increase in illumination are due to a decrease in the rate of release of transmitter from the L-neurones. In the cockroach, Mizunami and Tateda (1986) identified nine types of ocellar interneurones with arborizations in the ocellar tract where Lneurones terminate. When recordings were made in the ocellar tract, all types of neurones exhibited similar responses to a step stimulus given in the dark, a tonic hyperpolarization during illumination and one or a few

232

MAKOTO MlZUNAMl

transient depolarizations at the end of illumination. These neurones are classified into several physiological types from the responses recorded in their axons or terminal regions: some neurones exhibited spontaneous spike discharge, some neurones had no spontaneous discharge and others generated no spikes to either ocellar illuminations or extrinsic current injections into the neurones (Mizunami and Tateda, 1986, M . Mizunami, personal observation). Mizunami and Tateda (1988b) found that these neurones receive monosynaptic inputs from L-neurones. The synapses made from L-neurones to these neurones had similar properties to those reported for the synapses made from L-neurones to DNI neurones of locusts: excitatory (sign-conserving) synaptic transmission is tonically maintained under normal resting potentials, so that hyperpolarizing responses of L-neurones produce hyperpolarizations in third-order neurones. Mizunami (1990a) further studied the transfer characteristics of the synapses made from L-neurones to third-order ocellar neurones of the cockroach using simultaneous microlectrode penetrations and the application of tetrodotoxin. The stimulus used was a sinusoidally modulated light around a mean illuminance or an extrinsic current applied to the Lneurones. Although the waveform of the response of L-neurones to sinusoidally modulated light is almost sinusoidal, which indicates that the response is linear, the waveform of the response of third-order neurones deviates from the sinusoid and exhibits a half-wave rectification: the depolarizing response to light decrement is much larger than the hyperpolarizing response to light increment (Fig. 34A). Analysis of the synaptic FIG. 34 Synaptic transmission between second- and third-order ocellar neurones of the cockroach. (A) Responses of a second- and a third-order ocellar neurone evoked either by step-stimuli given in the dark o r by a sinusoidally modulated stimulus around a mean illuminance. The sinusoidal stimulus has a modulation depth of 0.7 and a modulation frequency of 2 Hz. Horizontal lines in the records are the steady (DC) potential levels maintained during steady illumination. The lowest trace indicates the stimulus light, monitored by a photodiode, Calibration: 5 mV for the third-order neurone; 9 mV for the second-order ocellar neurone. (B) Typical records of currentholtage relationship of a second-order ocellar neurone (L-neurone), measured by impaling the neurone with two electrodes. (C) Responses of a third-order ocellar neurone evoked by current stimuli applied to an L-neurone. Averaged current-voltage relationships from six L-neurones were used to estimate presynaptic potential changes during current stimuli for the experiments in B . Actual input resistance of L-neurones deviates slightly from cell to cell (+13%), thus, the estimated synaptic transfer curve may have slight errors. Lower traces in B and C indicate the magnitude of stimulus current. (D) Inputloutput voltage relationship of the synaptic transmission. Measurements were made at the steady-state value for 0.5 s current pulses. V , is the synaptic potential maintained in the dark. V,,, and V,,,, are the potentials of the second- and third-order neurone, respectively. The potentials were measured from the dark level, thus, VpOs, + V , is the actual postsynaptic potential. (E) Semilogarithmic plot of the inputloutput voltage relationship of the synaptic transmission. From Mizunami (1990a).

THE DIVERSITY OF INSECT OCELLAR SYSTEMS

233

3 rd

A

&L

"dep

;;_;I

I

I

0

2

4

"hY P

C

B

h

J10mV 9 nA

1 1 0 nA 1s

D

t6

1s

E

h

> E

v

+

5

2

-4

4

8

234

MAKOTO MlZUNAMl

transfer curve relating pre- and postsynaptic voltages (Fig. 34B ,C) showed that the synapses made from L-neurones to third-order neurones operate at an exponentially rising part of the overall sigmoidal transfer curve (Fig. 34D,E). Due to the non-linear characteristics of the synaptic transfer, the linear responses of presynaptic neurones are converted into half-wave rectified responses of postsynaptic neurones (Fig. 35A). The properties of synapses made from second- to third-order neurones (second synapses) of cockroach ocelli were compared to those reported for synapses made from photoreceptors to second-order neurones (first synapses) of other visual systems. In most visual systems so far studied, both photoreceptors and second-order neurones exhibit linear responses to changes in intensity (see Section 7.1.2), thereby suggesting a linear nature of signal transmission at first synapses. Indeed, studies of first synapses in turtle retina (Normann and Perlman, 1979), barnacle ocelli (Hayashi et al., 1985), dragonfly ocelli (Simmons, 1982a) and fly compound eyes (Laughlin et al., 1987) show that signal transmission occurs at the mid-region of the sigmoidal transfer curve where the transmission is linear. It is concluded, therefore, that operation ranges over the synaptic transfer curve differ between first synapses and second synapses (Fig. 35B), thus resulting in fundamental differences in signal transmission, i.e. transmission is non-linear and half-wave rectifying at second synapses, whereas it is linear at first synapses. Rectified responses have been noted in some, if not all, third-order neurones of a variety of visual systems including ganglion cells of goldfish retina (Spekreijse, 1969) and of cat retina (Hochstein and Shapley, 1976; Enroth-Cugell and Freeman, 1987), third-order neurones of locust compound eyes (Osorio, 1987, 1991; Jansonius and van Hateren, 1993), and locust ocelli (Simmons, 1981) and barnacle ocelli (Stuart and Oertel, 1978). The rectified responses seen in third-order neurones of these visual systems can be explained if their second synapses have a non-linear rectifying nature, as do second synapses of the cockroach ocellar system. The response of third-order neurones to brightening of cockroach ocelli is much smaller than that to dimming. In barnacle ocelli, Stuart and Oertel (1978) noted that the response to dimming is enhanced as signals are passed from second- to third-order neurones. Because the major function of barnacle ocelli is to detect dimming to facilitate a shadow-induced withdrawal of the animal into the shell (Gwilliam and Stuart, 1990), it is reasonable to use a large part of the dynamic range to code signals about dimming. In insect ocelli, Stange (1981) studied the role of ocelli in visual steering behaviour of dragonflies in flight and concluded that the decrease in illuminance of the ocelli has a strong effect on inducing a steering behaviour to avoid a nose-dive toward the ground, whereas the increase in illumination is less effective. It appears that, in these simple visual systems, the detection of dimming is more important than that of brightening, and thus dimmingspecific responses are formed. Functional roles of rectification in more

THE DIVERSITY OF INSECT OCELLAR SYSTEMS

235

B FIRST

SECOND

response response

, V pre

FIG. 35 (A) Signal rectification by non-linear synaptic transmission from second- to third-order ocellar neurones of the cockroach. Synaptic transmission occurs using an exponentially rising part of overall sigmoidal transfer curve, thus the depolarizing response to light decrement of presynaptic neurones is amplified, while the hyperpolarizing response to light increment is compressed. From Mizunami (unpublished). (B) Linear and non-linear signal transmission at graded synapses. The synapse between second- and third-order neurones (second synapse) of cockroach ocelli operates at an initially rising part of S-shaped input/output relationship; thus, the transmission is non-linear and rectifying. The synapse between photoreceptors and second-order neurones (first synapse) of visual systems presumably operates at a mid-region of the S-curve, where the transmission is essentially linear. From Mizunami (1990a).

MAKOTO MlZUNAMl

236

advanced visual systems, such as vertebrate visual systems and insect compound eyes, are discussed in Section 8.1.1. 7.5

MULTIMODAL INTEGRATION IN THIRD-ORDER AND HIGHER OCELLAR NEURONES

7.5.1 Multimodal third-order and higher neurones Milde (1988) studied the responses of interneurones in the median posterior slope (posterior protocerebrum) of the honeybee to stimulation of ocelli and compound eyes (Fig. 36). The interneurones recorded can be divided into two groups on the basis of their anatomy. These are: (1) interneurones associated with the L-neurone’s terminals, optic commissures and tracts originating from the medulla and lobula of the compound eyes; (2) interneurones belonging to the central complex. The dendritic arborization of the former neurones overlapped with the terminal arborization of the L-neurones (Fig. 36a,d) thus possibly receiving monosynaptic inputs from L-neurones. The light responses are generally complex, differing from neurone to neurone. However, many of the recorded interneurones from the former group exhibit antagonistic ocelli-compound eye responses. Responses to ocellar illumination in the neurones of the latter group demonstrate that some input to the central complex originates from the ocelli. Ohyama and Toh (1990b) examined response characteristics of the ipsilaterally and contralaterally descending ocellar third-order neurones, the DIO and DCO neurones, of the cockroach. Both DIO and DCO neurones showed no spontaneous spike discharges, and responded with a few off-spikes to ocellar illumination. They responded to various mechanical stimuli, like cercal stimulation, with a train of spikes which was suppressed by ocellar illumination. Current injections into DCO neurones caused spikes in probable leg motor neurones in N2, N2B, and N6 nerves of metathoracic ganglion and also in metathoracic leg muscles. Ohyama and Toh (1990b) proposed that, in the cockroach, the ocelli may contribute to course control during locomotion, an interesting hypothesis worthy of examination in the future. Mimura et al. (1969, 1970) found that some neurones in the brain of the fleshfly respond to both antenna1 and ocellar stimulations. Ocellar illumina-

FIG. 36 (a) Double marking showing an interneurone at the ocellar terminal region and an L-neurone of the honeybee. LOC, lateral ocellus; MOC, median ocellus; T, trachea. (b) Light responses to short (300ms) and long (3 s) stimulation. (c) Responsehntensity function for the ocellar response. (d) Staining of a similar interneurone and (e) its light responses (stimulus 300 ms). From Milde (1988).

THE DIVERSITY OF INSECT OCELLAR SYSTEMS

H

i

237

238

MAKOTO MlZUNAMl

tion produced transient or tonic excitation, or inhibition of the spontaneous discharge. The responses to olfactory or mechanical antennal stimulation were either facilitated or inhibited by ocellar input. Bimodal neurones have also been found in the protocerebrum of the honeybee with input from antennae and ocelli (Suzuki et al., 1976). Most of the units showed phasic excitation to olfactory stimuli of the antennae or to changes in light intensity over the ocelli. Many of the bimodal units also showed tonic excitation or inhibition in constant light. The response to antennal olfactory stimulus was either facilitated, inhibited or unaffected by constant illumination of the ocelli. It is interesting to examine the functional roles of the ocellar inputs to these units. It is possible that these ocellar effects are related to the setting of the general level of excitation of the brain as proposed by Mimura et al. (1969, 1970). Alternatively, they may be related to more direct behavioural actions. 7.5.2 Descending ocellar neurones as detectors of instability in flight Wilson (1978a) proposed that the ocelli are ideal for providing information regarding instability in flight by detecting movement of the horizon relative to the body. This hypothesis is supported by a number of behavioural studies of locusts and dragonflies (Section 4.3; Goodman, 1965; Stange and Howard, 1979; Stange, 1981; Taylor, 1981a,b). Simmons (1980) found that the compound eyes, the ocelli and the cephalic wind hairs detect different sensory consequences of flight instability, and three pairs of descending interneurones, DNI, DNM, and DNC neurones, bring this information to the thorax (Fig. 37a). Subsequent studies revealed an unexpected complexity and sophistication of feature detection by these multimodal descending neurones (Reichert et al., 1985; Griss and Rowell, 1986; Rowell and Reichert, 1986). DNI, DNM and DNC ignore small objects moving in the panorama, but they respond almost exclusively to movement of the animal in space about its three axes of rotations (Fig. 37b-e). All neurones are selective to the direction of movement and to the orientation of the body relative to the visual horizon (Fig. 37c). Movements in the preferred direction produce strong responses, while movements in the anti-preferred direction usually elicit no response at all. Movements in the preferred direction, but towards the normal flying position start to produce responses only as the animal approaches the normal flight position (Fig. 37d). The neurones function as feature detectors, responding only to specific sorts of deviation from course. DNI, DNM and DNC differ from one another principally in their directionality. DNI responds optimally to a diving banked turn to the ipsilateral side, DNM to downward pitch, and DNC to a diving banked turn to the contralateral side.

THE DIVERSITY OF INSECT OCELLAR SYSTEMS

239

Further studies revealed that the DNs are the subsystems of a set of neurones referred to as descending deviation detector neurones (DDNs), which detect course deviation by integrating information from the compound eyes, the ocelli, the wind-sensitive hairs on the head and the antennae, and the neck proprioceptors and transmit to the motor centres controlling the steering movements of head, wings, legs and abdomen, to induce steering movements (Hensler, 1992). Due to the sensitivity of DDNs to the orientation of the body relative to the horizon, the flight position is stabilized when the visual horizon is oriented horizontally (Hensler, 1988), thus leading to a dorsal light response, a response in which the flying insect orients the dorsal part of its body toward the brightest areas of the space, usually the sky. The main pathway between DDNs and wing motoneurones involves a population of thoracic interneurones, TINs (Fig. 38a,b). The TINS respond to simulated course deviation, and they receive this information either directly or indirectly from DDNs (Rowell and Pearson, 1983; Reichert et al., 1985; Reichert and Rowell, 1985; Rowell and Reichert, 1991). Typically DDN input is not sufficient to drive TINs above spiking threshold, but during flight, the central rhythm generator, which provides the basic motor pattern for the wingbeat, provides additional input (Fig. 38c,d). This latter input is usually rhythmic. It serves as a periodic gate for the information from DDNs, which then reaches wing motoneurones only during specific phases of the wingbeat (Fig. 38e; Rowell and Reichert, 1986). The finding of unusually sophisticated neural mechanisms subserving the flight course control of locusts is, no doubt, one of the major accomplishments of neurophysiology .

7.6

PRINCIPLESOF VISUAL PROCESSING IN AN OCELLAR SYSTEM

In Fig, 39, information processing in the cockroach ocellar system is schematically summarized. The light inputs which enter the ocellar photoreceptors, I ’ ( t ) , consist of two components, a time-varying component, Z ( t ) , and a steady mean, I,. The light-to-voltage transduction process is linear and has a lowpass filtering property. Its output, i.e. photoreceptor response, R’(t), consists of two components, a time-varying component, R(t), and a steady component, R,. R(t) is related to I ( t ) , and R, is related to I,. The magnitude of photoreceptor response, R’(t) = R, R(t), is approximately a logarithmic function of stimulus light intensity, Z ’ ( t ) = IO+Z(r), in a wide range of light intensity:

+

R‘(t) = k log Z‘(t) R, = k log I,

MAKOTO MlZUNAMl

240

where k > O . From eqns (1) and ( 2 ) , it can be shown that the relationship between Z ( t ) and R(t) is approximately linear when Z(t) is small:

R(t) = R'(t) - R, = k log (1 + Z(t)/lo) = k (Z(t)/lo- (Z(t)/Zo)2/Z+ (Z(t)/Z0)3/3-

(3)

When the stimulus contrast, Z(r)/Zo, is small or moderate, the contribution of second and higher terms in eqn (3) is small. Thus, R ( t ) = k Z(t)/Zo.

(4)

Equation (4) indicates a quasi-linear relationship between Z ( t ) and R(t). The gain, k , changes depending on the stimulus frequency, f. The frequency dependency of the system's response is not discussed here. The photoreceptor response then feeds into first synapses, i.e. synapses between receptors and second-order neurones. The first synapses are linear and sign-inverting, and have a highpass filtering property, the output of which, i.e. the response of second-order neurones, S'(f), can be divided into two components, a time-varying part, S ( t ) , and a steady mean, So. Because of the highpass nature of first synapses, So is small. Thus, Y ( t ) can be written as: S'(t) z=

S(t) = -k' Z(t)/Zo,

(5)

FIG. 37 The descending deviation detector neurones, DNI, DNM and DNC, of the locust. (a) Morphology of DNI, DNM and DNC neurones. Scale bar, 800pm. (b) Stimulus apparatus producing simulated course deviations. The animal's head was placed at the centre of a remotely steerable translucent hemisphere, on which was painted the opaque pattern shown, representing an artificial horizon and a structured visual field; the whole was illuminated diffusely with daylight from above and behind. (c) Response of the left D N C to simulated clockwise and anticlockwise rolls away from (upper two records) and towards (lower record) the normal flying position. The sketches show the initial and final positions of the horizon, and the arrows the direction of its rotation. Only clockwise rolls away from the normal position evoke graded responses. Upper traces: intracellular recording, scalebars 50mV, 200ms; lower traces: roll monitor, vertical scale bar 75". (d) Response of same unit as in (c) to three identical frontal wind stimuli (onset at arrows) during maintained simulated visual rolls to various positions, as shown in the sketches. Calibrations as in (c). Non-preferred orientations of the visual world inhibit the response to wind. (e) Response of a DNI (centre trace) to a visually simulated roll (lower trace, vertical scale bar 75") and to simultaneous ONiOFF stimulation of the ipsilateral ocellus by means of a separate light pipe. Ocellar OFF stimuli evoke only single spikes when presented alone (asterisk). These summate with the response induced by visually signalled roll. Ocellar O N stimuli are functionally incompatible with the visual information, as they imply a roll in the opposite direction; they strongly inhibit the visual response, which otherwise would consist of a long phasic-tonic burst, as in the middle record of (c). From Reichert et al. (1985).

THE DIVERSITY OF INSECT OCELLAFi SYSTEMS

z

n

241

MAKOTO MlZUNAMl

242

where k‘o> 0. Equation ( 5 ) indicates that the response of second-order neurones linearly codes for stimulus contrast, Z(f)/Zo. The response of second-order neurones feeds into second synapses. The second synapses are static (frequency-independent) and have an exponential input/output relationship. The response of third-order neurones, T‘(t), can be written as: T‘(t)= T(t) = a(exp (S’(t)/k”)- 1) = a(exp (4Z(t)/lo)- 1)

where T(t) is a time-varying component of the response of third-order neurones; k”U, > 0; a > 0; b o > 0. Because the exponential filter enhances the response to dimming and compresses the response to brightening, the response of third-order neurones mainly codes for dimming. This model suggests that dimming detection in the insect ocellar system is performed by a cascade of a few processing steps. Each step extracts an aspect of visual signals by removing the other aspect, so that the system can finally detect specific, biologically significant features. It is possible to speculate that this cascade reflects the evolutionary history of insect ocellar systems. That is, it may have originated from simple photoreceptors, and then first-order interneurones followed enabling stimulus contrast to be

FIG. 38 Flow of information from the deviation detector neurones (DN) to the flight motoneurones (FMN) is gated in the thoracic interneurone (TIN) by summation with a phasic signal derived from the central pattern generator for flight (CPG) in the locust. (a) Presynaptic DN (dotted) and representative postsynaptic TIN (solid) in mesothoracic ganglion (scale bar 300 pm). (b) Simultaneous intracellular recordings in a non-flying locust (scale bars 2 mV and 50 mV, respectively, 50 ms) from the TIN and DN shown in (a), and an extracellular recording from the dorsal longitudinal depressor muscle (‘emg’) which serves as a monitor of flight activity. A depolarizing current of 10 n A (‘cm’) injected into DN evokes 10 action potentials, which in turn evoke small 1:l excitatory postsynaptic potentials (e.p.s.p.s) in the TIN. (c, d) Same experiment as (b), but during flight motor output, indicated on the emg trace by periodic depressor muscle potentials. Scale bars 20 mV (TIN), 100 mV (DN) and 100ms. The TIN receives strong synaptic drive from the CPG, depolarizing in the depressor phase. Note that in (c) and (d) the current pulses come at different points in the wing-beat cycle. During the depressor phase, e.p.s.p.s evoked by spikes in the DN increase the number of action potentials in the depolarized TIN by -50%, which in turn influences the postsynaptic FMN (not shown). Visual stimuli can cause a spike rate in the DN 50% greater than that induced here by current injection, so that this figure underplays the potential effect. During the elevator phase, the same D N activity does not evoke spikes in the TIN, and thus has no effect on the FMN postsynaptic to that cell. (e) Summary diagram indicating the main relationships in the circuitry discussed here. Thickness of arrows is roughly proportional to the relative strength of the interaction. The TIN population is symbolized as an AND-gate, reflecting its role. C P G drive reaches the FMNs not only via the TINS, but also by other interneurones which d o not receive DN input. From Reichert et al. (1985).

THE DIVERSITY OF INSECT OCELLAR SYSTEMS 243

244

MAKOTO MlZUNAMl Photoreceptors

Light

Transduction

I'(t) = I,

+ I(t)

Second-Order Neurons

First Synapses

Third-Order Neurons

Second Synapses

R ( t ) = Ro + R(t) R'(t) = k log l'(t)

syt) = so+ S(t) S'(t) = S(t)

T'(t) = T o + T(t) T'(t) = T(t)

R(t) = k.I(t)/I,

S(t) = -k.I(t)/lo

T(t) = a (exp(-b.l(t)/l,)- 1)

Light Intensity

Temporal Contrast

Dimming

FIG. 39 Summary diagram of signal processing in cockroach ocellar system. The light-to-voltage transduction process in photoreceptors has a linear lowpass filtering property. The photoreceptor responses, which encode absolute light intensity, feed into first synapses. The first synapse has a sign-inverting, linear lowpass filtering property. The response of second-order neurones, which encodes contrast of intensity changes, feeds into second synapses. The second synapse is static (frequency-independent) and has an exponential input/output relationship. The response of third-order neurones encodes dimming. From Mizunami (unpublished).

detected, and finally second-order interneurones followed to specifically code for dimming. In other words, the present neural circuits of the cockroach ocellar system may involve ancestral neural circuits from which the ocellar system has evolved. The possible evolution of dimming detection circuits of insect ocellar systems will be discussed in Section 8.2.

8 Comparative approaches to the evolution of visual systems In this section, I will briefly discuss implications of insect ocellar research on the evolution of neural circuits subserving some basic visual functions. In the preceding sections, I discussed the generality, specificity and possible evolution of ocellar systems by comparing different ocellar systems and by comparing ocellar systems to other visual systems (e.g. Sections 5.1, 5.2.3, 5.6, 6.1, 7.1.2, 7.4). Such a comparison provides information not only about the evolution of ocellar systems, but also about the evolution of neural circuits subserving some basic visual functions. Revealing the evolution of neural circuits of animal brains is, no doubt, one of the most important research subjects of neurobiology (Bullock, 1984; Arbas et af., 1991), but,

THE DIVERSITY OF INSECT OCELLAR SYSTEMS

245

unfortunately, the luck of suitable model systems has hindered the progress in this research field. It is a practical view of evolution that complicated neural circuits subserving advanced visual functions have been evolved by step-by-step modification of simpler neural circuits subserving simpler visual functions. Such a change should accompany an improvement of visual performance, thus limiting the possible evolutionary pathways. A possible way to understand such evolutionary changes is to carefully analyse neural circuits subserving particular visual functions and to compare them to those subserving similar and different visual functions. Recently, Mizunami (1990a,b) found that some modifications of the neural circuit model which represents dimming detection in the cockroach ocellar system successfully serve to explain some visual functions of advanced visual systems, i.e. of insect compound eyes and vertebrate retinas. This finding suggests that a careful comparison of neural circuits of different visual systems may indeed be an effective measure of the possible evolutionary changes by which complex neural circuits subserving advanced visual functions have been formed from simpler circuits subserving simpler functions. Here I will first briefly summarize the findings of Mizunami (1990a,b). Then, I will discuss the evolution of neural circuits subserving some basic visual functions, on the basis of a few assumptions. 8.1

SOME HIGHER VISUAL FUNCTIONS ARE EXPLAINED BY MODIFYING NEURAL CIRCUITS FOR SIMPLER VISUAL FUNCTIONS

Mizunami (1990a,b) described two examples in which the duplications and modifications of a model which represents dimming detection in the cockroach ocellar system can explain the neural mechanism of higher visual functions. One is the neural mechanism of segregation of contrast signals into ON, OFF and ON-OFF channels (Mizunami, 1990a), and the other is the neural basis of the formation of directionally selective response to movement (Mizunami, 1990b), both of which are major components of visual processing in insect compound eyes and vertebrate visual systems. 8.1.1 Possible neural mechanism for segregating ON and OFF signals Some third-order neurones of vertebrates (retinal ganglion cells; Spekreijse, 1969; Sakai and Naka, 1987a,b; Sakuranaga et al., 1987) and insect compound eyes (Osorio, 1987) exhibit half-wave rectified, ON-depolarizing or OFF-depolarizing responses. These neurones form separate ON and OFF channels which specifically code signals about light increment and decrement, respectively. Some third-order neurones of vertebrate retinas (Sakai and Naka, 1987a,b) and insect compound eyes (Osorio, 1987, 1991; Jansonius and van Hateren, 1993) also show full-wave rectified, ON- and OFF-depolarizing responses, which form flicker-sensitivity, ON-OFF chan-

246

MAKOTO MlZUNAMl

A: Dimming Detection (Formation of OFF-channel)

B:

A Model of Formation of ON, OFF, ON-OFF Channels Light

Linear Filter

LO ff

OFF-channel

ON-OFF channel

LO"

ON-channel

FIG. 40 (A) A circuitry model of dimming detection in cockroach ocellar system, consisting of a sign-inverting bandpass linear filter followed by a synaptic rectifier with an exponential input-output relationship. The output of the linear filter, L O f f , and that of the rectifier, OFF-channel, represents the response of second- and third-order ocellar neurones, respectively. (B) A model of contrast detector which comprises the main structure of the movement detection circuit of Fig. 41A. Lo" and L O f f are on and off depolarizing linear responses, respectively. The outputs of the circuitry consist of three classes, on, off and on-ff depolarizing rectified responses. These specifically encode light increment, light decrement and flicker, respectively, and are indicated as O N , OFF and ON-OFF channels. Modified from Mizunami (1990b).

nels. The reasons that signals of intensity changes are not transmitted by linear contrast detectors but by separate ON and OFF channels have been explained by Shiller er af. (1986), who pointed out that, by having both ON and OFF channels, signals about both dimming and brightening can be transmitted as excitatory processes, thus without maintaining a high rate of spike discharge which require a high rate of metabolic activity. This allows for an economical coding of contrast. Mizunami (1990b) pointed out that the segregation of contrast signals into ON, OFF and ON-OFF channels can be explained if the synapses from second- to third-order neurones (second synapses) of these visual systems have a rectifying nature, as do those of cockroach ocelli (Fig. 40). Indeed, Toyoda (1974) and Miller (1979) proposed that the full-wave rectified response of ON-OFF amacrine cells of

THE DIVERSITY OF INSECT OCELLAR SYSTEMS

247

vertebrate retina can be explained if the cells receive half-wave rectified synaptic input from both O N and OFF bipolar cells. 8.1.2 Possible neural mechanisms for motion detection Advanced visual systems such as insect compound eyes and vertebrate visual systems have movement detectors which code visual motion in a directionally selective manner. Behavioural and psychophysical studies show that movement detection by insects and humans can be represented by a mathematical algorithm referred to as a correlation model (Fig. 41C; Hassenstein and Reichardt, 1956; Reichardt, 1961; Santen and Sperling, 1985; Adelson and Bergen, 1985; Reichardt, 1987). The neural mechanism of this motion computation, however, has not been revealed as yet. Mizunami (1990b) described a model mathematically equivalent to the correlation-type movement detector (Fig. 41A). The main structure of the model consists of contrast detectors made up from bandpass linear filters followed by synaptic rectifiers (see Fig. 40B). Linear, one-directional lateral interactions are assumed among the contrast detectors. Thus, the basic assumption of this model is that synapses between second- and third-order neurones of movement detection systems are non-linear and rectifying, as are those of cockroach ocellar systems. Mizunami (1990b) showed that synaptic rectifiers convert linear spatial interactions into a multiplication-like (quadratic) interaction, which is the core of the correlation-type movement detector. One of the neural models, which contain both excitatory (additive) and inhibitory (subtractive) lateral interactions among contrast detectors (Fig. 41B), well approximates the correlation model (Fig. 41D) in both time-averaged and dynamic (instantaneous) responses. Some of the basic features of the model agree with those of actual movement detector neurones of insects (see discussion in Mizunami, 1990b). The applicability of the model to the actual movement detectors of insects and vertebrates awaits future examination. 8.2

EVOLUTION OF NEURAL CIRCUITS BY MODIFYING SIMPLER CIRCUITS

Here I will discuss the evolution of neural circuits subserving some basic visual functions based on the assumptions that: (1) complex neural circuits subserving advanced visual functions have evolved through step-by-step modification of simpler neural circuits subserving simpler visual functions; (2) such step-by-step circuitry changes should accompany an improvement of visual performance; and (3) the present complex neural circuits contain, at least in part, simple neural circuits from which the present neural circuits have been evolved. It is not known whether the last assumption is valid. The validity of this assumption can be known only by examining whether the evolution of actual neural circuits can be explained by studies based on this

MAKOTO MlZUNAMl

248

B

A

Stationary

Nu1 I

Preferred

E-1 model

I

'8

RETINA ~

LAMINA

DONWF MEDULLA

LOBULA PLATE

C

-

D P

Preferred

D

744 Null

9%+&

FIG. 41 (A) A model of neural circuits for directional-selective motion detection. Possible location of each process in fly's compound eye is illustrated right. The model (E-I model) consists of one-directional linear interactions among contrast-detection circuities of Fig. 40B. The model involves both excitatory and inhibitory lateral interactions. IA and IB are inputs to left and right detectors. L- is a sign-inverting bandpass linear filter, the output of which approximates the response of lamina monopolar cells. E is a sign-conserving lowpass linear filter. R is a synaptic rectifier with exponential inputloutput relationship. Lateral interaction and rectification may take place in the medulla neuropil. On and off depolarizing linear responses feed into rectifiers, producing on and off depolarizing rectified responses, Ron and RO". Ron and Roff are assumed to represent responses of ON- and OFF-EMDs (elementary movement detectors). The final outputs of the models consist of three classes, Don, Doff and Don-off.The model outputs may represent responses of motion-sensitive neurones of the lobula plate. P is the preferred direction of motion.

THE DIVERSITY OF INSECT OCELLAR SYSTEMS

249

assumption. Here I will discuss first the possible evolution of neural circuits for dimming detection circuits from simple photoreceptors and then the possible evolution of neural circuits for the detection of movement, spatial contrast, and colour contrast from simple circuits for dimming detection. 8.2.1 Formation of dimming detectors Any of the sophisticated visual systems appears to be evolved out of single photoreceptors (L of Fig. 42a) whose functions are confined to detect the distribution of light intensities around the animal to perform photokinesis or phototaxis. Two essential changes may have occurred so that dimming detectors have been evolved from photoreceptors. First, photoreceptor responses need to be highpass filtered to detect temporal contrast, and thus second-order neurones need to follow the photoreceptors (H of Fig. 42b). Such a linear contrast detector, however, has fundamental deficits in performing behavioural reactions to self and object motion: it produces responses of the opposite sign to, for example, an approach of predator of dimmer and brighter than background. Thus, the linear contrast detector produces either inhibitory or excitatory responses to an approach of predators, so the detector may facilitate or suppress, for example, escape reactions of the animal. Non-linearity, thus, needs to be introduced into the visual system. If a linear contrast detector is followed by a rectifying process (R of Fig. 42b), unwanted responses to brightening can be eliminated. Third-order neurones are thus needed to follow second-order neurones to obtain dimming detectors (Fig. 42b) useful for performing, for example, escape reactions. 8.2.2 Local temporal contrast detector and movement detectors As has been discussed in Section 8.1, simple duplications and modifications of dimming detection circuits (Figs 42b and 40A) can explain a system for (B) Responses of the E-I model to a stationary flickering light and to a sinusoidal gratin moving in the preferred and null directions. The ON-OFF motion detector (Don*'f in A) exhibits a steady res onse to motion, whereas responses of ON and OFF motion detectors (Donand DO R) oscillate, depending on the spatial phase of the stimulus pattern. The stimulus parameters are: for the response to motion in the preferred direction, the phase lag due to the spatial separation of left and right input channels, C#Js, is 7d2 and the phase lag due to the delay in lateral interaction, c#+, is d 2 ; for the response to motion in the null direction, 4t = n12 and C#J1 = -7~12; for the response to a stationary flickering light, C#Js = 0 and 4I= 7r12. (C) The correlation model proposed by Hassenstein and Reichardt (1956) and Reichardt (1961). L+ and E are sign-conserving, lowpass linear filters. M is a multiplier. D is the final output. (D) Responses of the correlation model to a sinusoidal grating moving in the preferred and null directions. The stimulus parameters are the same as for (B). From Mizunami (1990b). '

250

MAKOTO MlZUNAMl

segregating temporal contrast signals into separate O N , OFF and ON-OFF channels (Figs 42c and 40B), which allow for economical contrast coding. Further addition of linear one-directional interactions between contrast detectors can explain directionally selective movement detectors (Figs 42d and 41A). 8.2.3 Spatial contrast detectors Because detection of self and object motion is more crucial for survival than detection of the form of objects (Horridge, 1987), it appears that systems for form vision have been evolved out of motion detection systems. A necessary pre-processing for form vision is to sharpen the contours of objects, i.e. to enhance the spatial contrast. This is performed in most visual systems by lateral inhibitions (reviewed by Laughlin, 1981). A lateral inhibition network is easily attained by replacing the one-directional linear lateral interaction of movement detectors with bidirectional linear inhibitory interactions (Fig. 42e). As the subsequent rectifying process converts linear lateral interactions into multiplicative interactions (Mizunami, 1990b), the model (Fig. 42e) is essentially similar to the multiplicative lateral inhibition model described by Pinter (1983) for explaining receptive field organization of locust DCMD (descending contralateral movement detector) neurones. 8.2.4 Colour-contrast detectors In many advanced visual systems, the efficacy of object detection is enhanced by unitizing information about colour contrast. In the circuits of Fig. 44f, mutual inhibitory interaction is assumed among colour-contrast detectors (c’) which receive inputs of the opposite sign from different spectral types of photoreceptors. Neurones which exhibit such colour opponent response have been reported from second- and third-order neurones of vertebrate retinas (Daw, 1968; Kaneko and Tachibana, 1983), possible third-order neurones of compound eyes of the bee (Kien and Menzel, 1977; Hertel, 1980) and of larval ocelli of the butterfly (Ichikawa, FIG. 42 The possible evolution of neural circuits subserving basic visual functions by step-by-step modifications of simpler neural circuits. (a) A photoreceptor (L) with a linear, lowpass-filtering property. (b) A dimming detector in which the output of a lowpass filter (L) feeds into a linear highpass filter (H) and then feeds into a rectifier (R). L represents photoreceptor, H represents the first synapse and R represents the second synapse. (c) A temporal contrast detector (CD) with separate ON, OFF and ON-OFF channels, where a brightening detection pathway is added to a dimming detector of (b). (c’) Colour contrast detector, where the outputs of photoreceptors of different spectral types feed into separate contrast detecting circuits (CDs), which make mutual sign-inverting (inhibitory) linear interactions to form colour-opponent responses. (d) A directionally selective movement detector where local contrast

251

THE DIVERSITY OF INSECT OCELLAR SYSTEMS

direction of motion

a

Optomotor response Fixation of moving objects

b intensity

Photokineses Phototaxis

dimming Shadow reflex Horizon detection

Detection of spatial contrast

+ local contrast Primitive Spatial Vision

Form vision

colour contrast Colour vision

d

-I+

e

CCD

-Fl

detectors (CDs) make one-directional, lowpass filtered (delayed), linear lateral interactions. Both sign-conserving and sign-inverting lateral interactions are assumed. (e) A spatial contrast detector where local contrast detectors (CDs) make sign-inverting linear lateral interactions to form, for example, centre-surround antagonistic receptive fields. (f) A colour and spatial contrast detector where local colour contrast detectors (CCDs) make sign-inverting, linear lateral interactions. From Mizunami (unpublished).

252

MAKOTO MlZUNAMl

1986, 1990). Horn (1974) showed that a model in which linear lateral interaction is followed by a non-linear thresholding operator is effective for explaining a phenomenon referred to as colour constancy, i.e. perception of colour in humans is independent of the spectral content of the illumination (Land and McCann, 1971). Interestingly, the model of Fig. 44f is essentially similar to that of Horn (1974). Colour constancy is also found in colour vision for bees (Neumeyer, 1981; Werner et al., 1988). The concept of neural circuitry evolution by gradual modifications of existing circuits briefly described here will provide a starting point for further advancement of the study of evolution of neural circuits. I conclude that careful comparison of existing neural circuits is indeed effective for understanding the evolution of neural circuits. Further advancement of comparative approaches may reveal the evolution of the most sophisticated biological systems-the brains of animals, including humans. Acknowledgements

I thank Junji Iketa, Hiroshi Masuda and Toshihiko Ohta for collecting insects in Plates 1 and 2, Dr Shunsuke Mawatari for lending equipment for photographing, J. Iketa for typing part of the manuscript, and Yuko Narita for secretarial assistance. Part of my work is supported by Sumitomo Science Foundation, Uehara Memorial Foundation and by grants from the Ministry of Education of Japan. References Adelson, E. H. and Bergen, J . R. (1985). Spatiotemporal energy models for the perception of motion. J. Opt. SOC. A m . A2, 284-299. Ammerrnuller, J. (1986). Passive cable properties of locust ocellar L-neurons. J. Comp. Physiol. 158, 339-344. Ammermuller, J . and Weiler, R. (1985). S-neurons and not L-neurons are the source of GABAergic action in the ocellar retina. J. Comp. Physiol. 157, 779-788. Ammermuller, J. and Zettler, F. (1986). Time- and voltage-dependent currents in locust ocellar L-neurons. J. C o m p . Physiol. 159, 363-376. Arbas, E. A., Meinertzhagen, I. A. and Shaw, S. R. (1991). Evolution in nervous systems. Annu. Rev. Neurosci. 14, 9-38. Autrum, H. (1950). Die Belichtungspotentiale und das Sehen der Insekten (Untersuchungen an Calliphora und Dixippus). Z . Vergl. Physiol. 32, 176-227. Baader, A. (1989). Sensitivity of ocellar interneurons of the honeybee to contrast and temporally modulated light. I . Neurobiol. 20, 519-529. Barlow, R. B. Jr., Bolanowski, S. J . Jr. and Brachman, M. L. (1977). Efferent optic nerve fibers mediate circadian rhythms in the Limulus eye. Science 197, 86-89. Barlow, R. B. J r . , Kaplan, E., Renninger, G. H. and Saito, T. (1987). Circadian rhythms in Limulus photoreceptors. I. Intracellular studies. J. Gen. Physiol. 89, 353-378.

THE DIVERSITY OF INSECT OCELLAR SYSTEMS

253

Baylor, D. A. and Hodgkin, A. L. (1974). Changes in time scale and sensitivity in turtle photoreceptors. J . Physiol. 242, 729-758. Baylor, D. A., Fuortes, M. G. F. and O’Bryan, P. M. (1971). Receptive field of single cones in the retina of the turtle. J . Physiol. 214, 256-294. Bicker, G. (1991). Taurine-like immunoreactivity in photoreceptor cells and mushroom bodies: a comparison of the chemical architecture of insect nervous systems. Brain Res. 560, 201-206. Bicker, G., Schafer, S., Ottersen, 0. P. and Storm-Mathisen, J. (1988). Glutamatelike immunoreactivity in identified neuronal populations of insect nervous systems. J . Neurosci. 8, 2108-2122. Browse-Gaury, P. (1968). Modification de la neurosecretion au niveau de la pars intercerebralis de Periplunetu americana L. en I’absence de stimuli ocellaires. Bull. Biol. 102, 481490. Brousse-Gaury, P. (1975). Projection des informations ocellaires dans le cerveau des Blattes: analyse et controle. Bull. Biol. 109, 125-163. Bullock, T. H. (1984). Comparative neuroscience holds promise for quiet revolutions. Science, 225, 473-478. Bullock, T. H. and Horridge, G. A. (1965). “Structure and Function in the Nervous Systems of Invertebrates”, Vols I and 11. W. H. Freeman & Co, San Francisco. Caesar, C. J. (1913). Die stirnaugen der ameisen. Zool. Jahrb. (Anat.) 35, 161-242. Cassier, P. (1965). Contribution a I’etude des reactions photomenotactique du criquet migrateur: Locusta migratoria migratoriodes (R. et F.) leur determinisme sensoriel. Insectes SOC. 12, 365-381. Chappell, R. L. and DeVoe, R. D. (1975). Action spectra and chromatic mechanisms of cells in the median ocelli of dragonflies. J . Gen. Physiol. 65, 399-419. Chappell, R. L. and Dowling, J . E. (1972). Neural organisation of the median ocellus of the dragonfly. 1. Intracellular electrical activity. J . Gen. Physiol. 60, 121-147. Chappell, R. L., Goodman, L. J . and Kirkham, J . B. (1978). Lateral ocellar nerve projections in the dragonfly brain. Cell Tissue Res. 190, 99-114. Chappell, R. L., Naka, K.-I. and Sakuranaga, M. (1985). Dynamics of turtle horizontal cell response. J . Gen. Physiol. 86, 423-453. Cooter, R. J. (1975). Ocellus and ocellar nerves of Periplunetu americuna (Orthoptera: Dictyoptera). Int. J . Insect Morphol. Embryol. 4, 273-288. Cornwell, P. B. (1955). The functions of the ocelli of Culliphora (Diptera) and Locustu (Orthoptera). J . Exp. Biol. 32, 217-237. Cowman, A. F., Zuker, C. S. and Rubin, G. M. (1986). An opsin gene expressed in only one photoreceptor cell type of the Drosophilu eye. Cell 44,705-710. Crowson, R. (1981). “The Biology of the Coleoptera.” Academic Press, London. Datum, K.-H., Weiler, R. and Zettler, F. (1986). Immunocytochemical demonstration of gamma-amino butyric acid and glutamic acid decarboxylase in R7 photoreceptors and C2 centrifugal fibers in the blowfly visual system. J . Comp. Physiol. 159, 241-249. Daw, N. W. (1968). Colour-coded ganglion cells in the goldfish retina: extension of their receptive fields by means of new stimuli. J . Physiol. 197, 567-592. Dickens, J . C. and Eaton, J . L. (1973). External ocelli in Lepidoptera previously considered to be anocellate. Nature 242, 205-206. Dixon, R., Kobilka, B., Strader, D., Benovic, J., Dohlman, H., Frielle, T., Bolanowski, M., Bennet, C., Rands, E., Diehl, R., Mumford, R., Slater, E., Sigal, I . , Caron, M. G., Lefkowitz, R. J . and Strader, C . D. (1986). Cloning of the gene and cDNA for mammalian p-adrenergic receptor and homology with rhodopsin, Nature 321, 75-79.

254

MAKOTO MlZUNAMl

Dow, M. A. and Eaton, J. L. (1976). Fine structure of the ocellus of the cabbage looper moth (Trichoplusia ni). Cell Tissue Res. 171, 523-533. Dowling, J . E. and Chappell, R. L. (1972). Neural organization of the median ocellus of the dragonfly. 11. Synaptic structure. J . Gen. Physiol. 60, 148-165. Dreisig, H. (1980). The importance of illumination level on the daily onset of flight activity in nocturnal moths. Physiol. Entornol. 5 , 327-342. Dubs, A., (1981). Non-linearity and light adaptation in the fly photoreceptors. J . Comp. Physiol. 144, 53-59. Eaton, J. L. (1971). Insect photoreceptor: An internal ocellus is present in sphinx moths. Science 173, 822-823. Eaton, J. L. (1976). Spectral sensitivity of the ocelli of the adult cabbage looper moth Trichoplusia ni. J . Comp. Physiol. 109, 17-24. Eaton, J. L. (1983a). Identification of dorsal ocellar primordia in larvae of Trichoplusia ni (Lepidoptera: Noctuidae). Ann. Entomol. SOC.Am. 76, 884-886. Eaton, J. L. (1983b). Development of dorsal ocelli and ocellar interneurons in Trichoplusia ni (Lepidoptera: Noctuidae). Ann. Entomol. SOC.Am. 76, 887-889. Eaton, J. L. and Pappas, L. G. (1977). Synaptic organisation of the cabbage looper moth ocellus. Cell tissue Res. 183, 291-297. Eaton, J . L. and Sprint, M. M. (1985). Moth ocellar interneurons show abnormal development in the absence of receptor innervation. Experientia 41, 111-112. Eaton, J. L., Tignor, K. R. and Holtzman, G. I. (1983). Role of moth ocelli in timing flight initiation at dusk. Physiol. Entomol. 8, 371-375. Enroth-Cugell, C. and Freeman, A. W. (1987). The receptive-field spatial structure of cat retinal Y cells. J . Physiol. 384, 49-79. Feiler, R., Harris, W.A., Kirschfeld, K., Wehrhahn, C. and Zuker, C. S. (1988). Targeted misexpression of a Drosophila opsin gene leads to altered visual function. Nature 333, 737-741. Fent, K. and Wehner, R. (1985). Ocelli: A celestial compass in the desert ant Cataglyphis. Science 228, 192-194. Fischbach, K . F. and Reichert, H. (1978). Interactions of visual subsystems in Drosophila melanogaster: A behavioural genetic analysis. Biol. Behav. 3, 305317. Fleissner, G. and Fleissner, G . (1978). The optic nerve mediates the circadian pigment migration in the median eyes of the scorpion. Comp. Biochem. Physiol. 61, 69-71. Fleissner, G . and Schliwa, M. (1977). Neurosecretory fibers in the median eyes of the scorpion, Androctonus australis L. Cell Tissue Res. 178, 189-198. French, A. S. and Jarvilehto, M. (1978). The transmission of information by first and second order neurons of the fly visual system. J . Comp. Physiol. 126, 87-96. Fuortes, M. G. F. and Hodgkin, A. L. (1964). Changes in time scale and sensitivity in the ommatidia of Lirnulw. J . Physiol. 172, 239-263. Goldsmith. T. H. and Ruck, P. R. (1958). The spectral sensitivities of the dorsal ocelli of cockroaches and honeybees. J . Gen. Physiol. 41. 1171-1 185. Goodman, L. J. (1965). The role of certain optomotor reactions in regulating stability in the rolling plane during flight in the desert locust Schistocerca gregaria. J . Exp. Biol. 42, 385407. Goodman, L. J. (1968). Interaction between ocellar and compound eye visual units in Schistocerca gregaria. Proc. 13th Int. Congr. Entomology Moscow 11, 17. Goodman, L. J. (1970). The structure and function of the insect dorsal ocellus. Adv. Insect Physiol. 7, 97-195. Goodman, C. S. (1974). Anatomy of locust interneurons: Constancy and variability. J . Comp. Physiol. 95, 185-201.

THE DIVERSITY OF INSECT OCELLAR SYSTEMS

255

Goodman, L. J. (1975). The neural organization and physiology of the insect dorsal ocellus. I n “The Compound Eye and Vision of Insects” (Ed. G. A. Horridge), pp. 515-548. Oxford University Press, Oxford. Goodman, C. S. (1976a). Constancy and uniqueness in a large population of small interneurons. Science 193, 502-504. Goodman, C. S. (1976b). Anatomy of the ocellar interneurons of acridid grasshoppers. I. The large interneurons. Cell Tissue Res. 175, 183-202. Goodman, C. S. (1977). Neuron duplications and deletions in locust clones and clutches. Science 197, 1384-1386. Goodman, L. J. (1981). Organisation and physiology of the insect dorsal ocellar system. In “Handbook of Sensory Physiology”, vol. VII 6C (Ed. H. Autrum), pp. 201-286, Springer-Verlag, Berlin. Goodman, C. S. and Williams, J. L. D. (1976). Anatomy of the ocellar interneurons of acridid grasshoppers. 11. The small interneurons. Cell Tissue Res. 175, 203-226. Goodman, L. J., Mobbs, P. G. and Guy, R. G. (1977). Information processing along the course of a visual interneuron. Experientia 33, 748-750. Goodman, L. J., Mobbs, P. G. and Kirkham, J. B. (1979). The fine structure of the ocelli of Schistocerca gregaria: The neural organisation of the synaptic plexus. Cell Tissue Res. 196, 487-510. Goodman, L. J., Fletcher, W. A., Guy, R. G . , Mobbs, P. G. and Pomfrett, C. D. J . (1987). Motion sensitive descending interneurons, ocellar LD and neck motoneurons in the bee: A neural substrate for visual course control in Apis mellifera. In “Neurobiology and Behaviour of Honeybees” (Eds R. Menzel and A. Mercer), pp. 158-171, Springer-Verlag, Berlin. Goodman, L. J., Ibbotson, M. R. and Pomfrett, C. J. D. (1990). Directional tuning of motion-sensitive interneurones in the brain of insects. In “Higher Order Sensory Processing” (Ed. D. M. Guthrie), pp. 4 3 4 3 , Manchester University Press, Manchester. Gould, J. L. (1975). Honey bee recruitment: The dance-language controversy. Science 189, 685-693. Griss, C. and Rowell, C. H. F. (1986). Three descending interneurons reporting deviation from course in the locust. I. Anatomy. J . Comp. Physiol. 158, 765-774. Guy, R. G., Goodman, L. J. and Mobbs, P. G. (1977). Ocellar connections with the ventral nerve cord in the locust, Schistocerca gregaria: Electrical and anatomical characteristics. J . Comp. Physiol. 115, 337-350. Guy, R. G., Goodman, L. J. and Mobbs, P. G. (1979). Visual interneurons in the bee brain: Synaptic organisation and transmission by graded potentials. J . Comp. Physiol. 134, 253-264. Gwilliam, G. F. and Stuart, A. E. (1990). Characteristics of neurones projecting from the supraoesophageal ganglion in the shadow reflex pathway of the barnacle. J . Exp. Biol. 151, 83-107. Hallberg, E., Nilsson, H. L. and Elofsson, R. (1980). Classification of amphipod compound eyes: the fine structure of the ommatidial units (Crustacea, Amphipoda). Zoomorphologie 94, 279-306. Hardie, R. C. (1987). Is histamine a neurotransmitter in insect photoreceptors? J . Comp. Physiol. 161, 201-213. Hardie, R. C. (1988). Effects of antagonists on putative histamine receptors in the first visual neuropile of the housefly (Musca domestica). J . Exp. Biol. 138, 221-241. Hardie, R. C. (1989). A histamine-activated chloride channel involved in neurotransmission at a photoreceptor synapse. Nature 339, 704-707. Hassenstein, B. and Reichardt, W. (1956). Systemtheoretische Analyse der Zeit-,

256

MAKOTO MlZUNAMl

Reihenfolgen- und Vorzeichenauswertung bei der Bewegungsperzeption des Riisselkafers Chlorophanus. Z . Narurforsch. l l b , 513-524. Hayashi, J . H., Moore, J . W. and Stuart, A. E. (1985). Adaptation in the input-output relation of the synapse made by the barnacle’s photoreceptor. J . Physiol. 368, 175-195. Heinzeller, T. (1976). Circadiane anderungen im endokrinen system der honigbiene, Apis rnellifera, effekt von haft und ocellenblendung. J . Insect Physiol. 22, 3 15-32 1. Hengstenberg, R. (1984). Roll-stabilization during flight of the blowfly’s head and body by mechanical and visual cues. In “Location and Orientation in Biology and Engineering” (Eds D. Varju and H. U. Schnitzler), pp. 121-134. SpringerVerlag, Berlin, Heidelberg, New York. Hengstenberg, R. (1993). Multisensory control in insect oculomotor systems. In “Visual Motion and its Role in the Stabilization of Gaze” (Eds F. A. Miles and J . Wallman), pp. 285-298. Elsevier, New York. Hengstenberg. R. and Hengstenberg, B. (1981). Intracellular staining of insect neurons with procion yellow. In “Neuroanatomical Techniques” (Eds N . J . Strausfeld and T. A. Miller), pp. 307-324. Springer-Verlag, Berlin, Heidelberg, New York. Hensler, K. (1988). The pars intercerebralis neurone PI(2)5 of locusts: convergent processing of inputs reporting head movements and deviations from straight flight. J . Exp. Biol. 140, 511-533. Hensler, K. (1992). Neural co-processing of course deviations and head movements in locusts. I. Descending deviation detectors. J . Cornp. Physiol. 171, 257-271. Hertel, H. (1980). Chromatic properties of identified interneurons in the optic lobes of the bee. J . Cornp. Physiol. 137, 215-231. Hesse, R. (1908). Das Sehen der neideren Tiere. G. Fischer, Jena. Hill, G. F. (1942). Termites (Isoptera) from the Australian Regions. (Referred to in Kalmus, 1945). Hochstein, S. and Shapley, R. M. (1976). Linear and non-linear spatial subunits in Y cat retinal ganglion cells. J . Physiol. 262, 265-284. Homberg, U. and Hildebrand, J. G. (1991). Histamine-immunoreactive neurons in the midbrain and subesophageal ganglion of the sphinx moth Munduca sexta. J . Comp. Neurol. 307, 647457. Horn, B. K . P. (1974). Determining lightness from an image. Computer Graphics and Image Processing 3, 277-299. Horridge, G. A. (1987). The evolution of visual processing and the construction of seeing systems, Proc. R . SOC.Lond. B. 230, 279-292. Hu, K. G. and Stark, W. S. (1977). Specific receptor input into spectral preference in Drosophila. J . Cornp. Physiol. 121, 241-252. Hu, K. G. and Stark, W. S. (1980). The roles of Drosophila ocelli and compound eyes in phototaxis. J . Cornp. Physiol. 135, 85-95. Hu, K. G., Reichert, H. and Stark, W. S. (1978). Electrophysiological characterization of Drosophila ocelli. J . Cornp. Physiol. 126, 15-24. Ichikawa, T. (1986). Color and spatial antagonism in the visual system of the larval swallowtail butterfly. Brain Res. 397, 381-385. Ichikawa, T. (1990). Spectral sensitivity of elementary color-coded neurons in butterfly larva. J . Neurophysiol. 64, 1861-1872. Jander, R. and Barry, C. K. (1968). Die phototaktische Gegenkopplung von Stirnocellen und Facettenaugen in der Phototropotaxis der Heuschrecken und Grillen (Saltatoptera: Locustu rnigratoria und Gryllus birnaculutus). Z . vergl. Physiol. 57, 432-458.

THE DIVERSITY OF INSECT OCELLAR SYSTEMS

257

Jansonius, N. M. and van Hateren, J. H. (1993). On-off units in the first optic chiasm of the blowfly. 11. Spatial properties. J . Comp. Physiol. 172, 467471. Kalmus, H. (1945). Correlations between flight and vision, and particularly between wings and ocelli, in insects. Proc. R. Entomol. SOC. Lond. A20, 84-96. Kaneko, A. and Tachibana, M. (1983). Double color-opponent receptive fields of carp bipolar cells. Vision Res. 23, 381-388. Kastberger, G. (1990). The ocelli control the flight course in honeybees. Physiol. Entomol. 15, 337-346. Kelly, D. H. (1971). Theory of flicker and transient responses. I. Uniform fields. J . Optical SOC. Am. 61, 537-546. Kerfoot, W. B. (1967a). Correlation between ocellar size and the foraging activities of bees (Hymenoptera: Apoidea). Am. Nat. 101, 65-70. Kerfoot, W. B. (1967b). Dorsal light receptors. Nature 215, 305-307. Kien, J . and Menzel, R. (1977). Chromatic properties of interneurons in the optic lobes of the bee. 11. Narrow band and colour opponent neurons. J . Comp. Physiol. 113, 35-53. Kirschfeld, K. and Lutz, B . (1977). The spectral sensitivity of the ocelli of Calliphora (Diptera). Z . Naturforsch. 32c, 439-441. Kirschfeld, K., Feiler, R. and Vogt, K . (1988). Evidence for a sensitizing pigment in the ocellar photoreceptors of the fly (Musca, Calliphora).J . Comp. Physiol. 163, 42 1-423. Klingman, A. and Chappell, R. L. (1978). Feedback synaptic interaction in the dragonfly ocellar retina. J. gen. Physiol. 71, 157-175. Knight, B. W. (1972a). Dynamics of encoding in a population of neurons. J . Gen. Physiol. 59, 734-766. Knight, B . W. (1972b). The relationship between the firing rate of a single neuron and the level of activity in a population of neurons. Experimental evidence for resonant enhancement in the population response. J . Gen. Physiol. 59, 767-778. Knight, B . W . , Toyoda, J.-I. and Dodge, F. A. (1970). A quantitative description of the dynamics of excitation and inhibition in the eye of Limulus. J . Gen. Physiol. 56, 421-437. Kondo, H. (1978). Efferent system of the lateral ocellus in the dragonfly: Its relationships with the ocellar afferent units, the compound eyes and the wing sensory system. J . Comp. Physiol. 125, 341-349. Koontz, M. A. and Edwards, J. S. (1984). Central projections of first-order ocellar interneurons in two orthopteroid insects Acheta domesticus and Periplaneta americana. Cell Tissue Res. 236, 133-146. Kral, K. (1979). Neuronal connections in the ocellus of the wasp (Paravespufa vulgaris L). Cell Tissue Res. 203, 161-171. Kral, K. (1980). Acetylcholinesterase in the ocellus of Apis mellifica. J . Insect Physiol. 26, 807-809. Kral, K. (1982). Large ocellar interneurons (L-neurons) in the brain of the wasp, Paravespula vulgaris L. and P. germanica L. (Hymenoptera: Vespidae). Int. .I. Insect Morphol. Embryol. 11, 307-318. Kral, K. (1983). Ultrastructural analysis of the ocellar tract of the wasp, Paravespula vulgaris L. and P . germanica L. (Hymenoptera: vespidae). Int. J. Insect Morphof. Embryol. 12, 313-320. Kral. K . , Bradacs, H. and Pabst, M. A. (1985). Freeze-fracture characteristics of the ocellar retina of the honeybee Apis mellifca carnica, Pollm. (Hymenoptera: Apidae). Int. J . Insect Morphol. Embryol. 14, 63-73. Kristensen, N. P. (1981). Phylogeny of insect orders. Annu. Rev. Entomol. 26, 135-157.

258

MAKOTO MlZUNAMl

Labhart, T. (1977). Electrophysiological recordings from the lateral ocelli of Drosophila. Naturwissenschaften 64,99-100. Lall, A. B. and Trouth, C. 0. (1989). The spectral sensitivity of the ocellar system in the cricket Gryllus firrnus (Orthoptera: Gryllidae). J . Insect Physiol. 35, 805-808. Land, E. H. and McCann, J. J. (1971). Lightness and retinex theory. J . Opt. SOC. A m . 61, 1-11. Land, M. F. and Fernald, R. D. (1992). The evolution of eyes. Annu. Rev. Neurosci. 15, 1-29. Laughlin, S. B. (1976). The sensitivities of dragonfly photoreceptors and the voltage gain of transduction. J . Cornp. Physiol. 111, 221-249. Laughlin, S. B. (1981). Neural principles in the peripheral visual systems of invertebrates. In “Handbook of Sensory Physiology”, vol. VII/6B (Ed. H. Autrum), pp. 135-280. Springer-Verlag, Berlin, Heidelberg, New York. Laughlin, S. B. and Hardie, R. C. (1978). Common strategies for light adaptation in the peripheral visual systems of fly and dragonfly. J . Cornp. Physiol. 128, 319-340. Laughlin, S. B., Howard, J. and Blakeslee, B . (1987). Synaptic limitations to contrast coding in the retina of the blowfly Calliphora. Proc. R. Sor. Lond. B. 231, 437467. Lin, J . T., Toh, Y., Mizunami, M. and Tateda, H. (1990). Putative neurotransmitter in the ocellar neuropil of American cockroaches. Zool. Sci. 7, 593-603. Littlewood, P. M. H. and Simmons, P. J . (1988). Hexamminecobaltic chloride provides a simple method for marking neurones for electron microscopy. Brain Res. 445, 165-170. Littlewood, P. M. H. and Simmons, P. J. (1992). Distribution and structure of identified tonic and phasic synapses between L-neurons in the locust ocellar tract. J . Cornp. Neurol. 325, 493-513. Metschl, N. (1963). Electrophysiologische untersuchungen an den ocellen von Calliphora. Z. vergl. Physiol. 47, 230-255. Meyer, H. W. (1978). Phototaxis in the walking male and female fly (Calliphora erythrocephala Meig.). Biol. Cybern. 31, 7-13. Meyer-Rochow, V. B. (1980). Electrophysiologically determined spectral efficiencies of the compound eye and median ocellus in the bumblebee Bornbus hortorum rarhakirnalainen (Hymenoptera, Insecta). J . Cornp. Physiol. 139, 261-266. Michener, C. D. (1944). Comparative external morphology, phylogeny, and a classification of the bee (Hymenoptera). Bull. A m . Mus. Nut. Hist. 82, 156-326. Milde, J. J . (1981). Graded potentials and action potentials in the large ocellar interneurons of the bee. J . Cornp. Physiol. 143, 427-434. Milde, J. J. (1984). Ocellar interneurons in the honey bee. Structure and signals of L-neurons. J . Cornp. Physiol. 154, 683493. Milde, J. J. (1986). Characteristics of interneurons in the ocellar system of the honeybee. Exp. Biol. 46, 1-9. Milde, J . J. (1987). The ocellar system of the honeybee. In “Neurobiology and Behaviour of Honeybees” (Eds R. Menzel and A. Mercer), pp. 191-200. Springer-Verlag, Berlin, Heidelberg, New York. Milde, J. J. (1988). Visual responses of interneurones in the posterior median protocerebrum and central complex of the honeybee Apis rnellifera. J . Insect Physiol. 34, 427-436. Milde, J. J. and Homberg, U. (1984). Ocellar interneurons in the honeybee. Characteristics of spiking L-neurons. J . Cornp. Physiol. 155, 151-160. Miller, R. F. (1979). The neuronal basis of ganglion-cell receptive-field organization and the physiology of amacrine cells. In “The Neuroscience: Fourth Study

THE DIVERSITY OF INSECT OCELLAR SYSTEMS

259

Program”. (Eds F. C . Schmitt and F. G . Worden), pp. 227-245. The MIT Press, Cambridge. Miller, G. V., Hansen, K. N. and Stark, W. S. (1981). Phototaxis in Drosophila: R1-6 input and interaction among ocellar and compound eye receptors. J . Insect Physiol. 27, 813-8 19. Mimura, K., Tateda, H., Morita, H. and Kuwabara, M. (1969). Regulation of insect brain excitability by ocellus. Z . vergl. Physiol. 62, 382-394. Mimura, K., Tateda, H., Morita, H. and Kuwabara, M. (1970). Convergence of antenna1 and ocellar inputs in the insect brain. Z . vergl. Physiol. 68, 301-310. Mismer, D . and Rubin, G. M. (1987). Analysis of the promoter of the ninaE opsin gene in Drosophila melanogaster. Genetics 116, 565-578. Mismer, D . , Michael, W. M., Laverty, T. R . and Rubin, G . M. (1988). Analysis of the promoter of the Ph2 opsin gene in Drosophila melanogaster. Genetics 120, 173-180. Mizunami, M. (1989). Dynamics of second-order neurons of cockroach ocelli. In “Neurobiology of Sensory Systems” (Eds N. R. Singh and N. J . Strausfeld), pp. 71-84, Plenum Press, New York. Mizunami, M. (1990a). Nonlinear signal transmission between second- and thirdorder neurons of cockroach ocelli. J . Gen. Physiol. 95, 297-317. Mizunami, M. (1990b). Synaptic rectification model equivalent to the correlationtype movement detector. Biol. Cybern. 64, 1-6. Mizunami, M. and Tateda H. (1986). Classification of ocellar interneurones in the cockroach brain. J . Exp. Biol. 125, 57-70. Mizunami, M. and Tateda, H . (1988a). Dynamic relationship between the slow potential and spikes in cockroach ocellar neurons. J . Gen. Physiol. 91, 703-723. Mizunami, M. and Tateda, H. (1988b). Synaptic transmission between second- and third-order neurones of cockroach ocelli. J . Exp. Biol. 140, 557-561. Mizunami, M., Yamashita, S. and Tateda, H. (1982). Intracellular staining of the large ocellar second order neurones in the cockroach. J . C o m p . . Physiol. 149, 2 15-219. Mizunami, M . , Tateda, H. and Naka K.-I. (1986). Dynamics of cockroach ocellar neurons. J . Gen. Physiol. 88, 275-292. Mizunami, M., Yamashita, S. and Tateda, H. (1987). Calcium-dependent action potentials in the second-order neurones of cockroach ocelli. J . Exp. Biol. 130, 259-274. Mizunami, M., Weibrecht, J. M. and Strausfeld, N . J. (1993). A new role for the insect mushroom bodies. In “Biological Neural Networks in Invertebrate Neuroethology and Robotics” (Eds R. D. Beer, R. Ritzmann, and T. Mckenna), pp. 199-225. Academic Press, San Diego. Mobbs, P. G . (1976). Development of the locust ocellus. Nature 264, 269-271. Mobbs, P. G. (1979). Development of the dorsal ocelli of the desert locust, Schistocerca gregaria Forsk. (Orthoptera: Acrididae). Int. J . Znsect Morphol. Embryol. 8: 237-255. Mobbs, P. G . (1985). Brain Structure. In “Comparative Insect Physiology and Biochemistry and Pharmacology vol. 5: Nervous system: Structure and motor function” (Eds G. A . Kerkut and L. I. Gilbert), pp. 300-370. Pergamon Press, Oxford. Mobbs, P. G . , Guy, R. G., Goodman, L. J . and Chappell, R . L. (1981). Relative spectral sensitivity and reverse Purkinje shift in identified L-neurons of the ocellar retina. J. C o m p . Physiol. 144, 91-97. Mote, M. I . and Wehner, R. (1980). Functional characteristics of photoreceptors in the compound eye and ocellus of the desear ant, Cataglyphis bicolor. J . Comp. Physiol. 137, 63-71.

260

MAKOTO MlZUNAMl

Naka, K.-I., Chan, R. Y. and Yasui, S. (1979). Adaptation in catfish retina. J . Neurophysiol. 42, 441-454. Naka, K.-I., Itoh, M.-A. and Chappell, R. L. (1987). Dynamics of turtle cones. J . Gen. Physiol. 89, 321-337. Naomi, S.-I. (1987). Comparative morphology of the Staphylinidae and the allied groups (Coleoptera, Staphylinoidea). I. Introduction, head structure, eyes and ocelli. Kontyu, Tokyo, 55, 45W.58. Nassel, D. R. and Hagberg, M. (1985). Ocellar interneurones in the blowfly Calliphora erythrocephala: Morphology and central projections. Cell Tissue Res. 242, 417-426. Nassel, D. R., Holmqvist, M. H., Hardie, R. C., Hakinson. R. and Sundler, F. (1988). Histamine-like immunoreactivity in photoreceptors of the compound eyes and ocelli of the flies Calliphora erythrocephala and Musca domestica. Cell Tissue Res. 253, 639-646. Nathans, J. and Hogness, D. S. (1984). Isolation and nucleotide sequence of the gene encoding human rhodopsin, Proc. Natl. Acad. Sci. USA 81, 48514855. Nathans, J., Thomas, D. and Hogness, D. S. (1986). Molecular genetics of human color vision: The genes encoding blue, green and red pigments. Science 232, 193-202. Neumeyer, C. (1981). Chromatic adaptation in the honeybee: successive color contrast and color constancy. J . Comp. Physiol. 144, 543-553. Nilsson, D. E. (1985). Optics and evolution of the compound eye. I n “Facets of Vision” (Eds D . G . Stavenga and R. C. Hardie), pp. 30-73. Springer-Verlag, Berlin, Heidelberg. Nishitsutsuji-Uwo, J . and Pittendrigh, C. S. (1968). Central nervous system control of circadian rhythmicity in the cockroach. 11. The pathway of light signals that entrain the rhythm. Z . vergl. Physiol. 58. 1-13. Normann, R . A. and Perlman, I. (1979). Signal transmission from red cones to horizontal cells in the turtle retina. J . Physiol. 286, 509-524. Ohno, S. (1970). “Evolution by Gene Duplication.” Springer-Verlag, Berlin. Ohyama, T. and Toh, Y. (1986). Multimodality of ocellar interneurones of the American cockroach. J. Exp. Biol. 125, 405-409. Ohyama, T. and Toh, Y. (1990a). Morphological and physiological characterization of small multimodal ocellar interneurons in the American cockroach. J . Comp. Neurol. 301, 501-510. Ohyama, T. and Toh, Y. (1990b). Morphological and physiological characterization of descending ocellar interneurons in the American cockroach. J . Cornp. Neurol. 301, 511-519. Okano, T . , Kojima, D., Fukada, Y., Schichida, Y. and Yoshizawa, T. (1992). Primary structures of chicken cone visual pigments: Vertebrate rhodopsins have evolved out of cone visual pigments. Proc. Natl. Acad. Sci. USA 89, 5932-5936. Osorio, D . (1987). The temporal properties of non-linear, transient cells in the locust medulla. J . Comp. Physiol. 161, 431440. Osorio, D . (1991). Mechanisms of early visual processing in the medulla of the locust optic lobe: how self-inhibition, spatial-pooling, and signal rectification contribute to the properties of transient cells. Visual Neurosci. 7 , 345-355. Pabst, M . A. and Kral, K. (1986). Cell junctions and other membrane specializations in the ocellus of the wasp, Paravespula germanica L. (Hymenoptera: Vespidae). Int. J . Insect Morphol. Embryol. 15, 331-342. Pabst, M. A . and Kral, K . (1989). Intracellular junctions and rhombic particle arrays in the developing and adult dorsal ocelli of the honeybee. Tissue and Cell 21, 199-21 0.

THE DIVERSITY OF INSECT OCELLAR SYSTEMS

261

Pan, K. C. (1981). The neural organization of the ocellar system and associated pathways in the central nervous system of the worker honeybee, Apis mellifera. Ph.D. Thesis, University of London. Pan, K. C. and Goodman, L. J. (1977). Ocellar projections within the central nervous system of the worker honey bee, Apis mellifera. Cell Tissue Res. 176, 505-527. Pappas, L. G . and Eaton, J . L. (1977). The internal ocellus of Manciuca sextu: Electroretinogram and spectral sensitivity. J. Insect Physiol. 23, 1355-1358. Patterson, J . A. and Chappell, R. L. (1980). Intracellular responses of procion filled cells and whole nerve cobalt impregnation in the dragonfly median ocellus. J. Comp. Physiol. 139, 25-39. Paulus, H. F. (1979). Eye structure and the monophyly of the arthropoda. In “Arthropod Phylogeny” (Ed. A. P. Gupta), pp. 299-383. Van Nostrand Reinhold, New York. Pinter, R. B . (1972). Frequency and time domain properties of retinular cells of the desert locust (Schistocerca gregaria) and the house cricket (Acheta domesticus).J. Comp. Physiol. 77, 383-397. Pinter, R. B . (1983). The electrophysiological bases for linear and for nonlinear product term lateral inhibition and the consequences for wide field textured stimuli. J. Theor. Biol. 105, 233-243. Pollock, J . A. and Benzer, S. (1988). Transcript localization of four opsin genes in the three visual organs of Drosophilu: RH2 is ocellus specific. Nature 333, 779-782. Purple, R. L. and Dodge, F. A. (1965). Interaction of excitation and inhibition in the eccentric cell in the eye of Limulus. Cold Spring Harbor Symp. Quant. Biol. 30,529-537. Reichardt, W. (1961). Autocorrelation, a principle for evaluation of sensory information by the central nervous system. In “Principles of Sensory Communication”. (Ed. W. A. Rosenblith), pp. 303-317, Wiley, New York. Reichardt, W. (1987). Evaluation of optical motion information by movement detectors. J. Comp. Physiol. 161, 533-547. Reichert, H. and Rowell, C. H. F. (1985). Integration of nonphaselocked exteroceptive information in the control of rhythmic flight in the locust. J. Neurophysiol. 53, 1201-1218. Reichert, H., Rowell, C. H. F. and Griss, C. (1985). Course correction circuitry translates feature detection into behavioural action in locusts. Nature 315, 142-144. Rence, B. G., Lisy, M. T., Garves, B. R. and Quinlan, B. J . (1988). The role of ocelli in circadian singing rhythms of crickets. Physiol. Entomol. 13, 201-212. Renner, M . and Heinzeller, T. (1979). Do trained honeybees with reliably blinded ocelli really return to the feeding site? J. Apicult. Res. 18, 225-229. Rivault, C. (1983). Role of photoreceptors in the synchronization of the circadian locomotor activity rhythms with the LD cycle in cockroaches (Periplanetu americana). J. Interdiscipl. Cycle Res. 14, 21-31. Roberts, S. K. (1965). Photoreception and entrainment of cockroach activity rhythms. Science 148, 958-959. Rotzler, S. (1989). Possible efference to the ocellar system of the locust and its effect on ocellar interneurons. J. Comp. Physiol. 165, 27-34. Rowell, C. H. F. and Pearson, K. G. (1983). Ocellar input to the flight motor system of the locust: structure and function. J. Exp. Biol. 103, 265-288. Rowell, C. H. F. and Reichert, H. (1986). Three descending interneurons reporting deviation from course in the locust. 11. Physiology. J. Comp. Physiol. 158, 775-794.

262

MAKOTO MlZUNAMl

Rowell, C. H. F. and Reichert. H. (1991). Mesothoracic interneurons involved in flight steering in the locust. Tissue and Cell 23, 75-139. Ruck, P. (1958). A comparison of the electrical responses of compound eyes and dorsal ocelli in four insect species. J . Insect Physiol. 2, 261-274. Sakai, H. and Naka, K.-I. (1987a). Signal transmission in catfish retina. IV. Transmission to ganglion cells. J . Neurophysiol. 58, 1307-1328. Sakai, H. and Naka, K.-I. (1987b). Signal transmission in catfish retina. V. Sensitivity and circuit. J . Neurophysiol. 58, 1329-1350. Sakuranaga, M., Ando, Y.-I. and Naka, K.-I. (1987). Dynamics of ganglion-cell response in the catfish and frog retina. J . Gen. Physiol. 90, 229-259. Salkoff, L. B. and Wyman, R. J. (1983). Ion currents in Drosophila flight muscles. J . Physiol. 337, 687-709. Santen, D. P. H. van and Sperling, G. (1985). Elaborated Reichardt detectors. J . Opt. SOC.A m . A2, 300-321. Schafer, S . , Bicker, G. (1986). Distribution of GABA-like immunoreactivity in the brain of the honeybee. J . C o m p . Neurol. 246, 287-300. Schafer, S., Bicker, G., Otterson, 0. P. and Storm-Mathisen, J. S. (1988). Taurine-like immunoreactivity in the brain of the honeybee. J . Comp. Neurol. 268, 60-70. Schlemermeyer, E., Schiitte, M. and Ammermiiller, J. (1989). Immunohistochemical and electrophysiological evidence that locust ocellar photoreceptors contain and release histamine. Neuroscience Lett. 99, 73-78. Schricker, B. (1965). Die Orientierung der Hongbiene in der Dammerung. Zugleich ein Beitrag zur Frage der Ocellenfunktion bei Bienen. Z . vergl. Physiol. 49, 420-458. Schuppe, H. and Hengstenberg, R. (1993). Optical properties of the ocelli of Calliphora erythrocephala and their role in the dorsal light response. J . Cornp. Physiol. 173, 143-149. Schiirmann, F. W. and Erber, J. (1990). FMRFamide-like immunoreactivity in the brain of the honeybee (Apis mellifera). A light- and electron microscopical study. Neuroscience 38, 797-807. Schwemer, J . (1989). Visual pigments of compound eyes: Structure, photochemistry and regeneration. I n “Facets of Vision” (Eds D. G. Stavenga and R. C. Hardie). pp. 112-133. Springer-Verlag, Berlin, Heidelberg, New York. Shaw, S. R. (1979). Signal transmission by graded slow potentials in the arthropod peripheral visual system. In “The Neurosciences: Fourth study program” (Eds F. 0. Schmitt and F. G. Worden), pp. 275-295. MIT Press, Cambridge. Shepherd, G. M. (1988). “Neurobiology” (Second Edn). Oxford University Press, New York. Shiller, P. H., Sandell, J. H. and Maunsell, J. H. R. (1986). Functions of the ON and OFF channels of the visual system. Nature 322, 824-825. Simmons, P. J. (1980). A locust wind and ocellar brain neurone. J . Exp. Biol. 85, 28 1-294. Simmons, P. J. (1981). Synaptic transmission between second- and third-order neurones of a locust ocellus, J . Cornp. Physiol. 145, 265-276. Simmons, P. J. (1982a). Transmission mediated with and without spikes at connexions between larger second-order neurones of locust ocelli. J . Comp. Physiol. 147, 401414. Simmons, P. J. (1982b). The operation of connexions between photoreceptors and large second-order neurones in dragonfly ocelli. J . Comp. Physiol. 149, 389-398. Simmons, P. J. (1985). Postsynaptic potentials of limited duration in visual neurones of the locust. J . Exp. B i d . 117, 193-213.

THE DIVERSITY OF INSECT OCELLAR SYSTEMS

263

Simmons, P. J . (1986). Interactions made by large, second-order neurones of the median ocellus of the locust. J . Comp. Physiol. 159, 97-105. Simmons, P. J. (1990). The effects of temperature on locust ocellar L-neurones and their interconnections. J . Comp. Physiol. 166, 575-583. Simmons, P. J. (1993). Adaptation and responses to changes in illumination by second- and third-order neurones of locust ocelli. J . Comp. Physiol. 173,635-648. Simmons, P. J. and Hardie, R. C. (1988). Evidence that histamine is a neurotransmitter of photoreceptors in the locust ocellus. J . Exp. Biol. 138, 205-219. Simmons, P. J. and Littlewood, P. M. H. (1989). Structure of a tonically transmitting synapse between identified interneurones in the locust brain. J. Comp. Neurol. 283, 129-142. Simmons, P. J. and Rind, F. C. (1993). Reponses to light signals by large, second-order ocellar neurones of the blowfly, Calliphora erythrocephala. J . Physiol. 473, 244p. Sontag, C. (1971). Spectral sensitivity studies on the visual system of the praying mantis, Tenodera sinensis. J . Gen. Physiol. 75, 93-112. Spekreijse, H. (1969). Rectification in the goldfish retina. Analysis by sinusoidal and auxiliary stimulation. Vision Res. 9, 1461-1472. Sprint, M. M. and Eaton, J. L. (1987). Flight behaviour of normal and anocellate cabbage loopers (Lepidoptera: Noctuidae). Ann. Entomol. SOC.A m . 80, 468-471. Stange, G. (1981). The ocellar component of flight equilibrium control in dragonflies. J . Comp. Physiol. 141, 335-347. Stange, G. and Howard, J . (1979). An ocellar dorsal light response in a dragonfly. J . Exp. Biol. 83, 351-355. Stark, W. S . , Sapp, R. and Carlson, S. D. (1989). Ultrastructure of the ocellar visual system in normal and mutant Drosophila melanogaster. J . Neurogenet. 5, 127-153. Stavenga, D. G., Bernard, G. D. , Chappell, R. L. and Wilson, M. (1979). Insect pupil mechanisms. 111. On the pigment migration in dragonfly ocelli. J . Comp. Physiol. 129, 199-205. Stockbridge, N. and Ross, W. N . (1984). Localized Ca2+ and calcium-activated potassium conductances in the terminals of a barnacle photoreceptor. Nature 309, 26&268. Stone, S . L. and Chappell, R. L. (1981). Synaptic feedback onto photoreceptors in the ocellar retina. Brain Res. 221, 374-381. Strausfeld, N. J . (1976). “Atlas of an Insect Brain.” Springer-Verlag, Berlin, Heidelberg, New York. Strausfeld, N. J. and Bassemir, U. K. (1985). Lobula plate and ocellar interneurons converge onto a cluster of descending neurons leading to neck and leg motor neuropil in Calliphora erythrocephala. Cell Tissue Res. 240, 6 1 7 4 0 . Strausfeld, N. J . and Campos-Ortega, J. A. (1977). Vision in insects: pathways possibly underlying neural adaptation and lateral inhibition. Science 195, 894-897. Strausfeld, N. J . , Bassemir, U. K., Singh, R. N. and Bacon, J. P. (1984). Organizational principles of outputs from dipteran brains. J . Insect Physiol. 30, 73-93. Stuart, A. E. and Oertel, D. (1978). Neural properties underlying processing of visual information in the barnacle. Nature 275, 287-290. Suzuki, H., Tateda, H. and Kuwabara, M. (1976). Activities of antenna1 and ocellar interneurones in the protocerebrum of the honey-bee. J . Exp. Biol. 64, 405-418. Taylor, C. P. (1981a). Contribution of compound eyes and ocelli to steering of locusts in flight. I. Behavioural analysis. J . Exp. Biol. 93, 1-18. Taylor, C. P. (1981b). Contribution of compound eyes and ocelli to steering of locusts in flight. 11. Timing changes in flight motor units. J . Exp. Biol. 93, 19-31.

264

MAKOTO MlZUNAMl

Taylor, C. P. (1981~).Graded interactions between identified neurons from the simple eyes of an insect. Bruin Res. 215, 382-387. Toh, Y. and Hara, S. (1984). Dorsal ocellar system of the American cockroach. 11. Structure of the ocellar tract. J . Ultrustruct. Res. 86, 135-148. Toh, Y. and Kuwabara, M. (1974). Fine structure of the dorsal ocellus of the worker honeybee. J . Morphol. 143, 285-306. Toh, Y. and Kuwabara, M. (1975). Synaptic organization of the flesh fly ocellus. J . Neurocytol. 4, 271-287. Toh, Y. and Sagara, H. (1984). Dorsal ocellar system of the American cockroach. I. Structure of the ocellus and ocellar nerve. J . Ultrustruct. Res. 86, 119-134. Toh, Y. and Tateda, H. (1991). Structure and function of the insect ocellus. Zool. Sci. 8, 395-413. Toh, Y. and Yokohari, F. (1988). Postsynaptic development of the dorsal ocellus of the American cockroach. J . Comp. Neurol. 269, 157-167. Toh, Y., Tominaga, Y . and Kuwabara, M. (1971). The fine structure of the dorsal ocellus of the fleshfly. J . Electron Microsc. 20, 56-66. Toh, Y., Sagara, H. and Iwasaki, M. (1983). Ocellar system of the insect: Comparison of dorsal ocellus and lateral ocellus. Vision Res. 233, 313-323. Tomioka, K . and Yamaguchi. T. (1980). Steering responses of adult and nymphal crickets to light, with special reference to the head rolling movement. J . Insect Physiol. 26, 47-57. Toyada, J.-I. (1974). Frequency characteristics of retinal neurons in the carp. J. Gen. Physiol. 63, 214-234. Tranchina, D., Gordon, J . and Shapley, R. (1983). Spatial and temporal properties of luminosity horizontal cells in the turtle retina. J . Gen. Physiol. 82, 573-598. Weber, G. and Renner, M. (1976). The ocellus of the cockroach Periplunetu umericuna (Blattariae). Receptory area. Cell Tissue Res. 168, 209-222. Weckstrom, M., Juusola, M. and Laughlin, S. B. (1992). Presynaptic enhancement of signal transients in photoreceptor terminals in the compound eye. Proc. R. SOC.Lond. B . 250, 83-89. Wehner, R. (1981). Spatial vision in arthropods. In “Handbook of Sensory Physiology” vol. VII/6C (Ed. H. Autrum), pp. 287-616. Springer-Verlag, Berlin, Heidelberg, New York. Wehner, R. (1987). ‘Matched filters’-neural models of the external world. J . Comp. Physiol. 161, 511-531. Wehner, R. and Strasser, S. (1985). The POL area of the honey bee’s eye: behavioral evidence. Physiol. Entomol. 10, 337-349. Wehrhahn, C. (1984). Ocellar vision and orientation in flies. Proc. R. SOC.Lond. B. 222. 409-411. Wellington, W. G. (1974). Bumblebee ocelli and navigation at dusk. Science 183, 550-551. Werblin, F. S. (1972). Functional organisation of a vertebrate retina: sharping up in space and intensity. Ann. N . Y . Acud. Sci. 193, 75-85. Werner, A., Menzel, R., Wehrhahn, C. (1988). Color constancy in the honeybee. J . Neurosci. 8, 156-159. Wilson, M. (1975). Autonomous pigment movement in the radial pupil of locust ocelli. Nature 258, 603-604. Wilson, M. (1978a). The functional organisation of locust ocelli. J. Comp. Physiol. 124, 297-316. Wilson, M. (1978b). Generation of graded potential signals in the second order cells of locust ocellus. J. Comp. Physiol. 124, 317-331. Wolsky, A. (1930). Optische Untersuchungen iiber die Bedeutung and Funktion der Insektenocellen. Z . Vergl. Physiol. 12, 783-787.

THE DIVERSITY OF INSECT OCELLAR SYSTEMS

LO3

Wolsky, A. (1933). Stimulationsorgane. Biol. Rev. 8, 37W17. Wunderer, H. and Kramer, J . J. D. (1989). Dorsal ocelli and light-induced diurnal activity patterns in the arctiid moth Creatonofos Transiens. J. Insect Physiol. 35, 87-95: Wunderer, H . , Weber, G. and Seifert, P. (1988). The fine structure of the dorsal ocelli in the male bibionid fly. Tissue Celi 20, 145-155. Yamashita, S. and Tateda, H. (1981). Efferent neural control in the eyes of orb weaving spiders. J. Comp. PhysioI. 143. 477-483. Yamazaki, S. and Yamashita, S. (1991). Efferent control in the ocellus of a noctuid moth. J. Cornp. Physiol. 169,. 647-652. Zipursky, S. L., Venkatesh, T. R., Teplow, D. B . and Benzer, S. (1984). Neuronal development in the Drosophila retina: Monoclonal antibodies as molecular probes. Cell 36, 15-26. Zuker, C. S . , Montell, C., Jones, K., Laverty, T. and Rubin, G. M. (1987). A rhodopsin gene expressed in photoreceptor cell R7 of the Drosophila eye: Homologies with other signal-transducing molecules. J. Neurosci. 7 , 1550-1557.

This Page Intentionally Left Blank

A1latostat ins: Identification, Primary Structures, Functions and Distribution Barbara Stay,a Stephen S. Tobeb and William G. Bendena‘ of Biology, University of Iowa, Iowa City, IA 52242-1324, USA Department of Zoology, University of Toronto, Toronto, ON M5S lA1, Canada Department of Biology, Queen‘s University, Kingston, ON K7L 3N6, Canada

a Department

1

2

3

4

5

6

Evidence that allatostatins, peptides inhibiting juvenile hormone synthesis, occur in the brain 268 1.1 Nerve severance 269 1.2 Brain implantation or removal 269 1.3 Inhibition of JH synthesis in vitro 270 1.4 Inhibition of JH synthesis by extracts of brain 270 Peptide identification by isolation 274 2.1 Amino acid sequences 274 2.2 Isolation procedure 274 2.3 Bioassay 277 2.4 Sequence and C-terminal analysis 278 2.5 Synthetic allatostatins 278 Structure-activity studies of Dipfoptera punctata allatostatins 279 3.1 The allatostatin ‘message’ sequence 279 3.2 The allatostatin ‘address’ sequence 280 3.3 Allatostatin analogues 282 3.4 Comparison of assays for inhibition of JH biosynthesis 286 Sensitivity of corpora allata to allatostatins 286 4.1 Redundancy in allatostatins 286 4.2 Developmental changes of corpora allata in response to allatostatins 287 4.3 Duality of responses to allatostatins 289 4.4 Responsiveness to analogues of allatostatins 290 4.5 Possible factors contributing to changes in responsiveness 291 4.6 Neural and humoral pathways for allatostatin action 291 4.7 Allatostatins and regulation of JH titre 292 Receptors for Diploptera punctata allatostatins 293 5.1 Approaches for the isolation and characterization of allatostatin receptors 294 5.2 The case for multiple allatostatin receptors 295 Distribution of allatostatin-immunoreactive cells 296 6.1 Manduca sexta 296 6.2 Cockroaches 299 6.3 Crickets 303

ADVANCES I N INSECT PHYSIOLOGY VOL. 25 ISBN U-12424225-7

Copyrighr 01994 Academic Press Limiied AN righrs of reproducrion in any form resewed

BARBARA STAY et a/.

268

Summary of distribution of allatostatins 303 6.5 Parallels with other neuropeptides 304 Other functions 306 7.1 Modulators of spontaneous and proctolin-induced contractions 308 7.2 Activity in interneuroneskhemical synapses 31 1 7.3 Neurohormone 311 Distribution in other insects and other phyla 312 8.1 Callatostatins, -Tyr-X-Phe-Gly-Leu/Met-amides from the blowfly 312 8.2 Immunoreactivity to allatostatin antisera in other insects 314 8.3 Crustacea, Decapoda 316 8.4 Arachnida, Acarina 316 8.5 Mollusca, Gastropoda 317 The gene for allatostatins 317 9.1 Diploptera punctata 317 9.2 Periplaneta americana 320 9.3 In situ hybridization 321 Metabolism and mode of action of allatostatins 324 10.1 Degradation of allatostatins 324 10.2 Mode of action of allatostatins 326 Factors regulating release of allatostatin 328 11.1 Nerve section, JH analogue, ovary 329 Conclusions 329 Acknowledgements 330 References 331 6.4

7

8

9

10

11 12

1 Evidence that allatostatins, peptides inhibiting juvenile hormone synthesis, occur in the brain

During insect development, juvenile hormone (JH) maintains the juvenile form and metamorphosis to the adult form appears to require a reduced titre of JH. In the adult stage, the same hormone regulates many reproductive functions including vitellogenesis (Steel and Davey, 1985). The titre of JH is largely determined by the rate at which corpora allata (CA) synthesize J H (Tobe and Stay, 1985; Tobe et al., 1985). Therefore, the production of JH must be modulated to permit the normal progress of development and reproduction. The regulation of JH production involves many factors, both extrinsic and intrinsic. The great diversity of conditions to which insects are adapted dictates that many different requirements and different systems for controlling the corpora allata exist. In all of them, the nervous system plays an important role (Tobe and Stay, 1985). In some species the brain regulates CA activity through stimulatory factors, in others inhibition appears to predominate (Scharrer, 1987; Tobe and Stay, 1985). This review will be concerned primarily with the neuropeptides of known amino acid sequence that inhibit JH production by the CA. Only a brief description with a few specific examples of the kinds of evidence that led to these identifications will be given here.

ALLATOSTATINS: IDENTIFICATION, PRIMARY STRUCTURES, AND FUNCTIONS

1.1

269

NERVE SEVERANCE

Axons with terminations in the CA originate in several regions of the brain and in the suboesophageal ganglion (reviewed in Tobe and Stay, 1985). Both light and electron microscopy have demonstrated that these fibres, which ramify in the CA, contain neurosecretory material (reviewed by Cassier, 1990). The conclusion that cerebral neuropeptides are responsible for inhibition of J H production came originally from observations following nerve severance. Scharrer (1952) demonstrated that severance of a nerve tract originating in the brain and terminating in the CA resulted in accumulation of neurosecretory material proximal to the cut, depletion of it distal to the cut, and structural changes in the CA that suggested increased CA activity. Such increased activity following nerve severance has long been inferred from experiments on several different species of insects by the induction of JH-dependent events, including supernumerary larval molts (e.g. Wigglesworth, 1936: Rhodnius prolixus; Luscher and Engelmann, 1960: Leucophaea maderae; Bhaskaran et al., 1980: Manduca sexta) and vitellogenesis at inappropriate times (Engelmann, 1959: Diploptera punctata; Hodkova, 1977: Pyrrhocoris apterus; Baehr, 1973: Rhodnius prolixus). Direct measurements of increased JH synthesis in vitro or J H titre, at some interval following severance of the C A nerves in vivo, has confirmed that the C A are released from inhibition in some species (e.g. Stay and Tobe, 1977: D . punctata; Khan et al., 1983: Leptinotarsa decemlineata; Poras et al., 1983: Locusta migratoria, Savio strain). 1.2

BRAIN IMPLANTATION OR REMOVAL

Experiments in which the brains were removed or implanted into test animals have also led to the conclusion that allatostatic material is produced by the brain. For example, based on the degree to which vitellogenesis was induced in diapausing adult P. apterus following implantation of diapausing brain-retrocerebral complexes, Hodkova (1979) concluded that in diapausing adult females, CA are inhibited in part by signals mediated by intact nerves to the C A and in part by humoral factors from the brain. These experiments were based on nerve severance or ablation of selected regions of the brain. In D . punctata last instar larvae, decapitation resulted in removal of humoral inhibition of implanted CA; this inhibition was restored following implantation of a larval protocerebrum (Paulson and Stay, 1987). In pregnant female D . punctata, removal of the brain (by decapitation) enhanced vitellogenesis and this release from inhibition was reversed by implantation of a protocerebrum (Rankin and Stay, 1985).

BARBARA STAY et al.

270

1.3

INHIBITION OF JH SYNTHESIS IN VITRO

In vitro measurements of JH synthesis by brain-corpora cardiaca (CC)-CA complexes have also provided evidence that axonal tracts arising from the neurosecretory cells of the brain deliver inhibitory factors to the CA and that these factors can also act humorally, i.e. through the incubation medium. At certain stages of the last two larval instars of M. sextu, synthesis of JH I by brain-CC-CA complexes was reduced by two-thirds, relative to isolated CA, demonstrating that the brain inhibited JH synthesis (Granger et af., 1981, 1982). J H synthesis by isolated CA incubated in the presence of the brain was not measured and thus it is not known whether an allatostatic factor was released into the medium. Such a factor was found subsequently from the wandering stage of last instar M. sexta (one of the brain-inhibitory stages demonstrated by Granger et al., 1981) by incubating isolated active CA with brains from these animals (Bhaskaran et al., 1990). The exposure time required for this inhibition was 12-16 h and the CA were inhibited until after metamorphosis as determined by an in vivo bioassay (Bhaskaran et af., 1990). Intact brain-CC-CA complexes from adult female D. punctata produced very little J H in vitro; more hormone was produced by CA in the presence of brain-CC, but isolated C A (without brain-CC) synthesized JH at the highest rate (Rankin et al., 1986). From these experiments, Rankin et al. (1986) concluded that intact nerves are normally responsible for inhibition of JH synthesis but an inhibitory factor released from the brain can also act humorally on the CA (Rankin et ul., 1986). Similar experiments have been carried out with brain-ring gland complexes of larval blowflies Sarcophuga bullata (Richard et al.. 1990) and Lucifia cuprina (T. D. Sutherland and P. D. East, personal communication). Intact complexes showed low rates of JH bisepoxide synthesis whereas ring glands co-incubated with brain or incubated alone showed 5 to 10 times greater rates of hormone production (Richard et al., 1990; T. D. Sutherland and P. D. East, personal communication). Intact brain-ring gland complexes of adult Drosophila melanogaster produced one-third of the JH produced by CA alone or those incubated with brains (Altaratz et al., 1991). No inhibition by detached brains was obvious in these experiments but extracts of brain revealed an inhibitory factor (see Section 1.4). 1.4

INHIBITION OF JH SYNTHESIS BY EXTRACTS OF BRAIN

The low rates of JH production by intact brain-CC-CA complexes in vitro relative to isolated CA suggested that release of the inhibitory material within the CA normally required the presence of intact nerves. However, release of inhibitor could be stimulated by treatment with high potassium, a

ALLATOSTATINS: IDENTIFICATION, PRIMARY STRUCTURES, AND FUNCTIONS

271

well-known depolarizing agent of neurosecretory cells and nerve endings in other insect systems (Maddrell and Gee, 1974). Treatment of isolated CA with high potassium medium did result in a marked inhibition of JH production (Rankin et a f . , 1986). Although an action of high potassium directly on CA cells could not be ruled out, extracts of C A tested on CA in vitro demonstrated that they contained a heat-stable, trypsin-sensitive inhibitory factor (Rankin et a f . , 1986). Extract of five pairs of CA inhibited JH synthesis (60%) to the same degree as 0.5 protocerebral lobe equivalents (Rankin et a f . , 1986). This study was thus the first to provide evidence of the peptidergic nature of the inhibitory material. Extract of brain tissue from 4-day last instar larvae of M . sexta, a stage at which J H synthesis is normally low, specifically decreased JH I synthesis in vitro in test CA (Granger and Janzen, 1987). This factor was heat-labile, pronase-sensitive and fast-acting, and its effect was readily reversible (Granger and Janzen, 1987). However, its apparent molecular size (Granger and Janzen, 1987) is larger than that of the now identified allatostatin from M . sexta (Kramer et a f . , 1991). Another heat-labile, pronase-sensitive factor has been extracted from brains and medium conditioned by brains of M. sexta that appears to be slow acting. Because its effect is irreversible until the completion of metamorphosis, it has been called allatinhibin to distinguish it from the fast-acting, reversible allatostatins (Bhaskaran et al., 1990; Unni et a f . , 1993). CA exposed to this brain factor for 12-16 h in vitro and subsequently tested in an in vivo bioassay were found to require metamorphosis before reversal of inhibition (Bhaskaran et al., 1990). The production of JH 111 bisepoxide by CA of larval and adult Drosophifa rnefanogaster is inhibited by a heat-stable extract of larval and adult brain tissues, respectively (Richard et a f . , 1990; Altaratz et a f . , 1991). About 80% inhibition of J H synthesis by larval glands was achieved with one to two larval brain equivalents (Richard et a f . , 1990) and 60% inhibition of adult CA with two adult brain equivalents (Altaratz et a f . , 1991). Larval S . buffata brain extract was effective on both S . bulfata CA (isolated from the rest of the ring gland) and on D . rnefanogaster ring glands (Richard et a f . , 1990). A factor was also extractable from CC-CA of adult D . mefanogaster that inhibited J H synthesis by the ring gland (dose not given) (Altaratz et a f . , 1991). The Drosophila mutant apterous that makes little JH also lacked extractable allatostatic factors in the brain (Altaratz et a f . , 1991). In support of this observation, the apterous gene product resembles a transcription factor (Cohen et a f . , 1992); this factor may regulate the expression of the D . rnelanogaster allatostatins. Aqueous extracts of larval brains of the blowfly Lucifia cuprina yielded a heat-stable factor that inhibited J H synthesis in a dose-dependent manner; four to five brain equivalents gave about 80% inhibition (T. D. Sutherland and P. D. East, personal communication). Partially purified extracts of brain tissue of the cricket Gryffusbimacufatus inhibited J H synthesis by cricket CA in vitro (Neuhauser et a f . , 1994).

TABLE 1 Primary structures of allatostatins determined by isolation and/or cDNA ~~

References for isolated peptides

Species Manduca sexta Diploptera punctata Isolation cDNA" designation 1 ASB2;V 2 3 VII 4 IV 5 6 I 7 111 8 I1 9 10 VI 11 12 13

pE-V-R-F-R-Q-C-Y-F-N-P-I-S-C-F-OH

L-Y-D-F-G-L-NH2 A - Y - S - Y - V - S - E - Y - K- R - L - P- V- Y - N - F - G L- N H 2 S-K-M-Y-G-F-G-L-NH2 D-G-R-M-Y-S-F-G-L-NHz D-R-L-Y-S-F-G-L-NH2 A-R-P-Y-S-F-G-L-NH2

-

A-P-S-G-A-8-R-L-Y-G-F-G-L-NH2 G-G-S-L-Y-S-F-G-L-NH2 G-D-G-R-L-Y-A-F-G-L-NH2 P-V-N-S-G-R-S-S-G-S-R-F-N-F-G-L-NH2 H - R- F - S F -G- L - N H 2 Y P-8-E P-F-N-F-G-L-NH2 I-P-M-Y-D-F-G-I-NH2

-

-

-

Kramer et al. (1991)

Pratt et a[. (1991a) Woodhead et al. (1994) Woodhead et al. (1989) Woodhead et al. (1989), Pratt e t a [ . (1989) Woodhead et al. (1989) Woodhead et al. (1989) Woodhead et al. (1994)

Periplaneta americana Isolation CDNA~ designation 1 2 3 4

5 I I1

6 7 8 9 10 11 12 13 14

L-Y-D-F-G-L-NH2 A-Y-S-Y-V-S-E-Y-K-R-L-P-V-Y-N-F-G-L-NH2 S-K-M-Y-G-F-G-L-NH2 S-G-N-D-G-R-L-Y-S-F-G-L-NHZ D-R-M-Y-S-F-G-L-NH, A-R-P-Y-S-F-G-L-NH, S-P-S-G-M-8-R-L-Y-G-F-G-L-NH~ G-G-S-M-Y-S-F-G-L-NH2 A-D-G-R-L-Y-A-F-G-L-NH, P-V-S-S-A-R-Q-T-G-S-R-F-N-F-G-L-NH2 S-P-Q-G-H-R-F-S-F-G-L-NH, S-L-H-Y-A-F-G-L-NH2 P-Y-N-F-G-L-NH2 I-P-M-Y-D-F-G-I-NH2

“Donly er al., 1993. ’(2. Ding, B. C. Donly, S. S. Tobe and W. G . Bendena, unpublished.

Weaver et a f . (1994) Weaver et a f . (1994)

BARBARA STAY et al.

274

2

2.1

Peptide identification by isolation AMINO ACID SEQUENCES

Once extract of brain tissue had been shown to inhibit JH synthesis, the obvious next step was to identify and characterize the active factors. This has been accomplished only for the lepidopteran Manduca sexta (Kramer et al., 1991) and two species of cockroaches Diploptera punctata (Woodhead et al., 1989; Pratt et al., 1991a) and Peripfaneta americana (Weaver et al., 1994). However, as noted in Section 9, amino acid sequences of as-yet unisolated members of the cockroach allatostatin families have been deduced from cDNA coding sequences of the allatostatin gene of D. punctata and of P. americana (Donly et al., 1993; Q . Ding, B. C. Donly, S. S. Tobe and W. G. Bendena, unpublished). The primary structures of these allatostatins are shown in Table 1. Except in its small size (15 residues) the M. sexta allatostatin shows no similarity to the allatostatins from the cockroaches; it is an acid with pyroglutamate at the N-terminus and only one has been found to date. Seven allatostatins that have been isolated from D . punctata are amidated and share the same three C-terminal residues (phenylalanine-glycineleucine). They also show strong conservation at the C-terminal 4, 5 and 6 positions. D . punctata allatostatins I, 11, 111 and IV were isolated simultaneously (Woodhead et al., 1989). Pratt et al. (1989), using a different isolation procedure, confirmed the structure of allatostatin I and later isolated the longest of the allatostatins, the tyrosine-rich allatostatin ASB2 or V (Pratt et al., 1991a). Subsequently D . punctata allatostatins VI and VII were isolated (Woodhead et al., 1994) as were P . americana allatostatins I and I1 (Weaver et al., 1994). The D . punctata and P. americana allatostatins I and I1 are highly conserved peptides that differ by only two and one residues, respectively. 2.2

ISOLATION PROCEDURE

2.2.1 Sources, extraction and separation steps The steps in the isolation of the allatostatins are shown in Table 2. The extracts were made from freshly dissected brains (Woodhead et al., 1989, 1994; Weaver et al., 1994), from brain-CC-CA complexes that were stored frozen in extraction solution (Pratt et al., 1989, 1991a) or from frozen trimmed heads (i.e. posterior part of head with brain - Kataoka et al., 1987). Acidic or acidic-ethanolic extractants were used except in the isolations by Woodhead et al. (1989, 1994) in which tissue was homogenized in saline and immediately heated at 100°C for 10 min. The starting material varied from 1000 brains to 30000 trimmed heads and the total yields

ALLATOSTATINS: IDENTIFICATION, PRIMARY STRUCTURES, AND FUNCTIONS

275

increased with increasing amounts of starting material. The yields per brain or trimmed head varied from 0.03pmol/trimmed head of M . sexta to 0.3pmol/brain for P. americana; these yields may reflect not only the differing amounts of allatostatin in the brains of these species but also the different isolation procedures. For the isolations of D. punctata allatostatin I, the different extraction and separation methods used by Woodhead et al. (1989) and Pratt er al. (1989) provided similar yields of this peptide/brain. Isolation of D. punctata tyrosine-rich allatostatin, ASB2 (V), accomplished with acid/ethanol extraction by Pratt et al. (1991a), was also realized with saline extraction of CA (Stay and Woodhead, 1993). As Table 2 indicates, the number and type of separation steps varied among investigators. In general, the larger the quantity of extract, the greater the number of steps employed. Separation of all allatostatins was accomplished by reverse-phase (RP) liquid chromatography (LC) on different types of columns, with different gradients of different organic solvents and various modifiers (Table 2). Only Weaver et al. (1994) achieved the isolation of P. americana allatostatins with a single type of column throughout with varied organic liquid phases, modifiers and gradients. Preparatory to RP chromatography, all of the samples were subjected to low pressure solid-phase extraction on RP CI8 cartridges (Sep-Pak, Waters) or C4packing (Vydac) with stepwise elution. Kramer et al. (1991) preceded the solid-phase extraction with a gel filtration step to reduce the large volume of material. Pratt et al. (1989) employed gel filtration as an analytical step to separate two size classes of allatostatins before a second solid-phase extraction with Diol Sep-Pak (Waters) to separate low and high molecular weight allatostatins. In two steps of the isolation of M . sexta allatostatin, preparative ion-exchange columns were utilized providing high capacity and good resolution for the large amount of material (Kramer et al., 1991). 2.2.2

Precautions to prevent loss of bioactiviry

Although dissection of insect brain tissue is not difficult, it is tedious and the number of animals of specific stage is limited. Therefore with limited material, minimizing the loss of biological activity in the samples is important. Various measures have been employed by different investigators toward this end. In the initial extraction, either small batches of brains were processed (100 or less for D. punctata and P. arnericana) or protease inhibitors were added (to batches of 10000 trimmed heads of M. sexta). Also, a volatile antioxidant was added to the fractions at each step to prevent oxidation of the peptide (Kramer et al., 1991; Weaver et a f . , 1994). These investigators also avoided possible loss of activity as a result of reduction of sample volume by evaporation between separations by pumping water-diluted samples containing organic solvents directly onto columns. Pratt et al. (1989, 1991a), and Weaver et al. (1994) also added a small

N

TABLE 2 Summary of steps in isolation of allatostatins species Peptide No. Reference Source

3OUW) trimmed heads. pharate adults

Extractant

Acetone defatted: 1 M H O A d 20 mM HCI Gel filtration (SP-Sephadex C-2.5) N H 4 0 A c ( p H 4;7). step elution RP C , cartidge (Vydac material) ACNfI'FA. step elution

Purification steps

Perrp/anero arnericuna

Diploprero pennara

Monduca s c m

Kramer el a/. (1991)

-J

Q)

I-IV VII VI I ASB?(V) Woodhead PI a / . Woodhead ei a / . (1994) Pratt el 01 (1989) Pratt era/.( I W l n ) (1989) -loo0 brains. -2000 brains. adult I200 brains. adult females II OOO b r a i n Z C CA virgin females females 1&12 day mated females HCUETH Saline. IIWC, 10 min

.

RP CIScartridge (Waters) ACNITFA, step elution

Ion exchange (TSK SP-5PW) LC NaCllMES buffer p H S.S. gradient RP C, (Vydac) LC I-propanoVHFBA, gradient

RP Cl* (Waters) LC A C N m A . gradient

RP C, (Vydac) LC M E T H K F A . gradient

RP C4 (Vydac) LC ACNITFA, gradient

RP CIH(Waters) LC A C N m A . gradient

RP Cln (Vydac) LC METFCTFA. gradient

RP CIS cartridge (Waters) ACNiTFA. step elution RP Diol cartridge (Waters) ACNIFA, step A C N K F A . step elution elution RP CIScartridge (Waters) ACNfI'FA. gradient RP CIH (Pierce)

I and I1 Weaver el a / . (1994) 4.500 brains. mated females

HCUETH

RP CIx cartridge (Wdters) A C N K F A . step elution

RP C:/Cix (Pharmacia) LC ACNITFA. gradient

I-Propanol/HCI. gradient

LC ACNiTEA ' F A . gradient ACN/NH,OAc.

Ion exchange (TSK SP-5PW) LC LiCllMES buffer p H 5.5. gradient RP C4 microbore (Vydac) LC I-propanoliTFA. gradient

Yield (estimated) Assay CA

-3OpmoUlOOO

0-1h 5th instar larvae. 2 pairs

RP Cs (Jones) LC ACNKFA, gradient

-100 pmol/llM)

IS0 pmo1/1000

RP CIK(Vydac) LC ACNfI'FA, gradient

100 pmol/lOOa

2 day virgin female, 1 C A 2 day virgin female, one pair

RP C I R(Pierce)

gradient

ACNITFA. gradient

LC ACNKFA, gradient 130 pmo1/1000 45 pmul/l(KKI

10 day mated female. one pair

W 3(xI pmol/llKYI 4 day virgin female. one pair

HOAc. acetic acid; NH,OAc, ammonium acetate; ACN, acetonitrile; RP, reverse phase; LC, liquid chromatography; TFA, trifluoroacetic acid; HFBA, heptafluorobutyric acid; MES, 2-(N-morpholino)ethanesulphonicacid; METH, methanol; ETH, ethanol; FA, formic acid; TEA, triethylamine.

% W %D v)

5 z

%

ALLATOSTATINS: IDENTIFICATION, PRIMARY STRUCTURES, AND FUNCTIONS

277

quantity of purified carrier protein to all steps except the last to prevent loss by adsorption to the surfaces of collection tubes. 2.3

BIOASSAY

Success in these isolations was made possible by use of a short-term in vitro radiochemical assay to monitor JH synthesis, the same assay used in most of the preliminary studies with crude extract. In these assays radiolabelled methionine is incorporated into the methyl ester moiety of JH without appreciable dilution (see review by Tobe and Stay, 1985) and the JH removed from the incubation medium with iso-octane (Feyereisen and Tobe, 1981). The incubation fluid was medium 199 fortified with 2% Ficoll (for M. sextu the pH was adjusted to 6.5). The biological activity of fractions of extract was determined by treating glands with aliquots of fractions and comparing their rates of J H synthesis over a 2-3 h period with untreated control glands. Alternatively, the percentage change in JH synthesis was determined by comparing rates of production during an initial 3-h period without extract with rates during a subsequent 3-h incubation in the presence of extract. The latter method depends upon constant rate of synthesis and this varies with the stage of the C A donor (Stay et ul., 1991a). However, for either type of bioassay, the selection of CA in a physiological state sensitive to the active allatostatic factors is important (see Section 4). In the isolation of the cockroach allatostatins, the addition of carrier protein to the tubes used in the test for allatostatic activity was important for recovery of activity, especially as the purification progressed. Pratt et al. (1989, 1991a) and Weaver et ul. (1994) used 0.1% bovine serum albumin with 0.03% of the protease inhibitor bacitracin. This concentration of bovine serum albumin was necessary for the adequate recovery of D. punctutu allatostatin V whereas a 20-fold lesser amount was used in isolation of D. punctutu allatostatins I-IV, VI, VII and none was used in isolation of M . sexta allatostatin.

2.3.1 Radioimmunoassay Using a polyclonal antibody to D. punctata allatostatin I (Yu et al., 1993), Weaver and Freeman (1992) developed a radioimmunoassay with which they detected the amount of allatostatin immunoreactivity in liquid chromatographic fractions of brain extract from male P . americana. Such assays could be used to detect immunoreactive allatostatins in unrelated species for which a suitable bioassay is not available (see Section 8.1). Also, immunoassays are useful to show quantitative changes in the occurrence of allatostatins (see Section 7.3).

278

2.4

BARBARA STAY et a/. SEQUENCE AND C-TERMINAL ANALYSIS

Automated Edman degradation sufficed to determine the amino acid sequences of D. puncfafa allatostatins I-IV, VI, VII (Woodhead er a f . , 1989, 1994) and I (Pratt et a f . , 1989) and the P. americana allatostatins I and 11 (Weaver et a[., 1994). The M . sexta allatostatin was reduced and ['H]carboxymethylated before enzymatic removal of the N-terminal pyroglutamate residue and was then sequenced (Kramer et a f . , 1991). For D. puncrara allatostatin ASB2 (V), sequencing of both the total peptide and a fragment resulting from a trypsin digest provided the sequence, except for a leucine/isoleucine uncertainty that was resolved by RP chromatography of the two forms; equal biological activity was found in both the leucine and isoleucine forms (Pratt et a f . , 1991a). C-terminal analysis was carried out in a variety of ways. The M. sexfu allatostatin was established to be the free acid by RP liquid chromatographic analysis of reduced carboxymethylated forms of tritiated native material and the synthetic carboxyl-terminal acid and amide forms. For the D. punctatu and P. americana allatostatins, with the exception of D. punctara allatostatin VI, the molecular mass of the native peptides was determined by mass spectrometry, either by fast atom bombardment positive mode (Pratt et af., 1989, 1991a; Weaver et a f . , 1994) or tandem mass spectrometry (Woodhead e f af., 1989, 1994). The D. puncrata allatostatin VI was determined to be the amide by chromatographic and bioactivity analysis of synthetic acid and amidated forms of the peptide (Woodhead et a f . , 1994). 2.5

SYNTHETIC ALLATOSTATINS

To confirm the structure of the allatostatins, it is necessary to demonstrate that the sequenced peptide possesses allatostatic activity. Allatostatins were synthesized according to the determined amino acid sequences, purified by liquid chromatography, and amino acid compositions confirmed. Their biological activity as well as chromatographic characteristics were then determined. All were found to have identical characteristics to the native peptides. Caution should be used in interpreting results from comparisons of synthetic and native peptides in the isolation of allatostatins. For example, the M . sexta allatostatin in either the acid (native) or amidated form showed indistinguishable biological activity (Kramer et a f . , 1991), and therefore only chromatographic characteristics were useful in deducing the native form. Similarly, the uncertainty about the leucineiisoleucine in ASB2 (V) could not be resolved by mass determination nor by biological activity, because both forms of the synthetic peptide had the same activity; therefore chromatographic characteristics were used to resolve this uncertainty (Pratt ef af., 1991a).

ALLATOSTATINS: IDENTIFICATION, PRIMARY STRUCTURES, AND FUNCTIONS

279

3 Structure-activity studies of Diproptera punctata allatostatins allatostatins

With the discovery of at least 30 different D. puncrutu allatostatin-like peptides in several species of insect, most showing the characteristic pentapeptide Tyr-Xaa-Phe-Gly-Leu-NH2 carboxyl (C)-terminus, it is now clear that these allatostatins represent a unique family of neuropeptides which probably serve several different functions in insects and related arthropods (see Sections 7 and 8). Although the C-terminal pentapeptide is highly conserved across the species studied to date, there is considerable variation in the N-terminal region of the molecules (the ‘address’ sequence); this variability provides for considerable differences in relative potency (defined here as EDSovalue) of the peptides in a variety of the bioassays, including their effect on JH biosynthesis, on endogenous and on stimulated muscle activity (see Section 7.1). In this section, we will consider the studies on truncated allatostatins and on modified allatostatins (analogues) from the perspective of biological potency, with a view to defining the critical residues conferring activity in the peptides and to providing conformational models of the peptides. Such information is essential for an understanding of ligand-receptor interactions.

3.1

THE ALLATOSTATIN ‘MESSAGE’ SEQUENCE

The high degree of conservation of the C-terminal pentapeptide implies an important physiological function. The pentapeptide, with Ser2, is a truncated version of allatostatin IV (dipstatin 5 according to position on the gene; see Section 9) and itself possesses considerable potency with respect to inhibition of JH release, although much less than the full peptide (EDSo= 3.2 x lo-‘ M vs 1.6 x lo-’ M (Stay er al., 1991b). Addition of the N-terminal Leu‘ increases biological potency by a further order of magnitude (EDSo= 5 x lo-’ M) (Stay et al., 1991b). It is noteworthy that the magnitude of the maximal response with either of these truncated peptides is similar to that of the authentic peptide, allatostatin IV (defined here as eficucy-maximal inhibition at apparent saturation). These data demonstrate: ( 1 ) that the conserved region of the peptide is critical for potency (i.e. signal transmission to target tissues); (2) that this portion of the molecule is likely to interact directly and be recognized by the receptor(s) for allatostatins; and (3) that this interaction is modified by the address sequences. The conclusion that the five residues of the C-terminus of allatostatin probably represent the core ‘message’ portion of the peptides, responsible for direct receptor interaction (Hayes et al., in press) is based on the following evidence: (1) the sequence Tyr/(Phe)-Xaa-Phe-Gly-Leu is highly conserved in all species studied to date (see Sections 8.1, 9); (2) deamidation of the C-terminal Leu abolishes activity with respect to

BARBARA STAY et al.

280

inhibition of J H biosynthesis, at least at 10 pM or less (Pratt et al., 1989, 1991b; Duve et al., 1993); (3) modification of the C-terminus, either through truncation (of -Leu or -Gly-Leu) or extension (addition of -Ala or -Ala-NH2), abolishes activity (Pratt et al., 1991a,b); (4) substitution of the Phe with Gly or of Leu with Ala abolishes potency (Pratt et al., 1991b); and ( 5 ) N-terminally truncated peptides retain efficacy (i.e. maximal magnitude of response at saturation), albeit at high doses, provided that the minima! pentapeptide sequence is employed (Stay et al., 1991b). 3.2

THE ALLATOSTATIN ‘ADDRESS’ SEQUENCE

The largest of the allatostatin peptides isolated to date is the octadecapeptide allatostatin V (dipstatin 2) (ASB2 of Pratt et al., 1991a) which is the most potent of the allatostatin family, with an EDSoof about 10-100 pM. This peptide possesses an unusual tyrosine-rich address sequence and was designated ASB2 by Pratt er al. (1991a) on the basis of its presumed differences in activity on C A at different developmental stages, physiological actions, potency and hydrophobicity from the four previously described allatostatins (Woodhead et al., 1989), as well as its presumed interaction with a different receptor subtype (Pratt et al., 1991b) (see Section 5 ) . However, based on the knowledge of the D.punctata allatostatin gene (see Section 9.1) and the 13 possible cleavage products (dipstatins), the rationale for the inclusion of allatostatin V (ASB2) in a different subfamily of peptides is unclear, since it is probably produced and cleaved at the same time as the other allatostatins, is only slightly more potent than other selected allatostatins and has not been demonstrated to require a different receptor subtype. Hence, it is premature to assign it to a different subfamily. ED,o values for several of the D. punctata allatostatins (Table 3 ) show that length of the peptide does not correlate well with biological p o t e n c y 4 n e of the smaller of the allatostatins, allatostatin IV, the octapeptide, shows potency only one order of magnitude less than that of allatostatin V. Prompted in part by the presence of a potential dibasic cleavage site within the octadecapeptide (dipstatin 2), Pratt et al. (1991a) have examined the potencies of several possible cleavage products. These authors were particularly concerned with demonstrating that the octadecapeptide, and not the cleavage products allatostatin V-( 1-9/10) and allatostatin V-( 11-18), was the natural peptide hormone. Several potential C-terminal cleavage products, including allatostatin V-(9-18) decapeptide amide, allatostatin V-( 1018) nonapeptide amide and allatostatin V-( 11-18) octapeptide amide all showed PICSO values (-log of concentration causing 1/2 maximal inhibition) in the 10-100 nM range, or about 100-200-fold less biological potency than authentic allatostatin V. Nonetheless, these products show appreciable potency. Na-acetylation of these potential cleavage products increased their activities/pIC50 values (by less than one order of magnitude). Pratt et al.

ALLATOSTATINS : IDENTIFICATION, PRIMARY STRUCTURES, AND FUNCTIONS

281

TABLE 3 Potencies of Diplopteru punctutu allatostatins for inhibition of juvenile

hormone biosynthesis Designation Allatostatin V IV I VI I1

I11 VII

Residues

Dipstatin

(n)

ED50 (MY'

2 5 7 6 11 9 8 4

18 8 13 8 11 10 9 9

1.0 x 1.6 X 4.1 X 2.3 X 3.2 X 7.2 X 9.4 x 2.0 x

lo-" lo-'' lo-''

lo-' lo-' lo-' lo-'

Order of potency 1 2 3 4

5 6 7 8

"Molar concentration of peptide required for 50% inhibition of JH release in a 3 h in v i m radiochemical assay with pairs of 2-day virgin CA compared to groups of controls ( n = 7). Peptide concentration was determined by amino acid analysis (S. S. Tobe, J .Zhang and W. G. Bendena, unpublished).

(1991a) suggest that this result is consistent with absence of such peptides in vivo, but is based on the assumption that if these peptides possessed specific receptors, N*-acetylation would reduce rather than enhance their potency. Much of the structure-activity work has focused on the larger peptides, allatostatin V (octadecapeptide) and allatostatin I (tridecapeptide). However, as noted above, the octapeptide-allatostatin IV-possesses levels of potency only slightly lower than the larger peptides and this suggests that the role of larger address sequences in defining potency may be of secondary importance. Alternatively, the different sized address sequences may interact with different types of receptor. Nonetheless, other octapeptides or nonapeptides with the allatostatin message sequence do show differing abilities to inhibit J H biosynthesis (Table k o m p a r e dipstatins 5 and 6: see also allatostatin VII (dipstatin 4) Woodhead et al., 1994; S. S. Tobe, J. Zhang and W. G. Bendena, unpublished) and hence, even a short address sequence can profoundly influence the potencies of the peptides. The address sequences of the allatostatins can range from 9-13 N-terminal residues in the case of allatostatin V, to 2-3 in the case of allatostatin IV. At present, it is uncertain if the C-terminal 6 residue (which in many of the allatostatins is Leu, the conservative substitution Val or Met) is part of the message sequence, but its presence does enhance potency (Stay et af., 1991b), suggesting that it is part of the address sequence. However, dipstatin 1 (Table 1, Fig. 9) is a hexapeptide of precisely this sequence (Donly et af., 1993) and its occurrence raises the question of potency and functional significance of such a core message peptide. Substitution of L-Arg' with D-Arg' in the octapeptide allatostatin IV resulted in only a slight

282

BARBARA STAY et al.

loss in potency (Hayes et al.. in press), whereas Pratt el al. (1991b) reported a 15-fold decrease in potency for the same substitution in allatostatin I. These observations highlight the importance of structure-activity studies on each of the 13 allatostatins for understanding the functions of the address sequences and receptor-ligand interactions. Coadministration of the allatostatin V-( 1-9) address sequence with selected allatostatins (allatostatins V, I or allatostatin V-( 11-18)) did not enhance potency or efficacy and hence, it is unlikely that the address region alone of this allatostatin possesses any direct, independent effect on JH biosynthesis (Pratt et al., 1991a). All of the D. puncfata allatostatin peptides (dipstatins 1-13) are capable of inhibition (efficacy), since they all contain the appropriate message sequence, but differ in potency (S. S. Tobe, J. Zhang and W. G. Bendena, unpublished). The diversity in the structure of the address regions probably accounts for this difference and may represent differences in degree of binding to the allatostatin receptor subtypes, all of which recognize the message sequence. 3.3

ALLATOSTATIN ANALOGUES

3.3.1 Analogues of allatostatin IV A study of the octapeptide allatostatin IV (dipstatin 5 ) has provided direct information on the importance of critical residues and their side chains in the message sequence (Hayes et al., in press). This peptide possesses relatively high biological potency but its address sequence is likely to be short (2-3 residues). The effect of single residue substitutions with either L-alanine or D-amino acids on inhibition of JH biosynthesis has been determined and the results are summarized in Fig. 1A and B respectively. Replacement with alanine for the study of side-chain importance was chosen because it has only a methyl group as a side chain and hence substitution would remove most of the chemical features of the replaced residue. Replacement of Tyr4, Phe6, Gly’ or LeuX with L-alanine all reduced the biological potency of the analogues significantly whereas replacements in the address region had a much reduced effect and decreased potency by an order of magnitude or less, except for the Asp’ substitution. This latter substitution of a hydrophobic residue for a charged residue at the Nterminus appears to be intermediate in effect and could be expected to cause appreciable loss in potency as a result of changes in the charge of the molecule. Substitution with the hydrophobic alanine in the message sequence for other hydrophobic residues dramatically reduced potency whereas substitution at the polar Sers position had little effect. It is also significant that this position (4 from C-terminal) is the only consistently variable residue in the message sequence of the allatostatins and occurrence of either polar (Asn in dipstatin 2) or hydrophobic residues (Gly in dipstatins 3 and 7

ALLATOSTATINS: IDENTIFICATION, PRIMARY STRUCTURES, AND FUNCTIONS

283

Dip 5 Leu

G lY Phe Ser

TY r Leu Arg

ASP mino acid Substitution

Dip 5

Leu GlY Phe Ser

TYr

Leu Arg ASP

10-'0

10-9

lo-'

IO-~

ED50 (Molar) FIG. 1 Potency (EDso) of single residue replacement analogues of allatostatin IV (dipstatin 5 ) . The EDSo (molar concentration required for 50% inhibition of JH release by D. puncrata CA) for each analogue is plotted at the sequence location of the replacement. Also shown is the EDso for allatostatin IV (dipstatin 5 ) . (A) Ala replacement. (B) D-amino acid replacing the native L-amino acid. From Hayes et at. (in press).

284

BARBARA STAY et al.

or Ala in dipstatin 9) does not appear to alter potency. The significant loss in potency following substitution at the GIy7 position in allatostatin IV is also predictable since this replacement would add bulk to this position (Gly has no side chain).

3.3.2 Conformational models of allatostatin IV Substitution of single residues of allatostatin IV with D-amino acids also resulted in significant loss of biological potency, particularly for substitutions in the C-terminal pentapeptide message sequence (see Fig. 1B) (Hayes et a f . , in press). Interestingly, sequential substitutions from the N-terminus resulted in sequential loss of potency, up to the C-terminal tripeptide Phe-Gly-Leu. These studies have provided insight into the regions of the peptide for which preferred conformations at the receptor are important, because substitution of D-amino acids will distort the structure of the peptide, through reversal of symmetry of either the backbone or the side chain, thus permitting the assessment of which residues are necessary for receptor interaction. D-replacements also provide additional information on the importance of side chains. On the basis of these data, Hayes et al. (in press) have suggested that both a and p elements are important structural features of the allatostatin IV molecule; it appears that an a-helical structure best describes the N-terminal address region (charged and polar) and a p-strand the C-terminal pentapeptide region (hydrophobic). The acceptance of ~ - A l arather ~ than L-Ala for Gly7 (compare Fig. 1A and lB), strongly suggests the occurrence of a Type I1 p-turn in the Phe6-Leu8 region (Hayes et al., in press). These authors concluded that Tyr4, Phe' and Leu8 were the three most important residues for both potency and efficacy and that the conformation of the C-terminal tripeptide is likely to be a Type I1 p-turn centred around the Phe6-Leu8 position (see Models, Fig. 2). The question of hydrophilicity of the various allatostatins as a basis for division of the family into two groups remains to be resolved. Certainly, the significance of the hydrophobicity of allatostatin V, particularly in the address sequence, cannot be resolved until it has been compared with hydrophobic address regions in other allatostatins. This will require the synthesis of appropriate analogues for all remaining 12 allatostatins. The implication from Pratt et al. (1991a) that the hydrophobic domain at Leu"-Pro12 in allatostatin V differs sufficiently from the hydrophilic Gln6-Arg7 of allatostatin I or Asp'-Arg' of allatostatin IV to dictate interaction with a different subclass of receptor thus remains to be tested and ultimately awaits the isolation and cloning of allatostatin receptors, since there is no evidence that the hydrophilicity at Leu' '-Pro12 is the principal active domain of the address sequence, recognized specifically by the receptor.

ALLATOSTATINS: IDENTIFICATION, PRIMARY STRUCTURES, AND FUNCTIONS

285

A.

B.

C.

FIG. 2 Schematic representation of a structure-activity model for allatostatin IV (dipstatin 5 ) . (A) Outlined residues have critical side chains for efficacy with respect to inhibition of J H biosynthesis and the relative size of the characters depicts relative importance of the side chain to activity. ( B , C) Stereo molecular models showin potential orientation of side chains in a type I1 p-turn centred around Phe6-GlyB. Only the C-terminal pentapeptide is shown and the three dots represent the continuation of the N-terminal sequence. (B) Torsional angles are for a type I1 p-turn and a p-strand. (C) Torsional angles for Tyr4 and Sers are for an a-helix and remainder of angles are as in (B). From Hayes er af. (in press).

BARBARA STAY et al.

286

3.4

COMPARISON OF ASSAYS FOR INHIBITION OF J H BIOSYNTHESIS

Examination of data from different laboratories assessing the relative potency of the various allatostatins and analogues shows some variability (compare, for example, data reported in Pratt et a[., 1990, 1991a,b; Stay et al., 1991a,b; present paper). This variation reflects not only slightly different modes of preparation of the peptides in solution and the use of animals of different age or reproductive state, but also slightly different assay methods (sequential assays versus parallel experimental and control assays), different vessels for incubation and different methods of assessing final concentration of peptides. i t is clear that the message portion of the molecule is active in the micromolar range and is fully capable of effecting maximal response in any of the relevant laboratories. Additionally, allatostatin V appears to be the most potent of t h e allatostatins. Thus, although there may be some variation in relative EDSovalues or pICso values, rank order of potency, in terms of inhibition of JH biosynthesis, is similar.

4

4.1

Sensitivity of corpora allata t o allatostatins R E D U N D A N C Y IN ALLATOSTATINS

With the elucidation of the many members of the D. punctata allatostatin family, the question of the functional significance of such a range of peptides arises. On what evolutionary basis would insects employ such an array of peptides, if the single or primary function is the inhibition of JH biosynthesis? This is particularly relevant if this family has evolved relatively recently. The high degree of conservation of the pentapeptide message segment and the high degree of homology of the allatostatin gene in the cockroaches studied (see Sections 9.1, 9.2) suggests that this family has either appeared recently or has been highly conserved. In either case, the preproallatostatin contains the full complement of peptides for that particular species and all the potential cleavage products display differences in potency but similar efficacy (S. S. Tobe, J . Zhang and W. G. Bendena, unpublished). Nonetheless, the problem of redundancy of peptides to regulate a single physiological function remains to be resolved, and although it is now clear that the allatostatins are capable of exerting other physiological effects, including modulation of spontaneous and proctolininduced muscle activity (Lange et af., 1993; Hertel and Penzlin, 1992; A. B. Lange, S. S. Tobe and W. G. Bendena, unpublished) (see Section 7.1), the isolation and identification of at least five different allatostatins from the corpora allata (Stay et al., 1991b; Stay and Woodhead, 1993) and the presence of the preproallatostatin in neurosecretory cells of the brain which probably innervate the CA (Fig. 10B), indicate that receptors on the cells of

HLLH I

u3 I A I INS: IDENTIFICATION, PRIMARY STRUCTURES, AND FUNCTIONS

287

these glands are likely to be exposed to the full complement of peptides. One possible explanation for this redundancy may reside in developmental differences in responsiveness of corpora allata to the range of allatostatins. Pratt et al. (1991a) differentiated allatostatin V from the other four allatostatins known at the time in part on the basis of qualitative differences in inhibition of JH biosynthesis in vitro. This conclusion was based only on apparent differences in responsiveness of Day 2 mated females to allatostatins I and V and it thus remains uncertain that such qualitative differences exist. A comparison of data for allatostatin I (10 pM) from Pratt et al. (1990) and for allatostatin V (10 nM) from Pratt et al. (1991a) demonstrates that although the dose of allatostatin employed differed by three orders of magnitude, the developmental profile of responsiveness of CA was virtually identical, suggesting significant quantitative differences but no qualitative difference. However, it is clear that the corpora allata of adult females differ appreciably in their responsiveness to selected allatostatins (Pratt et al., 1990, 1991a,b) and that developmental stage conditions this response (see Section 4.2). 4.2

DEVELOPMENTAL CHANGES OF CORPORA ALLATA IN RESPONSE TO

ALLATOSTATINS

4.2.1 Changes in adult and larval CA of D. punctata Corpora allata of adult female D. punctata respond to allatostatins in an age-dependent fashion. Studies have focused on developmental changes in responsiveness to allatostatin V (Pratt et al., 1991a) and allatostatin I (Pratt et al., 1990, 1991b; Stay er a f . , 1991a) and to a lesser extent, on allatostatin IV (Stay et al., 1991a). Day 6 mated females are extremely sensitive to these peptides, and Day 10 mated females as well as second reproductive cycle females, 3-4 days after parturition, show similar but slightly reduced sensitivity (Pratt et al., 1990, 1991a: Stay et al., 1991a). Young adult mated females, aged 2-5 days, and mid-pregnancy females show the lowest sensitivity to allatostatins (Stay et al., 1991a). In general, the responsiveness of CA from various aged females (mated unless designated virgin) to allatostatin I show the following rank order: Day 6 > D a y lO>Day 15 > Day 2 virgins > Day 39 > Day 2 >Day 4 = Day 5 (Pratt et al., 1990; Stay et al., 1991a; Woodhead et al., 1989). The sensitivity of CA of various ages similarly shows a broad range of effective dose for 50% inhibition of JH biosynthesis (ED,,)), from -0.1 nM for Day 6 mated females to >1 p M for Day 4 and 5 mated females. This latter change occurred over a 2-day interval (Stay et al., 1991a) and over only 1 day (Pratt et al., 1999) which is clearly rapid and remarkable; if this response is receptor-mediated, a very rapid turnover, upregulation/downregulation or unmasking of receptors should be expected.

288

BARBARA STAY et at.

Developmental changes in responsiveness of larval CA to allatostatins show a pattern distinct from that of adult female CA (Stay et al., 1991a). CA from both penultimate and final instar larvae early in the stadium also show very high sensitivity to allatostatin I (-80% inhibition at 10nM) but this sensitivity declines as the stadium progresses. In the penultimate instar the sensitivity increases toward the end of the stadium, whereas in the final stadium the glands have lost sensitivity to allatostatins at the same time that J H biosynthesis has declined. This loss of sensitivity in final instars suggests that allatostatins are not primary regulators at this time. Larval CA show a similar degree of sensitivity to allatostatins 11, I11 and IV (Stay et al., 1991a). However, it is clear that the importance of allatostatic inhibition of JH biosynthesis differs profoundly at different developmental stages. Our understanding of the changing sensitivity of corpora allata to the allatostatins remains rudimentary and will require full assessment of all of the known allatostatins for any given species.

4.2.2 Activity of C A as a function of sensitivity to allatostatins in D. punctata The developmental changes in sensitivity of CA to allatostatins have been proposed to be inversely related to the rates of JH biosynthesis during the first reproductive cycle, with the most biosynthetically active CA (Day 4-5 adult mated females) showing the lowest degree of inhibition by allatostatins (Pratt et al., 1990, 1991a; Stay et al., 1991a). Although this generalization may be applicable to females during the first and second vitellogenic cycles and in the first half of the penultimate stadium. it clearly is not applicable to either pregnant females or to Iate penultimate or final instars (Stay et a/., 1991a), in which responsiveness decreases as rates of JH biosynthesis decline. At these times, JH biosynthesis remains at low or undetectable leveis for extended periods, suggesting that alternative tonic mechanisms may be operative. This interval, particularly in last instar animals, represents a time of developmental reorganization and transition in the CA, and the glands appear to be incapable of producing even small amounts of JH, and cannot be rescued by treatment with the penultimate precursor farnesoic acid (Yagi et a/., 1991). It is known as well that J H biosynthesis can be inhibited in pregnant females by a humoral route, so long as the brain is present (Rankin and Stay, 1985). Although immunoreactive allatostatin has been found in the haemolymph of last instar animals particularly during the allatostatin-unresponsive period, as well as during pregnancy (Yu et a [ . , 1993) and bioactive allatostatins were demonstrated after HPLC separation in haemolymph of adult females (Woodhead et al., 1993). haemolymph allatostatins would not appear to be directly functional in the inhibition of JH biosynthesis. Rather, they may act to facilitate the inhibition of JH

ALLATOSTATINS: IDENTIFICATION, PRIMARY STRUCTURES, AND FUNCTIONS

289

biosynthesis, ensuring that JH biosynthesis remains at low levels and minimizing carbon flow early in the biosynthetic pathway. 4.2.3 Changes in adult CA of P. americana Interspecies effects of allatostatins have received limited study. For example, Woodhead er al. (1989) tested the effects of D. punctufa allatostatin I (dipstatin 7) on JH production by CA of immature female P. americana and observed a 92% inhibition at 10nM. Further studies by Weaver (1991) determined the developmental changes in the sensitivity of P. umericanu CA to dipstatin 7. During times of increasing JH biosynthesis (including the onset of vitellogenesis in both basal and penultimate oocytes), sensitivity to dipstatin 7 (at 10 pM) increases dramatically and reaches a maximum (inhibition of 97%) at the time corresponding to the onset of vitellogenesis in basal oocytes. The timing is thus quite distinct from that reported in female D. punctata, in which maximal sensitivity occurs at the conclusion of vitellogenesis (Pratt ef a f . , 1990; Stay et a f . , 1991a). Similarly, in D. puncfuta, sensitivity to dipstatin 7 remains stable or declines in early adult life, reaching a minimum at the peak of vitellogenesis, whereas in P. americana sensitivity increases until the onset of vitellogenesis. Weaver (1991) conjectures that this may be a function of the different modes of reproduction of the two cockroach species (viviparity vs oviparity) (see also Tobe, 1980). However until all allatostatins of the respective species have been assessed for activity at physiological doses, it is premature to attribute such differences to different reproductive modes. The homologous allatostatin I tridecapeptides in D. punctata (dipstatin 7) and P. americana (peastatin 7) differ in two residues in the address region (see Sections 2 and 9). The effect of the peptide on the same species is two to three orders of magnitude greater than its effect on the reciprocal species, as shown in Table 4 (Weaver, 1991; Weaver et al., 1994; Stay and Woodhead, 1993). On the basis of changes in activity as a result of substitutions in the address region for dipstatins 2, 5 and 7 (see Section 3.2) tested on D. punctata CA, the difference in response of P. americana CA to their own and D. punctata allatostatin is not surprising. Although potency is reduced, the efficacy of dipstatin 7 in P. americana CA is high, confirming the importance of the message sequence in related species. 4.3

DUALITY OF RESPONSES TO ALLATOSTATINS

The responsiveness of glands of specific ages to allatostatins appears to be a complex physiological phenomenon, as for example in the apparent duality of responses (pM-nM range) to allatostatin V (Pratt el a f . , 1990, 1991b). Such duality manifests itself in flattening or plateauing of dose-response

BARBARA STAY et al.

290

TABLE 4

Interspecific effects of allatostatin I on inhibition of JH synthesis Allatostatin I (M)"

Corpora allata tested

P. americana (peastatin 7)

D. punctata (dipstatin 7)

Diploptera punctatu ( 2 day virgin)

1.2 x 1OPh

2.0 x lo-"'

Periplanetu umericana (3 day virgin)

6.2 x 10-""

6.9 x 10-'"

"Concentration required to give SO% inhibition of JH synthesis "Weaver et al. (1904). 'Stay and Woodhead (1993). W e a v e r (1991).

curves, particularly at low concentrations of peptide, and can be resolved through statistical analysis (see Loftfield and Eigner, 1969). It is likely that such duality represents interaction of a single ligand with multiple allatostatin receptors (see Section 5.2), although we cannot exclude the possibility that modification or degradation of the peptide by peptidases associated with the corpora allata occurred during the course of the assays, resulting in multiple ligands in the incubation medium. The likelihood that interaction with multiple allatostatin receptors provides for the additivity of the dual responses requires further study since it has been shown that treatment of corpora allata with more than one type of allatostatin simultaneously gives similar flattened dose-response curves from CA of specific age (S. S. Tobe, unpublished) but is nonetheless consistent with this hypothesis. Such a possibility was considered feasible by Pratt et al. (1990) and was invoked to explain the flattened dose responses realized with brain extracts. This duality in response appears to be both stage- and allatostatin-specific, since coadministration of allatostatins I and V at 10 nM to CA of Day 2 mated females did not result in additive inhibition (Pratt et al., 1991a). Such so-called 'cross-reactivity' (Pratt et al., 1990) with other allatostatin receptors at high concentrations of mixtures of allatostatins demands the presence of multiple allatostatin receptors; the resolution of this important question will require the isolation/cloning of the allatostatin receptors. 4.4

RESPONSIVENESS T O ANALOGUES OF ALLATOSTATINS

Studies on responsiveness of CA to various analogues and truncated allatostatins have in general been performed only on CA from females of a limited range of age (Days 2, 6 and 10 mated females: Pratt et al., 1991a,b; Day 2 virgin females: Stay et al., 1991b). Because there are large changes in

ALLATOSTATINS: IDENTIFICATION, PRIMARY STRUCTURES, AND FUNCTIONS

291

responsiveness to the allatostatins themselves, it is likely that similar differences exist with respect to the truncated and modified allatostatins. If such changes in responsiveness are the result, in part, of changes in receptor number or subtype, it would appear prudent to examine the effects of the analogues on CA of different developmental and physiological states.

4.5

POSSIBLE FACTORS CONTRIBUTING TO CHANGES IN RESPONSIVENESS

The dramatic developmental changes in sensitivity of CA and responsiveness to allatostatins I and V in vitro (Pratt et al., 1990, 1991a,b) could be attributed to several factors, including: (1) changes in receptor subtype; (2) changes in number and turnover of receptors; upregulation and downregulation of receptors; ( 3 ) changes in ‘cross reactivity’ of receptors to allatostatins; (4) changes in ability of CA to degrade allatostatins in vitro; ( 5 ) changes in the basal lamina surrounding the CA and ability of allatostatins to penetrate it either in vivo or in vitro; (6) changes in the quantity of allatostatins within the CA and their release during incubation in vitro. Although all of these factors no doubt influence responsiveness of CA, it is likely that the primary regulator is modulation of receptor quantity and subtype.

4.6

NEURAL A N D H U M O R A L PATHWAYS FOR ALLATOSTATIN ACTION

The dynamic range of response of CA to allatostatin treatment generally ranges between 0% and 80% inhibition of J H biosynthesis (Pratt et al., 1990; Stay et al., lYYla,b). However, Pratt et al. (1990) show an inhibition of 96% at p M concentrations for CA from Day 6 mated females. Although precise titre values for individual allatostatins in the haemolymph or within the CA are not available, the immunoreactivity data of Yu et al. (1993) suggest that in the haemolymph, the concentrations of allatostatins are not likely to be greater than in the subnanomolar to low nanomolar range. This estimate of allatostatin concentration is conservative in part because the antibody to allatostatin I used to measure it has limited cross-reactivity to other allatostatins. Nevertheless, such concentrations of individual allatostatins provide inhibitions of 50-70% only at the most sensitive stages (i.e. Day 6 mated females) (Pratt et al., 1990, 1991a; Stay et a l . , 1991a) and considerably less inhibition of CA from females of other, less sensitive stages. Interestingly, the concentration of allatostatin I immunoreactive material in haemolymph of Day 6 mated females is at its lowest point at this age (-0.1 nM) and this concentration provides 4 0 % inhibition, even in CA from Day 6 females. Release of allatostatins within and from the CA is likely to create a localized elevated concentration of the peptides. It is known that at least some allatostatins are released from CA during incubation in v i m (Yu et

BARBARA STAY et al.

292

a / . , 1993) and this probably is a manifestation of release of these peptides

within the glands and its ‘leakage’ into the surrounding medium. The allatostatin concentration within the extracellular spaces of a pair of CA can be estimated by assuming that the quantity of allatostatins within the CA is -200 fmol per pair (based on allatostatin I immunoreactivity; Yu et al., 1993); that only about 5 1 0 % of that is released following nerve depolarization with high potassium (Yu et a / . , 1993); that the volume of the extracellular space in a pair of fresh CA is 2 n l (based on measurement of fixed CA (Johnson et al., 1993) and the fact that fixed CA volume, -20111 per pair, is half that of fresh (Szibbo and Tobe, 1981)). Although the actual concentration of the peptides in the extracellular spaces cannot be determined with accuracy, localized concentration might be as high as 10pM, sufficient to cause marked inhibition of JH biosynthesis in sensitive glands. Although some allatostatins are released into and from the CA, it cannot be assumed that all allatostatins from the prohormone are cleaved and released. Processing of the precursor may result in selective packaging of the products or in selective signalling for release from appropriate terminals, be they neurohaemal or within the CA. The acidic domains of the prohormone (Donly et a/., 1993) may be important to the processing of the precursor, to the protection of the peptides from proteolytic degradation and to the targeting of release sites. At present, it can only be stated with certainty that the first five isolated allatostatins (Woodhead et a/., 1989; Pratt et a/. , 1991a) are present within the CA (Stay et al., 1991b; Stay and Woodhead, 1993). 4.7

ALLATOSTATINS AND REGULATION OF J H TITRE

It remains to be determined if allatostatins are the principal regulators of JH biosynthesis and hence of JH titre (Tobe et al., 1985). In D. punctata at physiological haemolymph concentrations in vivo, allatostatins probably cannot inhibit JH biosynthesis to a level sufficient to reduce JH titre significantly at all developmental stages. Even a 50% inhibition in JH biosynthesis, a level unlikely to be achieved at physiological concentrations of allatostatins, would provide only a small reduction in J H titre, based on data relating JH biosynthesis to JH titre (Tobe et al., 1985) (in part because the slope of the line is <1 and the Y-intercept of the plot is >0). Indeed, injection of micromolar quantities of allatostatins into haemolymph of adult female D. punctata only reduced rates of JH biosynthesis marginally and oocyte growth slightly but significantly (Woodhead et a/. , 1993); similarly in P. americana, injection of allatostatin reduced J H titre only at some stages (Weaver et a/., 1994) (see Section 7.3). Associated with this question of whether JH biosynthesis in vivo is normally restrained by allatostatins is whether inhibition of JH biosynthesis in vitro truly represents any situation in vivo, although experimental

ALLATOSTATINS: IDENTIFICATION, PRIMARY STRUCTURES, AND FUNCTIONS

293

evidence suggests that it does (see Section 11). Clearly, the degree of inhibition of J H biosynthesis necessary to effect a significant change in JH titre requires further study and should prove an important area of research of practical significance. 5 Receptors for Diploptera punctata allatostatins

The occurrence of multiple allatostatin peptides has prompted the search for receptors of these important peptides. To date, only allatostatin-binding proteins (putative receptor) for D. punctata allatostatin I have been tentatively identified (Cusson et al., 1991, 1992) and further characterization of this molecular species has proven difficult, in part because of the lack of a reliable binding assay. The search for allatostatin receptors has also been complicated by the small amounts of tissues available (the maximal volume of a typical D. punctata CA is -20 nl and its weight is <20 pg), assuming that the C A are the primary target tissue for these peptides. However, allatostatins occur as well in brain and have also been reported in midgut tissue (Tobe et ul., 1994; B. Stay, C. G. Yu, W. G. Bendena and S. S. Tobe, unpublished; G. C. Unnithan, C. E. Summers, N. T . Davis and R. Feyereisen, unpublished) (see Section 6), and although it is uncertain if these tissues also represent targets of action, the ability of the allatostatins to modulate both myotropic activity (Lange et al., 1993) and neural activity (see Sections 7.1 and 7.2) suggests that these tissues must also possess receptors. Nonetheless, the receptors of the C A may differ from those of other tissues and it may be necessary to employ molecular expression cloning systems to ultimately isolate and characterize these molecules. In the search for allatostatin receptors, it is important that reliable binding assays be developed, using appropriate ligands. To this end, a highly specific radioactivity ligand with high affinity will be required. In parallel with such studies should be developmental profiles on allatostatin binding and receptors for the various target tissues since there is some indication that the number of receptors for allatostatin I does increase in CA of Day 6-7 mated females relative to Day 4 female CA (Cusson et al., 1992). As has been noted in Section 4, CA of Day 6 mated D. punctata are extremely sensitive to allatostatins, suggesting that CA of this age possess an increased number of physiologically relevant receptors. This dramatic shift in sensitivity, of several orders of magnitude, over a period of less than 24h (Pratt et al., 1990), might indicate that CA of this age similarly possess significantly more active receptors. In addition, the changes in sensitivity of CA to various allatostatins that have been documented (Pratt et al., 1990, 1991a; Stay et al., 1991a) may be largely attributable to changes in the quantity or activity of receptor subtypes.

BARBARA STAY et al.

294

5.1

APPROACHES FOR THE ISOLATION A N D CHARACTERIZATION OF ALLATOSTATIN RECEPTORS

The use of photoaffinity labels for the demonstration of specific binding proteins in the solubilized pellets of C A membranes has proven successful in revealing the presence of two allatostatin I-specific proteins of molecular size 59 and 39 kDa (Cusson et al., 1991, 1992). This approach employed the radioiodinated azidosalicylamide derivative of allatostatin I and the covalent linkage of this ligand to specific solubilized allatostatin-binding proteins by photolysis with UV light, in the presence or absence of specific competitors (allatostatin I or 111). Subsequent SDS polyacrylamide electrophoresis and autoradiography of the solubilized pellet containing the ligand-binding protein complex revealed the two proteins; binding of the ligand to these proteins showed competitive displacement since binding was reduced with increasing concentrations of authentic allatostatin I or 111. Interestingly, competitive displacement was also observed in the solubilized pellets of both brain and fat body, suggesting that these tissues also possessed allatostatinspecific binding proteins, although these proteins differed slightly in molecular size from those of the CA (38, 41 and 60 kDa for fat body and 41 kDa for brain). Other tissues, including ventral nerve cord, antennal muscle and antennal pulsatile organ muscle, did not show specific binding with the ligand. Since allatostatin-containing nerves are absent from antennal muscles this result was expected. However antennal pulsatile organ muscle does possess allatostatin-containing nerve endings (Lange et al., 1993), as does brain (Stay et af.,1992a) and ventral nerve cord (Lange et al., 1993) (see also Section 6.2). Thus allatostatin-specific binding proteins were expected in the brain (and may eventually be found in ventral nerve cord and pulsatile organ muscle) but the presence of putative receptors in the fat body is more puzzling, since neither immunocytochemistry, ELISA nor radioimmunoassay has revealed allatostatin-like peptides in this tissue. Nonetheless, the occurrence of allatostatin-binding proteins in fat body is consistent with the presence of immunoreactive and bioactive allatostatin in the haemolymph which exerts some undetermined effect via a humoral route (Yu et al., 1993; Woodhead et al., 1993; see Section 7.3). The photoaffinity ligand employed by Cusson et al. (1991, 1992) appears to display slightly reduced affinity since it is about one order of magnitude less potent than natural allatostatins, in terms of ability to inhibit JH biosynthesis. This is not surprising since the modification of the N-terminus and hence the associated address sequence is likely to affect binding to any putative receptors. Unfortunately, such ligands are not likely to permit the direct isolation of the receptors, because of both low affinity and low specific activity. Ligands with affinities equal to or greater than the natural allatostatins will ultimately be required for unequivocal isolation and characterization of allatostatin receptors.

ALLATOSTATINS: IDENTIFICATION, PRIMARY STRUCTURES, AND FUNCTIONS

295

The availability of relatively small amounts of tissue has complicated the quest for allatostatin receptors. For this reason, molecular approaches, including expression cloning, assuming the availability of high affinity ligands, signal sequence trapping (Tashiro er al., 1993) or cloning through homology with other G-protein-linked receptors, may prove more fruitful. Cusson et al. (1991, 1992) also reported that high amounts of competitor (either allatostatin I or 111) were required to realize a significant decrease in binding (up to 2 0 0 0 ~ ) In . only a few cases were these authors able to totally abolish binding of the photoaffinity analogue with high concentrations of authentic peptide competitor. These data suggest that either the photoaffinity ligand has a very high affinity for the receptor and is difficult to displace, that the authentic peptide competitors have a much higher affinity than the photoaffinity analogue, or, in the case of allatostatin 111 competition, that two different receptor subtypes are involved. 5.2

THE CASE FOR MULTIPLE ALLATOSTATIN RECEPTORS

The occurrence of multiple allatostatin species might a priori suggest that there must be individual receptors for each species of molecule. The evolutionary significance of such a diverse group of peptides with a common pentapeptide C-terminus may reside in a different function for each species, modified in particular by the N-terminal address sequences. The case for multiple allatostatin receptors is further strengthened by the following observations with respect to D. puncrara. 1. There are clear quantitative differences in the responsiveness of C A to those allatostatins studied to date and these differences are agedependent (Pratt et al., 1990, 1991b; Stay et al., 1991a). 2. Photoaffinity labelling has demonstrated two membrane proteins in membrane pellets of CA capable of binding the allatostatin I photoaffinity analogue, showing distinct differences in relative binding and competition (Cusson et al., 1991, 1992). 3 . There are distinct quantitative differences in labelling and competition with allatostatin 111 relative to allatostatin I in photoaffinity experiments (Cusson et al., 1991, 1992), with allatostatin I being the preferred competitor. 4. At micromolar concentrations, allatostatins cause strong inhibition of C A of many ages, even in many that are normally unresponsive at lower concentrations (Day 4-5 mated, mid-pregnancy and early second cycle females excepted), suggesting that at such pharmacological doses, allatostatins interact with multiple allatostatin receptors resulting in quasi-additive inhibition. 5 . The duality of responses at low and high doses of allatostatins and its resolution into two quasi-additive components (Pratt et a[., 1991b)

BARBARA STAY et al.

296

suggests two different receptor components, showing two different affinities. 6. Treatment of CA with all 13 allatostatins at once results in quasiadditive inhibition of JH biosynthesis (S. S. Tobe, J . Zhang and W. G. Bendena, unpublished). There also appear to be differences in responsiveness of D. punctata CA to allatostatins I , IV and V and at least at certain specific times (e.g. Day 2 mated, Day 4 mated, mid-pregnancy and second cycle females) (Pratt et al., 1990, 1991a: Stay et al., 1991a). Thus, CA from females of these specific ages may respond strongly to one allatostatin and only weakly to another. Although the statistical significance of these differences has not been evaluated at these specific times, these observations strongly suggest that there are at least three different receptors for these three allatostatins. The fact that 13 different allatostatins probably occur in D. punctata and that each shows distinct dose-response curves as well as different EDSo values (S. S. Tobe, J. Zhang and W. G. Bendena, unpublished) also argues for multiple allatostatin receptors. It remains to be determined, however, if each allatostatin species is associated with a different receptor and if each uses a common signal transduction mechanism.

6 Distribution of allatostatin-irnrnunoreactive cells The neurosecretory cells producing allatostatins were identified immunohistochemically with antibodies produced against synthetic allatostatins. 6.1

MANDUCA S E X T A

The neuronal elements of the M. sexfa larval brain as demonstrated by backfilling from the nerve that enters the CA, the nervus corporis allati (NCA) and their designations by Copenhaver and Truman (1986a), are shown in Fig. 3A. These neurosecretory cells are medial (IIa) and three

FIG. 3 Manduca sexta larval brain-retrocerebral complex, whole mounts. (A) Camera lucida drawing of central neurones as shown by cobalt backfilling of right nervus corporis allati. Lateral cells (Ia and Ib) extend axons in ipsilateral tract (A); lateral cells (111) extend axons in contralateral tract (C) and medial cells (Ha) extend axons in contralateral tract (B). Nervus corporis cardiaci 1 and 2 (NCC 1 + 2 ) ; dendritic fields (3, 4) (modified from Copenhaver and Truman, 1986a). (B) M. sexta allatostatin-immunopositive lateral la;, cells (arrowheads), Ib cells (arrows) and their ipsilateral axon, tract A (open arrows) to NCC 1 + 2, and (C) to their axon terminals in CC and CA; (B) and (C) are immunoreacted with antibody against M. sexta allatostatins visualized by horseradish peroxidase reaction; from fifth instar day 1. Photographs from Zitnan et al. (submitted). Bar = 0.1 mm.

ALLATOSTATINS: IDENTIFICATION, PRIMARY STRUCTURES, AND FUNCTIONS

1+2

297

298

BARBARA STAY et al.

groups of lateral cells, two ipsilateral, Ia (eight large) and Ib (a variable number, small), and the contralateral 111 cells (two large). Injection of single Ia cells with cobalt also demonstrated that these cells arborize terminally in the CC and CA (Copenhaver and Truman, 1986b). In the adult M. sexfa, Homberg et al. (1991) subdivided the Ia cells into two groups on the basis of differential immunoreactivity: Ia, (four cells immunoreactive with anti-leucine-enkephalin antiserum) and Ia2 (more medial in position, five cells immunoreactive with anti-eclosion hormone, anti-substance P and anti-corticotropin releasing factor antisera). As part of a study to compare accumulation of neuropeptides in brain cells of normal, starved and parasitized last instar larvae of M. sexfa, Zitnan et a f . (submitted) demonstrated allatostatin-imrnunoreactive cells in whole brains with rabbit polyclonal antiserum against M. sexfa allatostatin conjugated to thyroglobulin. Lateral cells of two different size classes were immunoreactive; eight small cells (Ib) and two to five larger more medial cells (Ia2) per hemisphere (Fig. 3B). In addition three to six small cells scattered among the Ia2 and Ib cells were immunoreactive and may be previously unidentified Ib cells. Ib cells send ipsilateral axons that bypass the corpora cardiaca and arborize in the inner part of the CA, whereas Ia2 cells arborize mostly in the CC (Fig. 3C). Allatostatin-immunoreactive material was evident in the ipsilateral tract in the brain (tract A of Copenhaver and Truman, 1986a) (Fig. 3A, B) and in the terminal arborizations in CC and CA (Fig. 3C). The intensity of immunoreactivity varied with age and condition; parasitized larvae showed the most intense immunoreactivity for allatostatin and the other neuropeptides studied (Zitnan et al., submitted). In normal larvae eclosion hormone does not occur in Ia2 cells, as described by Homberg et al. (1991) for adult brain but it was found in Ia2 cells in parasitized larvae (Zitnan et al., submitted). With respect to the possibility that multiple neuropeptides may exist within the allatostatinimmunoreactive cells, only proctolin-like peptide was found in Ib cells and eclosion hormone-like immunoreactivity was detected in Ia2 cells (Zitnan et al., submitted). However, the possibility of co-localization with allatostatin by means of double staining immunohistochemistry was not examined because all antisera were made in rabbits. Allatostatin-immunoreactive material in the brain was not limited to cells projecting to the CC-CA. Three interneurones located more posteriorly than the lateral cells showed immunoreactive axons branching in the medial tritocerebrum (Fig. 3B). Allatostatin-immunoreactive cells also occurred in the ventral nerve cord and in the midgut of parasitized larvae (Zitnan et al., unpublished observation). The allatotropin immunoreactive cells of M. sexta larval brain also project to the CC and CA but from different groups of cells, the ipsilateral Ia, cells and the contralateral 111 cells; these cells are obvious only in wandering larvae or parasitized ones (Zitnan et af., submitted).

ALLATOSTATINS: IDENTIFICATION, PRIMARY STRUCTURES, AND FUNCTIONS

299

From their immunohistochemical observations in M. sexta, Zitnan ef al. (submitted) suggested that allatostatin released from the Ia2 cells primarily at the CC could act neurohormonally whereas that released from the Ib cells within the CA could regulate JH synthesis in a direct way. The presence of allatostatin in interneurones, ventral ganglia and gut cells indicates multiple functions for this neuropeptide.

6.2

COCKROACHES

The neuroanatomy of the retrocerebral complex in cockroaches (Order Dictyoptera) differs slightly from that in Lepidoptera in that the NCC 1 and the NCC 2 leave the brain separately (Willey, 1961). Otherwise the lateral neurosecretory cells (in the pars lateralis) project ipsilaterally in the NCC 2 pathway as do most of the lateral cells in Lepidoptera, and the medial cells (in the pars intercerebralis) project to the contralateral NCC 1 and both of these nerves carry axons which terminate in the CC and CA as demonstrated by metal backfilling experiments in both P . americana (Pipa, 1978), and D. puncrata (Lococo and Tobe, 1984). Because the NCC 1 and 2 leave the brain in distinctly different places it was possible to sever these nerves separately and observe the effects on CA activity. Severance of NCC 1 relieved inhibition of the CA (Scharrer, 1946, 1952; Engelmann, 1957; Tobe et a f . , 1981) whereas severance of NCC 2 did not (Engelmann, 1957), thus implicating medial cells as a source of allatostatic neurosecretory material. However, radiofrequency cautery of either medial or lateral cells relieved inhibition of the CA, thus implicating both regions as a source of allatostatic material (Riiegg er al., 1983).

6.2.1 Diploprera punctata From the description of D . puncrara allatostatin-immunoreactive cells that project to the CC-CA in this cockroach (D. puncrata), it is clear that the lateral group of neurosecretory cells carry allatostatin to the CA (Fig 4A). This was demonstrated with a monoclonal antibody selective for D.puncrara allatostatin I, as compared with allatostatins II-VII (Stay et al., 1992a; A. P. Woodhead, unpublished observation, re VI and VII). The 30 or more immunoreactive lateral cells, mostly small (8 pM) and a few larger (15 pM), project in the NCC 2 and arborize terminally in the CC and CA. Immunoreactive neurosecretory vesicles occur within nerve terminals adjacent to CA cells but not all of the neurosecretory terminals within the CA are immunopositive (Stay er al., 1992a). Four large medial cells (two pairs in each hemisphere) in the ventral pars intercerebralis are intensely allatostatin immunoreactive (Fig. 4B); these cells do not project to the NCC 1 but rather are interneurones with collaterals that branch extensively in the anterior protocerebral neuropil; their large axons form a horizontal un-

300

BARBARA STAY et al.

branched loop in the brain (Fig. 4A) and terminate in extensive arborization in the lateral protocerebral neuropil (Stay et al., 1992a,b). Thus the imrnunocytochemical localization of D. punctutu allatostatins suggests that inhibition of the CA due to the identified allatostatin peptide has its origin in the lateral neurosecretory cells and reaches the CA through the NCC 2. However, allatostatin immunoreactivity is found in many other somata in the brain (in the protocerebrum adjacent to the optic lobes, between the protocerebrum and deutocerebrum, and in the tritocerebrurn), and abundant axonal tracts and arborizations are evident in the glomeruli of the antennal lobes, in the central body and in the optic lobes (Stay et ul., 1992a,b). As might be expected from this wide distribution in the brain, allatostatin irnmunoreactivity has been found in somata and axon pathways and fine terminal arborizations in the suboesophageal ganglion and all of the ventral ganglia (Stay et al., 1992a; Lange et ul., 1993; B. Stay, unpublished observations). Many visceral muscles are innervated by allatostatinimrnunoreactive nerves, including the antennal heart muscle and hindgut (Fig. 4C, D) (see Section 7). Neurosecretory cells of the midgut epithelium and innervation of rnidgut muscle are imrnunoreactive to polyclonal antiserum selective for allatostatin V (G. C. Unnithan, C. E. Summers, N. T. Davis and R . Feyereisen, unpublished observation) and to monoclonal antibody selective for allatostatin I (Fig, 4D, E) (B. Stay, personal

FIG. 4 Diploptera punctatu tissue from adult females immunoreacted with monoclonal antibody against D . punctatu allatostatin I . (A) Posterior view of whole brain-retrocerebral complex shows conspicuous immunoreactivity in paired NCC 2 (II), axons of lateral neurosecretory cells that extend to and arborize in the CC and CA. One of paired NCC 3 (111), also immunoreactive, is shown on left. Bilateral axon tracts forming loops in posterior of brain (arrowheads) originate from two pairs of medial cells and terminate in lateral protocerebral neuropil. NCC 1 nerves from medial pars intercerebralis lack immunoreactivity. Optic lobe (OL), antenna1 lobe (AL), tritocerebrum (Tri), antibody visualized by horseradish peroxidase reaction. x63. Bar = 0.2 mm. (From Stay and Woodhead, 1993). (B) Three of the four strongly immunoreactive medial interneurones of brain with axons which form the horizontal loop in (A). Collaterals of these cells branch around axon tracts of the NCC 1 (1). Anterior-posterior (A/P) axis of brain. Horizontal, 1 0 p m section of brain, horseradish peroxidase reaction; counterstain, haematoxylin. ~ 3 0 0 . Bar = 20 pm. (From Stay et al., 1992b.) (C) Electron micrograph of terminals near circular muscle of hindgut; two contain immunoreactive neurosecretory vesicles (indicated by lOnm gold particles). ~ 4 0 0 0 0 .Bar = 200nm. (From Lange et a f . , 1993). (D) Midgut, whole mount of cut open posterior part showing fine, immunoreactive branches of the proctodeal nerves (arrows). Immunoreactive neurosecretory cells appear as isolated dots and small streaks. Anterior-posterior axis as in (B). Horseradish peroxidase reaction. ~ 7 0 Bar . = 100 nm. (E) Electron micrograph of basal portion of epithelium in posterior midgut with an allatostatinimmunoreactive endocrine cell (EC) between two midgut cells (MC). Gold particles label electron lucent and dense vesicles (arrows), mitochondrion (m); basal lamina (bl). X20000. Bar = 400 nm.

ALLATOSTATINS: IDENTIFICATION, PRIMARY STRUCTURES, AND FUNCTIONS

301

302

BARBARA STAY et al.

observation). The allatostatin-immunoreactive material in the nerves of the antennal heart was extracted and shown on two successive chromatographic separations to elute at the same times as the synthetic allatostatins I , 11, 111, IV and V (Woodhead ef al., 1992). Also from the midgut an allatostatin was isolated chromatographically and shown by mass spectrometry to be identical to D. puncfata allatostatin ASB2 (V) (G. C. Unnithan, C. E. Summers, N. T. Davis and R. Feyereisen, unpublished observation). These findings suggest that the widespread immunocytochemical localizations of allatostatin indeed represent a wide distribution of these peptides and that all members of the allatostatin family occur together at each site. Such a distribution indicates functions for these peptides other than the regulation of juvenile hormone synthesis. 6.2.2 Periplane fa americana Antibodies against D. puncfafa allatostatins have been used to localize allatostatin-immunoreactive cells in the nervous system of P . americana. Since P. arnericana allatostatins are structurally similar to D. punctata allatostatins (Table l ) , the immunoreactivity described almost certainly represents the distribution of the native P . americana peptides and these are, as in D. puncfafa,apparently widespread in the central nervous system. The IgG fraction of a polyclonal antiserum against D . punctata allatostatin I conjugated to thyroglobulin and preabsorbed with thyroglobulin was used to localize allatostatins in the nervous system of P. americana (Agricola et al., 1992; Schildberger and Agricola, 1992). This antibody was shown by ELISA to react with D. puncfafa allatostatins I and 111 to a far greater extent than to D. puncfata allatostatin V (H. Agricola, personal communication). Numerous groups of cell bodies, many axon tracts and arborizations are immunoreactive in the brain including 20-25 cells in the pars lateralis, some or all of which project to the CC and CA (Fig. 5E, F) (Schildberger and Agricola, 1992). The monoclonal antibody to D . punctafa allatostatin I (Stay et al., 1992a) also shows immunoreactivity in the lateral cells, the NCC 2 and CC-CA and in the two pairs of strongly immunoreactive medial cells (Fig. 5A-C). In addition Schildberger and Agricola (1992) have shown both local and output elements in the antennal lobes, 16 pairs of cells projecting from the protocerebral bridge into the lower central body and one pair projecting to the lateral accessory lobe of the central body (Fig. 5D). Using the same polyclonal antibody, Agricola er al. (1992) described many cells in the suboesophageal and abdominal ganglia that are allatostatin immunoreactive, among them two dorsal unpaired medial neurones and two pairs projecting ipsilaterally in each ganglion, and a pair of neurones in the metathoracic ganglion that projects to all abdominal ganglia. Organs that are innervated by allatostatin-immunoreactive nerves included muscles of

ALLATOSTATINS: IDENTIFICATION, PRIMARY STRUCTURES, AND FUNCTIONS

303

the antenna1 heart, the hindgut, the midgut, and the hyperneural muscle, the salivary, the prothoracic and the male accessory glands and the abdominal perisympathetic (neurohaemal) organs and rnidgut neurosecretory cells (Agricola et al.. 1992). A fuller description of the allatostatinimrnunoreactive neurones will be useful to identify those neurones suitable for further characterization of the functions of the allatostatin peptides. 6.3

CRICKETS

Crickets, in order Orthoptera, are closely related to cockroaches and their relatives in order Dictyoptera, and therefore it is not surprising that D. punctata allatostatin is an effective inhibitor of J H synthesis by cricket C A in vitro (Neuhauser et al., 1994) and that antibodies against D. punctata allatostatins react with a similar distribution of cells as in the cockroaches. However, the allatostatin-like neuropeptides of the cricket have not yet been isolated. With polyclonal antiserum against D. punctata allatostatin I, 20 lateral cells were immunopositive in the cricket Gryllus birnaculatus; five to seven of these were shown by backfilling with lucifer yellow to project to the corpus allatum nerve 1 (NCA 1) and there was arborization of immunopositive axon terminals within the CA (Schildberger and Agricola, 1992). Also with monoclonal antibody against allatostatin I, lateral cells, the NCC 2, and CC-CA were found to be immunopositive in both G. birnaculutus and Acheta domesticus (Neuhauser et al., 1994) (Fig. 6A-C). And as in the cockroach the immunoreactivity was widespread in the brain (Schildberger and Agricola, 1992; Neuhauser et al., 1994). The two pairs of cells in the pars intercerebralis that were so strongly irnmunoreactive in D. punctata and P . americana are also conspicuous in the crickets G. bimacufatus and A. domesticus (Neuhauser et al., 1994) (cf. Figs 4A, 5B and 6A). 6.4

SUMMARY OF DISTRIBUTIONOF ALLATOSTATINS

D. punctata allatostatin-irnmunoreactive cells are found in all parts of the central nervous system in this cockroach and a similar distribution appears to exist in P. arnericana and in crickets. More detailed imrnunocytochemical studies will be required to determine whether similar identifiable neurones are immunopositive among closely related groups of insects. The lateral neurosecretory cells projecting to the CC-CA and the two strongly immunoreactive interneurones with somata in the pars intercerebralis are clearly homologous in the few insects studied to date. The distribution of M . sexta allatostatin-immunoreactive cells has been determined in detail only for the brain of larvae. Other parts of the nervous system have yet to be thoroughly examined. However, there is evidence that the irnmunoreactivity to M . sexta allatostatin is found in neurones throughout the nervous system and in midgut cells.

BARBARA STAY et al.

304

6.5

PARALLELS WITH OTHER NEUROPEPTIDES

T h e occurrence of M . sexta and D. punctata allatostatin immunoreactivity in different types of neurones within the nervous system is not unexpected because many other neuropeptides are found in such diverse distribution. For example proctolin, originally isolated from cockroach gut by its stimulating effect on visceral muscles, has now been demonstrated to occur in neurones of both visceral and skeletal muscle and in intra- and interganglionic neurones (reviewed by O’Shea and Adams, 1986; Orchard et al., 1989). Although proctolin occurs in a small number of neurones relative to the total population, it is distributed throughout the central nervous system a n d this distribution suggests multiple functions for this neuropeptide (e.g. in grasshopper: Keshishian a n d O’Shea, 1985; in blowfly: Nassel and O’Shea, 1987). Prothoracicotropic hormone (PTTH), the brain neuropeptide that stimulates the production of ecdysteroid which promotes insect development, now appears to be produced in locations other than the brain (Westbrook et al., 1993). T h e big PTTH of M . sexta has been demonstrated with a specific monoclonal antibody t o occur in lateral neurosecretory cells 111 that project t o the interior of the CA, presumably where the peptide is released into the haemolymph, in ventral medial cells that project t o the proctodeal nerve and what appear t o be interneurones within the brain a n d within t h e suboesophageal ganglion (Westbrook et a/., 1991, 1993). This suggests that PTTH probably has several functions other than the regulation of the production of hormone by the prothoracic glands (Bollenbacher et a[., 1993).

FIG. 5 Periplaneta americuna brain-retrocerebral complex immunoreacted with antibody against D. punctata allatostatin I. (Photographs D-F thanks to H. Agricola.) (A) Anterior view, right half of young larval brain. Numerous lateral cells (LC), a pair of medial cells (arrow), cells adjacent to and fibres in optic lobe (OL), antenna1 lobe (AL) and tritocerebrum (Tri) are immunoreactive. Monoclonal antibody visualized with horseradish peroxidase reaction. xSS. (B) Posterior view, right half of young larval brain. NCC 2 leaving brain (arrow), cells in tritocerebrum (Tri) and protocerebrum (P) are among conspicuously immunoreactive elements. Corpus cardiacum (CC), optic lobe (OL). Reaction and magnification as for (A). Bar = 0.3 mm. (C) Corpus cardiacum-corpus allatum from adult mated female. Immunoreactivity occurs in NCC 2 (2), NCC 3 (3) and in many axon terminals in CC and CA but not in NCC 1 (1). Reaction as in (A) and (B). ~ 9 0 Bar . = 0.1 mm. (D) Frontal section of adult male brain in region of central body shows immunoreactivity in neurones projecting from protocerebral bridge (PB) into lower central body (L). Immunoreactive fibres also present in upper division of central body (U). Polyclonal antibody, Texas red secondary antibody; 60 pm section. ~ 3 0 0 .Bar = 30 jm. (E) Portion of the corpus cardiacum with immunoreactivity in fine nerve fibres and varicosities. From adult male; whole mount; reaction as for (D). ~ 2 2 0 (F) . Whole immunoreacted corpus allatum from adult male. Reaction and magnification as for (E). Bar = 30 pm.

ALLATOSTATINS: IDENTIFICATION, PRIMARY STRUCTURES, AND FUNCTIONS

305

306

BARBARA STAY et at.

Also the presence of both M. sexta and cockroach allatostatin-like peptides in the midgut cells adds to the list of numerous neuropeptides already identified imrnunocytochemically in insect midgut. Antibodies against many vertebrate neuropeptides (e.g. pancreatic polypeptide, enkephalins, vasopressin) and several invertebrate ones (e.g. FMRFarnide, M. sexta allatotropin and diuretic hormone) react with midgut cells in a wide variety of insect species (Iwanaga et al., 1986; Brown et al., 1986; Montuenga ef af., 1989; Zitnan et af., 1993a). Little is known about the physiological effects of any of these neuropeptides in the gut. Those of the allatostatins have yet to be demonstrated and will likely involve complex actions and interactions such as may also be found for these peptides in the nervous system.

7 Other functions The irnmunocytochemical localization of allatostatin-like peptides in interneurones of the central nervous system, in neurones that innervate visceral muscle, glands other than the corpora allata, neurohaernal organs, and in midgut cells, suggests that allatostatins are multifunctional neuropeptides even though inhibition of JH synthesis by corpora allata in vitro was the only bioassay utilized in their isolations. Because nerves to the antennal pulsatile organ muscle and the proctodeal nerves to the muscles of hindgut showed strong imrnunoreactivity with antibody against D. punctata allatostatins, attention was directed to these organs with respect to whether the imrnunoreactivity indeed represented the allatostatin peptides and whether allatostatins would have an effect on the activity of these muscles (Woodhead ef al., 1992; Hertel and Penzlin, 1992; Lange et al., 1993). Crude extract of antennal pulsatile organ muscle not only showed an allatostatic effect on juvenile hormone synthesis by CA in vitro (50% inhibition at 2.6 organ equivalents) but also showed such

FIG. 6 Achetu domesticus brains and retrocerebral complex immunoreacted with monoclonal antibody against D.punctutu allatostatin I , visualized with horseradish peroxidase reaction. (A) Anterior view, whole brain of young larva. Note immunoreactive lateral cells (LC); two pairs of large medial cells (M), many cells between protocerebrum (P) and antennal lobes (AL) and numerous immunoreactive fibres near surface of protocerebrum and in antennal lobes. ~ 8 0 Bar . = 0.2mm. (B) Posterior view of portion of whole brain. Note NCC 2 tract (arrow) within the brain and numerous immunoreactive cells. X 80. (C) Dorsal view of adult retrocerebral complex (one CA removed). Immunoreactivity occurs in thin and varicose fibres on and in CC and CA, and in axons of NCC 2 entering CA (arrow). Also note immunoreactive fibres on surface of aorta (A) and NCC 1, 2 and 3 (1)(2)(3). X80.

ALLATOSTATINS: IDENTIFICATION, PRIMARY STRUCTURES, AND FUNCTIONS

307

BARBARA STAY et al.

308

bioactivity after separation on two different HPLC systems, and the activity occurred in fractions that eluted at the same times as synthetic D. puncram allatostatins I-V (Woodhead et al., 1992). Thus allatostatins almost certainly occur in the antennal heart nerves and probably also in those innervating hindgut musculature. The effect of D. puncrata allatostatins on the antennal pulsatile organ muscle and on the hindgut is discussed in Section 7.1. 7.1

MODULATORS OF SPONTANEOUS AND PROCTOLIN-INDUCED CONTRACTIONS

7.1.1 Antenna1 pulsatile organ muscle The antennal pulsatile organ is an accessory heart that aids circulation of haemolymph into the antennae. It consists of two elastic-walled ampullae at the base of each antenna that lead into vessels extending into the antennae and a transverse muscle inserting on each ampulla (Pass, 1985). This pulsatile organ muscle (antennal heart) contracts rhythmically, expanding the ampullae; as it relaxes the elastic ampullar wall forces haemolymph into the antennae (Pass, 1985). Studies suggest that the rhythmic contraction is myogenically generated (Hertel et al., 1985, 1988). Its rhythm can be increased by proctolin (Hertel et al., 1985) and inhibited by octopamine (Hertel et al., 1988), but the antennal heart nerves also exert neural control over the muscle (Hertel et al., 1985). These antennal heart nerves, whose somata originate in the suboesophageal ganglion and exit the tritocerebra in the NCC 3 (Pass et al., 1988a), are strongly immunoreactive to antibodies against D. puncrara allatostatin in both D. puncrata (Woodhead et al., 1992; Lange er al., 1993) and P. americana (Agricola et al., 1992). Application of allatostatin alone to the antennal heart muscle in vitro did not alter the contraction. This was found for D. puncrara allatostatin in P . americana (Hertel and Penzlin, 1992) and for D. puncfara allatostatins I and IV on antennal heart muscle of D. punctata (Lange et al., 1993). Also when allatostatin was applied along with proctolin it did not alter the stimulatory effect in either cockroach (Hertel and Penzlin, 1992; Lange e f al., 1993). However, in one set of experiments, when allatostatin was applied before proctolin and at a higher concentration (lo-’ M allatostatin and lO-’M proctolin), it clearly reduced the excitatory effect of proctolin (Hertel and Penzlin, 1992). Because proctolin was dominant when both substances were applied together and allatostatin antagonized proctolin action only when its concentration was higher than that of proctolin, Hertel and Penzlin (1992) suggested that both peptides affect the same binding sites. An alternative interpretation might be that the response of the antennal heart to allatostatin is slower than that of proctolin and therefore must be applied before proctolin in order to be effective. Proctolin probably occurs in the P .

ALLATOSTATINS: IDENTIFICATION, PRIMARY STRUCTURES, AND FUNCTIONS

309

americana antennal heart nerve even though it was not localized there immunocytochemically (Hertel et a f . , 1988) because proctolin was demonstrated in extract of D. punctata antennal heart by bioassay following separation on several different HPLC systems (Lange et a f . , 1993). Thus the action of allatostatins on the antennal heart muscle is probably to modulate the action of another peptide. most likely proctolin. Both allatostatinimmunopositive and immunonegative nerve endings were found in D. punctatu antennal heart (Lange et a f . , 1993) and a high concentration of octopamine was shown biochemically at the ampullar ends of the muscle (Pass et a f . , 1988b). What the content of the allatostatin-immunonegative terminals might be and whether proctolin, allatostatins, and octopamine are co-localized or are in different axons of the antennal heart nerves remains to be determined. Identification of the receptors for the allatostatins on the pulsatile organ muscle and how they affect the muscle will be necessary to clarify the modulatory role of allatostatins on this particular muscle.

7.1.2 Hindgut In contrast to the inability of allatostatins to alter spontaneous contractions of antennal heart muscle (Hertel and Penzlin, 1992; Lange et a f . , 1993), D . punctata hindgut muscle reacted to D. punctata allatostatins I and IV with a dose-dependent decrease in frequency and amplitude of spontaneous contraction that was reversible; the threshold concentration for response was between M and lo-’ M (Lange el a f . , 1993). It was also possible to show that when applied simultaneously with proctolin, allatostatin I or IV ( 5 X lop6M) antagonized the stimulatory effect of proctolin (5 X lo-’ M) on hindgut muscle (Fig. 7A). Recently this effect was shown to be dose dependent (Fig. 7B) and was also found to be a common property of all 13 members of the D. punctatu family of allatostatins (A. Lange, personal communication). Although the hindgut muscle clearly responds in a different way to allatostatin than does the antennal heart muscle, in both the effect is a decrease in contraction. It would appear therefore that allatostatins are among the few myoactive peptides that are inhibitory. Known inhibitors of muscular contractions in insects are leucomyosuppressin from the cockroach Leucophaea maderue (Holman et al., 1986), the related SchistoFLRF-amide from Schistocerca gregaria (Robb et a f . , 1989; Lange et af., 1991) and another related peptide from a fleshfly, Neobelfieriu buffata (Fonagy et a f . , 1992). Note, however, that a peptide can inhibit one type of muscle and stimulate another, e.g. SchistoFLRF-amide is a cardioinhibitor and a potentiator of skeletal muscle twitch (Robb et a f . , 1989). Much remains to be investigated with respect to allatostatins and muscle activity. How do other visceral muscles respond? Are skeletal muscles affected? Do M. sexta allatostatins also have myotropic actions?

BARBARA STAY et a/.

310

-

A

5 x 10-7M P

A

m

5x10~ MP 5 x ' 1 0 s M ASTI

r----X

v Allatostatin I 0 Allatostatin IV

Y

k

a (B)

-

IO-~

I O - ~

10-~

Peptide concentration (M) log scale

FIG. 7 Effect of allatostatins I and IV on proctolin-induced contraction of D. punctatu hindgut. (A) Proctolin (P) (applied at triangle) induces a large contraction

that is inhibited by application of D. puncfafaallatostatin (AST I) (applied at second Newtons; 1 min (from Lange et triangle). Washes denoted by bars. Scale = 5 x al., 1993). (B) Dose-dependent inhibition of 5 x lo-' M proctolin-induced contraction by D. punctata allatostatins I and IV. Each datum point is mean f SE of at least four preparations ( A . B. Lange, unpublished).

ALLATOSTATINS: IDENTIFICATION, PRIMARY STRUCTURES, AND FUNCTIONS

7.2

31 1

ACTIVITY I N INTERNEURONES/CHEMICAL SYNAPSES

Immunocytochemical identification of allatostatin-like material in interneurones in insects is the only evidence available to suggest that allatostatins do function as neurotransmitters and/or neuromodulators within the nervous system of insects. However, the effect of D. punctara allatostatins on the stomatogastric nervous system of a crab, Cancer borealis, indicates that allatostatins can act within the central nervous system. Many different neurotransmitters are present in the stomatogastric nerve entering the stomatogastric ganglion and each produces a different and specific change in the gastric/pyloric rhythms (Marder and Hooper, 198.5). Discharge of the motor neurones of the stomatogastric ganglion, which innervate the muscles of the gastric mill and pyloric regions of the foregut wall, regulates the rhythm of these muscles (Marder and Hooper, 198.5). D. punctara allatostatins (lo-' and lop6 M) applied to the stomatogastric ganglion in vitro decreased the frequency of the pyloric rhythm (Skiebe and Schneider, 1994) and the gastric rhythm (Skiebe-Corrette et a f . , 1993). Allatostatins I-IV had a similar dosedependent effect on the pyloric rhythm neurones, and allatostatin I11 (the least effective inhibitor of JH synthesis in D. punctata) was slightly more effective than the others (Skiebe and Schneider, 1994). These findings indicate that D. punctata allatostatins can act centrally to modify neuronal activity. In addition to their central action on the rhythm of the motor neurones D. punctata allatostatins act at the neuromuscular junction. Allatostatin 111 reduced the gain of gastric motor neurone to muscle interactions. The tension evoked in muscle by stimulation of motor nerves to two identified gastropyloric muscles was reduced by D. punctata allatostatin III( lo-' M) and the nerve-evoked excitatory junctional potentials in one of these muscles were reduced in amplitude and their facilitation was modified (Skiebe-Correte er al., 1993). Since D. punctata allatostatin-like immunoreactivity occurs in the stomatogastric nervous system of Cancer borealis (Skiebe and Schneider, 1994; see Section 8.3) and inhibits motor function at several sites, it is likely that a peptide(s) similar to the cockroach allatostatin(s) also occurs in decapod crustaceans. 7.3

NEUKOHOKMONE

The presence of allatostatin-immunoreactive material in the corpora cardiaca (Stay er d.. 1992a,b) and in abdominal perisympathetic organs (Agricola el al., 1992) indicates that allatostatins should be found in the haemolymph. The fact that JH synthesis by denervated CA is regulated also suggests that regulators can act through the haemolymph and that allatostatin might also function as a neurohormone.

BARBARA STAY et a/.

312

An ELISA using a polyclonal antibody to D. punctata allatostatin I demonstrated immunopositive material in haemolymph of D. punctata and the titre changed with age (Yu et af., 1993). In haemolymph of last instar larvae, concentration was low (<0.15 nM) in the first half of the stadium and increased to a peak (1.2 nM) in the second half of the stadium when JH synthesis is essentially turned off and adult development takes place. Also in adult females titre increased after oviposition and reached a peak early in pregnancy (2.4 nM on Day 13) and remained above 0.5 nM through the 50 days of pregnancy in which haemolymph allatostatin was monitored (Yu et al., 1993). Partial purification of several ml of haemolymph from adult female D. punctata yielded from an RP C18 column fractions with bioactivity that eluted at the same times as allatostatin I-V (Woodhead et al., 1993). The question of whether this haemolymph allatostatin could be inhibiting J H synthesis was addressed by injecting allatostatins repeatedly into the haemolymph of mated females during the reproductive cycle and demonstrating that compared with controls injected with inactive peptides, J H synthesis by the C A was reduced andlor egg development was delayed (Woodhead et al., 1993). The effect was not dramatic, probably because the half-life of the peptides in the haemolymph was short (see Section 10.1). Weaver el af. (1994) found that P. americana allatostatin injected into haemolymph of virgin female P. americana lowered the titre of J H 12 h after injection but they were unable to demonstrate a similar effect following injection into females with developing oocytes, a stage when JH titres are normally higher than in virgins (Edwards et al., 1990). In D . punctata the results suggest that haemolymph allatostatin could be delaying the normal progress of increased JH synthesis associated with egg development (Tobe and Stay, 1977) by a slight inhibition of the CA. 8 Distribution in other insects and other phyla

As a result of the relatively recent discovery of the allatostatins few studies have been published on the distribution of these peptides in diverse organisms. However, immunocytochemical evidence is accumulating to suggest that both moth and cockroach allatostatin-like materials occur in many orders of insects, in other classes of arthropods and in other phyla. The identity of the compounds responsible for this immunoreactivity is known only for the fly Calliphora vomitoria (Duve et af., 1993). 8.1

CALLATOSTATINS. -TYR-X-PHE-GLY-LEU/MET-AMIDES FROM THE BLOWFLY

Partially purified brain extract of C. vomitoria inhibited JH synthesis by CA of D. punctata (74% inhibition at 5 brain equivalents) (Duve et al., 1992).

ALLATOSTATINS: IDENTIFICATION, PRIMARY STRUCTURES, AND FUNCTIONS

TABLE 5

313

Primary structure of callatostatins from Cdfiphora vomitoriu (from Duve

et ul.. 1993)

No. 1.

2. 3. 4. 5.

Amino acids in single letter code D-P-L-N-E-E-R-R-A-N-R-Y-GF - G - L -NH2 L - N - E - E - R - R - A -N- R - Y - G - c F ) - G = - L ~ - N H : A-N-R-Y-GF - G - L -NH2 F -G - L a - N H Z (D/N)-R-P-Y-(S)G-P-P-Y-DF - G -M -NHzb

( ) Tentative assignment from sequence analysis. "Assignment based on recorded molecular weight and expected sequence similarity. "Suggested hy lack of activity of free acid form.

Subsequently, a hexadecapeptide, callatostatin 1, with C-terminal homology to cockroach allatostatins was found in extract of thoracic ganglia of the blowfly C . vomiroria during the purification of -Phe-Met-Arg-Phe-amide peptides (Duve ef af., 1993). Using an antibody against callatostatin 1 to identify the RIA-positive fractions from HPLC, a group of peptides was isolated from extracts of heads, thoracic ganglia or whole flies; callatostatin 1 was isolated from heads and brains as well as thoracic ganglia (Duve et af., 1993). The amino acid sequences of these five peptides are shown in Table 5 . The C-terminal residues are all -Y-X-F-G-L-NH? with the exception of methionine as the C-terminal amino acid in callatostatin 5. Whether callatostatins 2 and 3 are peptides determined by the gene or are cleavage products is not yet known. Both could be derived from callatostatin 1 by cleavage either by processing of prohormone or as an artefact of the purification (Duve et al.. l9Y3). None of the callatostatins is identical to the D. purzcrata allatostatins. The ability of callatostatin t o inhibit JH synthesis by CA of cockroaches and inability to inhibit that of flies indicates clearly that although the peptide structures are homologous their functions with respect to CA activity are not. None of the callatostatins, at concentrations between lo-' and lo-' M, showed inhibition of JH bisepoxide synthesis by CA from 6-10-day-old adult female C. vonzitoria; in contrast JH 111 biosynthesis by CA of D. punctatcr was inhibited (Duve et af., 1993). The most potent inhibitor was callatostatin 5 with C-terminal methionine which resulted in 50% inhibition of JH synthesis at lo-"' M. Higher concentrations were required to achieve similar inhibition with callatostatins 1 and 2 (lO-'M), callatostatin 3 (lo-' M) and M). It is likely that callatostatins have other Asn' callatostatin 4 (about functions in the fly and these may be similar to the other functions in D. punctatci. A different distribution of callatostatin 1-like material in CC-CA of C. vomitoriu and D. piincfata would be anticipated from the differential effect

BARBARA STAY et al.

314

of these peptides on CA activity in the two species. The following observations have been made by Duve et al. (1993) using antibodies to allatostatin I and callatostatin 1 on tissue from C . vornitoria and D. punctata. The CA of D. punctata were immunoreactive to callatostatin 1 whereas those of C. vomitoria were not. In D. punctata brain, lateral neurosecretory cells that are known to project axons to the CC-CA were immunopositive to both callatostatin I-specific antibody and allatostatin I-specific antibody, Thus callatostatin 1 antibody is reacting with peptides in the allatostatin-producing cells. Whether these are identified D. puncfata allatostatins or peptides identical to callatostatin remains to be determined. Callatostatin 1-immunopositive cells were also seen in the suboesophageal ganglion of D . punctata. In C. vornitoria the callatostatin 1 antiserum demonstrated cells and neuropil in brain, and suboesophageal, thoracic and abdominal ganglia. Also, in C. vomitoria the cardiac recurrent nerve, oviduct and hindgut nerves and the surface of the bundle of nerves to the metathoracic leg were immunopositive to callatostatin 1 antiserum. The callatostatin immunoreactivity in nerves to hindgut and oviduct suggests that these organs may respond to callatostatin and it will be interesting to learn whether the response is similar to that elicited by D. punctata allatostatin (see Section 7.1).

8.2

IMMUNOREACTIVITY TO ALLATOSTATIN ANTISERA IN O T H E R INSECTS

8.2.1 Drosophila rnelanogaster The occurrence of M . sexta allatostatin-immunoreactive cells in the central nervous system of D. rnelanogaster has been described in larvae, pupae and adults (Zitnan et al., 1993b). The M . sexta allatostatin imrnunoreactivity occurs in a larger set of neurones in D. rnelanogaster than in M . sexta. It is difficult to say whether some of the cells in D. rnelanogaster are homologous to the immunoreactive cells in M . sexta although M . sexta allatostatin immunoreactivity was found in the CA of the ring glands of D . melanogaster larvae. The effect of M . sexta allatostatin on D. rnelanogasfer CA has not yet been tested, therefore it is premature to speculate that M . sexta allatostatins may serve an allatostatic function in D. rnelanogaster. The distribution of immunoreactive cells in the brain, and suboesophageal, thoracic and abdominal ganglia suggests that these immunoreactive neurones are of several different types. In larvae (Fig. 8A) and adults some immunoreactive cells are in the position of medial and lateral neurosecretory cells in the protocerebrum, others occur in the optic lobes and the tritocerebra. In adults additional groups of immunoreactive cells appear in the optic and antenna1 lobes; some cells of the thoracic ganglia are immunoreactive only in the pupa. The projection of allatostatin-positive immunoreactive neurones to neurohaemal organs of the thoracic and

ALLATOSTATINS: IDENTIFICATION, PRIMARY STRUCTURES, AND FUNCTIONS

315

FIG. 8 Location of neurones in the central nervous system of larval Drosophila reacted with antibodies to: (A) M . sexra allatostatin (from Zitnan et at.. 1Y93b); and (B) D. puncrata allatostatin (J. G . Yoon and B. Stay, unpublished).

abdominal segments suggests that M. sexta allatostatin-like material is released into the haemolymph in D. melanogaster. The distribution of D. punctata allatostatin-like immunoreactive cells in the nervous system of D. melanogaster (J. G . Yoon and B. Stay, unpublished) is clearly different from the distribution of M. sexta allatostatin-like materials (Fig. 8B). As was found in the distribution of callatostatin-immunoreactive cells in adult C. vomitoria, in larval D. mefangaster immunoreactivity is not found in the C A , but is found in cells of brain, and suboesophageal and ventral ganglia and in nerves to the hindgut (J. G. Yoon and B. Stay, unpublished).

8.2.2

Manduca sexta

D. punctata allatostatin-like immunoreactive material is evident in the brain of the moth M. sexta (D. Zitnan, unpublished). The distribution again is different from that of M . sexta allatostatin and this would be expected since the allatostatins of moths and cockroaches are not reciprocally active as inhibitors of JH synthesis (Kramer et a f . , 1991; N. Granger, personal communication). It is known, however, that extract of brain of the moth Pseudafetia unipuncta inhibits JH synthesis by cockroach CA (M. Cusson

BARBARA STAY et al.

316

and S. S. Tobe, unpublished) and that immunoreactivity to D. punctata allatostatin occurs in the brain of this moth (B. Stay, unpublished observation). 8.2.3 Apis mellifera

D. punctata allatostatin-like immunoreactive material is present in the brain and suboesophageal ganglion of the honey bee. The most strongly reactive cells in the whole brain preparations were in cells at the base of the optic lobes (B. Stay, unpublished). 8.3

CRUSI'ACEA. D E C A P O D A

Since D. punctata allatostatins alter the rhythm-generating cells of the stomatogastric ganglion in the crab Cancer borealis (see Section 7.2), it was of interest to know whether D. punctata-like allatostatin could be localized immunocytochemically in this crab. With polyclonal antibody to D. puncrata allatostatin I (Agricola et a / . , 1992), Skiebe and Schneider (1994) have described the immunoreactivity in the stomatogastric nervous system. The motor neurones that generate the pyloric and gastric rhythms originate in the stomatogastric ganglion which contains no allatostatin-immunoreactive cell bodies but many imrnunoreactive fine processes and varicosities. These processes in the stomatogastric ganglion originate from bipolar peripheral cells, one in each lateral ventral nerve and each medial gastric nerve and also one cell (identity uncertain) in each commissural ganglion. The axons of the peripheral cells of the lateral ventral ar?d medial gastric nerves extend through the stomatogastric ganglion to the commissural ganglia, and sometimes the other pole of one of these cells could be traced to extensive branching in an identified gastropyloric muscle. The bipolar peripheral cells resemble the serotonergicicholinergic gastropyloric receptor cells (Katz er d.,1989). Simultaneous immunoreaction with antisera to serotonin and D. puncfrrtu allatostatin showed that these allatostatin-positive peripheral cells also contain serotonin. D. punctatu allatostatin immunoreactivity also occurs in numerous cells (29-36) in each commissural ganglion; in two cells of the oesophageal ganglion that project to the brain and in other processes. of unknown source, in the oesophageal ganglion, that project to the commissural ganglia and the brain. Thus allatostatin-like peptide occurs in interneurones and sensory neurones of the crab stomatogastric nervous system. 8.4

A K A C H N I D A . ACARINA

In the tick Dermacentor variabilis, D. punctatu allatostatin-immunoreactive cells were found in unfed virgin females by Zhu and Oliver (1993) using the monoclonal antibody against allatostatin I (Stay et al., 1992a). In the fused

ALLATOSTATINS: IDENTIFICATION, PRIMARY STRUCTURES, AND FUNCTIONS

317

central nervous system, the synganglion allatostatin-positive cells are located in numerous regions: protocerebral, cheliceral, stomodeal and opisthosomal. The immunoreactivity is not only in the cell bodies but also in varicosities in the subperineural areas of the cheliceral and opisthosomai ganglia. Whether the allatostatin-like material has multiple functions as the distribution suggests and what these functions might be need to be investigated. 8.5

MOLLUSCA. GASTROPODA

D. punctata allatostatin-like immunoreactivity was described in the ganglia of the freshwater snails Bulinus globosus and Stagnicola elodes (P. Rudolph, unpublished) with the monoclonal antibody to D. punctata allatostatin I . Cell bodies occur in all of the ganglia. The greatest concentration is found in the cerebral and pedal ganglia (50-60 and 80 cell bodies on each side, respectively) in B. globosus and in the pedal ganglia (60 cell bodies on the right side and 80 on the left) in S. elodes. Again, the degree to which this immunoreactive material resembles D. punctata allatostatin and the various functions of these peptides remain to be investigated. Extract of B . globosus central nervous system did not inhibit JH synthesis by C A of D. punctata virgin females when tested at 20 brain equivalents/CA (P. Rudolph, unpublished). 9 The gene for allatostatins

9.1

DIPLOPTERA PUNCTATA

The coding region that specifies an allatostatin gene in Diploptera punctata was identified and isolated by polymerase chain reaction (PCR) (Saiki et al., 1988) amplifications (Donly et al., 1993). Specific sequences were amplified from a complementary DNA (cDNA) representation of mRNA isolated from the brains of virgin females. The initial amplified sequence was the product of a pair of highly degenerate oligonucleotide primers (Donly et al., 1993) designed to produce an internal DNA sequence representing the peptide sequence of the octadecapeptide ASB2 (Pratt et af., 1991a) (Table 1, Fig. 9). The sequence of this amplified region permitted the design of a third oligonucleotide that had sufficient specificity to allow one-sided 5' end-specific PCR of the cDNA population. The cDNA population was first ligated to a plasmid vector to provide a known reference point beyond the cDNA terminus where opposing primers can anneal (Frohman et af., 1988). The sequence of the product of this latter amplification reaction was found to encode the appropriate N-terminus of ASB2 as well as a peptide of six amino acids that resembled the C-terminus of a true allatostatin (Fig. 9). This sequence was then used to design further specific oligonucleotide

BARBARA STAY eta/. Oipstatin #

1 2

3 4

5 6 7 8 9 1 0

,

0

#

,

I

,

'

#

I

0 0

8 8

) 0

I

I

8

I

I

I

I

4

0

I

,

,

,

I

I

I

,

#

I

,

@

I

I

k

I

t

3

0

0

)

4

1

i

8

:

0

I 1

8

'

I

)

I

@

12 13

11

IPMYDFGI IPMYDFGI

I

I

PFNFGL PYNFGL

~

I ~

, ,

A

SLHYPFG I (RK) SWYAFG L

I

YPQEHRFSFGL SPQGHRFSFGL

A

A

A

A

PVNSGRSSGSRFNFGL ' ' PVSSARQTGSRFNFGL : : AAAA GDGRLYAFGL ADGRLYAFGL : A GGS L YSFGL GGSMYSFGL

: I

A

APSGAQRLYGFGL SPSGMQRLYGFGL

A

A

ARPYSFGL ARPYSFGL DRLYSFGL DRMYSFGL A DGRMYSFGL SGNDGR L YSFGL I

A A A A SKMYGFGL SKMYGFGL AYSYVSEYKRLPWNFGL AY SYVSEYKRLPVYNFGL LYDFGL LYDFGL

FIG. 9 Schematic representation comparing the structures of the allatostatin polypeptide precursors of D. punctata and P. americana. The precursors begin with a hydrophobic leader region (cross-hatched) that is presumably cleaved by signal endoproteases. Black boxes represent the individual D. punctata (dipstatin, upper) and P. arnericana (peastatin, lower) allatostatin peptides which are numbered according to their position relative to the NH2 terminus in the precursor. The corresponding peptide sequences are listed below and amino acid differences

ALLATOSTATINS: IDENTIFICATION, PRIMARY STRUCTURES, AND FUNCTIONS

319

primers. These specific primers were used in combination with degenerate primers based on the amino acid sequences of the four isolated peptides (Woodhead et al., 1989). The prediction that these peptides were downstream of the 5' region initially sequenced was correct. The complete sequence encodes a 370 amino acid preproallatostatin polypeptide. Thirteen amidated peptides, which we designate here as dipstatins, should result from post-translational cleavage and modification. The peptide sequences derived from the DNA confirm the identity of the seven previously isolated peptides with in vitro allatostatic activity, namely dipstatins 2, 4, 5 , 7, 8, 9, 11 (Woodhead et al., 1989, 1994; Pratt et al., 1991a), and predict the existence of six new peptides with related sequences (Table 1, Fig. 9). The amino terminal address sequence of each peptide is unique, which contrasts with the tandem copies of several peptides found in the Drosophila melanogaster FMRFamide precursor (Schneider and Taghert, 1988; Taghert et af., 1990). These individual address sequences may regulate recognition and/or strength of binding to one or more receptors. The duration of the peptide activity might also be conferred by the stability imparted on each peptide by each unique amino terminus. Each of the allatostatins found within the precursor contains the carboxy terminal sequence Tyr/Phe-Xaa-Phe-Gly-Leu-NH2 with one exception, dipstatin 13, where Ile is substituted for the C-terminal Leu. Pratt et al. (1991a) have demonstrated that substitution of Ile for the C-terminal Leu in dipstatin 2 has limited effect on the ability of the peptide to inhibit the production of JH by CA of 10 day D. punctata females. Ten peptides (dipstatins 1-9, 13) contain Tyr in the fifth position from the carboxy terminus and three (dipstatins 1g-12) contain Phe (Fig. 9). The physiological consequences of the Phe substitution in this position remain to be studied. Three domains of 31 (aa 121-151), 16 (aa 236-251) and 78 (aa 268-345) amino acids in the precursor are acidic, having PI values of 2.14, 3.42 and 4.49, respectively. In the precursor, these acidic regions effectively neutralize the basic charge contribution of the peptides and their processing sites. Three potential dibasic cleavage sites within the third acidic domain would result in the release of two non-amidated peptides of 12 and eight amino acids (Fig. 9). These non-amidated peptides have a PI of 5.18 and 7.57 respectively and Ile as the C-terminal amino acid. Peptides with sequence similarity to these two non-amidated peptides were not found in GenBank or SwissProt databases. between the two cockroach species indicated with a filled triangle. Acidic regions are indicated by diagonal lines. Incorporated within the third acidic spacer region of the dipstatin precursor are sequences specifying two non-amidated peptides (shaded boxes). Each allatostatin, from both species, precedes a GKR sequence required for endoproteolytic cleavage and a-amidation. Where this processing signal is modified is indicated in parentheses.

BARBARA STAY e t a / .

320

9.2

PER I P L ANETA A M E R I C A N A

Restriction enzyme-digested genomic DNA of several cockroach species was tested for the presence of sequences related to the Diploptera punctata allatostatin coding region by Southern blot hybridization (W. G . Bendena, unpublished). In all species tested different strengths of hybridization signal were found to specific restriction fragments. DNA fragments from Periplaneta americana appeared to show weaker hybridization signals than those of D. punctata which suggested a more distant relationship. To characterize this sequence relationship and identify functionally important regions within the allatostatin precursor, the homologous gene coding region from P. americana was isolated and characterized and the gene products are Ding, B. C. Donly, S. S. Tobe and W. G. designated here as peastatins Bendena, unpublished). The allatostatin protein precursors of P. americana and D. punctata are similar in size (379 and 370 amino acids, respectively) and share 71% amino acid identity. As was suggested for the Drosophila melanogaster FMRFamide polypeptide precursor (Taghert et al., 1990), evolutionary pressure to preserve the precursor size may be required for optimal processing or packaging within secretory granules. The maximum conservation was found in regions that encode allatostatin-like peptides. The tripeptide sequence Gly-Lys-Arg required for amidation and processing (Eipper ef al., 1992) is also conserved. The coding regions of the allatostatin genes are remarkably similar in structure and organization. As described for the D. punctata precursor, the P. americana allatostatin-like peptides are also separated into groups by stretches of acidic amino acids. The functional significance of peptide grouping may reflect peptide groups targeted for discrete functions. Alternatively, synergism between individual peptides within groups or separate timing of peptide group release may occur. The P. americana precursor contains 14 allatostatin-like peptides, compared with the 13 peptides found in D. punctata. Peastatins 7 and 9 are identical in sequence to the two allatostatins purified from P. americana (Table 1; Weaver et al., 1994). Five P. americana peptides, peastatins 1, 2, 3, 6, 13 are identical to the D. punctata peptides, dipstatins 1, 2, 3, 6, 12, in both sequence and position within the precursor. The C-terminal amino acids in all the remaining peastatins have been conserved when compared with the dipstatin peptide in the equivalent position within the precursor. The amino-terminal address signal sequences contain either one (peastatins 5 , 8, 9, 13), two (peastatins 7, 11, 12) or four (peastatins, 4, 10) amino acid variations. Amino acid variations between the allatostatin peptides appear to have arisen through amino acid substitutions. One exception is peastatin 4 which varies from dipstatin 4 by a three amino acid extension at the amino terminal end in addition to a single amino acid substitution. Another exception is peastatin 12 where a single amino acid difference results in an amidated peptide. A non-amidated peptide in the equivalent position that

(a.

ALLATOSTATINS: IDENTIFICATION, PRIMARY STRUCTURES, AND FUNCTIONS

321

resembled an allatostatin was previously identified within the third acidic spacer of the D. punctata precursor (Donly et aI., 1993). As Periplaneta americana is the more ancient of the two cockroach species (McKittrick, 1964) the loss of amino acids (peastatin 4 to dipstatin 4) and a peptide function (peastatin 12) appear to have occurred through the evolutionary process. P. americana and D. purzctafa allatostatin precursors appear to be expressed from single copy genes. The coding region for each preproallatostatin is contained within approximately 1100 nucleotides and shares 67% homology at the neucleotide level. Northern blot analysis with poly-A+ extracted from the brains of mated females has revealed that the only detectable allatostatin mRNA transcripts in D. punctata and P. americana are 9.2 and 5.0 kb, respectively. Outside of the coding region, the nucleotide sequences appear to have rapidly diverged. In both organisms the sequences are extremely A/T rich and the variation in transcript size may be a consequence of differing numbers of repetitive sequences. It is tempting to speculate that non-coding sequences might have a regulatory function. Non-coding mRNA sequences might contribute to the timing of allatostatin expression by formation of higher order structures that regulate mRNA stability and/or translation. These sequences might also provide signals that tag the translational apparatus bearing allatostatin mRNA for transport into specific axonal processes and/or docking beneath post-synaptic sites (Steward and Banker, 1992). 9.3

I N SITU HYBRIDIZATION

Non-radioactive in situ hybridization has been used with digoxygeninlabelled allatostatin DNA probes (Tautz and Pfeifle, 1989), to localize cells of the brain that express allatostatin mRNA (Donly et a f . , 1993; Q. Ding, B. C . Donly, S. S. Tobe and W. G. Bendena, unpublished). Brains from mated female D. punctafa and P. americana were hybridized in situ with their respective allatostatin gene. In both species allatostatin mRNA is strongly expressed by two pairs of large medial cells in the pars intercerebralis of the protocerebrum (Fig. 10A and B). In P. americana (Fig. 10B) strong hybridization of allatostatin transcripts is also found in several lateral neurosecretory cells which are major sites of allatostatin-immunoreactive neurones that project to the corpora allata (Stay et al., 1992a; Section 6.2). Hybridization to lateral cells is also found in brains of D. puncfata but signal strength is consistently lower than found in medial cells. Numerous cells that hybridize to the allatostatin probe are also detected in cells of the tritocerebrum (Fig. 1OC). These cells are strongly immunoreactive and show axons that extend into the brain (Fig. 4A). In the back view of the P. americana brain the allatostatin probe also detects a cluster of four cells in each half of the brain near the optic lobe (Fig. 10D). These cells may

322

BARBARA STAY et al.

ALLATOSTATINS: IDENTIFICATION, PRIMARY STRUCTURES, AND FUNCTIONS

323

correspond to immunopositive cells that appear to send axons into the protocerebrum (B. Stay, personal observation).

10 Metabolism and mode of action of allatostatins

10.1

DEGRADATION OF ALLATOSTATINS

Allatostatins are very fast-acting peptides, whose effect on inhibition of JH biosynthesis is totally and rapidly reversible (Woodhead et al., 1989, 1994; Pratt et al., 1990). These observations suggest that there are mechanisms for the inactivation of the peptides in vivo, with the resulting reversal of effect, since the continued presence of allatostatins would result in prolonged inhibition. At present, it is unknown if CA become refractory to the peptides on chronic exposure, but available data from crude brain extract indicate that C A respond to repeated challenges of allatostatins, remain inhibited in the presence of the peptides for at least 46 h and recover within 30 min thereafter (Rankin and Stay, 1987). Inactivation mechanisms must exist in target tissues; however, it is uncertain if such mechanisms must also be present in haemolymph. Studies on the metabolism of allatostatins have been hampered by the lack of availability of appropriate radiolabelled allatostatins. Thus work has focused on degradation of selected unlabelled allatostatins in the presence of homogenates or tissue membrane preparations, or in some cases, whole organs in vitro, and the unavailability of radiolabelled substrates has made it necessary to use relatively large quantities of both the peptides (micromolar concentrations, and hence, pharmacological) as well as large amounts of the tissues, to permit their identification and separation by HPLC and subsequent amino acid analysis. These preliminary studies indicate that homogenates of several tissues, including CA, brain and gut, are capable of metabolizing allatostatins, generally from the N-terminus, suggesting the presence of arninopeptidases in these tissues (C. Garside and S. S. Tobe, unpublished). Half-life of allatostatins ranges between 20 and 40 min, and there appear to be two principal peptide products of degradation. Although FIG. 10 Digoxygenin-labelled DNA representing allatostatin sequences from the dipstatin or peastatin precursor was used for in situ hybridization to desheathed brains of Diploptera punctata (A) or Periplaneta americana (B-D). (A, B) Frontal brain views show two pairs of medial cells (M) and lateral neurosecretory cells (arrows), prominent in P. americana (B) and present but not clearly visible in this brain of D. punctata (A). ( C , D) Posterior views of left half brains of P. americana that show expression of allatostatin in mRNA of tritocerebral cells (arrows in C) and a cluster of four cells near the optic lobe (arrow in D). Bar = 86 pm in (A), 100 p m in (B-D). (A) is from Donly et al. (1993).

324

BARBARA STAY et a/.

such analyses provide some indication of primary routes of degradation and cleavage sites, they must ultimately be supported by studies at physiological concentrations (i.e. nM range), using radiolabelled substrates. 10.1.1 Degradation of allatostatins in the haemolymph

Haemolymph also possesses appreciable peptidase activity, and allatostatins can be metabolized by these enzymes. For example, dipstatin 7 (allatostatin I) is rapidly degraded to two products, with amino acid compositions corresponding to Leu-Tyr-Gly-Phe-Gly-Leu and to Tyr-Gly-Phe-Gly-Leu (C. Garside and S. S. Tobe, unpublished). These two products represent the C-terminal message sequence of allatostatins and their presence suggests that, as might be expected, degradation of the peptides occurs through the action of an arninopeptidase. The time course of degradation of dipstatin 7 (I) is shown in Fig. 11. Product 1 represents the hexapeptide and Product 2 the pentapeptide. Dipstatin 7 (I) is rapidily metabolized by haemolymph and the half-life of the full peptide is about 20 min. In vitro half-life for other allatostatins in haemolymph, including dipstatins and callatostatins, is shown in Table 6. The apparent half-life of dipstatin 5(IV) is significantly greater than that for other peptides, with a half-life of 787 min. Callatostatin 5 , an allatostatin-like peptide isolated from Calfiphora vomitoria (Duve et al. 1993; see Section 8.1), also shows a considerable degree of resistance to degradation, perhaps as a result of the methionine modification at the C terminus (C. Garside and S. S. Tobe, unpublished). The large differences in rates of degradation are surprising and indicate that the haemolymph peptidases show a high degree of specificity to the N-terminal address portion of the peptides. It is of interest that the major products of degradation are the message portions of the allatostatins (be they the hexapeptide or pentapeptide C termini; see Section 3.1). These peptides both show some potency and full efficacy in terms of inhibition of JH biosynthesis in vitro (Stay et al., 1991b). If these peptides persist in the haemolymph, they may modify the activity of the other allatostatins. The amino exopeptidase inhibitors, amastatin and bestatin, are both effective in inhibiting the breakdown of selected dipstatins (C. Garside and S. S. Tobe, unpublished). Studies on degradation of allatostatins in haemolymph have provided insights into routes of degradation, but such information is confounded by the occurrence of allatostatins in the haemolymph (Yu et a f . , 1993; Woodhead et a f . , 1993). For example, allatostatins occur in the haemolymph in the nanomolar range despite the existence of mechanisms for their degradation. This suggests that either the allatostatins are protected from degradation in the haemolymph or that such concentrations are far below those required for optimal enzyme catabolic activity. This question will only be resolved as radiolabelled substrates become available.

ALLATOSTATINS: IDENTIFICATION, PRIMARY STRUCTURES, AND FUNCTIONS

-dipstatin 7 - - -Product # 1 .----Product #2

80 h

bp

v

325

-

0

al

h

60

-

40

-

Y

0

aJ

a

20

0

0

20

40

60

80

100

120

Time (min)

FIG. 1 1 Time course of degradation of dipstatin 7 (allatostatin I) by haemolymph from Day 5 mated female Diploptera punctara and the appearance of its cleavage products. Product 1 amino acid composition was: 2 Leu, 1 Tyr, 2 Gly and 1 Phe, indicating -Leu-Tyr-Gly-Phe-Gly-Leu C-terminal sequence was the cleavage product. Product 2 amino acid composition was 1 Leu, 1 Tyr, 2 Gly and 1 Phe indicating the -Tyr-Gly-Phe-Gly-Leu C-terminal sequence as the cleavage product (from C. Garside and S. S. Tobe, unpublished). Dipstatin 7 (6pM) was incubated with haemolymph and the peptides extracted and separated by C18RP HPLC. Amino acid composition of the major products was determined by PTC amino acid analysis, using pre-column derivatization.

TABLE 6 Metabolism of allatostatins by haemolymph of Day 5 mated female Diploptera punctata" Allatostatin

Half-life (min)

Dipstatin 5 (IV) Dipstatin 7 (I) Dipstatin 9 (11)

787.5 13.3 7.2

Callatostatin 5

315.1

"Quantity of allatostatins remaining in the haemolymph incubations was determined at selected intervals by separation of extracts by HPLC followed by amino acid analyses of the appropriate HPLC fractions, to confirm the identity of the peaks. Methods as described in Fig. 11. Relative concentration was determined by comparison of relative peak area to peak area of authentic allatostatins at time 0.

BARBARA STAY et al.

326

10.2

MODE OF ACTION OF ALLATOSTATINS

The array of allatostatins and the likely occurrence of multiple receptor subtypes (see Section 5.2) raises the possibility of multiple signal transduction mechanisms for these peptides in terms of inhibition of JH biosynthesis. Several lines of evidence support this interpretation, particularly those obtained from experiments involving treatment of CA with either second messengers or agents that elevate levels of second messengers within the glands. This evidence is summarized below. 10.2.1 Agents which alter second messenger levels Treatment of CA in vifro with agents known to alter intracellular levels of second messengers exerts a pronounced effect on JH biosynthesis. Extracellular calcium, for example, directly influences rates of JH production (Kikukawa et al., 1987). In its absence, JH production virtually ceases whereas concentrations between 3 and 5 mM provide optimal rates of biosynthesis. Extracellular calcium no doubt influences intracellular levels of this second messenger; intracellular calcium concentration can be affected directly by calcium conductance and the demonstration of voltagedependent calcium channels in the cell membranes of the CA suggests strongly that the calcium concentrations within cells of the gland are regulated in part by these channels in concert with appropriate outward conductance, such as calcium-dependent potassium channels (Thompson and Tobe, 1986, 1990; McQuiston and Tobe, 1991a,b). For example, positive current injection or treatment of CA with calcium channel blockers (cobalt, cadmium, verapamil) evokes action potentials in CA cells (Thompson and Tobe, 1986; McQuiston et al., 1990). In addition, analogues of cAMP and cGMP directly affect the duration of these action potentials, with cAMP increasing and cGMP decreasing duration (McQuiston and Tobe, 1991b). Treatment of CA with forskolin and the subsequent elevation in cAMP levels results in an inhibition in JH biosynthesis; this effect is particularly evident at the end of the first vitellogenic cycle (Meller e f al., 1985). Similarly, treatment of CA with 8-bromo-cGMP strongly inhibits JH production during the vitellogenic cycle, although the developmental pattern of responsiveness to this agent is distinct from that of cAMP (Tobe, 1990; Z. F. Chen and S. S. Tobe, unpublished). This is not surprising in view of the very different effects which these agents have on action potential duration in CA cells (see above). Changes in receptor-coupled signalling mechanisms may thus be important components in effecting the dramatic changes in sensitivity of CA to allatostatins. As well, there are age-dependent changes in the effects of phorbol esters on JH production (Feyereisen and Farnsworth, 1987) and on action potential duration during the vitellogenic cycle

ALLATOSTATINS: IDENTIFICATION, PRIMARY STRUCTURES, AND FUNCTIONS

327

(McQuiston and Tobe, 1991b). In the latter case, the differences between Day 1 and Day 6 mated females are particularly striking and correspond to two ages which differ significantly in their responsiveness to allatostatins (Pratt et al., 1990; Stay et al., 1991a). Such data suggest that regulation of intracellular calcium concentration, by way of voltage-dependent calcium channels, outward conductances and regulation of release of calcium from intracellular stores (for example by inositides: Berridge, 1993), can alter JH production significantly. Modulation of intracellular calcium concentrations provides an obvious site of action for allatostatins. There appear to be multiple and parallel signalling mechanisms at work in the regulation of JH biosynthesis by allatostatins. These multiple mechanisms may function both at specific times, resulting in: (a) differences in responsiveness to different allatostatins, and (b) dose-dependent changes in responsiveness of CA to specific allatostatins and also throughout development, resulting in (c) changes in sensitivity during development to specific allatostatins (Pratt et a!. , 1990, 1991a; Stay et al., 1991a). This is consistent with the existence of multiple allatostatin receptors (see Section 5.2) Integration of the various allatostatic signals therefore could occur at the level of receptor subtypes on the cell membranes of CA cells (see also Li et al., 1993, for an example of parallel signalling systems in vertebrate neurones). 10.2.2 Measurement of cyclic nucleotides in CA Attempts to measure directly changes in either CAMP or cGMP following treatment with allatostatins have been unsuccessful (Cusson et al., 1992). Although this might indicate that second messengers are not involved in the signal transduction process for allatostatins, this appears unlikely in view of the major influence that these compounds, their analogues or agents have on JH biosynthesis (Tobe el a!., 1994). Rather, the inability to measure significant changes following treatment may be attributable to the rapidity of changes, to changes in flux or pool size of the second messengers and/or to the inability to detect relatively small changes in the quantities of these compounds. Small changes in the apparent turnover of second messengers could in fact represent very large changes in specific cellular compartments or specific populations of cells (see Yuen and Garbers, 1992). The quasi-additive response of CA to treatment with either different allatostatins or with widely different concentrations of the same allatostatin (Pratt et al., 1990, 1991b) indicates that there is likely to be an interaction between these multiple parallel signalling pathways. For example, modulation of intracellular calcium could influence the production of cyclic nucleotides or protein kinase C and subsequent phosphorylation reactions (Li et a f . , 1993). Such modulation could result in a considerable ampiifica-

328

BARBARA STAY et al.

tion of the allatostatic message beyond the initial signal transduction process (see Li et al., 1993). 11 Factors regulating release of allatostatin

Study of regulation of allatostatin release is possible now since the peptides have been identified and sensitive methods have been developed to quantitate them. However, this subject is just beginning to be investigated. Several relevant observations are mentioned in previous sections. One is the alteration in dynamics of neuropeptide accumulation in neurosecretory cells resulting from parasitization of M . sexta larvae by a braconid wasp (see Section 6.1). Neuropeptides of many kinds, including allatostatin, accumulated in the brain in parasitized larvae, very likely as a result of arrested release (Zitnan et al., submitted). A second observation concerns the apterous mutant of Drosophila (see Section 1.4). The corpora allata of apterous flies produce JH bisepoxide at a very low level and the brains fail to produce extractable allatostatic factor (Altaratz et al., 1991). These authors suggested that allatostatin production is under positive regulation by JH bisepoxide. However, the apterous gene product, a developmental regulatory protein (Cohen et al., 1992), may control the development of allatostatin-producing neurones. A third observation is the changing titre of allatostatins in brain and haemolymph of D. punctata (Yu et al., 1993) (see Section 7.3) which is indicative of regulation of allatostatin release, although it does not indicate the factors involved in this regulation. 11.1

NERVE SECTION, JH ANALOGUE, OVARY

Several factors that do influence allatostatin release were demonstrated by measuring haemolymph titre of allatostatin following experimental treatments of male D. punctata (Stay et al., 1994). Males treated with the JH analogue, hydroprene, showed a strong reduction in JH synthesis but a much less pronounced reduction occurred in animals in which nerves between the brain and the CA were severed. Measurement of haemolymph allatostatin showed a much higher concentration of allatostatin following J H analogue treatment in animals in which the nerves to the CA had been severed (>3.0 nM vs <0.04 nM). These results suggest that JH synthesis by the C A is normally inhibited as a result of feedback from JH titre by delivery of allatostatin directly to the CA and that when the nerves to CA are severed allatostatin is released into the haemolymph. Ordinarily males do not produce vitellogenin but as a result of the JH analogue treatment an implanted ovary accumulates vitellin yolk. The effect of implanting an ovary into males with denervated CA and treated with JH analogue was a reduction in the concentration of a haemolymph allatostatin compared with controls similarly treated but not implanted with an ovary. This observation

ALLATOSTATINS: IDENTIFICATION, PRIMARY STRUCTURES, AND FUNCTIONS

329

shows that growing ovaries inhibit the release of allatostatin and suggests that this may be one factor which results in increased JH synthesis in females with vitellogenic ovaries (Rankin and Stay, 1984).

12 Conclusions

The search for neuropeptide regulators of JH synthesis by the CA has led to the identification of a peptide from the moth Munducu sexfa and quite a different family of peptides from the cockroaches Diplopferu puncfufa and Periplunefu urnericuna. These were called allatostatins because they inhibit JH synthesis by CA in an in v i m bioassay. The moth and cockroach peptides each inhibit JH synthesis in their own species, but neither is effective in the reciprocal species. Thus although all insects have CA that produce very similar juvenile hormones they do not use similar peptides to regulate the CA. However, M . sexfa allatostatin is an effective inhibitor of JH synthesis in other Lepidopterans and D. puncfatu allatostatins inhibit cricket CA. Therefore structurally related peptides with similar allatostatic function will probably be found in closely related species. How the allatostatins effect the reduction in JH synthesis is an obvious question that remains to be answered. Isolation of the receptors for the allatostatins will be important in establishing how allatostatins act on the CA as well as other organs. It should be pointed out that allatostatins are surely only one of many regulatory factors acting on the CA. Homologous immunocytochemical localization of allatostatins suggests two obvious actions in addition to regulation of the CA. One is nervemuscle function and the other is nerve-nerve function. The former has been demonstrated in D . puncfutu; an inhibition of spontaneous and/or protolin induced visceral muscle contraction. The nerve-nerve action remains to be demonstrated in insects but has been demonstrated in crustaceans. Although the allatostatic function of the identified peptides may be limited to related species, allatostatin-like peptides appear to occur in both closely and distantly related insects, in other arthropods and in other phyla, as demonstrated by immunocytochemical localization. In the blowfly Culfiphoru vomitoriu several peptides with a terminal -Y-X-F-G-L/M-amide sequence have been isolated. In a crab the D. puncfufu allatostatins are effective inhibitory modulators of motor neurone endogenous discharge and also act at the neuromuscular junction. Similar studies in other organisms will contribute to our understanding of the function and evolution of these allatostatin-like peptides. There are undoubtedly many other allatostatic peptides to be discovered. It will be interesting to learn what the primitive and primary functions of these peptides are and how they came to regulate CA .

330

BARBARA STAY et al.

Acknowledgements

We would like to thank Dr Andrea Woodhead for critical reading of the manuscript and all of those investigators who have made their recent findings available for this review. We also thank Dean Abel, Kuen Chan and Marjorie Zalesky for their technical assistance. References Agricola, H., Schildberger, K . , Schmidt, A , , Naumann, W., Reissmann, S., Huber, F. and Penzlin, H. (1992). Immunocytochemical distribution of allatostatin in the nervous system of Periplaneta americana. In “Rhythmogenesis in Neurons and Networks” (Eds N. Elsner and D. W. Richter), p. 494. Geo. Thieme, Stuttgart. Altaratz, M., Applebaum, S. W., Richard, D. S . , Gilbert, L. I . and Segal, D. (1991). Regulation of juvenile hormone synthesis in wild-type and apterous mutant Drosophila. Mol. Cell. Endocrinol. 81, 205-216. Baehr, J. C. (1973). Controle neuroendocrine du fonctionnement du corpus allatum chez Rhodnius Drolixus. J . Insect Phvsiol. 19. 1041-1055. Berridge, M . J. (i993). Inositol trisphosphate ‘and calcium signalling. Nature 361, 3 15-325. Bhaskaran, C., Jones, G. and Jones, D. (1980). Neuroendocrine regulation of corpus allatum activity in Manduca sexta: Sequential neurohormonal and nervous inhibition in the last-instar larva. Proc. Natl. Acad. Sci. USA 7 7 , 4407-4411. Bhaskaran, G., Dahm, K . H., Barrera, P., Pacheco, J . L., Peck, K. E. and Muszynska-Pytel, M. (1990). Allatinhibin, a neurohormonal inhibitor of juvenile hormone biosynthesis in Manduca sexta. Gen. Comp. Endocrinol. 78, 12S136. Bollenbacher, W. E., Gray, R. S . , Muehleisen, D. P., Regan, S. A. and Westbrook, A. L. (1993). The biology of the prothoracicotropic hormone peptidergic neurons in an insect. A m . Zool. 33, 316-323. Brown, M. R., Crim, J. W. and Lea, A. 0. (1986). FMRFamide- and pancreatic polypeptide-like immunoreactivity of endocrine cells in the midgut of a mosquito. Tissue Cell 18, 419-428. Cassier, P. (1990). Morphology, histology, and ultrastructure of JH-producing glands in insects. In “Morphogenetic Hormones of Arthropods”, Vol. 1, Part 2 (Ed. A. P. Gupta), pp. 84-194. Rutgers Univ. Press, New Brunswick, NJ. Cohen, B., McGuffin, M. E., Pfeifle, G., Segal, D. and Cohen, S. M. (1992). apterous, a gene required for imaginal disc development in Drosophila encodes a member of the LIM family of developmental regulatory proteins. Genes Dev. 6, 715-729. Copenhaver, P. F. and Truman, J . W. (1986a). Metamorphosis of the cerebral neuroendocrine system in the moth Manduca sexta. J . Comp. Neurol. 249, 186-204. Copenhaver, P. F. and Truman, J. W. (1986b). Identification of the cerebral neurosecretory cells that contain eclosion hormone in the moth Manduca sexfa. J . Neurosci. 6, 1738-1747. Cusson, M . , Prestwich, G. D., Stay, B. and Tobe, S. S. (1991). Photoaffinity labeling of allatostatin receptor proteins in the corpora allata of the cockroach, Diploptera punctata. Biochem. Biophys. Res. Commun. 181, 736-742. Cusson, M., Prestwich, G. D., Stay, B. and Tobe, S. S. (1992). Evidence for allatostatin receptor proteins in the corpora allata of the cockroach Diploptera

ALLATOSTATI NS : IDENTIFICATION, PRIMARY STRUCTURES, AND FUNCTIONS

331

punctata, as demonstrated by photoaffinity labeling. In “Insect Juvenile Hormone Research: Fundamental and Applied Approaches” (Eds B. Mauchamp, F. Couillaud and J. C. Baehr), pp. 183-191. INRA, Paris. Donly, B. C., Ding, Q., Tobe, S. S. and Bendena, W. G. (1993). Molecular cloning of the gene for the allatostatin family of neuropeptides from the cockroach Diploptera punctata. Proc. Natl. Acad. Sci. USA 90, 8807-881 1. Duve, H., Thorpe, A., Yagi, K. J., Yu, C. G. and Tobe, S. S. (1992). Factors affecting the biosynthesis and release of juvenile hormone bisepoxide in the adult blowfly Calliphora vomitoria. J. Insect Physiol. 38, 575-585. Duve, H., Johnsen, A . H., Scott, A. G., Yu, C. G., Yagi, K. J., Tobe, S. S. and Thorpe, A. (1993). Callatostatins: Neuropeptides from the blowfly Caffiphora vomitoria with sequence homology to cockroach allatostatins. Proc. Natl. Acad. Sci. USA 90, 2456-2460. Edwards, J. P., Chambers, J., Short, J. E., Price, N. R., Weaver, R. J., Abraham, L. and Walter, C. M. (1990). Endogenous juvenile hormone 111 titres and in vitro rates of hormone biosynthesis by corpora allata during the reproductive cycle of adult female Periplaneta americana. In “Chromatography and Isolation of Insect Hormones and Pheromones” (Eds A. R. McCaffery and I. D. Wilson), pp. 3-8. Plenum Press, New York. Eipper, B. A., Stoffers, D. A . and Mains, R. E. (1992). The biosynthesis of neuropeptides: peptide a-amidation. Ann. Rev. Neurosci. 15, 57-85. Engelmann, F. (1957). Die Steuerung der Ovarfunktion bei der ovoviviparen Schabe Leucophaea maderae (Fabr.). J . Insect Physiol. 1, 257-278. Engelmann, F. (1959). The control of reproduction in Diploptera punctata (Blattaria). Biol. Bull. Woods Hole 116, 406-419. Feyereisen, R. and Farnsworth, D. E. (1987). Inhibition of insect juvenile hormone synthesis by phorbol 12-myristate 13-acetate. FEBS Lett. 222, 345-348. Feyereisen, R. and Tobe, S. S. (1981). A rapid partition assay for routine analysis of juvenile hormone release in insect corpora allata. Anal. Biochem. 111, 372-375. Fonagy, A , , Schoofs, L . , Proost, P., van Damme, J., Bueds, H. and de Loof, A. (1992). Isolation, primary structure and synthesis of neomyosuppressin, a myoinhibiting neuropeptide from grey fleshfly, Neobellieria bullata. Comp. Biochem. Physiol. 102C, 239-245. Frohman, M. A , , Dush, M. K. and Martin, G. R. (1988). Rapid production of full-length cDNAs from rare transcripts: Amplification using a single gene-specific oligonucleotide primer. Proc. Natf. Acad. Sci. USA 85, 8998-9002. Granger, N. A. and Janzen, W. P. (1987). Inhibition of Manduca sexta corpora allata in vitro by a cerebral allatostatic neuropeptide. Mol. Cell. Endocrinol. 49, 237-248. Granger, N. A , , Bollenbacher, W. E. and Gilbert, L. I. (1981). An in vitro approach for investigating the regulation of the corpora allata during larval-pupal metamorphosis. In “Current Topics in Insect Endocrinology and Nutrition” (Eds G. Bhaskaran, S. Friedman and J . G. Rodriguez), pp. 83-105. Plenum Press, New York. Granger, N. A., Niemiec, S. M., Gilbert, L. I. and Bollenbacher, W. E. (1982). Juvenile hormone synthesis in vitro by larval and pupal corpora allata of Manduca sexta. Mol. Cell. Endocrinol. 28, 587-604. Hayes, T. K., Guan, X-C., Johnson, V . , Strey, A. and Tobe S. S. (in press). Structure-activity studies of allatostatin 4 on the inhibition of juvenile hormone biosynthesis by corpora allata: the importance of individual side-chains and stereochemistry. Peptides. Hertel, W . and Penzlin, H. (1992). Function and modulation of the antenna1 heart of Periplaneta americana (L.). Acta Biol. Hung. 43, 113-125.

332

BARBARA STAY et al.

Hertel. W.. Pass, G. and Penzlin, H. (1985). Electrophysiological investigation of the antennal heart of Periplaneta arnericana and its reactions to proctolin. J . Insect Physiol. 31, 563-572. Hertel. W.. Pass. G. and Penzlin, H. (1988). The effects of the neuropeptide proctolin and of octopamine on the antennal heart of Periplaneta arnericana. Syrnp. Biol. Hung. 36, 351-362. Hodkova, M . (1977). Function of the neuroendocrine complex in diapausing Pyrrhocoris apterus females. J . Insect Physiol. 23, 23-28. Hodkova, M. (1979). Humoral and nervous inhibition of reproduction by brain in diapausing females of Pyrrhocoris apterus L. (Hemiptera). Zod. Jb. Physiol. 83, 126-136. Holman, G. M., Cook, B. J . and Nachman, R. J . (1986). Isolation, primary structure and synthesis of leucomyosuppressin, an insect neuropeptide that inhibits spontaneous contractions of the cockroach hindgut. Cornp. Biochern. Physiol. 85C, 329-333. Homberg, U . , Davis, N. T. and Hildebrand, J . G. (1991). Peptideimmunocytochemistry of neurosecretory cells in the brain and retrocerebral complex of the sphinx moth Manduca sexta. J . Cornp. Neurol. 303, 35-52. Iwanaga, T., Fujita, T.. Takeda, N.. Endo, Y . and Lederis, K. (1986). Urotensin I-like immunoreactivity in the midgut endocrine cells of the insects Gryllus bimaculatus and Periplaneta arnericana. Cefl Tissue Res. 244, 565-568. Johnson, G. D., Stay, B. and Chan. K . K. (1993). Structure-activity relationships in corpora allata of the cockroach Diploptera punctata: roles of mating and the ovary. CeN Tissue Res. 274, 279-293. Kataoka, H.. Troetschler, R . G.. Kramer, S. J . , Cesarin. B. J . and Schooley, D. A. (1987). Isolation and primary structure of the eclosion hormone of the tobacco hornworm, Manduca sexta. Biochern. Biophys. Res. Cornrnun. 146, 746-750. Katz, P. S., Eigg, M. H. and Harris-Warrick, R. M. (1989). Serotonergickholinergic muscle receptor cells in the crab stomatogastric nervous system. I . Identification and characterization of the gastropyloric receptor cells. J . Neurophysiol. 62, 558-570. Keshishian, H. and O’Shea, M. (1985). The distribution of a peptide neurotransmitter in the postembryonic grasshopper central nervous system. J . Neurosci. 5 , 992-1004. Khan, M. A , , Koopmanschap, A. B. and de Kort, C. A. D. (1983). The relative importance of nervous and humoral pathways for the control of corpus allatum activity in the adult Colorado potato beetle, Leptinotarsa decernlineata (Say). Gen. Cornp. Endocrinol. 52, 214-221. Kikukawa, S., Tobe, S. S., Solowiej, S., Rankin, S. M. and Stay, B. (1987). Calcium as a regulator of juvenile hormone biosynthesis and release in the cockroach Diploptera punctata. Insect Biochem. 17, 179-187. Kramer, S. J., Toschi, A , . Miller, C . A., Kataoka, H., Quistad, G. B., Li, J. P., Carney, R. L. and Schooley, D. A . (1991). Identification of an allatostatin from the tobacco hornworm Manduca sexta. Proc. Natl. Acad. Sci. USA 88,9458-9462. Lange, A. B., Orchard, I. and Te Brugge, V. A . (1991). Evidence for the involvement of a SchistoFLRF-amide-like peptide in the neural control of locust oviduct. J . Cornp. Physiol. A 168, 383-391. Lange, A. B., Chan, K. K. and Stay, B. (1993). Effect of allatostatin and proctolin on antennal pulsatile organ and hindgut muscle in the cockroach, Diploptera punctata. Arch. Insect Biochern. Physiol. 24. 79-92. Li, M., West, J. W., Numann, R . , Murphy, B. J., Scheuer, T. and Catterall, W. A. (1993). Convergent regulation of sodium channels by protein kinase C and CAMP-dependent protein kinase. Science 261, 1439-1442.

ALLATOSTATINS: IDENTIFICATION, PRIMARY STRUCTURES, AND FUNCTIONS

333

Lococo. D. J. and Tobe, S. S. (1984). Neuroanatomy of the retrocerebral complex, in particular the pars intercerebralis and partes laterales in the cockroach Diploptera punctata Eschscholtz (Dictyoptera: Blaberidae). Int. J. Insect Morphol. Ernbryol. 13, 65-76. Loftfield. R. B. and Eigner, E. A. (1969). Molecular order of participation of inhibitors (or activators) in biological systems. Science 164. 305-308. Liischer. M. and Engelmann, F. (1960). Histologische und experimentelle untersuchungen iiber die Auslosung der Metamorphose bei Leucophaea rnaderae (Orthoptera). J . Insect Physiol. 5 , 240-258. Maddrell, S. H. P. and Gee, J. D. (1974). Potassium-induced release of the diuretic hormones of Rhodnius prolixlrs and Glossina austeni: Ca dependence, time course and localization of neurohaemal areas. J . Exp. Biol. 61, 155-171. Marder, E. and Hooper, S. L. (1985). Neurotransmitter modulation of the stomatogastric ganglion of decapod crustaceans. I n “Model Neural Networks and Behavior” (Ed. A. I. Selverston), pp. 319-337. Plenum Press, New York. McKittrick, F. A. (1964). “Evolutionary Studies of Cockroaches.” Memoir 389. Cornell University Agricultural Experiment Station. McQuiston, A. R. and Tobe, S. S. (1991a). Evidence for a calcium-dependent outward conductance in endocrine cells of the cockroach, Diploptera punctata. Physiol. Entornol. 16, 429-433. McQuiston, A. R. and Tobe, S. S. (1991b). Developmental changes and second messenger modulation of the electrical properties of cells of the cockroach corpora allata. J. Insect Physiol. 37, 223-229, McQuiston, A. R . , Thompson, C . S. and Tobe, S. S. (1990). Voltage-dependent sodium and calcium channels of the corpora allata of the cockroach Diploptera punctata. I n “Insect Neurochemistry and Neurophysiology, 1989” (Eds A. B. Borkovec and E. P. Masler), pp. 321-324. Humana Press, Clifton, NJ. Meller, V. H., Aucoin, R. R., Tobe, S. S. and Feyereisen, R. (1985). Evidence for an inhibitory role of cyclic AMP in the control of juvenile hormone biosynthesis by cockroach corpora allata. Mol. Cell. Endocrinol. 43, 155-163. Montuenga, L. M., Barrenechea, M. A., Sesma, P., Lopez, J . and Vazques, J. J. (1989). Ultrastructure and immunocytochemistry of endocrine cells in the midgut of the desert locust, Schistocerca gregaria (Forskal). Cell Tissue Res. 258, 577-583. Nassel, D. R. and O’Shea, M. (1987). Proctolin-like immunoreactive neurons in the blowfly central nervous system. J. Comp. Neurol. 265, 437-454. Neuhauser, T., Sorge, D., Stay, B. and Hoffmann, K. H. (1994). Responsiveness of the adult cricket (Gryllus birnaculatus and Acheta dornesticus) retrocerebral complex to allatostatin-1 from a cockroach, Diploptera punctafa. J . Cornp. Physiol. B 164: 23-31. Orchard, I., Belanger, J. H. and Lange, A. B. (1989). Proctolin: a review with emphasis on insects. J. Neurobiol. 20, 470-496. O’Shea, M. and Adams, M. (1986). Proctolin: from “gut factor” to model neuropeptide. A d v . Insect Physiol. 19, 1-28. Pass, G. (1985). Gross and fine structure of the antennal circulatory organ in cockroaches (Blattodea, Insecta). J. Morphol. 185, 255-268. Pass, G., Agricola, H., Birkenbeil, H. and Penzlin, H. (1988a). Morphology of neurones associated with the antennal heart of Periplaneta americana (Blattodea, Insecta). Cell Tissue Res. 253, 319-326. Pass, G., Sperk, G., Agricola, H., Baumann, E. and Penzlin, H. (1988b). Octopamine in a neurohaemal area within the antennal heart of the American cockroach. J. Exp. Biol. 135, 495498.

334

BARBARA STAY et al.

Paulson, C. R. and Stay, B. (1987). Humoral inhibition of the corpora allata in larvae of Diploptera punctata: role of the brain and ecdysteroids. J . Insect Physiol. 33, 61HJ22.

Pipa, R. L. (1978). Locations and central projections of neurons associated with the retrocerebral neuroendocrine complex of the cockroach Periplaneta americana (L.). Cell Tissue Res. 193, 443-455. Poras, M., Baehr, J . C. and Cassier, P. (1983). Control of corpus allatum activity during the imaginal diapause in females of Locusta migratoria L. Int. J . Invert. Reprod. 6, 111-122. Pratt, G. E.. Farnsworth, D. E., Siegel, N. R., Fok, K. F. and Feyereisen, R. (1989). Identification of an allatostatin from adult Diploptera punctata. Biochem. Biophys. Res. Commun. 163, 1243-1247. Pratt, G. E., Farnsworth, D. E. and Feyereisen. R. (1990). Changes in the sensitivity of adult cockroach corpora allata to a brain allatostatin. Mol. Cell. Endocrinol. 70, 185-195. Pratt, G. E., Farnsworth, D. E., Fok, K. F., Siegel, N. R., McCormack, A. L., Shabanowitz, J . , Hunt, D. F. and Feyereisen, R. (1991a). Identity of a second type of allatostatin from cockroach brains: An octadecapeptide amide with tyrosine-rich address sequence. Proc. Natl. Acad. Sci. USA 88, 2412-2416. Pratt, G. E., Farnsworth, D. E., Siegel. N. R.. Fok, K. F. and Feyereisen, R. (1991b). Two types of allatostatic peptides from brains of the cockroach Diploptera punctata. In “Insect Neuropeptides” (Eds J . J . Menn, T. J . Kelly and E. P. Masler), pp. 177-192. ACS Symposium Series 453, American Chemical Society, Washington, DC. Rankin, S. M. and Stay, B. (1984). The changing effect of the ovary on rates of juvenile hormone synthesis in Diploptera punctata. Gen. Comp. Endocrinol. 54, 382-388. Rankin, S. M. and Stay, B. (1985). Regulation of juvenile hormone synthesis during pregnancy in the cockroach, Diploptera punctata. J . Insect Physiol. 31, 145-157. Rankin, S. M. and Stay, B. (1987). Distribution of allatostatin in the adult cockroach, Diploptera punctata and effects on corpora allata in vitro. J. Insect Physiol. 33, 551-558. Rankin, S. M., Stay, B., Aucoin, R. R. and Tobe, S. S . (1986). In vitro inhibition of juvenile hormone synthesis by corpora allata of the viviparous cockroach, Diploptera punctata. J . Insect Physiol. 32, 151-156. Richard, D. S.. Applebaum, S. W. and Gilbert. L. 1. (1990). Allatostatic regulation of juvenile hormone production in vitro by the ring gland of Drosophila melanogaster. Mol. Cell. Endocrinol. 68, 153-161. Robb, S . , Packman, L. C. and Evans, P. D. (1989). Isolation, primary structure and bioactivity of schistoFLRF-amide, a FMRF-amide-like neuropeptide from the locust, Schistocerca gregaria. Biochem. Biophys. Res. Commun. 160, 85G856. Riiegg, R. P., Lococo, D. J . and Tobe, S. S. (1983). Control of corpus allatum activity in Diploptera punctata: roles of the pars intercerebralis and pars lateralis. Experientia 39, 1329-1334. Saiki, R. K . , Gelfand, D. H., Stoffel, S., Scharf, S. J., Higuchi, R., Horn, G. T., Mullis, K. B. and Erlich, H. A. (1988). Primer-directed enzymatic amplification of DNA with a thermostable DNA polymerase. Science 239, 487491. Scharrer, B. (1946). Section of the nervi corporis cardiaci in Leucophaea rnaderae (Orthoptera). Anat. Record 94, 977. Scharrer, B. (1952). Neurosecretion. XI. The effects of nerve section on the intercerebralis-cardiacum-allatum system of the insect Leucophaea maderae. Biol. Bull. Woods Hole 102. 261-272.

ALLATOSTATINS: IDENTIFICATION, PRIMARY STRUCTURES, AND FUNCTIONS

335

Scharrer, B. (1987). Insects as models in neuroendocrine research. Annu. Rev. Entomol. 32, 1-16. Schildberger, K. and Agricola, H. (1992). Allatostatin-like immunoreactivity in the brains of crickets and cockroaches. In “Rhythmogenesis in Neurons and Networks” (Eds N. Elsner and D. W. Richter), p. 489. Geo. Thieme, Stuttgart. Schneider, L. E. and Taghert, P. H. (1988). Isolation and characterization of a Drosophilu gene that encodes multiple neuropeptides related to Phe-Met-ArgPhe-NH2 (FMRFamide). Proc. Nutl. Acad. Sci. USA 85, 1993-1997. Skiebe, P. and Schneider, H. (1994). Allatostatin peptide in the crab stomatogastric nervous system: inhibition of the pyloric motor pattern and distribution of allatostatin-like immunoreactivity. J . Exp. Biol. (in press). Skiebe-Corrette, P., Jorge-Rivera, J. C. and Marder, E. (1993). The allatostatins influence the gastric system of the crab. SOC. Neurosci. Abstr. 19, 931. Stay, B. and Tobe, S. S. (1977). Control of juvenile hormone biosynthesis during the reproductive cycle of a viviparous cockroach. I. Activation and inhibition of corpora allata. Gen. Comp. Endocrinol. 33. 531-540. Stay, B. and Woodhead, A. P. (1993). Neuropeptide regulators of insect corpora allata. Am. Zool. 33, 357-364. Stay, B., Joshi, S. and Woodhead, A. P. (1991a). Sensitivity to allatostatins of corpora allata from larval and adult female Diplopteru punctutu. J . Insect Physiol. 37, 63-70. Stay, B., Woodhead, A. P.. Joshi, S. and Tobe, S. S. (1991b). Allatostatins, neuropeptide inhibitors of juvenile hormone synthesis in brain and corpora allata of the cockroach Diplopteru punctutu. In “Insect Neuropeptides” (Eds J. J. Menn. T. J. Kelly and E. P. Masler), ACS Symposium Series 453, pp. 164-176. American Chemical Society, Washington, DC. Stay, B., Chan, K. K. and Woodhead, A. P. (1992a). Allatostatin-immunoreactive neurons projecting to the corpora allata of adult Diplopteru punctutu. Cell Tissue Res. 270, 15-23. Stay, B., Woodhead, A. P. and Chan, K. K. (1992b). Immunocytochemical distribution of allatostatin in the brain and retrocerebral complex of Diplopteru punctutu. In “Insect Juvenile Hormone Research: Fundamental and Applied Approaches” (Eds B. Mauchamp. F. Couillaud and J. C. Baehr), pp. 257-263. INRA, Paris. Stay, B., Sereg Bachmann, J. A , , Stoltzman, C. A , , Fairbairn, S. E., Yu, C. G. and Tobe, S. S. (1994). Factors affecting allatostatin release in a cockroach: nerve section, juvenile hormone analog and ovary. J . Insect Physiol. 40, 365-372. Steel, C. G. H. and Davey, K. G. (1985). Integration in the insect endocrine system. In “Comprehensive Insect Physiology, Biochemistry and Pharmacology”, Vol. 8 (Eds G. A. Kerkut and L. I. Gilbert), pp. 1-35. Pergamon Press, Oxford. Steward, 0. and Banker, G. A. (1992). Getting the message from the gene to the synapse: sorting and intracellular transport of RNA in neurons. Trends Neurosci. 15, 18CL186. Szibbo, C. M. and Tobe, S. S. (1981). Cellular and volumetric changes in relation to the activity cycle in the corpora allata of Diplopteru punctata. J . Insect Physiol. 27, 655465. Taghert, P. H., Schneider, L. E. and O’Brien. M. A. (1990). Structure, expression, and evolution of the Drosophilu DPKQDFMRF-amide neuropeptide gene. In “Molecular Insect Science” (Eds H. H. Hagedorn, J. G. Hildebrand, M. G. Kidwell and J. H. Law), pp. 213-221. Plenum Press, New York. Tashiro, K . , Tada, H., Heilker, R., Shirozu, M., Nakano, T. and Honjo, T. (1993). Signal sequence trap: a cloning strategy for secreted proteins and type 1 membrane proteins. Science 261, 60M03.

336

BARBARA STAY et al.

Tautz, D. and Pfeife, C. (1989). A non-radioactive in situ hybridization method for the localization of specific RNAs in Drosophila embryos reveals translational control of the segmentation gene hunchback. Chromosoma 98, 81-85. Thompson, C. S. and Tobe, S. S. (1986). Electrical properties of membranes of cells of the corpora allata of the cockroach Diploptera punctata: evidence for the presence of voltage-sensitive calcium channels. In “Insect Neurochemistry and Neurophysiology, 1986” (Eds A. B. Borkovec and D. B. Gelman), pp. 375-378. Humana Press, Clifton, NJ. Thompson, C. S. and Tobe, S. S. (1990). Innervation and electrophysiology of the corpus allatum. In “Cockroaches as Models for Neurobiology: Applications in Biomedical Research” (Eds I. Huber, E. P. Masler and B. R. Rao), pp. 89-101. CRC Press. Boca Raton, FL. Tobe, S. S. (1980). Regulation of the corpora allata in adult female insects. In “Insect Biology of the Future: VBW 8 0 ’ (Eds M. Locke and D. S. Smith), pp. 345-367. Academic Press, New York. Tobe, S. S. (1990). Role of intracellular messengers in the regulation of juvenile hormone biosynthesis in the cockroach, Diploptera punctata. In “Progress in Comparative Endocrinology”, Vol. 342 (Eds A. Epple, C. G. Scanes and M. H. Stetson), pp. 174-179. Wiley-Liss, New York. Tobe, S. S. and Stay, B. (1977). Corpus allatum activity in vitro during the reproductive cycle of the viviparous cockroach, Diploptera punctata (Eschscholtz). Gen. Comp. Endocrinol. 31, 138-147. Tobe, S. S. and Stay, B. (1985). Structure and regulation of the corpus allatum. Adv. Insect Physiol. 18, 305432. Tobe, S. S . , Stay, B., Friedel, T.. Feyereisen, R. and Paulson, C. (1981). The role of the brain in regulation of the corpora allata in female Diploptera punctata. In “Juvenile Hormone Biochemistry” (Eds G. E. Pratt and G. T. Brooks), pp. 161-174. Elsevier, Amsterdam. Tobe, S. S., Riiegg, R. P., Stay, B. A , , Baker, F. C., Miller, C. A. and Schooley, D. A. (1985). Juvenile hormone titre and regulation in the cockroach Diploptera punctata. Experientia 41, 1028-1034. Tobe, S. S., Yu, C. G. and Bendena, W. G. (1994). Allatostatins, peptide inhibitors of juvenile hormone production in insects. In “Perspectives in Comparative Endocrinology” (Eds K. G. Davey, R. E. Peter and S. S. Tobe), pp. 12-19. Natl. Res. Council of Canada, Ottawa. Unni. B., Barrera, P., Muszynska-Pytel, M., Bhaskaran, G. and Dahm, K. H. (1993). Partial characterization of allatinhibin, a neurohormone of Munducu sextu. Arch. Insect Biochem. Physiol. 24. 171-183. Weaver, R. J. (1991). Profile of the responsiveness of corpora allata from virgin female Periplaneta americana to an allatostatin from Diploptera punctara. J. Insect Physiol. 37, 111-1 18. Weaver, R. J . and Freeman, Z . A. (1992). Regulation of juvenile hormone biosynthesis by brain allatostatins in adult stages of the cockroach Periplaneta americana. In “Insect Juvenile Hormone Research: Fundamental and Applied Approaches” (Eds B. Mauchamp, F. Couillaud and J. C. Baehr), pp. 83-95. INRA, Paris. Weaver, R. J., Freeman, Z. A , , Pickering, M. G. and Edwards, J. P. (1994). Identification of two allatostatins from the CNS of the cockroach Periplaneta americana: novel members of a family of neuropeptide inhibitors of insect juvenile hormone biosynthesis. Comp. Biochem. Physiol. (B) 107C. 119-127. Westbrook, A. L., Haire, M. E., Kier, W. M. and Bollenbacher, W. E. (1991). Three-dimensional architecture of identified cerebral neurosecretory cells in an insect. J. Morphol. 208, 161-174.

ALLATOSTATINS: IDENTIFICATION, PRIMARY STRUCTURES, AND FUNCTIONS

337

Westbrook. A . L.. Regan, S. A . and Bollenbacher, W . E. (1993). Developmental expression of the prothoracicotropic hormone in the CNS of tobacco hornworm Manduca sexta. J . Comp. Neurol. 327, 1-16. Wigglesworth, V. B. (1936). The function of the corpus allatum in the growth and reproduction of Rhodnius prolixus (Hemiptera). Q.J. Microsc. Sci. 79, 91-121. Willey. R. B. (1061). The morphology of the stomodeal nervous system in Periplaneta americana (L.) and other Blattaria. J . Morphol. 108, 219-261. Woodhead. A . P., Stay, B., Seidel, S. L., Khan, M. A . and Tobe, S. S. (1989). Primary structure of four allatostatins: Neuropeptide inhibitors of juvenile hormone synthesis. Proc. Natl. Acad. Sci. USA 86. 59974001. Woodhead, A . P., Stoltzman, C. A . and Stay, B. (1992). Allatostatins in the nerves of the antenna1 pulsatile organ muscle of the cockroach Diplopteru punctata. Arch. insect Biochem. Physiol. 20. 253-263. Woodhead. A . P., Asano. W. Y. and Stay, B. (1993). Allatostatins in the haemolymph of Diploptera punctata and their effect in vivo. J . Insect Physiol. 39, 1001-1 005.

Woodhead. A . P.. Khan. M . A , , Stay, B. and Tobe, S. S . (1994). Two new allatostatins from the brains of Diploptera punctatu. insect Biochem. Mol. Biol. 24, 257-263. Yagi, K. J., Konz, K. G . , Stay. B. and Tobe. S. S. (1991). Production and utilization of farnesoic acid in the juvenile hormone biosynthetic pathway by corpora allata of larval Diploptera punctata. Gen. Comp. Endocrinol. 81. 284-294. Yu, C . G., Stay, B., Joshi, S. and Tobe, S. S. (1993). Allatostatin content of brain, corpora allata and haemolymph at different developmental stages of the cockroach. Diploptera punctata: quantitation by ELISA and bioassay. J . Insect Physiol. 39, 111-122. Yuen, P. S. T. and Garbers, D. L. (1992). Guanylyl cyclase-linked receptors. Annu. Rev. Neurosci. 15, 1Y3-225. Zhu, X. X. and Oliver, J . H. ( 1093). Allatostatin-immunoreactive neurosecretory cells in the synganglion of the tick Dermacentor variabilis (Acari: Ixodidae). J . A m . Trop. Med. Hyg. 49, 312, abstract. Zitnan, D., Sauman. I. and Sehnal. F. (1993a). Peptidergic innervation and endocrine cells of insect midgut. Arch. insect Biochem. Physiol. 22, 113-132. Zitnan, D., Sehnal, F. and Bryant, P. J. (1993b). Neurons producing specific neuropeptides in the central nervous system of normal and pupariation-delayed Drosophilti. Dev. Biol. 156, 117-135. Zitnan, D . , Kramer, S. J. and Beckage, N. E . (submitted). Accumulation of neuropeptides in the cerebral neurosecretory system of Manduca sexta larvae parasitized by the braconid wasp Cotesia congregata.

This Page Intentionally Left Blank

Index Abraxas grossulariata, 45 Abruptex, 86 Acarina, 316-17 acetylcholine (ACh), 210 acetylcholinesterase (AChE), 210 achaete, 88, 90, 92 achaete-scute complex (AS-C), 82 Acheta domestica, 202 Acheta domesticus, 175, 303 Actias selene, 48 Adoxophyes orana, 7, 8 , 3 6 Adoxophyes orana fasciata, 36 Aedes taeniorhynchus, 38 Aeschna, 159, 201, 212 Aeschna tuberculifera, 166 Agrotis segetum, 45 allatostatin-immunoreactive cells, 303 distribution of, 296-306 allatostatins, 267-338 activity in interneuroneskhemical synapses, 31 1 address sequence, 280-2 amino acid sequences, 278 analogues of allatostatin IV, 282-4 and regulation of JH titre, 292-3 bioassay, 277 C-terminal analysis, 278 cockroaches, 299-303 conformational models of allatostatin IV, 284 crickets, 303 degradation, 323-325 developmental changes in corpora allata in response to, 287-9 distribution of, 303 duality of responses, 289-90 evidence for occurrence in brain, 268-7 1 factors regulating release, 328-9 gene for, 317-23

immunoreactivity to antisera, 314 isolation and characterization of receptors, 294-5 isolation procedure, 274-7 message sequence, 279-80 metabolism of, 325 mode of action, 326-8 multiple receptors, 295-6 neural and humoral pathways for action of, 291-2 possible factors contributing to changes in responsiveness of CA to, 291 primary structures, 272-3 radioimmunoassay ,277 receptors for, 293-6 redundancy in, 286-7 responsiveness to analogues of, 290-1 sensitivity of corpora allata to, 286-93 structure-activity studies, 279-86 synthetic, 278 amino acid sequences, 274 Ampelisca, 158 Amsacfa moorei, 30, 31 Anadevidia peponis, 229 Anax, 159, 160,201 Anax junius, 159, 166, 202 Androctonus australis, 18 antenna1 pulsatile organ muscle, 308 Antennapedia, 108 Antheraea B virus, 45 Anticarsia gemmatalis, 15 Apis, 108, 110, 116, 131,201 Apis andreniformis, 131, 133, 135 Apis armbrusteri, 133 Apis cerana, 127, 132, 133-4 Apis dorsata, 127, 132, 133, 134-5 Apisflorea, 127, 128, 132, 135 Apis koschevnikovi, 131, 132, 133 Apis laboriosa, 134

340

Apismellifera, 108, 109, 114, 115, 117, 120, 122, 124, 127-9, 132, 133, 135-8, 158, 190, 200, 202, 210, 222,316 Apis mellifera capensis, 130 Apis mellifera carnica, 130, 158 Apis mellifera liguistica. 115, 120, 138 Apis mellifera mellifera, 126, 138 apoptosis. 1 4 1 5 apterous mutant, 329 Arachnida, 316-17 Argia vivida, 171 Artogeia rapae, 20 AS-C, 88-90 asense, 88 Austracris, 159 Austracris guitulosa, 159 Autographa californica, 3-5 Bacillus thuringiensis, 18 Baculoviridae, 15 classification, 4 baculoviruses, 2-29 assessing safety of genetically modified insecticides, 2&1 biological control of insect pests, 15-22 classification, 3-5 early gene expression, 9-10 expression of foreign genes in insect cell lines, 29 expression vectors, 22-9 future experiments, 22 gene promoters, 11-13 genetic modification of insecticides, 1&19 host range. 2-3 immediate-early (IE) genes, 9 insect cell lines, 28 insecticide improvement techniques, 19-20 isolation, 2-3 late genes, 10-1 1 multiple expression vectors, 25 non-occluded, 4 past field release experiments, 21-2 post-translational processing in insect cells, 27-8 replication in viiro, 9 replication in vivo, 5 4 selection of recombinant, 26-7 structure, 3 4

INDEX

transmission between hosts, 6 8 very late genes, 11 vs. chemical insecticides, 16 Barathra brassicae, 7-8, 8 bees ocellar tract of, 193 spiking and non-spiking L-neurones, 220-2 see also honeybee Bibionidae, 154 Big brain, 84 biological control of insect pests, 15-22 Biston petularia. 45 black beetle virus (BBV), 46, 47 Boettcherisca peregria, 158 Bombus lucorum, 132 Bombyx mori, 3, 4 , 7 , 14, 17,23, 36,45 Boolarra virus (BoV), 46 brain, allatostatins in, 26&71 Bulinus globosus, 317 callatostatins, 3 12-14 Calliphora, 157, 164-6, 171, 187, 194, 20 1 CaUiphora ery throcephala , 157, 202, 206, 220 Calliphora vomitoria, 3 12-1 5 , 324, 329 Calotermes, 156 Cancer borealis, 311, 316 Cataglyphis, 168, 177, 178 Chilo suppressalis, 38 Chironomus luridus, 30 chloramphenicol acetyl transferase (CAT), 10 Choristoneura, 32 Choristoneura fumiferana, 35, 41 Chrinomus luridus, 31 cockroaches allatostatins, 299-303 dynamics of spiking responses of L-neurones in, 222-5 ocellar tract of, 189 Colias euryiheme, 7 compound eyes, 152, 154, 164, 194, 201 corpora allata (CA), 268-71, 275 developmental changes in response to allatostatins, 287-9 measurement of cyclic nucleotides in, 327 sensitivity to allatostatins, 286-93 coxsackievirus B3 (cVB3) , 47 Creaiotos transiens. 175

INDEX

cricket paralysis virus, 45 crickets, allatostatins, 303 Crustacea, 316 Culex tritaeniorhynchus, 44 cyclic nucleotides, measurement in CA. 327 Cydia pomonella, 15 cytoplasmic (C)PV infections, 7

Dacus oleae, 53 Darna B virus, 45 Darna trima, 50 Dasychira B virus, 45 Daughterless, 82, 84,89 Decapoda, 316 Delta, 86, 94 physical interactions, 85-7 Dermacentor variabilis, 316 Dilophus febrilis, 158 Diploptera punctata, 269,270, 274, 275, 277-80,282,283,287,289,292, 293,295,296,299,300,302-4, 306,308-21, 322, 325, 328,329 D N A replication, 13-14 DNI neurones, 231 Drosophila, 45, 47, 50, 51, 53, 76, 77, 79,80, 82, 106-10, 113, 117, 125, 129, 152, 161, 166. 169, 180, 2046, 210, 211, 219, 315, 328 Drosophila apterous, 271 Drosophila melanogaster, 17, 51, 53, 76, 80, 106-8, 124, 158, 270,271, 31415,319, 320 genetic mechanisms of early neurogenesis, 75-103 Drosophila ninaE, 206 Drosophila yakuba, 114-17, 120 Enterovirus, 52 Entomopoxvirinae, 31 entomopoxviruses (EPVs), 29-38 biological control agents, 38 classification, 30-3 host range, 29-30 isolation, 29-30 molecular studies, 34-6 replication cycle in insects, 33-4 replication in vitro, 36-8 structure, 30-3 virions, 31-2 epidermal decision. 87-8

341

epidermal growth factor (EGF), 85 Eristalis tenax, 171 Escherichia coli, 21, 26, 126 E(SPL)-C, 87-8, 9 2 4 Estigmene acrea, 28, 30, 32,33,35-7,41 Eubaculovirinae, 3, 4 Eurythenes, 158 extracellular virus (ECV), 4, 5, 11 Flock House virus (FHV), 46 Fushi tarazu, 108 G A B A (y-aminobutyric acid), 206, 209, 218 Galleria mellonella , 5 , 4 1 Gastropoda, 317 Gonometa, 45, 50 Gonometa podocarpi, 5 1,53 granulosis virus (GV), 3 Gryllus bimaculatus, 171, 186,271, 303 Gryllus firmus, 166 gypsy moth virus (GMV), 46

Heliothis armigera, 7 Heliothis virescens, 8, 17 Heliothis zea, 4, 8, 20, 36 Hemicordulia tau, 166 Heteronychus arator, 44 hindgut muscle, 309 Homona magnanima, 36 honeybee Africanized bee problem, 138-40 dwarf, 135 elongation factor 1 (EF-I), 107-8 gene activity in embryonic development, 124 genes and sequences, 107-14 genes coding 'for venom compounds, 110-12 genetic variability among species, 131-3 genetical research, 1067 in situ hybridization, 113-14 mitochondrial D N A markers, 129-30 mitochondrial genes, 116-17 mitochondrial genome, 114-22 length variation, 117-22 non-coding sequences, 117-22 molecular biology, 10549 molecular phylogeny, 130-1

342

honeybee-contd. molecular variability within species, 1334 nuclear DNA markers, 125-8 nuclear genes, 107-14 population variability, 125-30 segmentation genes, 108-10 variability at population level, 126-7 variability within colony, 127-8

Inachis io, 45 information processing in ocellar system see ocellar system insect pests, biological control of, 15-22 insect virology, 1-73 in situ hybridization, 3 2 1 4 intrinsic factors, 80 iridescent viruses (IVs), 38-43 biological control agents, 42-3 classification, 38-9 host-range, 38-9 isolation, 38-9 molecular studies, 42 replication cycle, 4&1 structure, 3 9 4 0

juvenile hormone (JH), 268 inhibition of biosynthesis, 281, 286 inhibition of synthesis, 290 by extracts of brain, 270-1 in vitro, 270 titre regulation, 292-3 juvenile hormone esterase (JHE), 17-18

large monopolar cells (LMCs), 164 Latoia viridissirna , 5 1 Leptinotarsa decernlineata, 269 Lesteva, 154 Lestidae, 154 lethal of scute, 88, 90,92 Leucophaea rnaderae, 269, 309 Limulus, 154,213-15,225,229 Locusta, 160,210 Locusta rnigratoria, 269 locusts anterior ocellar focus, 189-93 ocellar tract of, 189-93 passive and active membrane properties of L-neurones, 218-20

INDEX

synaptic interactions among Lneurones, 225-8 Lucilia cuprina, 270, 271 Lymantria dispar, 6, 20, 36, 37, 53

Marnestra brassicae, 8, 9, 20, 28 Manawata virus (MwV), 46 Manduca sexta, 17, 156, 165, 206, 269, 270,271, 274, 275, 277, 278, 29& 9,303, 304, 306, 309, 31k16, 328, 329 Mansonia uniformis, 5 1 Megapis, 131 Melanoguin sanguinipes, 30 Melolontha rnelolontha, 30, 31 Metridium senile, 116 Micrapis , 131 Mollusca, 317 multiple nucleocapsids per envelope (MNPV), 3,8 Musca, 166, 201 Musca dornestica, 202, 206 Mytilus, 129 Neobellieria bulluta, 309 Neodiprion sertijer, 6 neural circuits in ocellar systems, 199-203 modification of simpler circuits, 24752 modifications for visual functions, 245 see also visual systems neural decision, 88-9 neural organization of ocellar pathways, 179-204 neuralizing signals, 80 neuroblasts, segregation of, 77-9 neuroectoderm, 75, 76 cell interactions in, 79-80 neurogenesis cellular basis of, 76-7 genetics of, 8 0 4 neurogenic genes functionally interrelated, 84-5 interactions, 89 neurohormones, 206,311-12 neurones L-neurones, 164, 1914,200,201,211, 213,214, 229, 231,232,234 bees, 220-2 cockroaches, 222-5

INDEX

dynamics and sensitivity of lightadapted, 215-18 genetic determination of, 183-S locusts, 218-20, 225-8 morphology of, 182-3 S-neurones, 188, 201 second-order, 189 see also ocellar neurones neuropeptides, 304 neurotransmitters, 206 Nodamura virus, 46 Nodaviridae, 43-8 bipartite RNA genome, 47 classification, 44-7 host range, 44 isolation, 44 molecular studies, 47-8 RNA genome, 47 structure, 44-7 Notch, 82, 86 physical interactions, 85-7 nuclear polyhedrosis virus (NPV), 3, 6-7 Nudaurelia P virus, 48-50 Nudaurelia P virus, see also Tetraviridae Nudaurelia capensis /3 virus (NPV), 48-50 Nudaudrelia capensis virus (NV), 49 Nudaurelia cytharea capensis, 48 Nudibaculovirinae, 3 Nudoraurelia cytherea capensis, 50 ocellar neurones absolute sensitivity, 163-4 as detectors of instability in flight, 238-9 descending interneurones, DNI, DNM and DNC, 238-9 detection of absolute intensity levels, 164 functional properties, 162-8 information processing in secondorder, 218-28 linear and non-linear signal transmission at graded synapses, 235 modulatory roles of efferent neurones, 228-3 1 morphology of second-order, 181-8 morphology of third-order, 195-8 multimodal integration in third-order and higher, 236-9 polarization sensitivity, 168

343

responses of small second-order, 228 second-order, 181-8 signal processing between second- and third-order, 231-6 signal rectification by non-linear synaptic transmission from second- to third-order, 235 small-diameter, 186-8 spatial properties, 162-3 spectral sensitivity, 165-8 speed of signal transmission, 164-5 synaptic transmission between secondand third-order, 232 ocellar pathways, neural organization of, 179-204 ocellar photoreceptors, 21 1-15 ocellar plexus information processing, 21 1-18 synaptic organization of, 1 8 6 1 ocellar system, 151-265 bee-type, 199 behavioural roles, 169-79 ‘cockroach-type’, 199 contribution to phototactic orientation, 169-70 control of neuroendocrinic secretion, 177-9 detection of polarized light, 177 distribution, 154-7 dorsal ocelli, 152, 157, 162 information processing in, 2 1 1 4 light intensity perception for control of diurnal activity, 173-4 locust type, 199 molecular basis, 204-1 1 neural circuits in, 199-203 neural organization, 153 ontogenetic development, 161-2 orientation toward edges, 171 phylogeny, 154-7 stimulatory role, 169 structure, 157-9 visual course control in flight, 170-1 visual processing in, 23944 ocellar tract bees, 193 cockroaches, 189 locusts, 189-93 wasps, 193 ocellar tract neuropil, synaptic organization of, 189-93 Orgyia pseudotsugata, 4 Othnonius batesi, 34

344

Panolisflamrnea, 6, 8, 20 Pantatoma, 154 Paracentrotus lividus, 115, 130 Parasa viridissirna, 51, 53 Paravespula, 201 Paravespula gerrnanica, 191, 193, 202 Paravespula vulgaris, 158, 193, 202, 222 peptides identification by isolation, 274-8 immunocytochemical localization of allatostatin-like, 306 in diverse organisms, 312-17 inhibiting juvenile hormone synthesis, 268 Periplaneta, 154, 158, 171, 201 Periplaneta americana, 158, 161, 166, 182, 186, 192, 196, 200, 202, 274, 275, 277, 278, 289, 292, 299, 3024, 308-9, 312, 318, 32@3, 329 Philosarnia B virus, 45 Phormia, 201 Picornaviridae, 43, 45, 50-3 classification, 5@2 host range, 50-2 isolation, 50-2 picornaviruses biological control agents, 53 molecular studies, 53 virion structure, 52 virus replication, 53 Pierb rapae, 7, 17 Plodia interpunctella, 4 polyhedra-derived virus (PDV), 4 Porthetria dispar, 7 posterior slope nueropil, synaptic organization of, 193-5 poxviruses, 29 procephalic neuroectoderm, 75 proctolin-induced contractions, 308 proneural genes, 82, 88-9 interactions, 89 proneural products, 94 Prothoracicotropic hormone (PTTH), 304 Pseudaletia separata, 36 Pseudaletia unipuncta, 315 Pyernotes tritici. 19 Pyrrhocoris apterus, 269 Rachiplusia ou, 5 random amplified polymorphic D N A technique (RAPD), 128

INDEX

Reoviridae, 45 restriction fragment length polym;rphisms (URFLP), 126, 128, 130, 133 Rhodnius prolixus, 269 Rhopalosiphum padi, 51, 52 RNA polymerase, 10-11 RNA viruses of insects, 43-53 Sarcophage bullata, 270, 271 Schistocera gregaria , 200 Schistocerca, 125, 227 Schistocerca gregaria, 159, 161, 165, 192, 202,309 Schistocerca nitens, 186 scute, 88 Shaggy function, 82 Sirnulium, 43 Single nucleocapsids per envelope (SNPV), 3, 7 split, 86 Spodoptera exernpta, 7,45 Spodoptera exigua, 45 Spodoptera frugiperda, 6, 7 , 9, 14, 15, 23, 24, 36, 41 Spodoptera littoralis, 7 Stagnicola elodes, 3 17 Striped Jack Nervous Necrosis Virus (SJNNV), 46 synaptic organization ocellar plexus, 180-1 ocellar tract neuropil, 189-93 posterior slope neuropil, 193-5 T A A G motif, 11-12 Tachycines, 201 Teleogryllus cornrnodus, 175, 176 Teleogryllus oceanicus, 50 Tetraviridae, 43, 45, 48-50 biological control, 50 classification, 48 host range, 48 isolation, 48 molecular studies, 49-50 replication, 49-50 virion structure, 49 Thosea asigna, 50 Thosea B virus, 45 Tipula IV type 1 (TIV), 38 Tribolium castaneum, 124 Trichoplusia, 201

INDEX

Trichoplusia B virus, 45 Trichoplusia ni, 5 , 9, 14, 19, 21, 22, 28, 41, 45, 53, 158, 159, 165, 174, 175, 180, 181, 202, 228 Triphena pronuba, 45 type I viroplasm, 33 type I1 viroplasm, 33 Valanga nigricornis , 159 ventral nervous system condensation defective (vnd), 82 ventral neuroectoderm (VNE), 75-84 Vespa, 201 visual pigment, 204-5 visual processing, in ocellar system, 239-44 visual systems colour-contrast detectors, 25&2

345

dimming detectors, 249 evolution of, 244-52 local temporal contrast detector, 249-50 movement detectors, 249-50 neural mechanisms for motion detection, 247 neural mechanisms for segregating ON and OFF signals, 245-7 spatial contrast detectors, 250 see also Neural circuits Wasps, ocellar tract of, 193 Wiseana, 38 Wiseana cervinata, 38 Wiseana granulosis, 38 Wiseana signata, 38 Wiseana umbraculata, 38

This Page Intentionally Left Blank