Rhodopsin and G-Protein Linked Receptors, Part A,

BIOMEMBRANES A Multi- Volume Treatise Volume 2A • 1996 RHODOPSIN AND G-PROTEIN LINKED RECEPTORS This Page Intentio...

Author: A.G. Lee

16 downloads 736 Views 15MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

BIOMEMBRANES A Multi- Volume Treatise

Volume 2A •

1996

RHODOPSIN AND G-PROTEIN LINKED RECEPTORS

This Page Intentionally Left Blank

BIOMEMBRANES A Multi-Volume Treatise

RHODOPSIN AND G-PROTEIN LINKED RECEPTORS Editor: A. G. LEE Department of Biochemistry University of Southampton Southampton, England

VOLUME 2A

1996

@)JAl PRESS INC. Greenwich, Connecticut

London, England

Copyright © 1996 by JAI PRESSINC 55 Old Post Road, No. 2 Greenwich, Connecticut 06836 JAI PRESSLTD. The Courtyard 28 High Street Hampton Hill, Middlesex TWl 2 1PD England All rights reserved. No part of this publication may be reproduced, stored on a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, filming, recording, or otherwise, without prior permission in writing from the publisher. ISBN: 1-55938-659-2 Manufactured in the United States of America

CONTENTS (Volume 2A)

LIST OF CONTRIBUTORS

ix

PREFACE A.G. lee

...

Xlll

RHODOPSIN STRUCTURE AND FUNCTION Burton J. Litrnan and Drake C. Mitchell CHARACTERIZATION OF THE PRIMARY PHOTOCHEMICAL EVENTS IN BACTERIORHODOPSIN AND RHODOPSIN Jeffrey A. Stuart and Robert R. Birge

1

33

LIGHT-INDUCED PROTEIN-PROTEIN INTERACTIONS ON THE ROD PHOTORECEPTOR DISC MEMBRANE Klaus Peter Hofmann and Martin Heck

141

MICROBIAL SENSORY RHODOPSINS John L. Spudich and David N. Zacks

199

ALPHA-ADRENERGIC RECEPTORS David B. Bylund

227

V

This Page Intentionally Left Blank

CONTENTS (Volume 2B) [3-ADRENERGIC RECEPTORS

Susan M. Pellegrino, Norman H. Lee, and Claire M. Fraser

G PROTEIN-COUPLED SEROTONIN RECEPTORS

Jean C. Shih and Timothy K. Gallaher

THE MUSCARINIC ACETYLCHOLINE RECEPTORS

Petra HOgger, Wolfgang Sad6e, and Jelveh Lameh

ADENOSINE RECEPTORS

David R. Luthin, John A. Auchampach, and Joel Linden

245 281 301 321

METABOTROPIC GLUTAMATE RECEPTORS

Patrick J. O'Hara

349

GLYCOPROTEIN HORMONE RECEPTORS: NEW MOLECULAR SUPPORTS FOR OLD BIOLOGICAL FUNCTIONS

Roland Salesse and Jean Gamier

PLATELET ACTIVATING FACTOR RECEPTOR

Shivendra D. Shukla

INDEX

387 463 481

vii

This Page Intentionally Left Blank

LIST OF CONTRIBUTORS

John A. Auchampach

Departments of Internal Medicine and Molecular Physiology and Biological Physics University of Virginia Charlottesville, Virginia

Robert R. Birge

Department of Chemistry Syracuse University Syracuse, New York

David B. Bylund

Department of Pharmacology University of Nebraska Medical Center Omaha, Nebraska

Claire M. Fraser

Department of Molecular and Cellular Biology The Institute for Genomic Research Gaithersburg, Maryland

Timothy K. Gallaher

Department of Molecular Pharmacology and Toxicology School of Pharmacy University of Southern California Los Angeles, California

Jean Garnier

Unite d'lngenierie des Proteines IN RA-Biotech nologies Cedex, France

Martin Heck

Institut fLir Medizinische Physik und Biophysik Humboldt-Unlversit~it zu Berlin Berlin, Germany

Klaus Peter Hoffman

Institut for Medizinische Physik und Biophysik Humboldt-Universit~it zu Berlin Berlin, Germany

LIST OF CONTRIBUTORS Petra H6gger

Departments of Pharmacy and Pharmaceutical Chemistry University of California San Francisco, California

Jelveh Lameh

Departments of Pharmacy and Pharmaceutical Chemistry University of California San Francisco, California

Norman H. Lee

Department of Molecular and Cellular Biology The Institute for Genomic Research Gaithersburg, Maryland

Joel Linden

Departments of Internal Medicine and Molecular Physiology and Biological Physics University of Virginia Charlottesville, Virginia

Burton J. Litman

Laboratory of Membrane Biochemistry and Biophysics National Institute of Health Rockville, Maryland

David R. Luthin

Departments of Internal Medicine and Molecular Physiology and Biological Physics University of Virginia Charlottesville, Virginia

Drake C. Mitchell

Laboratory of Membrane Biochemistry and Biophysics National Institute of Health Rockville, Maryland

Patrick J. O'Hara

DNA Chemistry and Computer Science ZymoGenetics, Inc. Seattle, Washington

Susan M. Pellegrino

Department of Molecular and Cellular Biology The Institute for Genomic Research Gaithersburg, Maryland

List of Contributors Wolfgang Sadie

Departments of Pharmacy and Pharmaceutical Chemistry University of California San Francisco, California

Ronald Salesse

Unite d'lngenierie des Proteines IN RA-Biotech nologies Cedex, France

Jean C. Shih

Department of Molecular Pharmacology and Toxicology School of Pharmacy University of Southern California Los Angeles, California

Shrivendra D. Shukla

Department of Pharmacology School of Medicine University of Missouri Columbia, Missouri

John L. Spudich

Department of Microbiology and Molecular Genetics University of Texas Medical School Houston, Texas

Jeffrey A. Stuart

Department of Chemistry Syracuse University Syracuse, New York

David N. Zacks

Department of Microbiology and Molecular Genetics University of Texas Medical School Houston, Texas

This Page Intentionally Left Blank

PREFACE

The quantity of information available about membrane proteins is now too large for any one person to be familiar with anything but a very small part of the primary literature. A series of volumes concentrating on molecular aspects of biological membranes therefore seems timely. The hope is that, when complete, these volumes will provide a convenient introduction to the study of a wide range of membrane functions. Application of the techniques of molecular biology has provided the sequences of a very large number of membrane proteins, and has led to the discovery of superfamilies of membrane proteins of related structure. The classic example of the superfamily is the seven helix receptor superfamily, all related in structure to bacteriorhodopsin, and named after the seven trans-membrane or-helices identified in bacteriorhodopsin. This volume explores the structures and functions of this superfamily. As editor, I wish to thank all the contributors for their efforts and the staff of JAI Press for their professionalism in seeing everything through to final publication. A.G. Lee Editor

xiii

This Page Intentionally Left Blank

RHODOPSIN STRUCTURE AND FUNCTION

Burton J. Litman and Drake C. Mitchell

I. II.

III.

IV. V. VI.

VII.

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

2

Rhodopsin Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Introduction

3

A.

General Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

3

B.

Structural C o m p a r i s o n with Bacteriorhodopsin

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

5

C. Helix-Linking Loops and Functional Domains Rhodopsin Photointermediates . . . . . . . . . . . A. Overview . . . . . . . . . . . . . . . . . . . . B. Early Photointermediates . . . . . . . . . . . . C. Late Photointermediates . . . . . . . . . . . . D. Structural Changes in MII . . . . . . . . . . .

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

7 9 9 9

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

12

O p s i n - C h r o m o p h o r e Interactions . . . . . . . . . . . . . . . . . . . . . . . . . Rhodopsin-Transducin Interactions . . . . . . . . . . . . . . . . . . . . . . . . Rhodopsin-Lipid Interactions . . . . . . . . . . . . . . . . . . . . . . . . . . .

13 15 17

11

A.

Disk M e m b r a n e Lipid Composition

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

17

B.

Lipid Modulation o f MII Formation . . . . . . . . . . . . . . . . . . . . .

18

C.

M e m b r a n e Lipid D o m a i n Model . . . . . . . . . . . . . . . . . . . . . . .

22

Conclusions

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

Biomembranes Volume 2A, pages 1-32 Copyright © 1996 by JAI Press Inc. All rights of reproduction in any form reserved. ISBN: 1-55938-659-2

24

BURTON J. LITMAN and DRAKE C. MITCHELL

!. I N T R O D U C T I O N The visual pigment, rhodopsin, is localized in the retinal rod outer segment (ROS) disk membrane and is a prototypical member of the seven transmembrane helical, superfamily of G protein-coupled receptors. Photoactivation of rhodopsin results in the formation of an active conformation, R*, which binds the rod cell G protein, transducin (Gt) and activates it by catalyzing the exchange of Gt-bound GDP for GTP (Figure 1, for recent reviews see Stryer, 1991; Khorana, 1992; Nathans, 1992; Hargrave et al., 1993). A single R* can activate several hundred Gt molecules during its lifetime. The ct subunit of Gt, Gta, activates the effector enzyme, a cGMP phosphodiesterase (PDE), by binding its inhibitor subunit, thus initiating cGMP hydrolysis. The reduction in cGMP concentration results in a hyperpolarization of the ROS plasma membrane, due to the reduced flux of Na ÷ ions through the cGMP-gated channels, with concomitant change in neurotransmitter release at the synaptic end of the rod cell, which generates the neuronal response to light. Rhodopsin is unique as a receptor in that it contains a covalently bound spectroscopic marker, the retinal chromophore. As will be discussed later in RhodopsinTransducin Interactions, several different investigations point to identifying R* with the metarhodopsin II (MII) photointermediate. Thus, monitoring the time-depend-

R.. G, (GTP). j f=~ GTP R.~. hV

~ .

G,,1,

GDP

R*- G, (GDP)

MII = R* +

Gt=(GTP)~%%

PDIE i

G,=(GTP)*-PDEi ] G,= (GTP)*- PDE, + PDF_=*~

//i

G, (GDP)

PDEi

Figure 1. Visual sig,nal transduction in rod cell. Light (hv) converts rhodopsin (R) to an activated form (R), which binds and activates Gt(GDP) by catalyzing the exchange of bound GDP for GTP. Gt(GTP)* then dissociates and Gt~(GTP) binds to the inactive form of the cGMP-phosphodiesterase (PDEi). This complex dissociates to yield the active subunit complex PDE~.I3and Gto~(GTP)*. PDE~, where PDE~ is the inhibitory subunit of the PDE, initiating the hydrolysis of cGMP by PDE~.I3. The lowered concentration of cGMP induces closure of cGMP-gated Na+ channels in the plasma membrane, hyperpolarizing the cell.

Rhodopsin Structure and Function

ent changes and peak intensity of the MII spectrum allows one to determine the kinetic and equilibrium properties associated with the formation of R*, which is the functional equivalent of the activated agonist-bound conformation in liganded receptor systems. A number of studies suggest that MII exists as several structural substates, which are spectrophotometrically indistinguishable, but differ both with respect to the configuration of the hydrophilic loops connecting the transmembrane helices and their interaction with the peripheral proteins of the visual transduction pathway. Various lines of research indicate, however, that the interaction of these transduction proteins with R* are reflected by the spectrophometrically monitored behavior of MII. As such, MII and R* will be used interchangeably in this review. The following topics associated with the visual transduction pathway will be discussed: rhodopsin structure, the interaction of opsin, the protein portion of rhodopsin, with the retinal chromophore, rhodopsin photochemistry, and the interaction ofphotoactivated rhodopsin with Gt, which is the first stage of amplification in the visual transduction pathway and is initiated by the formation of a R*.Gt complex. Rhodopsin is an integral membrane protein with about 50% of its mass in the phospholipid bilayer. In addition, the retinal binding site, where the initial event of photon absorption occurs in the form of the cis-trans isomerization of retinal, is at about the midpoint of the membrane bilayer. Hence, the effect of membrane lipid composition and associated physical properties on the photoactivation of rhodopsin will also be discussed.

!!. R H O D O P S I N STRUCTURE A. General Structure

Bovine opsin consists of 348 amino acids. Hydropathy profile analysis of the amino acid sequence suggests seven transmembrane regions connected on both the cytoplasmic and intradiskal membrane surfaces by hydrophilic loops of varying lengths (Ovchinnikov et al., 1983; Hargrave et al., 1983; Nathans and Hogness, 1983). A variety of physical measurements, including infrared linear dichroism on oriented rod outer segments (Michel-Villaz et al., 1979) and circular dichroism (Shichi and Shelton, 1974; Stubbs and Litman, 1976) have demonstrated that the transmembrane segments are (z-helices, oriented essentially perpendicular to the membrane surface. Palmitoyl acyl chains, covalently bound to Cys 322 and Cys 323, are thought to anchor part of the carboxyl terminal tail to the membrane forming a putative fourth loop consisting of 11 amino acids on the cytoplasmic surface of the rod outer segment disk membrane (Ovchinnikov et al., 1988), as shown in Figure 2. An additional posttranslational covalent modification is found in the amino terminal tail at Asn 2 and Asn 15, where N-linked oligosaccharides are added (Fukuda et al., 1979; Liang et al., 1979). Glycosolation at Asn 15 appears to be required for normal binding and activation of Gt, although blocking glycosolation at Asn 2 has no effect on interaction with G t (Kaushal et al., 1994). Glycosolation does not

BURTON J. LITMAN and DRAKE C. MITCHELL H, Ac-N

-

INTRADISKAL SURFACE

~o~)I

VII

i,. | _ ~.-~ ,lu

Palmitoyl Cysteines

=q_,,U.}7=.o; ~'y2p vY~ ~. N.__)F..~"~. ! " ~ V

T,,~

CYTOPLASMIC

SURFACE

~.~..~.~, T>~ = " ~ o, ~ ~ - "ooc-~.~.~-~=~ "-'° ~:~"~ Regions of Transducin Binding

Figure 2. Model for the topography of rhodopsin in the rod outer segment disk membrane. N-linked oligosaccharides are shown attached to Asn 2 and Asn15. The heavy line connecting Cys 11° and Cys 187 denotes the single disulfide bond found in rhodopsin. Palmitic acid moieties attached to Cys322 and Cys323 anchor part of the carboxyl tail, forming the putative 4th cytoplasmic loop, i4 (from Hargrave et al., 1993). appear to be required for formation of the final, folded structure of opsin. Nonglycosolated opsin is normally palmitoylated, inserts properly into the membrane, and combines with 11-cis retinal to form the characteristic rhodopsin pigment, although it has a diminished ability to activate Gt (Kaushal et al., 1994). Opsin and 11-cis retinal combine via a protonated Schiff base linkage at Lys296 to form rhodopsin (Bownds, 1967; Wang et al., 1980; Mullen and Akhtar, 1981; Findlay et al., 1981). Linear dichroism measurements on oriented ROS show that the retinal transition dipole moment is tilted about 15° with respect to the plane of the disk membrane (Liebman, 1962; Chabre and Breton, 1979; Michel-Villaz et al., 1982). The chromophore is located about equidistant from the cytoplasmic and intradiskal surfaces of the disk membrane, as shown by fluorescence energy transfer measurements, in which a water soluble donor transferred energy to retinal with equal efficiency when placed on either side of the membrane (Thomas and Stryer, 1982). Thermal denaturation studies show that bound retinal confers structural

Rhodopsin Structure and Function

stability to the final, folded protein structure, as evidenced by the fact that rhodopsin is 43 kcal/mole more stable than opsin (Khan et al., 1991). The basic topography of rhodopsin in the disk membrane, first suggested by a variety of spectroscopic techniques, and later by hydropathy profile analysis of the primary structure, has been refined by a variety of experiments. In intact ROS disks, points of proteolytic cleavage are found in loops i 1, i2, and i3, and in the carboxylterminal sequence. This indicates that these regions are exposed to the cytosol of the rod cell (Trayhurn et al., 1974; Pober and Stryer, 1975; Albert and Litman, 1978; Fung and Hubbell, 1978a,b). Proteolytic cleavage sites in the amino-terminal tail and in loop e 2 are made accessible by freezing and thawing ROS disks, indicating that these regions are exposed to the intradiskal space (Martynov et al., 1983). The sites of lectin binding to the N-linked oligosaccharides in the amino-terminal tail (Rohlich, 1976), antibody binding (Molday and Molday, 1979), and chemical modification (Barclay and Findlay, 1984) are all consistent with the topography deduced from proteolysis experiments. Circular dichroism measurements indicate that rhodopsin tertiary structure is approximately 55% or-helix (Shichi and Shelton, 1974, Stubbs and Litman, 1976), consistent with the transmembrane regions being in the form of or-helices. A series of high resolution FTIR measurements on rhodopsin in ROS disk membranes confirms a helix content of about 50%, and suggests that as much as 15% of this structure is in the form of more extended 310-type helix (Garcia-Quintana et al., 1993). On the basis of FTIR measurements made in 1H20 and 2H20, these authors concluded that 61% of the protein is solvent accessible and most of the inaccessible domains are in the transmembrane helices located in the hydrophobic portion of the bilayer. They also report evidence for [3-strands and reverse turns in the hydrophilic linking regions. The consistent picture which emerges from biochemical and spectroscopic studies is a structure consisting of seven transmembrane or-helices, with the remaining 50% of the structure about evenly divided between the intradiskal and cytoplasmic membrane surfaces.

B. Structural Comparison with Bacteriorhodopsin It has generally been supposed that the three-dimensional structure ofrhodopsin, and the entire superfamily of seven transmembrane helix, G protein-coupled receptors, is similar to that ofbacteriorhodopsin, the light-driven proton pump from Halobacterium halobium, whose three-dimensional structure is known (Henderson et al., 1990). The presumed structural homology between rhodopsin and bacteriorhodopsin springs from several striking similarities between the two proteins. Both have large hydrophobic domains consisting of seven, transmembrane (x-helices, their carboxyl terminal ends are exposed to the cytoplasm and amino terminal ends are exposed extracellularly, and both have chromophores consisting of an isomer of retinal linked via a protonated Schiffbase to a lysine, approximately in the middle of the a-helix nearest the carboxyl terminus. However, there are

BURTON J. LITMAN and DRAKE C. MITCHELL

significant functional differences between the two proteins. In rhodopsin, retinal isomerization drives the formation of binding sites for transducin and other regulatory enzymes on the cytoplasmic surface, while in bacteriorhodopsin, retinal isomerization is coupled to vectorial proton transport. The bacteriorhodopsin ground state chromophore is in the linear all-trans configuration, while in rhodopsin, it is 11-cis retinal. In rhodopsin, photoisomerization of the chromophore from 11-cis to all-trans retinal ultimately destabilizes the opsin-retinal Schiff base linkage and the chromophore is released by opsin. In contrast, the 13-cis chromophore ofphotoexcited bacteriorhodopsin remains bound and returns to the all-trans configuration via a purely thermally driven process. Bacteriorhodopsin has no significant sequence homology with rhodopsin (Hargrave et al., 1983), or any other member of the G protein-coupled receptor superfamily (Baldwin, 1993). The two-dimensional projection structure of bovine rhodopsin at 9,h. resolution has been obtained by electron crystallography of two-dimensional rhodopsin crystals (Schertler et al., 1993). The density map, shown in Figure 3A, consists of four well defined peaks, which are interpreted as a-helices perpendicular to the membrane, and an arcshaped feature which is interpreted as three somewhat tilted helices. This structure differs from that ofbacteriorhodopsin in several respects, as shown in Figure 3B. At 9.& resolution, only two sharp density peaks are visible in the projection structure of bacteriorhodopsin, indicating that two of the rhodopsin helices are more perpendicular to the plane of the membrane than in bacteriorhodopsin. The cross section of rhodopsin (in the plane of the membrane) is less ellipsoidal and more circular than that of bacteriorhodopsin. The region between the helices, which forms the retinal chromophore binding pocket, is somewhat S-shaped in rhodopsin, while in bacteriorhodopsin it is essentially straight, a

A

B

:0

Figure 3. Projection structure of rhodopsin. A. Projection density map of a single rhodopsin molecule at 9 ,g, resolution. B. Projection density map of a single bacteriorhodopsin molecule at 9 ,g, resolution, using the same scale as in A (from Schertler et al., 1993).

Rhodopsin Structure and Function

Figure 4. Proposed assignment of the putative transbilayer oc-helices of rhodopsin to the projection structure (adapted from Baldwin, 1993).

probable consequence of ground state rhodopsin having a bent, 11-cis retinal chromophore, while ground state bacteriorhodopsin has a straight all-trans retinal chromophore. Baldwin (1993) has recently proposed an arrangement of the helices for rhodopsin and other G protein-coupled receptors. Following a detailed analysis of the amino acid sequences of 204 G protein-coupled receptors, the structural constraints imposed by sequence homology were used to assign particular hydrophobic stretches in the primary sequence to peaks in the projection structure of Schertler et al. (1993). Based on sequence alignment of the entire G protein-coupled receptor superfamily, three constraints on the arrangement of the helices were determined: (1) successive helices in the sequence must be adjacent in the three dimensional structure, (2) helices I, IV, and V must be most exposed to the lipid bilayer, and (3) helix III must be least exposed to the lipid bilayer. These three constraints along with the requirement for the Schiffbase counterion, Glu ~13(Zhukovsky and Oprian, 1989; Sakmar et al., 1989; Nathans, 1990a), to be near the end of the lysine side chain of Lys 296, lead to a single, best assignment of helices to the projection structure. The resulting assignment of the helices to the two dimensional structure, shown in Figure 4, makes no predictions regarding structure of the hydrophilic loops connecting the helices, but represents the most detailed three-dimensional model for rhodopsin proposed to date.

C. Helix-Linking Loops and Functional Domains Compared to the transmembrane domain, little structural information is available regarding the hydrophilic intradiskal and cytoplasmic regions of rhodopsin. Gt, rhodopsin kinase, arrestin, and phosphatases bind to regions exposed to the cyto-

BURTON J. LITMAN and DRAKE C. MITCHELL

plasm at various times following photolysis, thus, this surface has attracted more experimental attention than the intradiskal domain. The sites ofphosphorylation by rhodopsin kinase are located in the carboxy terminal sequence. The first and second phosphorylation sites are at Ser338 and Ser343, respectively, and higher levels of phosphorylation occur at Thr335,Thr 336,Thr349, and Thr342 (McDowell et al., 1993; Papac et al., 1993). Most of the structural information regarding the cytoplasmic surface has focused on the requirements for Gt binding and activation; this information will be covered in the section on Rhodopsin-Transducin Interactions. Rhodopsin's single disulfide bond, between Cys 110 and Cys 187, is found on the intradiskal surface (Karnik and Khorana, 1990). Although, this pair of cysteines is conserved in all visual pigments and G-protein coupled receptors, it is not strictly required for formation of functional rhodopsin. Replacement of Cys 11° and Cys 187 by alanine residues results in a properly folded, functional form ofopsin (Davidson et al., 1994), however, replacement of these cysteines with serines prevented proper folding and insertion into the bilayer (Karnik et al., 1988). A detailed mutagenesis study of the loops exposed on the intradiskal surface and the N-terminal tail indicates that this region is necessary for proper folding of opsin (Doi et al., 1990). Deletion of as few as two amino acids in any one of the three intradiskal loops or the N-terminal tail caused mutant opsins expressed in COS-1 cells to remain in the endoplasmic reticulum, indicating that the mutant proteins could not form a properly folded structure. Experiments with proteolyzed rhodopsin have shown that the retinal binding domain, formed by the transmembrane helices, and the cytoplasmic surface of rhodopsin are functionally distinct domains. Treatment of intact ROS disks with papain results in cleavage of the hydrophilic loop i3, which connects helices V and VI. This produces a large fragment, F 1, consisting of the carbohydrate-containing 5 amino terminal helices and a smaller fragment, F2, consisting of the 2 carboxyl terminal helices. These two fragments copurify on Concanavalin A-Sepharose as a tight complex, which is stabilized by helix-helix interactions alone (Albert and Litman, 1978). Proteolysis was found to have no effect on the visible CD spectrum, indicating that the retinal binding site remained intact, while changes in the far-ultraviolet CD spectrum showed that proteolysis reduced the helical content of rhodopsin by only a few percent (Albert and Litman, 1978). In addition, papain proteolysis of loop i3 had very little effect on the MI---~MII transition kinetics (Litman et al., 1982), which is somewhat surprising since retinal is attached to F2, while the Schiffbase counterion, Glu 113(see, sections opsin-chromophore interactions), is in F 1. These results suggest that a basic feature of rhodopsin's structurefunction relationship involves a one-way flow of information from the retinal binding domain out to the cytoplasmic surface. Conformation changes in the retinal binding domain, triggered by retinal isomerization, produce conformation changes at the cytoplasmic surface, which lead to formation of the Gt binding site. In contrast, proteolytic cleavage of the carboxyl terminus and the cytoplasmic loop connecting the F 1 and F2 fragments has little effect on opsin conformation changes

Rhodopsin Structure and Function

involving the transmembrane helices which form the retinal binding domain (Litman et al., 1982).

III.

RHODOPSIN PHOTOINTERMEDIATES A. Overview

In rhodopsin, nature has provided investigators with a G protein-coupled receptor with an intrinsic reporter of the interactions between the binding pocket of the receptor and both agonist and antagonist ligands. Absorption of a photon by rhodopsin converts 11-cis retinal, the antagonist, to all-trans retinal, the agonist, with a quantum efficiency of 0.67 (Kropf, 1967; Dartnall, 1968). Hubbard and Kropf (1958) identified retinal photoisomerization as the initial event in visual transduction, and this step is functionally equivalent to the binding ofagonist in the liganded G protein-coupled receptors. In rhodopsin, photoisomerization induces a series of charge rearrangements in the retinal binding pocket, which spawn a series of rapid absorption changes known as the photobleaching sequence. Thus, UV/VIS spectroscopy can be utilized to temporally and energetically resolve the initial series of intramolecular conformation changes, which ultimately create the conformation in the cytoplasmic surface which binds G c The sequence ofrhodopsin photointermediates is shown in Figure 5. Blue shifted intermediate (BSI) and metarhodopsin II' (MII') are not observed in low temperature equilibrium studies, but are observed in time-resolved experiments at room temperature. The wavelength maxima and transition times given in Figure 5 are those observed in time resolved measurements for rhodopsin at ambiem temperature (~-20 °C) and neutral pH. Photoproducts such as hypsorhodopsin, which form only under high light intensities or low temperatures, have been excluded.

B. Early Photointermediates Uncertainty persists regarding the exact nature of photorhodopsin, as to whether it is a single species, and whether it can fall back to rhodopsin or must complete the photointermediate cascade (Lewis and Kliger, 1992). One or more red-shifted precursors of bathorhodopsin were observed at low temperatures (Peters et al., 1977) and in a number of ambient temperature kinetic measurements (Monger et al., 1979; Kandori et al., 1989; Yan et al., 1991). Following Lewis and Kliger (1992), the term photorhodopsin is used here to refer to one or more red-absorbing photointermediates, which are detected on the picosecond timescale. Bathorhodopsin (batho) is the first photointermediate which was demonstrated to have an all-trans chromophore (Mathies et al., 1976; Hug et al., 1988). Batho stores about 36 kcal/mol (Cooper, 1979; Boucher and Leblanc, 1985; Shick et al., 1987), or about 60% of the 57 kcal/mol of incident photon energy for photons of

10

BURTON J. LITMAN and DRAKE C. MITCHELL R h o d o p s i n (498)

P h o t o r h o d o p s i n (.--.560) 4 ps B a t h o r h o d o p s i n (540) 35 ns Blue Shifted I n t e r m e d i a t e (477)

+

200 ns

Lumirhodol~sirt (497)

~ M e t a II' (380)

ff

90 ~ts

Meta I (478)

(?) I ms M e t a r h o d o p s i n II (380) -- 5 rain M e t a r h o d o p s i n III (465) ,~

--- 10 rnin

all trans retinal + opsin

Figure 5. The rhodopsin photoreaction cascade. Numbers in parenthesis next to individual species indicate their approximate wavelength of maximum absorbance in nanometers. -limes for transitions between species correspond to conditions of 20-25 °C, pH 7.0 in ROS disk membranes. Transitions involving Lumirhodopsin, Metarhodopsin I, Metarhodopsin II, and Metarhodopsin I1' that have been proposed, but not fully characterized are marked (?).

500 nm wavelength. This energy drives all subsequent transitions, and is stored via distortion of the all-trans chromophore (Birge et al., 1988; Smith et al., 1991). It has been demonstrated in native rhodopsin at room temperature that batho rapidly establishes an equilibrium with BSI (Hug et al., 1990; Lewis et al., 1990; Randall et al., 1991). BSI is higher in both enthalpy and entropy than batho (Hug et al., 1990). Thus, at low temperature, the rate of back reaction to batho is sufficiently large that BSI never accumulates and at room temperature, where the batho ~-~ BSI equilibrium constant is 1.4, decay to lumirhodopsin is fast enough that a spectral peak due to BSI is never observed. A study of BSI in rhodopsin, where retinal was replaced with a variety of retinal analogs, showed that the rate of BSI formation is strongly chromophore dependent, but the rate of BSI decay is largely chromophore independent (Randall et al., 1991). This suggests that the rate

Rhodopsin Structure and Function

11

limiting step in BSI decay involves a protein relaxation, indicating that the BSI to lumirhodopsin (lumi) transition is the earliest step in the photoreaction cascade that is associated with a protein conformation change. C. Late Photointermediates

In contrast with the events leading up to lumirhodopsin (lumi), the sequence of events following formation of lumi are not as well characterized, as shown by the transitions marked (?) in Figure 5. The 'classical' scheme, based on low-temperature measurements, of lumi decaying to MI, which then forms a quasi-stable equilibrium with MII (Mathews et al., 1963) is unable to account for ambient temperature kinetic measurements conducted with a high level of signal to noise on rhodopsin in disk membrane suspensions (Stewart et al., 1977; Lewis et al., 1981; Straume et al., 1990; Thorgeirsson et al., 1992, 1993). To adequately describe absorbance changes on the time scale of lumi, MI, and MII, ambient temperature kinetic measurements require three or more discrete exponential decays, rather than the two exponential decays predicted by the 'classical' scheme. The kinetic behavior of photointermediates, which occur on the submicrosecond timescale, is identical for rhodopsin in ROS disk membranes or solubilized in detergent, however all transitions beyond lumi are dramatically accelerated in detergents (Applebury et al., 1974, Litman et al., 1981). Artifacts due to light scattering, that are inherent in UV/VIS measurements on membrane suspensions, have made it difficult to acquire data of sufficient signal-to-noise to test detailed photoreaction models. A series of recent investigations utilizing multichannel detection with microsecond time resolution makes it clear that the kinetic complexity observed on the lumi to MII timescale is due to a decay product of lumi, which is isochromic with MII (Thorgeirsson et al., 1992, 1993). The existence of two, isochromic species of MII was previously suggested in studies ofMI ~-~ MII kinetics (Hoffmann et al., 1978; Straume et al., 1990), pH dependence (Parkes and Liebman, 1984), and the kinetics of proton uptake by photoactivated rhodopsin (Arnis and Hofmann, 1993). The work by Kliger and coworkers indicates that the fastest absorbance increase at 380 nm is due to a photointermediate with an absorbance maximum at 380 nm, which forms directly from lumi (Thorgeirsson et al., 1992, 1993). They have proposed several kinetic models to link the new intermediate, designated MII' in Figure 5, with lumi, MI, and MII. The decay scheme beyond lumi, shown in Figure 5, summarizes the essential features of the models which have been tested to date, and points out the transitions which are still unresolved. A definitive model, which accounts for both the kinetics and the observed equilibrium between 480 nm and 380 nm absorbing species, over a range of temperatures and pH values, has not been unambiguously determined. The sequence of events beyond MII is well understood in terms of the scheme shown in Figure 5. Below pH 7.7, MII can decay either to MIII or directly to opsin + retinal387 (Blazynski and Ostroy, 1984). Stabilization of MII by G t decreases

12

BURTON J. LITMAN and DRAKE C. MITCHELL

formation of MIII from MII, and favors decay of MII to opsin + retinal via a pathway which bypasses Mill (Pfister et al., 1983; Hofmann et al., 1983). An equilibrium between MII and Mill was first suggested to account for linear dichroism measurements of the slow intermediates (Chabre and Breton, 1979) and is supported by kinetic measurements of MII decay in the presence of G t (Pfister et al., 1983; Hofman et al., 1983). The equilibrium between MII and MIII was directly demonstrated by showing that photolyzed rhodopsin, which appeared to have decayed to MIII, could be converted into a mixture of MII and MIII by the addition of G t (Kibelbek et al., 1991). Complex formation between MII and G t is well documented (see Rhodopsin-Transducin Interactions) and the formation of the MII.G t complex shifted the equilibrium, which strongly favors Mill, towards MII. At neutral pH, MIII decays to opsin + retinal (Blazynski and Ostroy, 1984).

D. Structural Changes in MI! MII is considered the most important photointermediate species because of its identification with R* (Emeis et al., 1982; Bennett et al., 1982; Kibelbek et al., 1991). A wide variety of physical measurements show that MII is the most structurally, as well as functionally, unique rhodopsin photointermediate. In MII, the Schiff base linking the retinal chromophore to tys 296 becomes deprotonated (Doukas et al., 1978), and in the native protein this deprotonation is required for formation of MII (Longstaff et al., 1986). The first evidence for conformational changes in the MI-MII transition was the susceptibility of the retinal chromophore in MII to attack by hydroxylamine (Falk and Fatt, 1966) and the reduction of the Schiffbase by sodium borohydride (Akhtar et al., 1965). The MI-MII equilibrium is pressure dependent, with high pressure favoring MI (Lamola et al., 1974; Attwood & Gutfreund, 1980). In purified ROS disks, the pressure dependence was used to calculate a reaction volume of 108 ml/mol for the MI to MII transition (Attwood and Guffreund, 1980). Several unique structural features of MII have been elucidated by a number of investigations utilizing FTIR difference spectroscopy. MII appears to have a more open, less rigid structure than MI, and appears to return to a more rhodopsin-like structure during the transition to Mill (DeGrip et al., 1985, Rothschild et al., 1987). The result is that MII has a unique peptide backbone conformation, resulting from a few peptide bonds undergoing a secondary structural rearrangement (Rothschild et al., 1987; Klinger and Braiman, 1992). This structural rearrangement and the expansion of MII relative to MI is consistent with a small movement of helices IV and VI towards the cytoplasmic surface during the MI to MII transition (Pellicone et al., 1985). Acid pH favors MII over MI (Mathews et al., 1963) and photolyzed rhodopsin was reported to take up protons from the aqueous medium with kinetics similar to the formation ofMII (Wong and Ostroy, 1973). Amore recent study of the kinetics of proton uptake and MII formation for rhodopsin in octyl glucoside and dodecyl

Rhodopsin Structure and Function

13

maltoside demonstrates that proton uptake from the aqueous medium occurs after deprotonation of the retinal Schiffbase and formation ofMII (Arnis and Hofmann, 1993). The authors postulate that this proton-dependent transition from MII a to MII b corresponds to formation of the loop conformation on the cytoplasmic surface which binds G t. Conformation changes in the cytoplasmic loops in MII have been observed by a variety of spectroscopic techniques. A comparison of the FTIR spectra of proteolyzed and unmodified rhodopsin showed that loops i2 and i 3 adopt a well-defined conformation in rhodopsin, but become more random and flexible in MII (Ganter et al., 1992). Conformation changes in loops i2 and i4 have been directly observed using nitroxide spin labels attached to cysteine residues at positions 140 in loop i2 and 316 in loop i4. Formation of MI produced no changes in the EPR spectra of these two attached nitroxides, while formation of MII produced significant changes in their EPR spectra (Resek et al., 1993). Conformational changes in loop i2 in MII were also indicated by a time-resolved EPR study which detected conformational changes at Cys 14° with a time constant and activation energy consistent with MII formation (Farahbakhsh et al., 1993).

IV. OPSIN-CHROMOPHORE INTERACTIONS Interactions between opsin and the retinal chromophore may be divided into factors which modulate the static, or ground state, spectrum of rhodopsin, and those that modulate the dynamic, or post-photolysis, spectral characteristics. The static absorption spectrum ofrhodopsin is determined by the unique electronic environment that opsin provides for the retinal chromophore. In ground state bovine rhodopsin, the opsin-chromophore interactions lead to an opsin-induced spectral shift of 58 nm to the red, from the benchmark of 440 nm for a protonated retinylidene Schiff base in methanol (Blatz et al., 1972), to a maximum absorbance at 498 nm. The specific amino acids responsible for this opsin shift have not been identified. In recent years site-directed mutagenesis has been used to gain much information about interactions between specific amino acids and the retinal chromophore. Several charged amino acids have been eliminated from consideration by an extensive site-directed mutagenesis study. Six amino acids from helices II, III, or IV of bovine rhodopsin, Asp 83, Met 86, G1u 122, GIH134, Arg 135, and His 211, were altered by a total of 14 different mutations (Nathans, 1990b). All of the mutant proteins had an absorbance maximum near 498 nm, indicating that none of these six charged residues contribute to the opsin shift. The central question of rhodopsin's structure-function relationship is the mechanism by which retinal isomerization in the interior of the protein triggers conformational changes on the cytoplasmic surface. This process necessarily begins with dynamic, or transitory, interactions between retinal and the amino acid side chains which make up the retinal binding pocket. Thus, knowledge of the identity and disposition of the amino acids, which interact directly with retinal, is necessary to

14

BURTON J. LITMAN and DRAKE C. MITCHELL

understanding the molecular mechanism which couples retinal isomerization to conformation changes of the entire protein. A major advance in this area is the identification of G l u 133 as the retinylidene Schiff base counterion (Sakmar et al., 1989; Zhukovsky and Oprian, 1989; Nathans, 1990b). This determination principally rests on the fact that replacement of Glu 133with Gln causes a blue-shift of the chromophore spectrum by 120 nm to a maximum at 380 nm and, lowers the pK of the Schiff base to such an extent that at neutral pH, the Schiff base nitrogen is no longer protonated (Sakmar et al., 1989; Zhukovsky and Oprian, 1989). Robinson et al. (1992) proposed that rhodopsin with an 11-cis chromophore is constrained in an inactive conformation, unable to bind and activate transducin, by a salt bridge between the retinylidene Schiffbase nitrogen and Glu 113.They reasoned that retinal isomerization moves the Schiff base nitrogen away from Glu 113, breaking a salt bridge, and thus allowing deprotonation of the retinylidene Schiff base nitrogen and adoption of the active MII conformation. The issue of whether or not disruption of the Schiffbase proton, Glu 113charge pair is required for formation of R* will be discussed further in the Rhodopsin-Transducin Interactions section. In one, detailed study, a series of rhodopsin mutants, each with a single amino acid replacement, was tested for their ability to bind 11-cis retinal, generate chromophore with a native-like absorption spectrum, and activate G t (Nakayama and Khorana, 1991). The results of this study, along with those of an earlier study using a photoactivatible analog of retinal (Nakayama and Khorana, 1990), indicated that Glu 122,Trp 126,Trp265and Tyr268interact strongly with the retinal chromophore. In the structure proposed by Baldwin (1993) Glu 122and Tyr268 face the intrahelical cleft, which forms the retinal binding pocket, and Trp 265 is positioned in a way that would allow it to form part of the pocket for the [3-ionone ring, as suggested by Nakayama and Khorana (1991). Two mutations, D83N and W 161L, had no effect on the absorption spectnnn of unbleached rhodopsin, suggesting that Asp 83 and Trp 161 have no interaction with the retinal chromophore, and in the proposed structure these two residues do not face the intrahelical cleft; Asp 83 faces helix I and Trp161 faces helix III. In a second study, each of rhodopsin's six histidine residues was replaced one at a time by phenylalanine and/or cysteine (Weitz and Nathans, 1992). All of the resulting mutant proteins formed pigments with normal ground state spectral properties, but following photolysis the mutants H211F and H211C were unable to form the photointermediate MII. Photobleaching of these two mutants produced only MI at pH 5.8 and pH 7.8. The authors proposed that protonation of His 211 destabilizes the ionic interaction between the protonated Schiffbase of retinal and its counterion, Glu 113,thereby favoring formation of MII. However, in Baldwin's (1993) proposed rhodopsin structure His 211in helix IV is more than a helix diameter away from Glu 113 in helix III. Experiments in which the retinylidene Schiffbase was methylated, and thereby made unable to undergo deprotonation, demonstrated that deprotonation of the retinylidene Schiffbase is required for formation of MII and binding of Gt in native

Rhodopsin Structure and Function

15

rhodopsin (Longstaff et al., 1986). Thus, Schiff base deprotonation is directly involved in coupling retinal isomerization to overall protein conformational changes, and underlines the importance of identifying the amino acid which acts as an acceptor for the Schiff base proton. The identity of this residue has not been unambiguously determined. However, a comparison of the FTIR spectra of native rhodopsin and the site-directed mutants D83N, E 122Q, and D83N/E 122Q lead to the suggestion that Glu 113, which serves as the Schiffbase counterion, is also the acceptor for the net proton transfer from the Schiff base to opsin in MII, possibly with a water molecule acting as an intermediary (Fahmy et al., 1993). Recent site-directed mutagenesis studies show that Schiff base deprotonation is structurally distinct from formation of R* (Zvyaga et al., 1993, 1994). The Schiff base counterion was moved one helix turn by construction of the triple replacement mutant G 113A/A 117G/G 122N. The resulting mutant rhodopsin formed native-like pigment, and in response to light activated G t and formed MI and MII in a quasistable equilibrium. However, the pH dependence of the MR--~MII equilibrium was reversed, with alkaline pH favoring MII (Zvyaga et al., 1993). The altered pH dependence was interpreted as demonstrating that movement of the counterion had uncoupled retinylidene Schiffbase deprotonation from the influence of one or more histidines. These results suggest that while a specific acceptor for the retinylidene Schiffbase proton probably controls the transition from MI to MII in the presence of the Glu 113 counterion, it is not strictly required for MII formation when the counterion is absent.

V. RHODOPSIN-TRANSDUCIN INTERACTIONS Several independent lines of evidence associate the Gt-activating conformation of photoactivated rhodopsin, R*, with the photointermediate MII. These include the following: (1) when rhodopsin is bleached in the presence of G t and the absence of GTP, formation of an excess of MII is observed relative to that amount of MII that would be present in the absence of G t. The excess MII is proportional to the amount of added G t and the effect saturates at a 1:1 ratio of bleached rhodopsin t o G t (Emeis et al., 1982). (2) IfG t is added to a bleached rhodopsin sample, which has progressed to the MIII photointermediate, then the presence of a MII spectral component is observed (Kibelbek et al., 1991). The preceding observations can be attributed to the formation of a stable MII.G t complex in the absence of GTP, which has an absorption spectra identical to that of uncomplexed MII, and is thus observed spectrally as an additional MII contribution. Monitoring the amount of excess MII as a function of Gt, concentration allows one to determine the binding constant of the MII.G t complex (Kahlert and Hofmann, 1991). (3) The spectrophotometric decay of MII was found to occur with essentially the same decay constant as the loss of Gt-activating activity associated with R*, strongly identifying R* with MII (Kibelbek et al., 1991). Given the preceding evidence, MII is generally assumed to be synonymous with R*.

16

BURTON J. LITMAN and DRAKE C. MITCHELL

In the previous section on opsin-chromophore imeractions, those structural features important for MII formation were discussed. The initial stage of signal amplification in the visual transduction pathway is the interaction of MII with G t to form a MII.G t complex (Figure 1). Interactions of G t with MII occur on the cytosolic surface of the disk membrane, where G t is bound as a peripheral protein. The cytosolic surface ofrhodopsin is formed from the four hydrophilic loops il, i2, and i3 connecting helices I and II, IV and V, and VI and VII, respectively, and 14, the putative fourth loop in the carboxyl terminus formed by the palmitoyl chains at Cys 322 and Cys 323 partitioning into the lipid of the disk membrane, as depicted in Figure 2. Various segments of the cytosolic surface of rhodopsin have been identified as being involved in G t interactions. Limited proteolysis experiments, which removed segments of the carboxyl terminus beyond the putative i4 loop, had little effect on G t binding to the disk membrane (KOhn and Hargrave, 1981). Similar limited proteolysis was reported to produce a 40% increase of PDE activity, presumably as the result of more efficient G t activation (Aton and Litman, 1984). Phosphorylation of a number of ser and thr residues in this region is thought to be responsible for the down regulation ofrhodopsin's receptor function (Miller et al., 1986; Wilden et al., 1986). G t binding studies demonstrate that the affinity of MII for G t decreases with increasing phosphorylation level, while the amount of MII changes very little (Mitchell et al., 1992a). Hence, the region of the carboxyl terminus beyond the putative i4 loop appears to sterically hinder G t binding and when phosphorylated, plays a major role in down regulating G t binding. Small peptides simulating the sequence of regions of the i2, i3, and i4 loops were found to reduce G t binding to MII, suggesting that these loops are involved in the formation of the G t binding site (Konig et al., 1989). Studies of a variety of mutant rhodopsins, having both single amino acid mutations and sequence deletions or replacements, have further identified several amino acid residues and hydrophilic loops as sites of G t interaction. Franke et al. (1990) examined three mutants, all of which form pigments with native-like absorption spectra, but differed greatly in their interaction with G t. When residues 140-152 in loop i 2 were replaced with a sequence of residues from the amino terminus, G t binding was detected, but no light stimulated GTPase activity, which is an indirect measure of G t activation, was observed. A deletion of residues 237-249 in i 3 also showed G t binding without activation. Residues Glu TM and Arg 135, Which form a charge pair at the hydrophilic interface of helix III, appear to be critical for G t binding. When the order of these residues is changed to Arg TM and Glu 135, no G t binding was observed. Replacement of these two amino acids by Val also resulted in a lack of G t activation (Franke et al., 1992). These studies show that G t binding and activation represent separate steps involving different segments in the rhodopsin structure and suggest that the charged pair, Glu TM and Arg 135, is necessary for G t binding, while loops i2 and i3 are required for G t activation.

Rhodopsin Structure and Function

17

One of the most interesting results obtained from mutagenesis experiments was the observation of a constitutively active rhodopsin molecule in the absence of any bound retinal (Robinson et al., 1992). In these experiments, mutants, which prevented the formation of the charge pair between the protonated Schiff base at the Lys 296 retinal attachment site and Glu 113, were expressed. An El l3Q mutant was produced, which recombined with a propylamine derivative of vitamin A aldehyde to form a bleachable pigment capable of activating G r The striking finding was that in the absence of any bound retinal the protein was still capable of activating G t. A K296G mutant showed the same properties as the E 113Q mutant. These observations led to a proposal that the charge pair formed by Glu 113 and the protonated Schiff base restricted the formation of the active conformation of rhodopsin. After the release of retinal, a charge pair between Glu 113 and Lys 296 fulfilled this role. Removal of the constraint imposed by the charge pair in the mutants allowed the formation of a constitutively active opsin. Although removal of the Schiff base proton during the activation step appears to be required in rhodospins which can form a charge pair, it may not be necessary when the charge pair is not formed. A mutagenesis experiment, in which the double mutant E 113A/A117E was formed, shows that a MI-like photointermediate, whose spectrum indicates the presence of a protonated Schiffbase, is capable of activating G t (Zvyaga et al., 1994). It would appear that the lack of charge pair formation, negates the need to deprotonate the Schiff base in order to form a Gt-activating conformation.

VI.

R H O D O P S I N - L I P I D INTERACTIONS A. Disk Membrane Lipid Composition

The native disk membrane is relatively simple from a compositional perspective. Approximately 90-95% of the integral membrane protein of the disk is rhodopsin (Papermaster and Dreyer, 1974; Smi.th et al., 1975). The lipid composition consists primarily of about 42% phosphatidylethanolamine (PE), 40% phosphatidylcholine (PC), 16% phosphatidylserine (PS), some minor amounts of phosphatidylinositol, and 12-14 mol% cholesterol (Anderson et al., 1976; Stone et al., 1979). The amino-containing lipids are thought to be preferentially distributed towards the cytoplasmic surface of the disk (Litman, 1982). Cell membranes in the retina, as well as those of other neuronal tissue, are unique in their high content of polyunsaturated phospholipid acyl chains (Salem, 1989). In most of these cells, the dominant polyunsaturate is 20:4n6, which is confined primarily to PE and PS. The disk membrane has about a 75/1 phospholipid to rhodopsin ratio (Stubbs and Litman, 1978) and is exceptional in that approximately 47 mol% of its phospholipid acyl chains are derived from docosahexaenoic (22:6n3) acid, with about 16%, 20%, 5% and 3.6 mol% of 16:0, 18:0, 18" ln9, and 20:4n6 respectively (Stone et al., 1979). In addition, the 22:6n3 is prevalent in all three major phospholipid classes.

18

BURTON J. LITMAN and DRAKE C. MITCHELL

Consistent with the high level of acyl chain unsaturation, Cone and coworkers determined rhodopsin to have a high degree of both rotational (Cone, 1972) and translational (Poo and Cone, 1974) freedom, indicating that the disk membrane is a prime example of the Singer-Nicholson fluid mosaic membrane (Singer and Nicholson, 1972).

B. Lipid Modulation of Mll Formation Early studies ofrhodopsin photochemistry demonstrated that MII formation was dependent on both the presence and type of lipid associated with rhodopsin. In delipidated ROS membranes, photolysis of rhodopsin did not progress beyond the MI photointermediate (Applebury et al., 1974). MII formation was also found to be faster in more unsaturated PCs than in disaturated PCs (O'Brien et al., 1977; Mitchell et al., 1992b). The rate of meta II formation increased dramatically when disk membranes were detergent solubilized and was shown to be dependent on the amount ofphospholipid in the rhodopsin-containing micelles (Litman et al., 1981). More recent studies, in which the MI-MII equilibrium constant, Keq, was measured, show that the amount of MII formed is increased by higher levels of acyl chain unsaturation (Wiedmann et al., 1988; Mitchell et al., 1992b; Gibson and Brown, 1993) and decreased by increased bilayer cholesterol contem (Mitchell et al., 1990). Early reports of the lack of MII formation in disaturated PCs led to the speculation that acyl chain unsaturation was a requirement for meta II formation (Baldwin and Hubbell, 1985a,b). Subsequently, it was shown that MII not only formed in di- 14:0 PC (DMPC), but was capable of activating Gt (Mitchell et al., 1991). Thus, it would appear that no requirement of unsaturation exists for either MII formation or subsequent Gt activation. The question then remains as to the role of the preponderance of 22:6 acyl chains in the disk membrane phospholipids and what special properties are imparted to membranes by these polyunsaturated phospholipids. Several properties of phospholipids, related to both head group and acyl chain content, have been examined relative to their ability to influence the function of integral membrane proteins. Head group effects are in general associated with either a surface potential imparted by charged lipids or the force exerted on a membrane protein, when phospholipids such as PE, which have a propensity to form an inverted tubular hexagonal type II (Hn) phase, are confined to a planar bilayer (Gruner, 1985), whereas acyl chain effects are expressed as variation in membrane thickness (Moufitsen and Bloom, 1984) and acyl chain packing free volume (Mitchell et al., 1990, 1992b).

Phospholipid Head Groups In mixtures of 16:0, 18:1 PC (POPC) and POPE, the Keq for MII formation increases as a function of the mole fraction of POPE (Gibson and Brown, 1993; Mone and Litman, 1991). At 37 ° C, Keq increases by about 60% in a 1:1 POPCPOPE relative to pure POPC in reconstituted vesicles (Mone and Litman, 1991).

Rhodopsin Structure and Function

19

Whether this effect is related to the tendency of PE to form HII phase remains to be determined. The presence of PS also enhances the formation of MII (Gibson and Brown, 1991; Mone and Litman, 1991). In a 1:1 POPS-POPC mixture at 37 ° C, Keq increased by about 4 fold, relative to a pure POPC vesicle system (Mone and Litman, 1991). The influence on Keq of 16:0, 18:1 phosphatidic acid, another negatively charged phospholipid, was similar to that of POPS. The effect of PS on Keq is explained by the predicted effect of varying the vesicle surface potential on the pK a of an amino acid side chain of rhodopsin, which is critical with respect to the MI to MII conversion (Mone and Litman, unpublished). Thus studies incorporating PE into PC bilayers suggest that the presence ofphospholipids with a strong tendency to form an H H phase in the bilayer may influence protein conformation, while studies incorporating PS into PC bilayers clearly show that vesicle surface potential can have a marked effect on protein conformation. By affecting the level of activated receptor formed, both these membrane properties can strongly influence the level of function in a signalling pathway.

Phospholipid Acyl Chains The best correlation between PC acyl chain composition and MII formation obtained to date is with a parameter referred to as the fractional volume, fv; this is derived from the dynamic fluorescence anisotropy properties of the hydrophobic probe, DPH (1,6-diphenyl-l,3,5-hexatriene) (Straume and Litman, 1987a,b). F v was used to characterize the relative acyl chain packing free volume of a variety of lipid bilayers. Acyl chain packing free volume is related to bilayer compressibility and appears to be a primary factor in modulating conformational changes of integral membrane proteins. F v relates to phospholipid acyl chain packing properties in the following way. In an ordered ensemble ofphospholipid molecules, the tight packing of the acyl chains limits the solid angle over which a free tumbling probe, such as DPH, can distribute; this is typical of the gel state. However, when the chains are more disordered, such as in the liquid crystalline phase, DPH molecules can distribute over a much larger solid angle. Fractional volume is defined as the volume over which DPH can distribute in the anisotropic structure of the bilayer relative to the volume it distributes over in an isotropic medium such as a liquid hydrocarbon (Straume and Litman, 1987a,b). In phospholipid bilayers in the liquid crystalline phase, fv was found to increase with increasing temperature (Figure 6A), and decrease with increasing cholesterol content (Figure 6B), reflecting the temperature-induced disordering and cholesterol-induced ordering of the acyl chains, observed by other physical techniques. F v also increases going from di(14:0)PC to di(20:4)PC (Figure 6C). This indicates that higher levels of polyunsaturation result in decreased restriction of the angular distribution of DPH, which is associated with increased acyl chain packing free volume. This is in agreement with the weaker interactions between acyl chains of polyunsaturated phospholipids observed in differential scanning calorimetry

20

BURTON J. LITMAN and DRAKE C. MITCHELL

B

0.16 -

0.14 -

rv 0.12 -

0.10 -

0.08

I

10

20

30

40

Temperature

(°C)

0

10

20

Mol% cholesterol

30

I

di(14:0)

I

I

16:0,22:6

16:0,18:1

di(20:4)

Figure 6. Illustration of changes in fractional volume, fv, in response to factors known to modify phospholipid acyl chain packing properties. A. Temperature vs. fv for rhodopsin-containing (16:0,22:6)PC (PDPC) vesicles. B. Bilayer mol% cholesterol vs. fv at 37 °C for rhodopsin-containing egg PC vesicles. C. Acyl chain composition vs. fv at 37 °C. Data in A, B, and C. (DMPC only) are taken from Mitchell et al., 1992b. The remaining data in C is unpublished data of Mitchell and Litman.

(Keough and Kariel, 1987) and bilayer compressibility (Needham and Nunn, 1990) measurements. The temperature dependence of Keq for the MI-MII equilibrium and fv were determined in parallel spectrophotometric and fluorescence measurements made on reconstituted vesicles containing approximately a 100/1 ratio of phospholipid to rhodopsin. A linear relationship was found between Keq and fv for each lipid bilayer examined. The slope of these correlation lines were found to increase with increasing degree of sn-2 unsaturation (Figure 7A). The slopes of the correlation lines for 16:0, 20:4 PC (PAPC) and 16:0, 22:6n3 PC (PDPC) are 18 and 41 respectively, demonstrating that the more polyunsaturated PC has a 2.3 fold greater propensity to promote the formation of MII. These results demonstrate that phospholipids containing polyunsaturated acyl chains have an enhanced ability to Utilize acyl chain packing free volume to promote the formation ofMII. PAPC and PDPC are reported to have very similar hydrophobic thickness (Mclntosh and Brown, unpublished). Therefore, this observation indicates that membrane protein conformational equilibria can be modulated by variation of acyl chain unsaturation under conditions of constant hydrophobic thickness. The experiments summarized in

Rhodopsin Structure and Function

21

Figure 7 demonstrate that the extent of MII formation shows very good correlation with the increase in acyl chain packing free volume induced by increasing levels of unsaturation in the sn-2 chain.

A Ke q

4 3

1.0

B

Koq 0.5

0.0

0.00

0.05

0.10

0.15

fv-fv ° Figure 7. The MI <-->MII equilibrium constant, Keq, as a function of the normalized fractional volume, fv - f v°. The parameter f v° is fv intercept value of the Keq vs fv correlation line. Subtraction of fv ° shifts all lines to a common origin, making a comparison of the variation in slopes more direct. A. Effect of phospholipid acyl chain composition for rhodopsin in di(14:0)PC (DMPC)(o), egg PC (rq), (16:0, 20:4)PC (PAPC) (A), (16:0, 22:6)PC (PDPC) (,), di(20:4)PC (DAPC) (o), and di(22:6)PC (DDPC) (m) vesicles at a rhodopsin to lipid ratio of 100/1. All points resulted from measurements 10, 20, 30, and 37 °C, except for DMPC which was studied at 26, 30, 37, and 45 °C. B. Effect of bilayer cholesterol for 30 mol% cholesterol in (16:0, 20:4)PC (o), 15 mol% cholesterol in egg PC (m), and 30 mol% cholesterol in egg PC (A). Solid lines are the correlation lines due to variation of temperature for rhodopsin in the host phospholipid, as shown in A. Panel A is reproduced from Litman and Mitchell (1996), data in panel B is from Mitchell et al., 1992b.

22

BURTON J. LITMAN and DRAKE C. MITCHELL

Cholesterol Cholesterol is one of the most studied molecules in lipid bilayers. The dependence of Keq and fv on bilayer cholesterol content in egg PC and PAPC is shown in Figure 7B. The Keq, fv data points are for the cholesterol-containing systems and lie on the correlation lines obtained for cholesterol-free egg PC-and PAPC by temperature variation. Thus, the cholesterol-induced decrease in fv induces a proportional decrease in Keq, and the extent of change in Keq for an incremental decrease in fv is determined by the properties of the host phospholipid. These data demonstrate that cholesterol modulates membrane properties by decreasing acyl chain packing free volume within the constraints imposed by the acyl chain composition. In PC systems, the acyl chain composition appears to be the primary factor in determining the bilayer physical properties, which regulate MII formation (Mitchell et al., 1992b). The lipid bilayer serves as the solvating medium for integral membrane proteins and will therefore determine the conformational degrees of freedom available to these proteins. The conversion of MI to MII has associated with it a positive volume change of 100cc/mole (Attwood and Gutfreund, 1980). Cholesterol, which has been shown to decrease acyl chain packing free volume, decreases the amount of MII formed relative to a cholesterol free membrane. Hence, access to conformations with larger volumes is not favored by factors which reduce acyl chain packing free volume. Given that denatured structures are also associated with larger volume conformations, one would predict that cholesterol would increase the thermal stability of an integral membrane protein by stabilizing the native conformation. In experiments where 30% cholesterol was incorporated into egg PC vesicles containing rhodopsin, cholesterol induced a shift in the T m for rhodopsin from 68 ° C to 71.6 ° C (Mone and Litman, 1990). Thus, factors which are favorable to the stabilization of a protein, do not necessarily promote the function of the protein, which may require a certain degree of conformational flexibility to form its active state. The properties of the lipid bilayer must be optimized to promote both protein stability and function.

C. Membrane Lipid Domain Model The Keq, fv correlation data can be explained by the model shown in Figure 8. Raman vibrational spectroscopy data indicate a propensity for phospholipids, which contain a saturated sn-1 chain and a polyunsaturated sn-2 chain, to form microdomains as the level of unsaturation in the sn-2 position increases (Litman et al., 1991). This tendency is driven by the stronger sn-1, sn-1 saturated chain interactions relative to the weaker sn-2, sn-2 polyunsaturated chain interactions. As a result of these interactions sn-1 chains are oriented towards the interior of the domains and preferentially interact with one another rather than with the sn-2 chains. Rhodopsin and its surrounding layer of phospholipids can be viewed as an independent domain within the membrane.

/ sn'l

t

b D

(

CDq3~

D

Figure 8. Top view of the microdomain model for acyl chain packing in a polyunsaturated phospholipid bilayer containing rhodopsin. The small ellipses represent PC molecules in which the sn-1 chain and sn-2 chain are designated by the filled and unfilled sectors, respectively. In the liquid crystalline phase, the domains have a characteristic lifetime and size, which can be influenced by the partitioning of lipid soluble molecules, such as cholesterol, into the bilayer. Rhodopsin is shown surrounded by a rapidly exchanging boundary layer of lipid, creating a protein-containing domain. The sn-1 chains are oriented towards the tightly packed interior of the domain, while the sn-2 chains are at the domain boundary and determine the lateral compressibility properties of the bilayer. 23

24

BURTON J. LITMAN and DRAKE C. MITCHELL

Domain-domain interactions depend on the nature of the sn-2 chain and will strongly influence the lateral compressibility of the membrane. If some portion of the volume expansion associated with the MI to MII conversion is associated with the conformation change in rhodopsin, then the magnitude of Keq will depend on the membrane compressibility. As such, the slope of t h e Keq - fv correlation lines will depend on the acyl chain at the sn-2 position, which is in good agreement with experimental observations. Cholesterol shows a preferential interaction with the saturated sn- 1 chain relative to the polyunsaturated sn-2 chain. This is demonstrated by observations that cholesterol has little effect on either the enthalpy of the gel to liquid cr3/stalline phase transition of dipolyunsaturated lipids (Kariel et al., 1991) or the compressibility modulus ofdi 20:4 PC (Needham and Nunn, 1990). In our model, cholesterol is expected to partition into the interior of the microdomain so as to interact with the saturated sn-1 chain, where it can effect the stability, size, and volume of the domains, thus altering the physical properties of the bilayer. However, if the basic domain structure persists, the lateral packing properties of the system will continue to depend on the sn-2 chain interactions. Cholesterol is observed to shift the system along the correlation line determined by the sn-2 acyl chain, in good agreement with the model.

VII. CONCLUSIONS The unique ability to monitor the conformation changes associated with the functional activation of rhodopsin provide an insight into the forces modulating this process. The extent of formation of MII is influenced by both intramolecular and intermolecular interactions. G l u 113 appears to form a salt bridge with the protonated Schiff base in unbleached rhodopsin and with Lys296 in retinal-free opsin. This salt bridge prevents formation of the active conformation of opsin. In mutagenesis experiments, changing either G l u 113 o r Lys296 so as to prevent formation of the salt bridge was found to produce a constituitively active form of opsin, which can activate Gt in the absence of retinal. The role of photoisomerization of retinal is to induce a conformation change in the retinal binding domain of opsin, which leads to the loss of the Schiffbase proton and disruption of the salt bridge, resulting in formation of MII. These findings demonstrate the importance of intramolecular charge interactions in regulating function in this receptor superfamily. Among intermolecular forces, compositionally determined acyl chain packing properties were found to play a major role in modulating MII formation. The ROS disk membrane phospholipid acyl chain composition is approximately 50% 22:6n3. MII formation was found to increase with increasing phospholipid polyunsaturation. An explanation of the cholesterol and phospholipid acyl chain dependence of MII formation is found in a model of membrane lipid packing. This model suggests that lipids in biological membranes pack in lateral microdomains, whose formation

Rhodopsin Structure and Function

25

is driven by the interaction of saturated sn-1 chains, which is stronger than the interaction of polyunsaturated sn-2 chains. It is significant that the vast majority of the molecular diversity of naturally occurring phospholipids is a result of substitution at the sn-2 position, whereas the sn-1 position shows little heterogeneity and is generally either 16:0 or 18:0. The critical nature of the high content of 22:6n3 acyl chains in ROS disk phospholipids is demonstrated in experiments to determine the effects of a deficiency in this acyl chain. In newborn rhesus monkeys, whose mothers were maintained on an 18:3n3 deficient diet, plasma 18:3n3 was absent and relative to controls, plasma 22:6n3 fell to about 9% in 8 weeks and 6% after 12 weeks. The visual acuity, compared to control infants, fell to about one-half in the 8 to 12 week time period (Neuringer et al., 1984). The role of the 22:6n3 acyl chains appears to be structural, in that higher levels of polyunsaturation promote a higher yield of MII, which would produce a higher level of activity in the visual transduction pathway.

REFERENCES Akhtar, M., Blosse, R T., & Dewhurst, P. B. (1965). The reduction of a rhodopsin derivative. Life Sci. 4, 1221-1226. Applebury, M. L., Zuckerman, D. M., Lamola, A. A., & Jovin, T. M. (1974). Rhodopsin. Purification and recombination with phospholipids assayed by the metarhodopsin I leads to metarhodopsin II transition. Biochemistry 13, 3448-3458. Albert, A. D., & Litman, B. J. (1978). Independent structural domains in the membrane protein bovine rhodopsin. Biochemistry 17, 3893-3900. Anderson, R. E., Landis, D. J., & Dudley, P. A. (1976). Essential fatty acid deficiency and renewal of rod outer segments in the albino rat. Inv. Ophthal. 15, 232-236. Amis, S., & Hofmann, K. P. (1993). Two different forms ofmetarhodopsin II: Schiffbase deprotonation precedes proton uptake and signaling state. Proc. Natl. Acad. Sci. USA 90, 7849-7853. Aton, B., & Litman, B. J. (1984). Activation of rod outer segment phosphodiesterase by enzymatically altered rhodopsin: a regulatory role for the carboxyl terminus of rhodopsin. Exp. Eye Res. 38, 547-559. Attwood, P. V., & Guffreund, H. (1980). The application of pressure relaxation to the study of the equilibrium metarhodopsin I and II from bovine retinas. FEBS Lett. 119, 323-326. Baldwin, J. (1993). The probable arrangement of the helices in G protein-coupled receptors. EMBO J.I 4, 1693-1703. Baldwin, P. A., & Hubbell, W. L. (1985a). Effects of lipid environment on the light-induced conformational changes of rhodopsin 1. Absence of metarhodopsin II production in dimyristoylphosphatidylcholine recombinant membranes. Biochemistry 24, 2624-2632. Baldwin, P. A., & Hubbell, W. L. (1985b). Effects of lipid environment on the light-induced conformational changes of rhodopsin. 2. Roles of lipid chain length, unsamration, and phase state. Biochemistry 24, 2633-2639. Barclay, P. L., & Findlay, J. B. (1984). Labelling of the cytoplasmic domains of bovine rhodopsin with hydrophilic chemical probes. Biochem. J. 220, 75-84. Bennett, N., Michel-Villaz, M., & Kfihn, H. (1982). Light-induced interaction between rhodopsin and the GTP-binding protein. Metarhodopsin II is the major photoproduct involved. E. J. Biochem. 127, 97-103.

26

BURTON J. LITMAN and DRAKE C. MITCHELL

Birge, R. R., Einterz, C. M., Knapp, H. M., & Murray, L. E (1988). The nature of the primary photochemical events in rhodopsin and isorhodopsin. Biophys. J. 53, 367-385. Blatz, P. E., Mohler, J. H., & Navangul, H. V. (1972). Anion-induced wavelength regulation of absorption maxima of Schiffbases of retinal. Biochemistry 11,848-855. Blazynski, C., & Ostroy, S. E. (1984). Pathways in the hydrolysis of vertebrate rhodopsin. Vis. Res. 24, 459-470. Boucher, E, & Leblanc, R. M. (1985). Energy storage in the primary photoreaction of bovine rhodopsin. A photoacoustic study. Photochem. Photobiol. 41,459-465. Bownds, D. (1967). Site of attachment of retinal in rhodopsin. Nature 216, 1178-1181. Chabre, M., & Breton, J. (1979). The orientation of the chromophore of vertebrate rhodopsin in the "meta" intermediate states and the reversibility of the meta II-meta III transition. Vis. Res. 19, 1005-1018. Cone, R. A. (1972). Rotational diffusion of rhodopsin in the visual receptor membrane. Nature New Biol. 236, 39-43. Cooper, A. (1979). Energy uptake in the first step of visual excitation. Nature 282, 531-533. Dartnall, H. J. A. (1968). The photosensitivities of visual pigments in the presence of hydroxylamine. Vis. Res. 8, 339-358. Davidson, F. F., Loewen, P. C., & Khorana, H. G. (1994). Structure and function in rhodopsin: replacement by alanine of cysteine residues 110 and 187, components of a conserved disulfide bond in rhodopsin, affects the light-activated metarhodopsin II state. Proc. Natl. Acad. Sci. USA 9 l, 4029-4033. DeGrip, W. J., Gillespie, J., & Rothschild, K. J. (1985). Carboxyl group involvement in the meta I and meta II stages in rhodopsin bleaching. A Fourier transform infrared spectroscopic study. Biochim. Biophys. Acta 809, 97-106. Doi, T., Molday, R. S., & Khorana, H. G. (1990). Role of the intradiscal domain in rhodopsin assembly and function. Proc. Natl. Acad. Sci. USA 87, 4991-4995. Doukas, A. G., Aton, B., Callender, R. H., & Ebrey, T. G. (1978). Resonance Raman studies of bovine metarhodopsin I and metarhodopsin II. Biochemistry 17, 2430-2435. Emeis, D., Kuhn, H., Reichert, J., & Hofmann, K. P. (1982). Complex formation between metarhodopsin II and GTP-binding protein in bovine photoreceptor membranes leads to a shift of the photoproduct equilibrium. FEBS Lett. 143, 29-34. Fahmy, K., Jager, F., Beck, M., Zvyaga, T. A., Sakmar, T. P., & Siebert, F. (1993). Protonation states of membrane-embedded carboxylic acid groups in rhodopsin and metarhodopsin II: a Fourier-transform infrared spectroscopy study of site-directed mutants. Proc. Natl. Acad. Sci. USA 90, 10206-10210. Falk, G., & Fatt, P. (1966). Rapid hydrogen ion uptake of rod outer segments and rhodopsin solutions on illumination. J. Physiol.(Lond) 183, 211-224. Farahbakhsh, Z. T., Hideg, K., & Hubbell, W. L. (1993). Photoactivated conformational changes in rhodopsin: a time-resolved spin label study. Science 262, 1416-1419. Findlay, J. B., Brett, M., & Pappin, D. J. (1981). Primary structure of C-terminal functional sites in bovine rhodopsin. Nature 293, 314-317. Franke, R. R., Konig, B., Sakmar, T. P., Khorana, H. G., & Hofmann, K. P. (1990). Rhodopsin mutants that bind but fail to activate transducin. Science 250, 123-125. Franke, R. R., Sakmar, T. P., Graham, R. M., & Khorana, H. G. (1992). Structure and function in rhodopsin. Studies of the interaction between the rhodopsin cytoplasmic domain and transducin. J. Biol. Chem. 267, 14767-14774. Fukuda, M. N., Papermaster, D. S., & Hargrave, P. A. (1979). Rhodopsin carbohydrate. Structure of small oligosaccharides attached at two sites near the NH2 terminus. J. Biol. Chem. 254, 82018207.

Rhodopsin Structure and Function

27

Fung, B. K., & Hubbell, W. L. (1978a). Organization of rhodopsin in photoreceptor membranes. 1. Proteolysis of bovine rhodopsin in native membranes and the distribution of sulfhydryl groups in the fragments. Biochemistry 17, 4396-4402. Fung, B. K., & Hubbell, W. L. (1978b). Organization of rhodopsin in photoreceptor membranes. 2. Transmembrane organization of bovine rhodopsin: evidence from proteolysis and lactoperoxidase-catalyzed iodination of native and reconstituted membranes. Biochemistry 17, 4403-4410. Ganter, U. M., Charitopoulos, T., Virmaux, N., & Siebert, F. (1992). Conformational changes ofcytosolic loops of bovine rhodopsin during the transition to metarhodopsin-II: an investigation by Fourier transform infrared difference spectroscopy. Photochem. Photobiol. 56, 57-62. Garcia-Quintana, D., Garriga, P., & Manyosa, J. (1993). Quantitative characterization of the structure of rhodopsin in disc membrane by means of Fourier transform infrared spectroscopy. J. Biol. Chem. 268, 2403-2409. Gibson, N. J., & Brown, M. F. (1991). Role of phosphatidylserine in the MI-MII equilibrium of rhodopsin. Biochem. Biophys. Res. Comm. 176, 915-921. Gibson, N. J., & Brown, M. F. (1993). Lipid headgroup and acyl chain composition modulate the MI-MII equilibrium of rhodopsin in recombinant membranes. Biochemistry 32, 2438-2454. Gruner, S. M. (1985). Intrinsic curvature hypothesis for biomembrane lipid composition: A role for nonbilayer lipids. Proc. Natl. Acad. Sci. USA 82, 3665-3669. Hargrave, P. A., McDowell, J. H., Curtis, D. R., Wang, J. K., Juszczak, E., Fong, S. L., Rao, J. K., & Argos, P. (1983). The structure of bovine rhodopsin. Biophys. Struc., & Mech. 9, 235-244. Hargrave, P. A., Hamm, H., & Hofmann, K. P. (1993). Interaction of rhodopsin with the G-protein, transducin. Bioessays 15, 43-50. Henderson, R., Baldwin, J. M., & Ceska, T. A. (1990). Model for the structure of bacteriorhodopsin based on high-resolution electron cryo-microscopy. J. Mol. Biol. 213, 899-929. Hoffmann, W., Siebert, F., Hofmann, K. P., & Kreutz, W. (1978). Two distinct rhodopsin molecules within the disc membrane of vertebrate rod outer segments. Biochim. Biophys. Acta 503,450-461. Hofmann, K. P., Emeis, D., & Schnetkamp, P. P. (1983). Interplay between hydroxylamine, metarhodopsin II and GTP-binding protein in bovine photoreceptor membranes. Biochim. Biophys. Acta 725, 60-70. Hubbard, R., & Kropf, R. (1958). The action of light on rhodopsin. Proc. Natl. Acad. Sci. USA 44, 130-139. Hug, S. J., Lewis, J. W., & Kliger, D. S. (1988). Evidence for a common batho intermediate ofrhodopsin and isorhodopsin. J. Am. Chem. Soc. 110, 1998-1999. Hug, S. J., Lewis, J. W., Einterz, C. M., Thorgeirsson, T. E., & Kliger, D. S. (1990). Nanosecond photolysis of rhodopsin: evidence for a new, blue-shifted intermediate. Biochemistry 29, 14751485. Kahlert, M., & Hofmann, K. P. (1991). Reaction rate and collisional efficiency of the rhodopsintransducin system in intact retinal rods. Biophys. J. 59, 375-386. Kandori, H., Shichida, Y., & Yoshizawa, T. (1989). Absolute absorption spectra of batho- and photorhodopsins at room temperature. Picosecond laser photolysis of rhodopsin in polyacrylamide. Biophys. J. 56, 453-457. Kariel, N., Davidson, E., & Keough, K. (1991). Cholesterol does not remove the gel liquid crystalline phase transition ofphosphatidylcholines containing two polyenoic acyl chains. Biochim. Biophys. Acta 1062, 70-76. Karnik, S. S., Sakmar, T. P., Chen, H. B., & Khorana, H. G. (1988). Cysteine residues 110 and 187 are essential for the formation of correct structure in bovine rhodopsin. Proc. Natl. Acad. Sci. USA 85, 8459-8463. Karnik, S. S., & Khorana, H. G. (1990). Assembly of functional rhodopsin requires a disulfide bond between cysteine residues 110 and 187. J. Biol. Chem. 265, 17520-17524. Kaushal, S., Ridge, K. D., & Khorana, H. G. (1994). Structure and function in rhodopsin: the role of asparagine-linked glycosylation. Proc. Natl. Acad. Sci. USA 91, 4024-4028.

28

BURTON J. LITMAN and DRAKE C. MITCHELL

Keough, K., & Kariel, N. (1987). Differential scanning calorimetric studies of aqueous dispersions of phosphatidylcholines containing two polyenoic chairs. Biochim. Biophys. Acta 902, 11-18. Khan, S. M., Bolen, W., Hargrave, P. A., Santoro, M. M., & McDowell, J. H. (1991). Differential scanning calorimetry of bovine rhodopsin in rod-outer-segment disk membranes. E. J. Biochem. 200, 53-59. Khorana, H. G. (1992). Rhodopsin, photoreceptor of the rod cell. An emerging pattern for structure and function. J. Biol. Chem. 267, 1-4. Kibelbek, J., Mitchell, D. C., Beach, J. M., & Litman, B. J. (1991). Functional equivalence of metarhodopsin II and the Gt-activating form of photolyzed bovine rhodopsin. Biochemistry 30, 6761-6768. Klinger, A. L., & Braiman, M. S. (1992). Structural comparison of metarhodopsin II, metarhodopsin III, and opsin based on kinetic analysis of Fourier translbrm infrared difference spectra. Biophys. J. 63, 1244-1255. Konig, B., Arendt, A., McDowell, J. H., Kahlert, M., Hargrave, P. A., & Hofmann, K. P. (1989). Three cytoplasmic loops of rhodopsin interact with transducin. Proc. Natl. Acad. Sci. USA,86, 68786882. Kropf, A. (1967). Intramolecular energy transfer in rhodopsin. Vis. Res. 7, 811-818. Ktihn, H., & Hargrave, P. A. (1981). Light-induced binding of Guanosinetriphosphate to bovine photoreceptor membranes: Effect of limited proteolysis of the membranes. Biochemistry 20, 2410-2417. Lamola, A. A., Yamane, T., & Zipp, A. (1974). Effects of detergents and high pressures upon the metarhodopsin I ~ metarhodopsin II equilibrium. Biochemistry 13, 738-745. Lewis, J. W., & Kliger, D. S. (1992). Photointermediates of visual pigments. J. Bioenerg. Biomem. 24, 201-210. Lewis, J. W., Winterle, J. S., Powers, M. A., Kliger, D. S., & Dratz, E. A. (1981). Kinetics of rhodopsin photolysis intermediates in retinal rod disk membranes. I. Temperature dependence of lumirhodopsin and metarhodopsin I kinetics. Photochem. Photobiol. 34, 375--384. Lewis, J. W., Hug, S. J., Wallace-Williams, S. E., & Kliger, D. S. (1990). Direct evidence for an equilibrium between early photolysis intermediates of rhodopsin. J. Am. Chem. Soc. 112, 6711-6712. Liang, C. J., Yamashita, K., Muellenberg, C. G., Shichi, H., & Kobata, A. (1979). Structure of the carbohydrate moieties of bovine rhodopsin. J. Biol. Chem. 254, 6414--6418. Liebman, P. A. (1962). In situ microspectrophotometric studies on the pigments of single retinal rods. Biophys. J. 2, 161-178. Litman, B. J., Kalisky, O., & Ottolenghi, M. (1981). Rhodopsin -phospholipid interactions: dependence of rate of the meta I to meta II transition on the level of associated disk phospholipid. Biochemistry 20, 631-634. Litman, B. J., Aton, B., & Hartley, J. B. (1982). Functional domains of rhodopsin. Vis. Res. 22, 1439-1442. Litman, B. J. (1982). Methods for determining rod outer segment disk phospholipid transmembrane topology. Meth. In Enzymol. 81,309-315. Litman, B. J., Lewis, E. N., & Levin, I. W. (1991). Packing characteristics of highly unsaturated bilayer lipids: Raman spectroscopic studies of multilamellar phosphatidylcholine dispersions. Biochemistry 30, 31 3-319. Litman, B. J., & Mitchell, D. C. (1996). A role for phospholipid polyunsaturation in modulating membrane protein function. Lipids. In press. Longstaff, C., Calhoon, R., & Rando, R. (1986). Deprotonation of the Schiffbase ofrhodopsin is obligate in the activation of the G protein. Proc. Natl. Acad. Sci. 83, 4209-4213. Martynov, V. I., Kostina, M. B., Feigina, M. I., & Miroshnikov, A. I. (1983). Study of the molecular organization of visual rhodopsin in photoreceptor membranes by limited proteolysis. Bioorganich. Khim. 9, 734-745.

Rhodopsin Structure and Function

29

Mathews, R., Hubbard, R., Brown, P., & Wald, G. (1963). Tautomeric forms of metarhodopsin. J. Gen. Phys. 47, 215-240. Mathies, R., Oseroff, A. R., & Stryer, L. (1976). Rapid-flow resonance Raman spectroscopy of photolabile molecules: rhodopsin and isorhodopsin. Proc. Natl. Acad. of Sci. USA 73, 1-5. McDowell, J. H., Nawrocki, J. P., & Hargrave, P. A. (1993). Phosphorylation sites in bovine rhodopsin. Biochemistry 32, 4968--4974. Michel-Villaz, M., Saibil, H. R., & Chabre, M. (1979). Orientation ofrhodopsin alpha-helices in retinal rod outer segment membranes studied by infrared linear dichroism. Proc. Natl. Acad. Sci. USA 76, 4405-4408. Michel-Villaz, M., Roche, C., & Chabre, M. (1982). Orientational changes of the absorbing dipole or retinal upon the conversion of rhodopsin to bathorhodopsin, lumirhodopsin, and isorhodopsm. Biophys. J. 37, 603-616. Miller, J. L., Fox, D. A., & Litman, B. J. (1986). Amplification ofphosphodiesterase activation is greatly reduced by rhodopsin phosphorylation. Biochemistry 25, 4983--4988. Mitchell, D. C., Straume, M., Miller, J. L., & Litman, B. J. (1990). Modulation of metarhodopsin formation by cholesterol-induced ordering of bilayer lipids. Biochemistry 29, 9143-9149. Mitchell, D. C., Kibelbek, J., & Litman, B. J. (1991). Rhodopsin in dimyristoylphosphatidylcholine-reconstituted bilayers forms metarhodopsin II and activates Gr Biochemistry 30, 37-42. Mitchell, D. C., Kibelbek, J., & Litman, B. J. (1992a). Effect ofphosphorylation on receptor conformation: the metarhodopsin I in equilibrium with metarhodopsin II equilibrium in multiply phosphorylated rhodopsin. Biochemistry 31, 8107-8111. Mitchell, D. C., Straume, M., & Litman, B. J. (1992b). Role of sn-l-saturated, sn-2-polyunsaturated phospholipids in control of membrane receptor conformational equilibrium: effects of cholesterol and acyl chain unsaturation on the metarhodopsin I in equilibrium with metarhodopsin II equilibrium. Biochemistry 31,662-670. Molday, R. S., & Molday, L. L. (1979). Identification and characterization ofmultiple forms ofrhodopsin and minor proteins in frog and bovine rod outer segment disc membranes. Electrophoresis, lectin labeling, and proteolysis studies. J. Biol. Chem. 254, 4653-4660. Mone, A. P., & Litman, B. J. (1990). Cholesterol increases the stability of rhodopsin in egg PC vesicles to thermally induced unfolding. Biophys. J. 57, 74a. Mone, A. P., & Litman, B. J. (1991). Variation of phospholipid headgroup composition modulates the formation of metarhodopsin II. Biophys. J. 59, 621 a. Monger, T. G., Alfano, R. R., & Callender, R. H. (1979). Photochemistry ofrhodopsin and isorhodopsin investigated on a picosecond time scale. Biophys. J. 27, 105-115. Mouritsen, O. G., & Bloom, M. (1984). Mattress model of lipid-protein interactions in membranes. Biophys. J. 46, 141-153. Mullen, E., & Akhtar, M. (1981). Topographic and active-site studies on bovine rhodopsin. FEBS Lett. 132, 261-264. Mullen, E., & Akhtar, M. (1983). Structural studies on membrane-bound bovine rhodopsin. Biochem. J. 211, 45-54. Nakayama, T. A., & Khorana, H. G. (1990). Orientation of retinal in bovine rhodopsin determined by cross-linking using a photoactivatable analog of 11-cis-retinal. J. Biol. Chem. 265, 15762-15769. Nakayama, T. A., & Khorana, H. G. (1991). Mapping of the amino acids in membrane-embedded helices that interact with the retinal chromophore in bovine rhodopsin. J. Biol. Chem. 266, 4269-4275. Nathans, J., & Hogness, D. S. (1983). Isolation, sequence analysis, and intro-econ arrangement of the gene encoding bovine rhodopsin. Cell 34, 807-8 14. Nathans, J. (1990a). Determinants of visual pigment absorbance: identification of the retinylidene Schiff's base counterion in bovine rhodopsin. Biochemistry 29, 9746-9752. Nathans, J. (1990b). Determinants of visual pigment absorbance: role of charged amino acids in the putative transmembrane segments. Biochemistry 29, 937-942. Nathans, J. (1992). Rhodopsin: Structure function, and genetics. Biochemistry 31, 4923-4931.

30

BURTON J. LITMAN and DRAKE C. MITCHELL

Needham, D., & Nunn, R. S. (1990). Elastic deformation and failure of lipid bilayer membranes containing cholesterol. Biophys. J. 58, 997-1009. Neuringer, M., Condor, W. E., Van Patten, C., & Barstad, L. (1984). Dietary omega-3 fatty acid deficiency and visual loss in infant Rhesus monkeys. J. Clin. Invest. 73,272-276. O'Brien, D. F., Costa, L. F., & Ott, R. A. (1977). Photochemical functionality ofrhodopsin-phospholipid recombinant membranes. Biochemistry 16, 1295-1303. Ovchinnikov, Y. A., Abdulaev, N. G., Feigina, M. I., Artamonov, I. D., & Bogachuk, A. S. (1983). Visual rhodopsin. III. Complete amino acid sequence and topography in a membrane. Bioorgan. Khim. 9, 1331-1340. Ovchinnikov Y. A, Abdulaev, N. G., & Bogachuk, A. S. (1988). Two adjacent cysteine residues in the C-terminal cytoplasmic fragment of bovine rhodopsin are palmitylated. FEBS Lett. 230, 1-5. Papac, D. I., Oatis, J. E., Jr., Crouch, R. K., & Knapp, D. R. (1993). Mass spectrometric identification of phosphorylation sites in bleached bovine rhodopsin. Biochemistry 32, 5930-5934. Papermaster, D. S., & Dreyer, W. J. (1974). Rhodopsin content in the outer segment membranes of bovine and frog retinal rods. Biochemistry 13, 2438-2444. Parkes, J. H., & Liebman, P. A. (1984). Temperature and pH dependence of metarhodopsin I--metarhodopsin II kinetics and equilibrium in bovine rod disk membrane suspensions. Biochemistry 23, 5054-5061. Pellicone, C., Nullans, G., & Virmaux, N. (1985). Localization of light-induced conformational changes in bovine rhodopsin. FEBS Lett. 181,179-183. Peters, K., Applebury, M. L., & Rentzepis, P. M. (1977). Primary photochemical event in vision: proton translocation. Proc. Natl. Acad. Sci. USA 74, 3119-3123. Pfister, C., Kuhn, H., & Chabre, M. (1983). Interaction between photoexcited rhodopsin and peripheral enzymes in frog retinal rods. Influence on the postmetarhodopsin II decay and phosphorylation rate of rhodopsin. E. J. Biochem. 136, 489-499. Pober, J. S., & Stryer, L. (1975). Light dissociates enzymatically-cleaved rhodopsin into two different fragments. J. Mol. Biol. 95, 477-481. Poo, M., & Cone, R. A. (1974). Lateral diffusion of rhodopsin in the photoreceptor membrane. Nature 247, 438--441. Randall, C. E., Lewis, J. W., Hug, S. J., Bjorling, S. C., Eisner-Shanas, I., Friedman, N., Ottolengi, M., Sheves, M., & Kliger, D. S. (1991). A New photolysis intermediate in artificial and native visual pigments. J. Am. Chem. Soc. 113, 3473-3485. Resek, J. F., Farahbakhsh, Z. T., Hubbell, W. L., & Khorana, H. G. (1993). Formation of the meta II photointermediate is accompanied by changes in the cytoplasmic surface ofrhodopsin. Biochemistry 32, 12025-12032. Robinson, P. R., Cohen, G. B., Zhukovsky, E. A., & Oprian, D. D. (1992). Constitutively active mutants of rhodopsin. Neuron 9, 71 9-725. Rohlich, P. (1976). Photoreceptor membrane carbohydrate on the intradiscal surface of retinal rod disks. Nature 263, 789-791. Rothschild, K. J., Gillespie, J., & DeGrip, W. J. (1987). Evidence for rhodopsin refolding during the decay of Meta II. Biophys. J. 51,345-350. Sakmar, T. P., Franke, R. R., & Khorana, H. G. (1989). Glutamic acid-ll3 serves as the retinylidene Schiff base counterion in bovine rhodopsin. Proc. Natl. Acad. Sci. USA 86, 8309-8313. Salem, N., Jr. (1989). Omega-3 fatty acids: Molecular and biochemical aspects. In: New Protective Roles for Selected Nutrients (Spiller, G. A., & Scala, J., eds.), pp. 109-228, Alan R. Liss Inc., New York, NY. Schertler, G. F., Villa, C., & Henderson, R. (1993). Projection structure of rhodopsin. Nature 362, 770-772. Schick, G. A., Cooper, T. M., Holloway, R. A., Murray, L. P., & Birge, R. R. (1987). Energy storage in the primary photochemical events of rhodopsin and isorhodopsin. Biochemistry 26, 2556-2562.

Rhodopsin Structure and Function

31

Shichi, H., & Shelton, E. (1974). Assessment of physiological integrity of sonicated retinal rod membranes. J. Supramol. Struct. 2, 7-16. Singer, S. J., & Nicholson, G. L. (1972). The fluid mosaic model of cell membranes. Science 175, 720-731. Smith, H. G., Jr., Stubbs, G. W., & Litman, B. J. (1975). The isolation and purification of osmotically intact discs from retinal rod outer segments. Exp. Eye Res. 20, 211-217. Smith, S. O., Courtin, J., de Groot, H., Gebhard, R., & Lugtenburg, J. (1991). 13C magic-angle spinning NMR studies of bathorhodopsin, the primary photoproduct of rhodopsm. Biochemistry 30, 7409-7415. Stewart, J. G., Baker, B. N., & Williams, T. P. (1977). Evidence for conformeric states of rhodopsm. Biophys. Struc. Mech. 3, 19-29. Stone, W. L., Farnsworth, C. C., & Dratz, E. A. (1979). A reinvestigation of the fatty acid content of bovine. Rat and frog retinal rod outer segments. Exp. Eye Res. 28,387-397. Straume, M., & Litman, B. J. (1987a). Influence of cholesterol on equilibrium and dynamic bilayer structure of unsaturated acyl chain phosphatidylcholine vesicles as determined from higher order analysis of fluorescence anisotropy decay. Biochemistry 26, 5121-5126. Straume, M., & Litman, B. J. (1987b). Equilibrium and dynamic structure of large, unilamellar, unsaturated acyl chain phosphatidylcholine vesicles. Higher order analysis of 1,6-diphenyl- 1,3,5hexatriene and 1-[4-(trimethylammonio)phenyl]- 6-phenyl-l,3,5-hexatriene anisotropy decay. Biochemistry 26, 5113-5120. Straume, M., Mitchell, D. C., Miller, J. L., & Litman, B. J. (1990). Interconversion of metarhodopsins I and II: a branched photointermediate decay model. Biochemistry 29, 9135-9142. Stryer, L. (1991). Visual excitation and recovery. J. Biol. Chem. 266, 10711-10714. Stubbs, G. W., & Litman, B. J. (1976). Microviscosity of the hydrocarbon region of the bovine retinal rod outer segment disk membrane determined by fluorescent probe measurements. Biochemistry 15, 2766-2772. Stubbs, G. W., & Litman, B. J. (1978). Effect of alterations in the amphipathic microenvironment on the conformational stability of bovine opsin. 1. Mechanism of solubilization of disk membranes by the nonionic detergent, octyl glucoside. Biochemistry 17, 215-219. Thomas, D. D., & Stryer, L. (1982). Transverse location of the retinal chromophore of rhodopsin in rod outer segment disc membranes. J. Mol. Biol. 154, 145-157. Thorgeirsson, T. E., Lewis, J. W., Wallace-Williams, S. E., & Kliger, D. S. (1992). Photolysis of rhodopsin results in deprotonation of its retinal Schiff's base prior to formation ofmetarhodopsin II. Photochem. Photobiol. 56, 1135-1144. Thorgeirsson, T. E., Lewis, J. W., Wallace-Williams, S. E., & Kliger, D. S. (1993). Effects of temperature on rhodopsin photointermediates from lumirhodopsin to metarhodopsin II. Biochemistry 32, 13861-13872. Trayhurn, P., Mandel, P., & Virmaux, N. (1974). Removal of a large fragment of rhodopsin without changes in its spectral properties, by proteolysis of retinal rod outer segments. FEBS Lett. 38, 351-353. Wang, J. K., McDowell, J. H., & Hargrave, P. A. (1980). Site of attachment of l l-cis-retinal in bovine rhodopsin. Biochemistry 19, 5111-5117. Weitz, C. J., & Nathans, J. (1992). Histidine residues regulate the transition of photoexcited rhodopsin to its active conformation, metarhodopsin II. Neuron 8, 465-472. Wiedmann, T. S., Pates, R. D., Beach, J. M., Salmon, A., & Brown, M. E (1988). Lipid-protein interactions mediate the photochemical function of rhodopsin. Biochemistry 27, 6469-6474. Wilden, U., Hall, S. W., & Kuhn, H. (1986). Phosphodiesterase activation by photoexcited rhodopsin is quenched when rhodopsin is phosphorylated and binds the intrinsic 48-kDa protein of rod outer segments. Proc. Natl. Acad. Sci. USA 83, 174-1178. Wong, J. K., & Ostroy, S. E. (1973). Hydrogen ion changes of rhodopsin I. Proton uptake during the metarhodopsin 1 478 metarhodopsin II 380 reaction. Arch. Biochem. Biophys 154, 1-7.

32

BURTON J. LITMAN and DRAKE C. MITCHELL

Yan, M., Manor, D., Weng, G., Chao, H., Rothberg, L., Jedju, T. M., Alfano, R. R., & Callender, R. H. (1991). Ultrafast spectroscopy of the visual pigment rhodopsin. Proc. Natl. Acad. Sci. USA 88, 9809-9812. Zhukovsky, E. A., & Oprian, D. D. (1989). Effect of carboxylic acid side chains on the absorption maximum of visual pigments. Science 246, 928-930. Zvyaga, T. A., Min, K. C., Beck, M., & Sakmar, T. P. (1993). Movement of the retinylidene Schiffbase counterion in rhodopsin by one helix turn reverses the pH dependence of the metarhodopsin I to metarhodopsin II transition. J. Biol. Chem. 268, 4661-4667. Zvyaga, T. A., Fahmy, K., & Sakmar, T. P. (1994). Characterization ofrhodopsin-transducin interaction: a mutant rhodopsin photoproduct with a protonated Schiffbase activates transducin. Biochemistry 33, 9753--9761.

CHARACTERIZATION OF THE PRIMARY PHOTOCHEMICAL EVENTS IN BACTERIORHODOPSIN AN D RHODOPSIN

Jeffrey A. Stuart and Robert R. Birge

I.

II.

III.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Bacteriorhodopsin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Rhodopsin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Bacteriorhodopsin and Rhodopsin Active Sites . . . . . . . . . . . . . . . A. The Bacteriorhodopsin Active Site . . . . . . . . . . . . . . . . . . . . . . B. Chromophore Orientation and Location . . . . . . . . . . . . . . . . . . . C. Characterization o f the Opsin Shift . . . . . . . . . . . . . . . . . . . . . . D. The Rhodopsin Active Site . . . . . . . . . . . . . . . . . . . . . . . . . . The Primary Event in Bacteriorhodopsin . . . . . . . . . . . . . . . . . . . . . A. The J Intermediate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

B. C. D.

The K Intermediate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

34 34 37 41 42 43 45 56 64 66 66

Ultrafast Spectroscopic Studies . . . . . . . . . . . . . . . . . . . . . . . Quantum Efficiency o f the Primary Event . . . . . . . . . . . . . . . . . .

70 72

Biomembranes Volume 2A, pages 33-139 Copyright © 1996 by JAI Press Inc. All rights of reproduction in any form reserved. ISBN: 1-55938-659-2 33

JEFFREY A. STUART and ROBERT R. BIRGE

34

IV.

V.

VI. VIII.

E. Photoelectric Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75 F. Molecular Dynamics of the Primary Event . . . . . . . . . . . . . . . . . 75 G. Energy Storage in the Primary Event . . . . . . . . . . . . . . . . . . . . 78 The Primary Event in Rhodopsin . . . . . . . . . . . . . . . . . . . . . . . . 80 A. Photorhodopsin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82 B. Bathorhodopsin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84 C. Ultrafast Spectroscopic Studies . . . . . . . . . . . . . . . . . . . . . . . 88 D. Quantum Efficiency of the Primary Event . . . . . . . . . . . . . . . . . 96 E. Molecular Dynamics of the Primary Event . . . . . . . . . . . . . . . . . 98 F. Energy Storage in the Primary Event . . . . . . . . . . . . . . . . . . . . 103 G. Molecular Origins of Photoreceptor Noise . . . . . . . . . . . . . . . . . 103 H. Parallels between Rhodopsin and Bacteriorhodopsin Photochemistry . . . 114 Response of the Protein During the Primary Event . . . . . . . . . . . . . . . 115 A. Bacteriorh0dopsin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115 B. Rhodopsin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117 Protein Control of Isomerization . . . . . . . . . . . . . . . . . . . . . . . . . 118 Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120 I.

INTRODUCTION A. Bacteriorhodopsin

Bacteriorhodopsin is a 26 kD (248 amino acids) molecular weight protein contained within the cell membrane of Halobacterium salinarium (informally called Halobacterium halobium and reclassified in accordance with (Tindall, 1992)), a halophilic archaebacterium that grows and prospers in salt marshes and lakes where the concentration of NaC1 can exceed 4 M; roughly six times that of sea water (-0.6, M). Its primary and approximate secondary structure is shown in Figure 1. As the name implies, this unique protein possesses many similarities to the visual protein rhodopsin, although the physiological functions of the two proteins are quite different. Rhodopsin acts as the primary photoreceptor that enables dim light vision for most vertebrate animals. Bacteriorhodopsin (BR) enables the bacteria to act as a facultative anaerobe when the oxygen tension of its environment is low. The protein accomplishes this function by acting as a light driven proton pump which causes the formation of an electrochemical gradient that can be harnessed by the bacterial cell to do work (thus converting sun light into chemical energy). Absorption of light by the retinal cofactor in this protein initiates a primary photochemical event followed by a series of thermal relaxations back to the initial state. Each of the intermediates is characterized by a different absorption maximum, and decays with time constants ranging from picoseconds to milliseconds (Figure 2). The formation of the earliest well defined intermediate, K, involves a trans to cis isomerization about the C13=C14 bond (Figure 3). The primary work done by the

Cytoplasmic Surface

(INSIDE)

35

(OUTS1 DE) Surface N I

figure 1. Amino acid sequences and putative membrane spanning regions of bacteriorhodopsin. Amino acid abbreviations are as follows: (alanine (Ala, A); arginine (Arg, R); Asparagine (Asn, N);Aspartic Acid (Asp, D); Cysteine (Cys, C); Glutarnate (Glu, E); Glutamine (Gln, 0);Glycine (Gly, G); Histidine (His, H); lsoleucine (Ile, I);Leucine (Leu, L); Lysine (Lys, K); Methionine (Met, M); Phenylalanine (Phe, F); Proline (Pro, P); Serine (Ser, S); Threonine (Thr, T); Tryptophan (Trp, W); Tyrosine (Tyr, Y); Valine (Val, V). Amino acids which under nominal conditions carry a charge are shown in thickened circles (positively charged) or boxes (negatively charged).

36

JEFFREYA. STUART and ROBERT R. BIRGE

kv/~(620)~

~

(5 Sis (>150K)]

I hV

\ ,~

-7ms

~V

hv

.umirhod~opsi (497)) n( lOOns (>I30K____)I t

<

!

tl

(I~~I

t

('0s(p"-7' ]

iletarhodopsin II(380)

etarhodopsiI11n(465)~

hV "~P(490))

(Rhodopsin (498)~k

(a)

athorhodopsin (540)~

|l-trans Retinal(380)"~ & Opsin (280) 2

(b)

Figure 2. Photocycle of light-adapted bacteriorhodopsin (left) and photobleaching sequence of vertebrate rhodopsin (right). Individual species are indicated within the polygons and the numbers in parentheses following the labels indicate the approximate absorption maximum in nanometers extrapolated to ambient temperature. Only selected intermediates are shown, and species that cannot be trapped at low temperatures are not shown. Accordingly, J (an unstable species that precedes K) and photorhodopsin (an unstable species that precedes bathorhodopsin) are not included. Relative free energies are related approximately to vertical position. Temperatures required for observing the formation of subsequent intermediates and formation times extrapolated to ambient temperature are indicated for selected reactions, and are very approximate. A few key thermal and photochemical branching reactions are shown. gradient (ApH .~ 1) is the synthesis of ATP via anaerobic oxidative phosphorylation, and as such was one of the first systems demonstrated to support Mitchell's chemiosmotic hypothesis (Racker and Stoeckenius, 1974). When light is absent and the oxygen tension is high, the bacteria reverts to the more common aerobic oxidative phosphorylation. Despite many years of research on this protein, the nature of the proton pump mechanism is still unknown. The lack of a definitive mechanism is due in part to a lack of concrete information about the chromophore active site interactions. It is evident from this description, albeit brief, why this protein is so profoundly interesting to study (especially in light of the similarities to its visual analogs). For an excellent review of bacteriorhodopsin's photocycle and its functioning as a proton pump see the recent review by Ebrey (1993). We will use the symbol BR to represent the protein, bacteriorhodopsin, and the symbol bR to represent the light-adapted protein which contains only the all-trans retinyl chromophore. In the dark, a mixture of 13-cis and all-trans

Characterization of the Primary Photochemical Events

(~

37

8q(x)= -0.325 &l(y)= 0.713 Aq(rms)= 0.784 A~= 3.1 °~

K

bFI

Figure 3.

A model for the primary event in light-adapted bacteriorhodopsin, including the change in the transition dipole moment between bR and K associated with the uu <--So transition. The bR state chromophore is low-lying strongly-allowed ,,1~,+,, all-trans, and is shown with dark shading. The K state chromophore is 13-cis, and is shown with lighter shading. The contours indicate Iocalizations of positive (dark) and negative (light) charge shifts as a result of the photoisomerization. These'charge shifts are responsible for the fast photovoltaic signal that accompanies the primary event.

chromophores coexists in thermal equilibrium, but the population containing 13-cis does not pump protons and is not the biologically active form. Our discussion is limited to an analysis of the primary photochemical event in light-adapted bacteriorhodopsin.

B. Rhodopsin Rhodopsin is the primary photoreceptor in dim light for most higher organisms. It is a 348 amino acid protein with a molecular weight of about 40,000 D, and is the most studied of a broad class of receptors that interacts with GTP-binding proteins (Figure 4). All of these proteins share a common structural motif of seven transmembrane spanning c~-helices bundled together to form a central binding site. The central function of this group of proteins is one of signal transduction as the result of an initial stimulus. In rhodopsin, this stimulus is the absorption of a photon, which results in a series ofphotochemically induced transformations in the protein (Figure 2), eventually eliciting a chemical cascade response and the generation of a signal in the optic nerve (see below). In the other proteins of this class of receptors, the stimulus typically involves the binding of a hormone to the receptor, thus initiating the familiar cascade mechanism of signal amplification that is charac-

38

JEFFREYA. STUART and ROBERT R. BIRGE

c(~U

v~®e"~-

Cytoplasmic

~

Surface

~.,~o.,,u,or

(~ ~/~

- - - . c ~ ........ 2. .......... ©

:~--

~

"-

~

"-"

F

~:'; ~(::~ (~ ~ ~J,~.C~(~) ,k~~ ,~ ,~ AI,~) ,~

...... [...e.. ....~ . !

~

:.-.~:':.':----~[

~

........ ~

~"., ~

..... i

-(~)

_.~1~

~

........ ~

~_,~'

.....

(INSIDE)

(ou,~,o~

%

%.:,_ ,a~_ ® ~ a ) ® , ~ ~ .~..~ '" ,~o® ~ " " ~®©(E~"~"

Figure 4. Amino acid sequence, and putative membrane spanning regions of human rhodopsin. Amino acid abbreviations are as given in Figure 1. The chromophore is covalently linked to the protein via a protonated Schiff base linkage to Lys296. teristic of nearly all biological signal transduction systems. Until recently, bacteriorhodopsin's similarity to this family of proteins, and to rhodopsin in particular, seemed to be purely coincidental. However, work by Bibikov and coworkers was shown that BR may actually have a role as a proton-motive force sensor that initiates a photophobic response, in addition to its role in solar energy conversion (Bibikov et al., 1991; Skulachev, 1992). A recent article by Henderson and Schertler compares and contrasts the properties of the rhodopsins and their bacterial analogues (Henderson et al., 1990b). The key features ofthe binding sites ofbactefiorhodopsin and rhodopsin are shown in Figures 5 and 6. Rhodopsin is responsible for generating an optic nerve impulse in the visual receptors of the three phyla that possess image-resolving eyes: mollusks, arthropods and vertebrates. Despite what is believed to be independent evolutionary development, these three phyla have converged on a remarkably similar protein structure and an identical light absorbing chromophore, 11-cis-retinal (Figure 6). The generation of a nerve impulse following excitation of rhodopsin involves a complex series of reactions which ultimately hyperpolarize the plasma membrane of the rod cells in the retina (Liebman et al., 1987; Stryer, 1986). The plasma membrane contains numerous cation-specific channels which are open to sodium ion flow in

,<

39 r.~

r.~

Figure 5. The membrane spanning segments of the protein backbone of bacteriorhodopsin along with the all-trans retinal chromophore and selected amino acids viewed from the cytoplasmic side (data from Henderson et al. 1990). The chrornophore is attached to lysine 21 6 (Lys~i6)and the amino acids shown were selected based on their importance to the photochemical properties of the protein (ASP = aspartic acid; TYR = tyrosine). The helix labels (A-G) correspond to those used to identify the membrane spanning regions in Figure 1.

40

JEFFREY A. STUART and ROBERT R. BIRGE

Rhodopsin

\:

:/

Bathorhodopsin Figure 6. Our model of the binding site of rhodopsin (top) and the primary photochemical event that generates bathorhodopsin (bottom) based on the available spectroscopic, photocalorimetric and quantum efficiency data. Key chromophorecounterion electrostatic interactions are indicated with dashed lines. The Glu113 counterion on helix C is represented by a fixed CH3-CH2-CO2- moiety. The position of the counterion was optimized to reproduce as closely as possible the absorption maxima of rhodopsin and bathorhodopsin as well as the energy storage in bathorhodopsin (Birge et al., 1988). Note that the binding site is neutral (Birge et al., 1985). Key features of the model are from references cited within the text. the dark. Light sets offa series of biochemical reactions which block these channels, and the resultant hyperpolarization generates a more negative potential inside the cell. The key feature of this event is that a single photon of light can generate a hyperpolarization of close to 1 mV (~106 Na + ions blocked), which is sufficient to activate a nerve impulse in a dark adapted retina. (In practice, ~ 100 photons must enter the eye in order to generate an observed reproducible physiological response. Light adaptation decreases this sensitivity level further.) The significant amplification factor in the hyperpolarization of the plasma membrane is accomplished via a

Characterization of the Primary Photochemical Events

41

complex biochemical reaction cycle involving transducin, a peripheral membrane protein o f the G-protein family. A single molecule o fphotoactivated rhodopsin (R*) catalyzes the activation of up to 1,000 transducin molecules (T~cGDP + R* --~ R*-Ta~v-GDP + GTP --~ R*-Ta~v -GTP + GDP), and represents the initial stage in the amplification process. The second stage of amplification involves a splitting off of the a subunit of transducin from the 13 and 7 subunits, and activation of phosphodiesterase (PDE) by the a subunit (PDE + R*-Ta~v-GTP--~ PDE + T~-GTP + R* + Tf3v --~ PDE*-Ta-GTP + R* + T~v). (The binding of GTP to transducin releases activated rhodopsin (R*) for continued catalytic activity via the initial stage.) The activated phosphodiesterase complex (PDE*-T~-GTP) hydrolyzes cyclic GMP (c-GMP) to 5'-GMP which closes the sodium ion channels [c-GMP + H20 + open channel(s) --4 5'-GMP + H + + closed channel(s)]. The transducin cycle returns to the starting point through deactivation ofphosphodiesterase via hydrolysis of GTP bound to T~ and recombination of the 13 and ~/ subunits with T~ (PDE*-Ta-GTP + T~v --+ PDE + Ta~v-GDP). Two mechanisms operate to close down the cascade and regenerate the resting state in preparation for reactivation by a subsequent photon absorption event. Activated rhodopsin (R*) is removed through phosphorylation followed by the binding of arrestin (An) (R* + ATP + An --~ R*-P + ADP + An ~ An-R*-P). Arrestin is an inhibitory protein than blocks the binding oftransducin to photoactivated rhodopsin. The second mechanism involves restoration of the open channels via catalysis of GTP with guanylate cyclase (GC) followed by hydrolysis ofpyrophosphate (PP) (closed channels + GTP + GC + H20 c-GMP + PP + H20 --~ c-GMP + 2P + open channels).

II. THE BACTERIORHODOPSIN A N D RHODOPSIN ACTIVE SITES Any detailed discussion of the primary event in bacteriorhodopsin and rhodopsin must include a description of the retinal chromophores, the binding sites, and their associated interactions. It is these interactions, after all, that will ultimately determine the nature of the primary event. In the case of the rhodopsins, several factors must be taken into account: these include structural considerations of the retinal chromophores, interactions with the surrounding protein, and the chromophore's electronic properties. The lack of high resolution diffraction data on rhodopsin and bacteriorhodopsin, however, has precluded definitive assignment of the binding site geometries in either protein. However, the contributions made by Henderson and coworkers with their model of bacteriorhodopsin based upon high resolution electron cryo-microscopy have had a significant impact in the field (Henderson et al., 1990a). A projection structure of rhodopsin to 9 .&resolution has recently been reported by the same authors (Schertler et al., 1993). Nevertheless, spectroscopic, site-directed mutagenesis, and chromophore analog studies all have provided insight into the nature of these binding sites. Research in many fields, and with many techniques, has contributed to the bulk of knowledge on both proteins. The

42

JEFFREYA. STUART and ROBERT R. BIRGE

scope of what follows is by no means intended to represent a comprehensive review, but to present the reader with models of the binding sites that are at present generally accepted.

A. The Bacteriorhodopsin Active Site The chromophore in light-adapted bacteriorhodopsin (bR) has been shown to be an all-trans-retinal that is covalently linked via a protonated Schiff base to the e-amino group ofLys-216 on helix G (see Figures 1, 3, and 5) (Lewis et al., 1974). Upon absorption of light, the chromophore undergoes an isomerization from the all-trans to a 13-cis configuration (discussed in more detail below). One of the most interesting features of the protein binding site is its influence on the optical properties of the chromophore, which has come to be known as the "opsin shift" (Nakanishi et al., 1980). This term refers to the red shift in the visible absorption maximum of the chromophore that is observed upon its incorporation into the protein (opsin), as compared to that of protonated retinal Schiff base in solution. There are two main differences between the chromophore in solution and in the protein: (1) the Schiff base linkage to Lys-216, and (2) the environment of the binding site provided by the protein. The linkage to the protein alone cannot account for the magnitude of the observed red shift (Ottolenghi and Sheves, 1989). There has been much work done in attempting to decipher the structural and electronic mechanisms by which the apo-protein is able to regulate the absorption maximum of the chromophore. The picture that is emerging appears to be a complex combination of contributions from structural (protein and chromophore), electrostatic, and weak- and nonbonding interactions.

Configuration of Retinal within the Binding Site The two spectroscopic techniques that have provided the most detailed information about the configuration of retinal within the bacteriorhodopsin binding site are NMR and resonance Raman. NMR techniques indicate that the chromophore exists in a 6-s-trans conformation (Harbison et al., 1985; van der Steen et al., 1986; Smith et al., 1989). This is unusual when compared to the energetically more favorable 6-s-cis conformation in solution (by about 4 kcal/mol) (Honig et al., 1971, 1975; Birge et al., 1982), but allows for more extensive n-electron delocalization along the polyene chain (Albeck et al., 1992; Harbison et al., 1985; Lugtenburg et al., 1986; Smith et al., 1987; Spudich et al., 1986; Honig et al., 1976). In addition, a high resolution 13CNMR study on retinal derivatives done by Harbison et al. (1985) has shown that the C 8 chemical shift is sensitive to the C6-C7 conformation: a shift from 138.9 ppm to 130.9 ppm was seen upon going from 6-s-cis to the 6-s-trans isomer of retinoic acid. A shift of 131.6 ppm was seen for bacteriorhodopsin (Harbison et al., 1985), thus indicating the 6-s-trans conformation. Ab initio NMR-shielding calculations done by Wada et al. (1994) also indicated a planar 6-s-trans chromophore, as did deuterium solid state NMR studies that examined

Characterization of the Primary Photochemical Events

43

the dynamics of the chromophore methyl groups (Copi6 et al., 1994). This latter study, however, supports a 6-s-trans configuration slightly perturbed by the protein matrix, based upon comparison of the methyl rotors kinetics of model compounds (Copi6 et al., 1994). Recently, rotational resonance NMR (a new solid-state technique, see (Cruezet et al., 1991; Thompson et al., 1992)) confirmed NMR chemical shift (Harbison et al., 1984b) and resonance Raman vibrational data (Smith et al., 1984, 1987; Fodor et al., 1988a, b; Ames et al., 1989) that the C15--N bond of the protonated retinal Schiffbase linkage to Lys-216 has an anti (all-trans) conformation in bR (Thompson et al., 1992). FTIR studies on isotopically labeled chromophores by Livnah and Sheves (1993) reaffirmed the C15=N configuration as all-trans. Their results showed that the coupling between the C14-C15 stretching frequency and the N-H rock is very weak in all-trans isomers (for both syn and anti configurations). It is much stronger, however, in the 13-cis isomer (for both the syn and anti C15=N configurations), thereby allowing the C14-C15 stretching frequency to act as a marker for the all-trans and 13-cis chromophores. When applied to bR, no significant CI4-C15/N-H coupling was observed, indicating the all-trans configuration that is currently accepted. Studies on dark-adapted bR exhibited significant coupling, indicative of the presence of 13-cis. Smith et al., examined the hydrogen out-of-plane vibrational modes as a probe of chromophore structure (Smith et al., 1987); HOOP modes are expected to lack Raman intensity if the chromophore is planar. This is indeed found to be the case in bR (there is little torsional distortion from planarity), although these modes do gain significant intensity with some of the photocycle intermediates (Braiman and Mathies, 1982).

B. Chromophore Orientation and Location Aside from the general structural considerations of the light adapted chromophore (Figure 3: a planar all-trans retinal with an anti conformation about the C15-=N bond and a 6-s-t.rans ring-chain conformation) there is the question of orientation in the membrane. Several studies have addressed this issue, with respect to tilt angle (Urabe et al., 1989; Hauss et al., 1990; Ulrich et al., 1992), orientation of the chromophore molecular plane (Lin and Mathies, 1989; Urabe et al., 1989), N-H bond direction (Lin and Mathies, 1989; Hauss et al., 1990), and [3-ionone ring placement (Boehm et al., 1990; Ulrich et al., 1992). The results indicate a fairly consistent model of an essentially planar chromophore with a tilt angle of about 200-25 ° with respect to the membrane plane (Bogomolni et al., 1977; Heyn et al., 1977; Korenstein and Hess, 1978; Clark et al., 1980; Lin and Mathies, 1989; Urabe et al., 1989; Boehm et al., 1990; Hauss et al., 1990; Henderson et al., 1990a; Ulrich et al., 1992), a molecular plane nearly perpendicular to that of the membrane (Lin and Mathies, 1989; Urabe et al., 1989), and the N-H bond direction pointing towards the exterior of the cell (Lin and Mathies, 1989; Hauss et al., 1990). Work by Fahmy et al. (1989) also reported an average angle of 90+_20° for the retinal plane with that of the purple membrane. Furthermore, they reported the chromo-

44

JEFFREYA. STUART and ROBERT R. BIRGE

phore to exhibit small twists of 15-30 ° around the C-C single bonds of the retinal polyene chain, but displaying no large overall helicity (< 15°) (Fahmy et al., 1989). Recent deuterium NMR studies by Ulrich et al. (1992) however, indicate that the chromophore may actually be bent in the molecular plane. This slight curvature was posited in order to account for the discrepancy between the commonly accepted average retinal tilt angle of 67 ° (with respect to the membrane normal from spectroscopic studies (Heyn et al., 1977; Korenstein and Hess, 1978; Earnest et al., 1986; Lin and Mathies, 1989; Schertler et al., 1991)) or 65°+_12° as found via neutron diffraction studies (Hauss et al., 1990), and their reference angle of 47 °. This angle was determined by assuming that the bond to the deuterated methyl group on C 5 should make a perfect right angle with the extended conjugated system of the polyene in a planar 6-s-trans conformation. If this is the case, the angle made by the long axis of the chromophore to the membrane normal (as a function of the location of the three deuterated methyl groups) is 47 ° (Ulrich et al., 1992). The only way to account for the difference between the two angles is that retinal must have a slight in-plane bending, and that the angles determined spectroscopically and otherwise must represent an average chain orientation (Ulrich et al., 1992). Indeed, the authors go on to cite several other cases where curved retinals or retinal analogues have been postulated, and that the distortion may aid in relieving steric crowding of the methyl groups on the polyene chain. In all of the cases examined, the curvature is at least 10° between the average tilt axis and the local reference angle as defined above, implying that the chromophore tilt may actually be steeper than what is commonly accepted at present. It is also important to note that the spectroscopic studies alluded to above all rely on the assumption that the electronic transition moment is parallel to the long chromophore axis. This is not necessarily the case, and may further skew the average angle away from the actual (see Schertler et al., 1991; Ulrich et al., 1992 and the references cited therein). Although the landmark bacteriorhodopsin structural paper by Henderson et al., (1990a) was able to locate the cyclohexene ring within the binding pocket (Figure 5), no information on its actual orientation and conformation was evident. The cyclohexene ring is now generally accepted to lie in the same plane as the polyene (vertical to the membrane plane), spatially situated in the vicinity of helices D and E, as shown by photoaffinity labeling studies (Huang et al., 1982; Boehm et al., 1990). It is tilted towards the exterior membrane surface, as suggested by second harmonic studies (Huang and Lewis, 1989). The aforementioned deuterium NMR studies by Ulrich and coworkers confirm the vertical ring orientation, as well as the 6-s-trans conformation, due to the spatial constraints imposed upon the polyene by the locations of deuterated methyl groups on carbons 1 and 5 (Ulrich et al., 1992). The methyl groups on carbons 9 and 13 are reported to point towards the intracellular surface (with the Schiff base proton oriented toward the exterior cell membrane) (Lin and Mathies, 1989; Hauss et al., 1990; Henderson et al., 1990a; Ulrich et al., 1992), although tentative findings by Boehm et al. indicate the opposite inclination (Boehm et al., 1990; Park et al., 1989), not only for the C-5, -9, and -13

Characterization of the Primary Photochemical Events

45

methyl groups, but for the N-H bond as well (pointing toward the interior membrane surface). In a study that attempted to examine the shape of the ionone ring binding pocket, retinal analogues with a fixed 6-s-cis conformation were incorporated into bacteriorhodopsin. It was demonstrated that the 6-s-trans ring configuration was not necessary for the function of light-induced proton pumping (Iwasa et al., 1992). Although this ability was unaffected, the synthetic pigments displayed different absorption maxima, indicating altered steric interactions. Based on these findings, the authors suggest the existence of amino acids that interact with the chromophore near C 4 and C 9 (Iwasa et al., 1992). Several attempts have been made to determine the depth of the chromophore within the membrane. Electron density maps generated by Henderson et al. (1990), reveal that the Schiffbase resides in the center of the protein, with the cyclohexene ring positioned about 5 A below, toward the extracellular surface. Most studies place the center of the chromophore at about 10 A from the membrane surface (with the Schiffbase and the cyclohexene ring at about 15 and 9 A respectively) (Otomo et al., 1988; Leder et al., 1989; Hauss et al., 1990). Discrepancies exist, however, in the determination of whether it lies closest to the extracellular or cytoplasmic membrane surface. Fluorescence energy transfer studies indicated the location to be closer to the cytoplasmic surface (Otomo et al., 1988). These results are contradicted by other fluorescence energy transfer studies done by Leder et al. (1989), and the second harmonic generation studies of Huang and Lewis (1989). Both of the latter authors came to the conclusion that the retinal moiety lies closest to the external membrane surface. Park et al. (1989) place retinal C 9 15 A distant from the extracellular side. These results provide further conformation of the conclusions obtained from photoaffinity labeling studies (Huang et al., 1982), which place the retinal ring in the vicinity of amino acid residues Ser-193 and Glu- 194 (helix F). Both of these residues are predicted to be closer to the extracellular surface. More recent photoaffinity labeling studies by Boehm et al. (1990) (using ring-labeled retinals), identified Trp-137 and Trp-138 (helix E), and Ala- 126 and Leu-127 (helix D) as photoaffinity labeling sites (also closer to the extracellular side). In Henderson's structure, the ionone ring is located about 10 A above Trp-189, roughly level with Pro-186. Henderson notes, however, that if retinal is allowed to pivot downwards about the Schiff base, it is able to reach Ser-193 and Glu-194, which are located in such a way to ideally allow for reaction with a ring-labeled retinal that may be too large to take its normal configuration (Henderson et al., 1990a). It is also reported that retinal is in contact with Trp-138, but not Trp- 137 (Henderson et al., 1990a).

C. Characterization of the Opsin Shift The binding site interactions with retinal are what largely determine the intriguing properties of the chromophore and its associated photochemistry. Indeed, it is the protein-retinal interactions that produce the "opsin shift" referred to earlier, and

46

JEFFREYA. STUART and ROBERT R. BIRGE

determine the nature of the primary event. Despite the fact that this topic is covered more completely in a separate chapter of this book, some mention of these interactions is required herein for the sake of subsequent discussion. Of the experimental techniques used to examine the retinal binding site, several have made extensive contributions including electron microscopy, neutron diffraction, UVVIS electronic spectroscopy, two-photon electronic spectroscopy, NMR, resonance Raman and FTIR, and finally, analog chromophore and site-directed mutagenesis studies (see Birge, 1990a for a review). The current view indicates that the binding site is a very ionic (possibly charged) complex environment. Although still an issue of heated debate, the following properties of the retinal binding site in bR (that result in the opsin shift) are for the most part commonly accepted. The suggestion that the red shift (opsin shift) observed in visual pigments might be due to electrostatic interactions between the chromophore and the protein was made by Kropf and Hubbard in 1958 (Kropf and Hubbard, 1958). The mechanism by which a shift is produced involves a stabilization of delocalized electronic structures due to the location of protein charges near the conjugated plane of the retinylidene chromophore. Theoretical studies later confirmed this conjecture (for a review see Ottolenghi and Sheves 1989). It is now known that there are at least three factors that contribute to the observed opsin shift of about 5100 cm-1 (Ottolenghi and Sheves, 1989) in bacteriorhodopsin (Lugtenburg et al., 1986; Baasov et al., 1987a; Smith et al., 1987; Albeck et al., 1992; Ebrey, 1993). Protonated retinal Schiff base absorption maxima can be shifted by (1) weak hydrogen bonding to the Schiff base proton or a weak interaction with a positive counterion, (2) a 6-s-trans conformation about the C6--C7 single bond (the ringchain conformation), and (3) a dipole or negative counterion in the vicinity of the ionone ring (or a highly polarizable retinal binding pocket). All of these perturbations have in common the effect of increasing n-electron delocalization along the retinal polyene chain (Albeck et al., 1992; Ebrey, 1993). For some time now it has been well established that the retinal Schiff base in bacteriorhodopsin is protonated (Waddell et al., 1973; Lewis et al., 1974; Aton et al., 1977; Das et al., 1979; Eyring and Mathies, 1979; Bagley et al., 1982; Rothschild et al., 1982; Harbison et al., 1983; Hildebrandt and Stockburger, 1984; Birge et al., 1985, 1987; Smith et al., 1985b; Birge and Zhang, 1990a). Electrostatic interactions between the protonated chromophore and the surrounding protein are responsible for the shift of the visible absorption maximum, changes in the pK a values of the Schiff base and active site amino acids, and certainly play a role in the actual biological function of the protein. Evidence for the interaction of the protonated retinal Schiff base with a counterion, and/or a weak hydrogen bonding situation, comes from several sources, including NMR, FTIR, and resonance Raman. The most direct evidence comes from the latter: the Schiffbase C--N bond stretching frequency has proven to be a sensitive probe of electrostatic perturbations in its immediate environment (Baasov et al., 1987a). Based upon comparative studies with model compounds, it was determined that this frequency was lowered

Characterization of the Primary Photochemical Events

47

due to diminished ion pairings and decreased solvation of the positive nitrogen. Interactions in the vicinity of the ionone ring or near C 9 affect the C--C stretching frequency, but not the C--N stretching mode, and therefore cause little change in the x-electron delocalization of the C = N bond (Baasov et al., 1987a,b; Ottolenghi and Sheves, 1989). Electrostatic interactions between the protonated nitrogen and nonconjugated positive charges shift the frequency of the C--N mode, which is particularly sensitive to hydrogen bonding with the N-H bond. This indicates that there must be a mixing of the C--N stretching and C = N - H bending modes. Additionally, the C = N stretching frequency of 1640 cm-1 in bR and its 17 cm-1 deuterium shift can attributed to weak hydrogen bonding between the Schiff base and its coumerion and/or poor solvation of the positive nitrogen (Baasov et al., 1987a; Ebrey, 1993). Such interactions at the Schiff base result in enhanced x-electron delocalization along the polyene chain by reducing the stability of the positive charge at the Schiff base (Smith et al., 1987). This is thought to be the primary factor contributing to the opsin shift in bR (Lugtenburg et al., 1986; Spudich et al., 1986). From 15N NMR chemical shift anisotropies of model compounds as compared to bR, de Groot et al. (1989) found evidence to support a very weak hydrogen bond. The latter displayed an isotropic shift 11 ppm further down than any of the model compounds examined (de Groot et al., 1989). This confirmed earlier 15N NMR studies of the Schiff base (Harbison et al., 1983). Furthermore, the two-photon studies of Birge and Zhang (1990a) do not support the existence of strong hydrogen bonding between a negative counterion and the imine proton. They suggest in one of their models that the interaction may be mediated by one or more water molecules (see discussion below, Birge and Zhang, 1990a). Distortions of the conjugated retinal plane will also result in a shift in the absorption spectrum. In general, for the case of conjugated polyenes, a twist about a double bond will induce a red shift, whereas a twist of a single bond produces a blue shift in the absence of any external electrostatic perturbations, in which case a red shift can be induced via single bond twists (Kakitani et al., 1985). In the specific case of retinal, a planar conformation of the C6-C 7 bond increases the opsin shift, whereas nonplanar configurations lead to a consequent decrease. The question of bacteriorhodopsin chromophore conformation (the second contribution to the opsin shift) was addressed earlier: the protonated retinal Schiff base exists with a 6-s-trans ring-chain configuration. Retinoids in solution and crystal structures exist predominantly in the energetically more favorable 6-s-cis conformation (Honig et al., 1971; Simmons et al., 1981). Such a cis to trans isomerization is predicted by theoretical means to red shift the absorption maximum of retinal by about 25 nm (Honig et al., 1976; Harbison et al., 1985; Lugtenburg et al., 1986; Spudich et al., 1986), and would allow for a more fully delocalized x-electron system (due to the addition of the C5=C 6 bond) and a consequently overall higher C-C bond order (reflected in the C--C stretching frequencies) (Baasov et al., 1987a, b; Smith et al., 1987). The mechanism by which this conformation is able to shift the absorption

48

JEFFREY A. STUART and ROBERT R. BIRGE

maximum is largely through excited state stabilization: the addition of another double bond into the conjugated system has the effect of further enhancing the stabilization (Albeck et al., 1992). The third contribution to the opsin shift is electrostatic in nature, and postulates the existence of some type of ionic perturbation, presumably an ion pair, dipole, or isolated negative charge, in the vicinity of the ionone ring and specifically at C 5. The existence of this perturbation is somewhat controversial, but could contribute between 500 to 1,000 cm-1 to the observed shift (Lugtenburg et al., 1986; Spudich et al., 1986; Ottolenghi and Sheves, 1989; Mathies et al., 1991). The presence of a charge in the location of the C5=C 6 bond as part of the mechanism responsible for giving bacteriorhodopsin its purple color was first examined by Nakanishi et al. (1980) using dihydroretinal derivatives. Later work on the same analog chromophores by Lugtenburg and coworkers went a step farther by hypothesizing the need for an additional positive charge near C 7 (Lugtenburg et al., 1986). NMR experiments had already predicted this, and also lent support to the C5--C 6 negative charge (from C 5 and C 7 13C chemical shift data; Harbison et al., 1985). It is important to note that the negative charge will have the effect of preferentially stabilizing the excited state over the ground state (thus producing the red shift), whereas a positive point charge will have the opposite effect (resulting in a blue shift). It is therefore necessary to conclude that the electrostatic perturbation near the ring is dipolar in nature, and there is much evidence to support this conclusion (Harbison et al., 1985; Lugtenburg et al., 1986). Several plausible candidates for this interaction include the cooperative effects of Trp-138 and Pro-186 for the respective negative and positive dipolar contributions, or the protonated carboxylic group of Asp-115 (Braiman et al., 1988; Mathies et al., 1991). Recent 13C experiments done by Albeck et al. (1992) however, have shown on the basis of their chemical shift measurements that it is not necessary to include a nonconjugated charge or dipole in the ionone ring region in order to successfully model the binding site. A model that accounts for the full opsin shift can be constructed solely by the use of a weak Schiff base counterion interaction enforced by strong hydrogen bonding to the counterion. An increase in the distance between the Schiffbase and its counterion can largely account for the red shift, but additional counterion stabilization forces are needed to account for the change in the pK a of the Schiff base nitrogen in bR (Albeck et al., 1992). Such a diminished ion pairing can contribute as much as 2500 crn-1 to the observed opsin shift (Ottolenghi and Sheves, 1989). Additional counterion stabilization forces may be provided by highly structured water molecules, which have been shown to have significant effects upon calculated pK a values (Sampogna and Honig, 1994). In addition, a model that incorporates these features can also account for the high pK a value (13.3 + 0.3) of the protonated Schiff base (Druckmann et al., 1982; Baasov et al., 1987a; Ottolenghi and Sheves, 1989). It is important to note, however, that this data does not rigorously exclude the possibility of a dipolar group near the retinal ring. Any charge in the region would have to be situated in such a way as to comply with the

Characterization of the Primary Photochemical Events

49

Table 1. Contributions to the Opsin Shifts of Bacteriorhodopsin and Rhodopsin* Opsin Shift Contribution -+ Bacteriorhodopsin (total opsin shift ~ 5,100 cm- ) Rhodopsin (total opsin shift = 2,650 cm -1)

Hydrogen-Bonding or Ionic Interaction with the PSB 3000 c m -1 (weak Hydrogen bond to PSB) =1200 cm -1 from interaction of PRB with 1° counterion

Ring-Chain Configuration 1500 cm -1

(6-s-trans) =1450 cm -1 (twisted 6-s-cis)

~3-Ionone Ring and~or Polyene Chain Interactions 500-1,000 cm -1 (ionic or dipolar at [3-Ionone ring) Undetermined contribution

(see text)

Note." *Datacompiled from references cited in text.

C 5 chemical shift data. Recent two-photon studies also support the existence of a dipolar group near the ring, and are not consistent with a charged species in that region (Birge and Zhang, 1990a). Trp-138 and Pro-186 have been suggested to be responsible for the respective negative and positive perturbations (Mathies et al., 1991). The protonated carboxylic group of Asp-115 might also be a possibility: a slight rotation of helix D would place this group in the vicinity ofC 5 and C 7 of the chromophore (Braiman et al., 1988; Mathies et al., 1991). The energetic contributions to the opsin shift due to each of the factors described above have been nicely reviewed and summarized by Ottolenghi and Sheves (see Ottolenghi and Sheves, 1989). Weak hydrogen bonding between the protonated Schiff base and its counterion (which is probably involved in a strong hydrogen bond with water or some other moiety) accounts for about 3,000 cm-1 (of the total 5,100 crn-1). The planar 6-s-trans ring-chain conformation of the chromophore contributes an additional 1,500 cm-~, and a small contribution of about 500 cm -~ is assigned to the effect of a protein dipole in the vicinity of the ionone ring in a symmetric arrangement with the negative dipolar end near the C5=C 6 bond and the positive end at the C7-C 9 bond. A recent report by Hu et al. (1994a) postulates the existence of a synergistic effect on the electronic structure of the chromophore that stems from the 6-s-trans conformation and the counterion to the protonated Schiff base; based on their results, these are the only two contributions needed to account for the full opsin shift (Hu et al., 1994a). The various contributions to the opsin shift are summarized in Table 1.

Active Site Interactions The primary counterion. The identity of the primary counterion to the imine proton of the Schiff base remains an issue of debate. Among the candidates are Asp-85 and Asp-212, with the remote possibility ofTyr- 185. Both Asp residues are known to be deprotonated in bR (Metz et al., 1992), but the state ofTyr-185 is less sure. Instead of a lone amino acid residue acting as primary counterion, it is more

50

JEFFREYA. STUART and ROBERT R. BIRGE

likely that the binding site has a complex ionic structure with hydrogen bonding and electrostatic interactions between one or more counterions, water, polar amino acid side groups, and the imine proton. There is even the possibility of the presence of one or more divalent cations (Jonas and Ebrey, 1991; Sweetman and E1-Sayed, 1991; Zhang et al., 1993). Recently there has been a great deal of evidence in support of the assignment of Asp-85 as the primary counterion. Site-directed mutagenesis studies done on Asp-85 and -212 revealed that replacement of Asp-85 by Asn caused much larger red shifts in the absorption maximum than equivalent substitutions at residue 212 (Marti et al., 1991). In the same study it was shown additionally that at least one carboxylate group is required at one of these positions for retinal binding, indicating that a counterion is necessary for formation and stabilization of the protonated Schiffbase. Because of the overall larger effects seen upon replacements of Asp-85, it was determined that this residue, and not Asp-212, is the primary retinylidene Schiff base counterion (Marti et al., 1991). Further studies by Subramaniam et al. (1992) (involving replacement of Asp-85 by the uncharged proton acceptor histidine) determined that it is not enough that the residue at position 85 be a proton acceptor for functioning of the proton pump, but it must be negatively charged as well (Subramaniam et al., 1992). This result was confirmed by studies done by the same group, in which the Asp-85 side chain was replaced by various analogs. It was shown that the most important attribute of position 85 was the presence of a carboxylate group: an increase of the side chain by two bond lengths had no effect, as long as the terminal carbon was a carboxylate (Greenhalgh et al., 1992). In fact, neutral substitutions in general at position 85 resulted in complete inactivation of proton pumping activity by bR (Mogi et al., 1988; Stern and Khorana, 1989; Greenhalgh et al., 1992; Marti et al., 1991). Substitutions for Asp-85 also lead to a larger decrease in the pK a of the Schiffbase than for Asp-212 (Marti et al., 1991; Subramaniam et al., 1992). From a purely phenomenological point of view, the role of Asp-85 as primary counterion makes good intuitive sense, as it has been shown to act as the proton acceptor from the protonated Schiffbase after initiation of the bR photocycle (Henderson et al., 1990a; Ceska et al., 1992). It should be noted that prior to the studies described above, most evidence seemed to indicate Asp-212 as the primary counterion. In fact, Khorana and coworkers, who now agree on Asp-85, previously assigned Asp-212 to that role (Hackett et al., 1987; Mogi et al., 1988), and there are several models proposed by Birge that assume Asp-212 to be the primary counterion as well (Birge et al., 1987, 1989a, 1989b; Birge and Pierce, 1983; Zhang, 1989). These models were proposed, however, based partially on the above cited work by Khorana, and on the fact that Asp-212 is located on the same helix (G) as Lys-216, to which the chromophore is bound. This places Asp-212 in correct proximity to the protonated Schiffbase (four amino acids distant, or approximately one helix turn away). Additionally, placement of an alanine at position 212 resulted in a light unstable pigment that regenerated the active opsin-chromophore complex over fifty times more slowly than wild-type

Characterization of the Primary Photochemical Events

51

bR (Hackett et al., 1987; Mogi et al., 1988). It is now clear from the electron diffraction data of Henderson that both Asp residues are roughly equidistant (~4 A) from the chromophore, below the protonated Schiff base, toward the protein exterior (between Lys-216 and the chromophore) (Henderson et al., 1990a). In the absence of any unaccounted-for stabilizing influences, the fact that these two residues can exist in such close proximity while in an ionized state is surprising. Finally, a recent study on Ca 2+ binding to deionized bacteriorhodopsin posits that Tyr-185 and Asp-212 may be associated with two independent high-affinity Ca 2+ binding sites in the active site (Zhang et al., 1992, 1993). The authors go on to say that Asp-212 seems to be more strongly coupled with these cations than Asp-85; this would leave the latter as the more likely candidate for the primary counterion to the protonated Schiffbase. Sampogna and Honig (1994) recently performed calculations of the pK a values for charged residues in the bacteriorhodopsin active site, giving some intriguing results. The ionization states of Asp-85 and the protonated Schiffbase were found to be highly coupled to each other, whereas Asp-212 seems to be stabilized by a strong hydrogen-bonding network (Sampogna and Honig, 1994). This helps to explain the experimental determinations of Asp-85 as the primary counterion rather than Asp-212; the former has no accessible hydrogen bonding groups to insulate it, and therefore no compensatory mechanisms upon its replacement, thus resulting in the large changes that are seen in the Asp-85 mutants (with respect to wild-type) (Sampogna and Honig, 1994). The interactions of Asp-212 with the protein, however, are more immune to genetic manipulation due to the hydrogen-bonding network. The net result of the strong coupling that exists between the protonated Schiffbase and Asp-85 is an extreme sensitivity to even the slightest changes in the active-site environment, causing large shifts in their relative pK a values (Sampogna and Honig, 1994). The trans-to-cis isomerization is also expected to produce large changes in the pK a values of charged species in the binding site. This study also hoped to give insight into the location of the Arg-82 side chain, which can be directed either toward the exterior of the protein, or towards the interior of the protein and still remain consistent with Henderson's diffraction data. Models were examined with the side chain in both positions, and the resulting pK a values were much the same when the inherent uncertainties in the calculations are considered (Sampogna and Honig, 1994). When the side chain was directed in such a way so as to stabilize Asp-85, the pK a calculated for Asp-85 agreed with experimental determinations. However, this result was nearly duplicated in a model in which the Arg-82 side chain was given the opposite position, and water molecules were allowed to complex with Asp-85. Generally speaking, highly structured water molecules were found to be very effective in stabilizing ionized residues based upon these calculations, so much so, in fact, that electroneutrality of the binding site may not be mandatory (Sampogna and Honig, 1994). The question of calcium in the binding site was not addressed, but the inclusion of specific Ca 2+ ions would surely affect these calculations, especially in light of the purported interaction of Asp-212

52

JEFFREYA. STUART and ROBERT R. BIRGE

with a strongly bound calcium ion (as compared to the suggestion of a complex hydrogen bonding network). It seems clear that the Asp-212 plays a complex role in the binding site. It would also be reasonable to assign Asp-85 and Asp-212 as cocounterions with roughly equivalent electrostatic participation in the stabilization of the protonated Schiffbase.

The role of tyrosine-185. Tyrosine-185 has also been suggested by some authors as a possibility for a primary coumerion (Roepe et al., 1987, 1988a,b; Rothschild et al., 1990). This, however, is a matter of some controversy. More specifically, there is a great deal of disagreement on the protonation state oftyrosine residues throughout the photocycle. There is much evidence for the existence (or formation) of tyrosinate residues in bR and some of the intermediate states, especially K and M. Evidence for the existence of tyrosinate in these states comes from FTIR (Rothschild et al., 1990), site directed mutagenesis (Ahl et al., 1987), and UV-VIS spectroscopic studies (Dufiach et al., 1990). The ultraviolet resonance Raman work by Harada et al. (1990) supports this conclusion. Evidence against tyrosinate existence is provided by NMR (Herzfeld et al., 1990; McDermott et al., 1991) and Ultraviolet resonance Raman studies (Ames et al., 1990, 1992). One model currently being examined consists of the existence of a tyrosinate in light adapted bR (absent in dark adapted bR550) which protonates upon or before the formation of the K intermediate, and consequently deprotonates again during the formation of the M intermediate. Roepe and coworkers (Roepe et al., 1988) have speculated that the 13-trans, 15-trans configuration of the protonated Schiffbase chromophore leads to a stable tyrosinate group due to electrostatic interactions with the positive charge on the Schiff base in bR. Furthermore, isomerization about either bond disrupts this stabilization causing protonation of the tyrosinateresidue (Roepe et al., 1988). It should be noted that although it is thought that the tyrosine in question is Tyr-185, it is possible that more than one tyrosine is involved in the above protonations and deprotonations. The Ca 2+ binding studies previously cited argue against Tyr-185 as the primary coumerion due to its involvement with the high affinity Ca 2÷ binding sites (Zhang et al., 1993). There have been attempts to merge the conflicting experimental results by various authors, one of whom has suggested the existence of an extremely polarizable hydrogen bond between a tyrosine or tyrosinate with some acceptor group such as a carboxylate residue or the protonated Schiffbase (Rothschild et al., 1990). In such a bond the proton would be shared between the two groups on a time scale faster than what is measurable by certain spectroscopies, especially NMR. The bond would also be characterized by a double well potential minimum, making it very stable (Rothschild et al., 1990). The proposed hydrogen bond has also been suggested to be in a situation that helps to stabilize the relative orientation of the putative F and G helices, as well as to affect the isomerization of retinal in a regioselective manner about the C13--'C14 double bond (Rothschild et al., 1990). Although recent solid-state NMR studies by McDermott et al. (1991) have demon-

Characterization of the Primary Photochemical Events

53

strated against the existence of any tyrosinate, they do not rule out the possibility of this type of hydrogen bonding situation, providing that the equilibrium strongly favors tyrosine. Indeed, their results do support the concept oftyrosine participation in a delocalized deprotonation "defect" which would favor tyrosine rather than tyrosinate, especially in the M state. Data obtained from electron diffraction studies indicate that it is more likely for Tyr-185 to be involved in a hydrogen bond with Asp-212 rather than the Schiff base (Henderson et al., 1990b). It has also been suggested that the reason for the lack of proton pumping activity by dark adapted bR550 is the absence of an initial tyrosinate (Roepe et al., 1988). Although one would expect that the changes in the absorption spectrum (electronic, vibrational or nuclear) upon formation of tyrosinate from tyrosine would be enough to make a positive determination, the above conflicting research has demonstrated otherwise. Role of water in the active site. The role of water in the active-site/chromophore environment was first examined via resonance Raman spectroscopy (Hildebrandt and Stockburger, 1984). It is now generally accepted that water plays a vital role as an intermediary in protein-chromophore interactions of the binding site pocket, often being postulated as having the effect of stabilizing the interaction between the protonated Schiff base and its counterion. Using resonance Raman, Hildebrandt and Stockburger (1984) observed a considerable narrowing of the bR C--N stretching band when the solvent was changed from water to D20. They interpreted this as a resonance energy transfer from the C = N stretching vibration to the bending vibration of water molecules associated with the Schiff base nitrogen. These water molecules are tightly bound, and could be removed only by prolonged application of a strong vacuum. The resulting dehydrated protein was deprotonated at the Schiff base and had experienced a change in the retinal configuration. They concluded that water plays an important role in the composition of the binding site by solvating the counterion and Schiffbase, and consequently stabilizing the structure. It is noted, however, that there cannot be a relatively large amount of water because the Schiffbase bond would not be stable. Solid state NMR studies by Harbison et al. (1988) were able to monitor the kinetic transfer of magnetization from the solvent water to the Schiff base nitrogen, which was interpreted as occurring via chemical exchange of the Schiff base proton with solvent water. Because the protons responsible for the transfer of magnetization had to be from the solvent (due to the nature of the experiment), the exchange must be directly between the Schiff base nitrogen with bulk water, implying that the retinal binding site has access to the solvent. This concurs with the structural model proposed recently by Henderson, which indicates that the region between the chromophore Schiffbase and the extracellular membrane surface is hydrophilic and relatively open to water exchange (Henderson et al., 1990a). Perhaps the most direct evidence for the presence of water in the binding site comes from neutron diffraction studies, which determined the presence of four (+1) bound water molecules at the Schiff base end of the chromophore (Papadopoulos et al., 1990). These four

54

JEFFREYA. STUART and ROBERT R. BIRGE

water molecules are very strongly bound, and are present even at 0% relative humidity. Only by application of prolonged high vacuum could they be removed. The presence of water in the retinal binding site also seems to be a prerequisite for proton transfer steps during the photocycle (Maeda et al., 1992), as well as for internal hydration to stabilize transition states which involve separation of a proton from its negative coumerpart in an ion pair (Cao et al., 1991). Most recently, a resonance Raman study of Schiffbase hydrogen/deuterium exchange by Deng and coworkers suggested the existence of an oriented water molecule (or perhaps multiple water molecules) in the vicinity of the protonated Schiff base that would allow for the high rates of H-D exchange that they observed (higher than what was observed for model compounds in solution) (Deng et al., 1994). The proposed water molecule(s) would be held in position by a specific hydrogen bonding environment that would have the effect of increasing the effective water concentration at the Schiffbase (i.e., a direct hydrogen bond to the protonated Schiffbase) (Deng et al., 1994). There is also a report of direct interaction of a water molecule with the primary counterion Asp-85; FTIR studies demonstrated that a water molecule which interacted weakly with Asp-85 in the wild-type protein disappeared upon the removal of the negative charge by substitution of the primary counterion with asparagine (Maeda et al., 1994). This provides direct evidence that the primary counterion is hydrated.

Description of the Active Site The only realistic description of the binding site is probably one which incorporates all of the above described characteristics into an integrated complex counterion environment interacting with the protonated Schiff base. This would include contributions primarily from Asp-85 and to a lesser extent Asp-212, as well as Tyr-185, Arg-82 (positioned below Asp-85 (Henderson et al., 1990a)) and Arg-227 (Stem and Khorana, 1989). In Henderson's model of the active site, the Arg-82 side chain was placed extending toward the extracellular surface, although a conformation where it is directed toward the interior near Asp-85 and-212 is also possible. A recent examination ofpK a constants in the Henderson binding site was unable to make a concrete prediction (Bashford and Gerwert, 1992; Sampogna and Honig, 1994). Experimental evidence for the interaction of Arg-82 and Asp-85 indicates that Arg-82 is responsible for stabilization of the ionized form of Asp-85 (Balashov et al., 1994), implying a conformation directed toward the interior of the protein. Tyr-57 is also reported to be involved in a hydrogen bond network with Try-185 and Asp-212 (Henderson et al., 1990a). Asp-ll5 is situated near the ionone ring, and may be involved as a nonconjugated dipolar interaction. The immediate region around the Schiffbase would be stabilized by interactions with tightly bound water molecules (which certainly play some role in proton transport, and may help to mediate other electrostatic interactions), and perhaps several divalent cations (reportedly associated with Asp-212 and Tyr-185 (Zhang et al., 1993)). Such a complex counterion would help to explain why it has been difficult to definitively

Characterization of the Primary Photochemical Events

55

assign a single primary counterion by site-directed mutagenesis studies. Indeed, as pointed out by de Groot et al. (1989) if the active site environment were complex enough, removal of one or more of the individual counterions may have only a slight effect on the overall counterion properties. This type of coumefion environment has been proposed by several authors (de Groot et al., 1989; Ebrey, 1993). A plausible model for the bacteriorhodopsin binding site is shown in Figure 5. Evidence from site-directed mutagenesis and Henderson's electron diffraction studies indicate that the retinal binding pocket is bounded by tryptophan (Trp-86, -182,-89, and-137) and tyrosine residues (Tyr-57 and-185) (Mogi et al., 1989a; Henderson et al., 1990b). Several authors have proposed models of the binding site in which Trp-182 and -189 sandwich the retinal polyene chain (Lin and Mathies, 1989; Oesterhelt and Tittor, 1989; Rothschild et al., 1989a); Trp-86 is also thought to be involved (Rothschild et al., 1989b). These models are in general supported by the low resolution electron diffraction structure of the bR active site (Henderson et al., 1990b), and may also contribute to the electronic properties of the chromophore binding site. A recent site-directed mutagenesis study by Greenhalgh et al. (1993) examined the roles ofhydrophobic amino acids in the binding site, including Met-20, Val-49, Ala-53, Met- 118, Gly- 122, and Met- 154. These specific residues were chosen based upon the assertion by Henderson's structure that they are part of the retinal binding pocket (Henderson et al., 1990a). It was found that for all of those examined, with the exception of Met-20, substitution resulted in altered rates of chromophore regeneration (that varied from 20 times faster to 45 times slower than the wild-type protein), blue shifts in the visible spectra of up to 80 nm, and changes in the retinal isomer composition of the mutant chromophores (Greenhalgh et al., 1993). Provided that there are no cooperative effects, these results indicate that (with the exception of Met-20) all of the above residues are involved in the structure of the binding pocket and interact with the chromophore.

D. The Rhodopsin Active Site The chromophore in all visual rhodopsins has been shown to be 11-c/s-retinal (Figure 6). This chromophore has been conserved in all visual pigments, from the color resolving retinal cone pigments in humans, to the rhodopsin-like pigments in chicken (Fukuda et al., 1990), insect (Pande et al., 1987), and octopus (Ohtani et al., 1988). One of the incredible attributes of the rhodopsin family of proteins is that despite the fact that each rhodopsin studied to date has the same highly conserved chromophore and structure combination, they exhibit wide range of spectral properties. Human and bovine rhodopsins absorb maximally at about 500 nm (Birge, 1990a), chicken at --562 nm (Fukuda et al., 1990), insect at 345 nm (Pande et al., 1987), and octopus at 472 nm (Deng et al., 1991a). The human cone pigments responsible for color vision have maxima that range from 440 nm (blue) to 530 nm (green) to 560 nm (red) (Sakmar et al., 1991). For purposes of comparison, the protonated retinal Schiff base in methanol absorbs at about 440

56

JEFFREYA. STUART and ROBERT R. BIRGE

nm. Given that these proteins do exhibit some homology in their primary structure, the highly conserved secondary structure consisting of a seven a-helical bundle as a shared structural motif, as well as an identical chromophore moiety, how is this wide range of absorption maxima possible? In other words, by what mechanism are the respective opsins able to regulate, or "spectrally tune" (Nathans, 1990a), the properties oftheir common chromophore? This has proven to be the central problem in the study of visual pigments (especially for the cone pigments), and it is still not well understood. One of the main hindrances in the study of the opsin shift of the visual pigments has been the inability to make suitable quantities ofrhodopsin and rhodopsin mutants. The large size of the protein also inhibits progress toward a structural determination. Some analogies have been drawn from bacteriorhodopsin, for which these interactions are much better understood, but these have provided only limited insight. Only recently, with the advent of new and more sensitive techniques that allow examination of much smaller quantities, have detailed results been available. Unfortunately, there has been no detailed structural information based upon a crystal structure reported as of yet. However, a 9 A resolution projection structure has recently been presented by Henderson's group which was able to resolve the seven trans-membrane a-helices (Schertler et al., 1993). These helices are oriented approximately perpendicular to the membrane plane, and are tilted differently than what is found in bacteriorhodopsin. As in the previous section, chromophore configuration and chromophore-protein active site interactions will be examined, with emphasis on how they contribute to the opsin shift. Unless otherwise stated, the discussion will center on bovine rhodopsin, which is the most commonly studied.

Configuration of Retinal within the Binding Site As for bacteriorhodopsin, NMR and resonance Raman spectroscopies have provided the most detailed information about the configuration of rhodopsin's l l-cis-retinal chromophore. Low-temperature solid-state 13C NMR studies by Smith et al. have shown that the cylclohexene ring of the retinal Schiffbase has a twisted 6-s-cis conformation, as opposed to the planar 6-s-trans configuration found in bacteriorhodopsin (Smith et al., 1987a, 1990). Similar studies by Mollevanger et al. (1987) (done on lyophilized protein) confirm this. These results were based upon comparisons with C 5 chemical shift data obtained from model retinyl Schiff base compounds. Neither of the studies, however, were able to discern any information with regard to the C6-C 7 dihedral angle. The latter authors only referred to the C6--C7 bond as being nonplanar and unperturbed (by interactions with the protein) (Mollevanger et al., 1987). Knowledge of this angle is helpful in determining its contribution to the opsin shift. As discussed previously for bacteriorhodopsin, the planar 6-s-trans ring-chain conformation contributes about 1,500 cm-1 to this shift. In model retinal Schiffbase compounds (in solution) the C6-C 7 bond is in the 6-s-cis configuration and additionally twisted by 40-70 ° out of the plane defined by the retinal chain (Honig et al., 1971). This bond configuration can result

Characterization of the Primary Photochemical Events

57

in a 25-35 nm red shift in the absorption spectrum (Honig et al., 1976; van der Steen et al., 1986), which constitutes a substantial contribution to the overall opsin shift for rhodopsin (approximately 1,450 cm-1 of the total 2,650 cm-1, assuming a 30 nm protein induced distortional red shift [Smith et al., 1987a]). Smith and coworkers, however, did find evidence for a steric interaction at C 8 which points to a conformation for the retinal Schiff base similar to that of model compounds (Smith et al., 1987a). A recent attempt to estimate the ring-chain conformation of the rhodopsin chromophore using an ab initio method based on 13C NMR chemical shifts was inconclusive as to whether a skewed 6-s-cis or 6-s-trans configuration exists (Wade et al., 1994). Resonance Raman spectroscopy has provided a great deal of information regarding the state of the Schiff base and single bond configurations along the retinal polyene chain. Studies by Palings et al. in 1987 allowed for the assignment of the C10-Cll and the C14--C15 bonds as s-trans in rhodopsin, bathorhodopsin, and isorhodopsin. In addition they were able to determine that the Schiff base is protonated, and has a trans (anti) configuration (Palings et al., 1987). The low temperature FTIR studies done by Bagley and co-workers gave similar results, namely that the Schiff base is protonated (in rhodopsin, isorhodopsin, and bathorhodopsin as well), the C--N bond is trans (in all three species), and a 10-s-trans bond configuration exists for bathorhodopsin (which also supports an all-trans conformation for the retinal chromophore in bathorhodopsin) (Bagley et al., 1985). Later Raman studies by Loppnow et al. (1990) on rhodopsin regenerated with 7,9-dicis-retinal confirmed these results on the basis of deuterium induced shifts (Loppnow et al., 1990). Due to the presence of two cis bonds in this chromophore, it was assumed that the N-H bond of the 7,9-dicis-analog would point in a direction opposite to that of the native chromophore. Based on the assumption that the ionone ring should occupy a similar binding site as that found in the 11-cis- and 9-cisrhodopsins, it was expected that the C--N bond for this analog would exhibit a syn geometry. The unanticipated result of a trans (anti) geometry for this chromophore can be rationalized in light of comparisons to a model binding site proposed by Liu and Mirzadegan, (1988) which indicated that the syn configuration would result in a placement of the ionone ring that is incompatible with the binding site geometry. In addition, this arrangement would not allow for proper alignment of the protonated Schiff base with its proposed counterion (Liu and Mirzadegan, 1988; Loppnow et al., 1990). In bacteriorhodopsin it has been determined that a significant portion of the opsin shift results from a weakly hydrogen bonded protonated Schiff base (see above). In contrast, results from Raman data show that the rhodopsin Schiffbase is strongly hydrogen bonded, based on C = N stretching modes which are sensitive to the local environment (Baasov et al., 1987a; Deng and Callender, 1987; Gilson et al., 1988). This conclusion is based primarily on the observation of a large deuterium shift of --30 cm-1 (Gibson and Cassum, 1987), although the observed stretching frequency of 1,660 cm-~ also supports this conclusion: the protonated Schiffbase in the protein

58

JEFFREYA. STUART and ROBERT R. BIRGE

is more strongly solvated than that of model compounds in methanol (Baasov et al., 1987a,b). Neither the observed stretching frequency or the large deuterium shift change on phototransformation to bathorhodopsin, indicating that the Schiff base environment is not significantly changed (Deng and Callender, 1987). Other factors that affect the C--N stretching frequency have already been discussed (for a good review see Deng and Callender, 1987).

Chromophore Orientation and Location There is little direct information available on the orientation of the chromophore in the binding site. Early polarized light studies on individual rod cells viewed through a microscope gave a chromophore orientation in rhodopsin of 16°-18 ° with respect to the membrane plane (Chabre et al., 1982). This roughly parallel orientation to the plane of the membrane is close to ideal for the absorption ofunpolarized light (Nathans, 1992). Chromophore location in the protein was determined via fluorescence energy transfer to be roughly in the center of the membrane, at about 22 ,~ from the interior of the disc membrane and 28 A from the exterior aqueous space (Thomas and Stryer, 1982). The chromophore for iodopsin, a red sensitive pigment in chicken retinal, is closer to the membrane surface as determined by binding site accessibility studies (Fukuda et al., 1990). Recent site directed mutagenesis studies indicate that the protonated Schiffbase is located between residues 113 and 117 (Zhukovsky et al., 1992), and ESR (electron spin resonance) studies indicate that the ring moiety of the protonated retinal Schiffbase lies sequestered deep within the membrane in a hydrophobic pocket (Renk et al., 1987). Chemical cross-linking studies using a photoactivatable analog of 11-cis-retinal indicated ring interactions with helices C and F (Nakayama and Khorana, 1990). Cross linking sites on helix C included Phe- 115, Ala- 117, Glu- 122, Trp- 126, and Ser- 127, while Trp-126 was the only significant site on helix E These results are consistent with the ring moiety being buried in a binding pocket formed by helices C and F, serving to anchor the chromophore in a specific orientation. Such an orientation, with the ring placed toward Phe-115 and Ala-117, yields results consistent with those obtained from linear dichroism studies by Chabre (Chabre, 1975; Chabre et al., 1982; Chabre and Worcester, 1982).

Characterization of the Opsin Shift From the data available, the opsin shift of about 2,650 c m -1 in rhodopsin seems to have only two major contributing factors (see Table 1): electrostatic interactions along the polyene chain (Ottolenghi and Sheves, 1989) and a twisted 6-s-cis ring-chain configuration (Smith et al., 1987a). Unlike bacteriorhodopsin, the protonated Schiff base is strongly hydrogen bonded (as shown for solvated model compounds), and the chromophore configuration is close to that of model compounds in solution. In addition, no protein induced perturbations in the vicinity of the ring moiety are known to exist, although recent circular dichroism studies done on rhodopsins incorporating 6-s-cis- and 6-s-trans-locked bicyclic retinals did lend

Characterization of the Primary Photochemical Events

59

some support to the contrary (Ito et al., 1992). The analog proteins displayed shifts smaller than that of bovine rhodopsin, suggesting that protein interactions in the cyclohexene ring area are responsible for a portion of the opsin shift (see below) (Ito et al., 1992). The picture of the binding site that emerges is one that, at least for bovine rhodopsin, is much less complicated than bacteriorhodopsin. The studies discussed above are consistent with the existence of a negatively charged counterion that is hydrogen bonded to the protonated Schiff base nitrogen. In addition, some sort of protein induced perturbation is present near C12 (see below). As alluded to earlier, there is some controversy as to the nature of this perturbation to the polyene chain. Most studies seem to indicate the presence of an electrostatic interaction, but recent two-photon studies on rhodopsin with a locked 11-cis chromophore by Birge et al. (1985) indicate that the binding site is overall electrically neutral. Additional support for an overall neutral binding site comes from recent resonance Raman microprobe studies (Lin et al., 1992), and sitedirected mutagenesis studies (Zhukovsky and Oprian, 1989; Nathans et al., 1990b). These studies present strong arguments against the existence of additional counterions in the active site. Studies by Lin et al. (1992) presented evidence for a direct interaction between the protonated Schiff base and its counterion, Glu-113. Based upon the comparison of the vibrational spectra from a number of mutant rhodopsins and native rhodopsin, they concluded that not only is there a strong interaction between Glu- 113 and the protonated Schiffbase, but Glu- 113 additionally perturbs the retinal polyene chain at position C12 (Lin et al., 1992) (as evidenced by weakened hydrogen out-of-plane vibrations in rhodopsin with mutations at position 113). In systematic replacements of all of the negatively charged amino acids in rhodopsin, Nathans demonstrated that only Glu-113 interacted directly with the chromophore (Nathans, 1990b). In previously done similar studies, Zhukovsky and Oprian provide additional evidence for the existence of only one counterion in the active site (Zhukovsky and Oprian, 1989).

Active Site Interactions The primary counterion. As discussed for bacteriorhodopsin, the protein chromophore interactions in the rhodopsin active site are responsible for the observed opsin shift of 2,650 cm-1 (Smith et al., 1987a). Contributions to the shift from chromophore configuration are discussed above, leaving the remainder of the shift to be accounted for by electrostatic and polar interactions. In bovine rhodopsin both types of interactions may exist. Of the two types, the largest effect can be attributed to the primary counterion, whose identity has recently been confirmed as Glu-113 by numerous site directed mutagenesis studies (Sakmar et al., 1989, 1991; Zhukovsky and Oprian, 1989; Nathans, 1990a,b,c, 1992; Zhukovsky et al., 1992). Previously thought to be Glu- 122, replacement of Glu- 113 by Gin leads to a more dramatic shift in the absorption spectrum (500 nm to 380 nm), as well as a drop in the pK a of the protonated Schiff base to about 6 from its native value of

60

JEFFREYA. STUART and ROBERT R. BIRGE

greater than 16 (Zhukovsky and Oprian, 1989; Gat and Sheves, 1993; Steinberg et al., 1993). Replacement with an aspartate residue (E 113D) resulted in a photoactive pigment that was only slightly red shifted (505 nm): this is consistent with a red shift that is produced by movement of the primary counterion away from the PSB (aspartate has a shorter chain length than glutamate) (Sakmar et al., 1989). In a number of SDM studies by Nathans, a total of more than 20 mutants were examined, where each negatively charged aspartate and glutamate was replaced by a neutral residue (Nathans, 1990a,b, 1992). With the exception of replacements at Glu-113, every one of the mutants exhibited an absorption spectrum with properties identical, or very similar, to the wild type protein. Disallowing any cooperative effects, Glu-113 is the primary counterion. An additional result of his studies indicates that none of the other negatively charged residues plays any significant role in the spectral tuning of the chromophore. Any other active site perturbations would therefore have to be nonelectrostatic in nature. In addition to the primary counterion, there is a large body of evidence for an additional perturbation along the retinal polyene chain. There has been some controversy as to whether the perturbation is electrostatic or the result of a dipole interaction, but currently the former type is in favor. Evidence supporting the existence of a second point charge comes primarily from NMR spectroscopy. Solid state ~3C NMR chemical shift data of the retinal chromophore done by Smith et al. (1987a) indicated no charge perturbation at C14, but similar studies by Mollevanger et al. (1987) in the same year did indicate a negative charge perturbation in the vicinity of C12. Subsequent solid state 13C NMR studies by Smith et al. (1990) indicated a deshielding of C13, presumably the result of a charge perturbation. Resonance Raman studies on the factors that affect the C---N stretching frequency (Baasov et al., 1987a) placed the perturbation closer to carbons 9 and 10 than to 12 and 14. A charge in the vicinity of C12--C14would be expected to modulate the C--N stretching frequency, which was not evident. Finally, synthetic analog studies have provided strong evidence for the existence of an electrostatic perturbation along the polyene chain. Regeneration studies of bovine and octopus rhodopsins with a series of 11-cis-dihydro chromophores supported a model incorporating an active site with two negative charges: one in the vicinity of C13 and a second directly interacting with the protonated retinal Schiffbase (Koutalos et al., 1989). In a more recent study by Han and coworkers, semi-empirical molecular orbital calculations and 13C NMR chemical shifts were used to locate the counterion in bovine rhodopsin (Han et al., 1993). Based on comparisons to model compounds, it was determined that the primary counterion must be located in the vicinity of C12; additional counterions were unable to accurately model the observed chemical shifts, nor could they be modeled by varying the distance between counterion and chromophore (Han et al., 1993). The authors propose a model that places a carboxylate counterion (presumably Glu-113) interacting specifically with C12. As any attempt to introduce additional ionic perturbations along the polyene chain resulted in models that did not accurately reproduce the chemical shifts seen in the

Characterization of the Primary Photochemical Events

61

native protein, only one counterion must be present, allowing for a neutral binding site. The protonated Schiffbase does not directly interact with the primary counterion, but is probably hydrogen bonded to one or more water molecules. This assertion is supported by the work of Ganter et al. (1988b) who found that the protonated Schiffbase was stabilized by water (see below). Furthermore, only one of the carboxylate oxygens can interact directly with the polyene chain, and it must be closest to C12 (Han et al., 1993). This binding site model is in many respects similar to that proposed by Birge and coworkers based on two-photon studies, the main difference being in the orientation of the carboxylate group with respect carbons 13-15 (see below) (Birge, 1990a; Birge and Zhang, 1990b; Han et al., 1993). It must be noted, however, that a recent study of pK a values calculated for bacteriorhodopsin ac.tive site residues concluded that highly structured water molecules buried deep within the protein can stabilize ionic species to such an extent that electroneutrality is no longer required for the maintenance of a binding site with charged residues (such as Asp-85, -212 and the protonated Schiffbase in bR (Sampogna and Honig, 1994). A comparative analog binding study ofrhodopsin and iodopsin (a red-sensitive pigment from chicken retina) indicated that the location of the polyene charge perturbation was similar in both proteins, but location of the primary counterion to the PSB had to be different to account for the larger red shift of the chicken pigment (Fukuda et al., 1990). It is interesting to note that a resonance Raman study of blue visual pigments has shown that for these pigments, an unperturbed protonated Schiffbase alone is sufficient to account for the observed 440 nm absorption maximum, without the presence of any additional protein-chromophore interactions (Loppnow et al., 1989). Role of water in the active site. As in bacteriorhodopsin, there is strong evidence for the presence of water in the rhodopsin chromophore binding site. The first such evidence came from dehydration studies of air-dried films of bovine rod outer segments (Rafferty and Shichi, 1981). Dehydration of these films produced a pigment that absorbed maximally at 390 nm. The transition to this pigment was found to be fully reversible, as evidenced by the reversal of the changes in the absorption spectrum. The authors attributed these spectral changes to the reversible deprotonation of the retinylidene Schiff base in more than 50% of the rhodopsin molecules. Furthermore, the dehydrated species underwent photochemistry to form a red shifted product (479 nm) similar to metarhodopsin I, which is protonated. This lead the authors to postulate that there are different mechanisms for SB protonation in rhodopsin and meta-I. Water in the form of a hydronium ion hydrogen bonded to the SB provides the proton for rhodopsin, while that for meta-I must come from some unidentified donor within the active site via a light-induced proton transfer (Rafferty and Shichi, 1981). Later analogous studies on the photoreaction of vacuum-dried samples indicated that rhodopsin, isorhodopsin, and lumirhodopsin exist with the SB in an equilibrium ofprotonated and nonprotonated forms (Ganter et al., 1988b). Addition of water had the effect of stabilizing the

62

JEFFREYA. STUART and ROBERT R. BIRGE

protonated forms of each; in metarhodopsin-I the protein itself was able to stabilize the protonated Schiffbase. Conformation of the involvement of water was obtained by difference FTIR spectroscopy of rhodopsin and bathorhodopsin hydrated with 180 labeled water (Ganter et al., 1988b). The difference spectra (between rhodopsin with and without 180 labeled water) indicated a small but reproducible line at 1684 crn-1 which can be attributed to the change in the environment of a water molecule tightly bound to the Schiff base during the phototransition. Their data are also consistent with a model of the primary event in which the Schiffbase environment is largely unchanged during the transition to bathorhodopsin (the line at 1684 cm -1 remains unchanged in the phototransition from rhodopsin to bathorhodopsin). Additional support for the roll of water in the active site comes from recent PSB pK a studies (Gat and Sheves, 1993; Steinberg et al., 1993). Both of these studies indicate an apparent pK a of greater than 16 for rhodopsin, which is posited to result from a specific arrangement between the protonated Schiff base linkage and a carboxylate group that incorporates one or more water molecules that act to bridge the two groups in a strongly hydrogen bonded configuration (Gat and Sheves, 1993). Finally, a recent study by Deng et al. (1994) reports evidence for a water molecule tightly bound to the retinal Schiff base in both bacteriorhodopsin and bovine rhodopsin as indicated by deuterium-hydrogen exchange studies on the Schiffbase. Their studies indicated that the exchange is very fast for both proteins 01/2 = 6.9 ms and 1.3 ms for rhodopsin and bacteriorhodopsin, respectively), and is pH dependent, indicating water catalysis. The presence of one or more bound water molecules with a specific orientation would increase the effective water concentration around the Schiff base, thereby accelerating the exchange process (relative to model compounds in aqueous solution) (Deng et al., 1994). These conclusions concur with those arrived at by the pK a studies mentioned above. It is also noted that a recent FTIR study indicates that the rhodopsin protein on the whole has a generally larger water accessibility than previously thought (Garcia-Quintana et al., 1993).

Description of the Active Site The active site model that is most consistent with the experimental evidence currently available is that proposed by Birge in 1985 (Birge et al., 1985), in which a single counterion interacts not only with the chromophore, but also with the C12-C15 region of the retinal polyene chain. The counterion carboxylate is oriented in such a way as to allow interaction of the negatively charged oxygen atoms with carbons 12 through 15. Raman, FTIR, one-photon, two-photon, photocalorimetric, and chromophore analog studies are consistent with the model, which is shown in Figure 6 (Birge et al., 1985, 1988; Kakitani et al., 1985). Models which move the counterion one or two atoms closer to the center of the polyene chain and simultaneously closer to the polyene atoms are also possible. However, the positive charge is primarily localized on atoms C13 , C15 and the imine proton (Birge and Hubbard, 1980; Birge et al., 1987), and one oxygen atom must be within --4 A of this section

Characterization of the Primary Photochemical Events

63

in order to accommodate the transition energy and oscillator strength data (Birge et al., 1988). As discussed above, a similar model has recently been proposed by Han et al. (1993) which differs in that only one of the carboxylate oxygens is permitted to interact directly with the chromophore at C12. The second oxygen in their model points away from the chain. As discussed earlier, water most certainly plays an important role in the active site. There is evidence to suggest that one or more water molecules are present within the binding site (Rafferty and Shichi, 1981), and the observation of an ND shift of 33 cm-1 suggests that one water molecule is hydrogen bonded to the imine proton. This is a somewhat controversial assignment, because many investigators assume that the imine proton is hydrogen bonded to a negatively charged counterion based on deuterium isotope effects. It is generally recognized that a large Vc=N deuterium (ND shift) isotope shift is characteristic of strong hydrogen bonding to the imine proton (Bagley et al., 1985; Deng and Callender, 1987; Palings et al., 1987). Based on this argument, hydrogen bonding is strongest in R (ND shift--33 cm-1), strong in B (ND shift ~31 cm -1) and moderately strong in I (ND shift--24 cm-1). As a comparison, the all-trans protonated Schiffbase (ATRPSB) in methanol exhibits a ND shift of~26 cm-1. Spectroscopic studies of ATRPSB indicate that the counterion is intimately associated with the imine proton in nonpolar environments (Birge et al., 1987). The work ofBlatz (Blatz and Mohler, 1972) suggests that in highly polar, strongly hydrogen bonding solvents, the counterion is highly solvated and the imine proton is hydrogen bonded with the solvent. The ND shift has been measured for ATRPSB in both environments and differs by only 3 cm-~ (Deng and Callender, 1987" Palings et al., 1987; Smith et al., 1985b). One concludes that the ND shift, while diagnostic of hydrogen bonding, is not sensitive to the nature of the hydrogen bond. A quantitative relationship between the ND shift and the strength of a hydrogen bond to the imine proton has not been firmly established (for more details see Kakitani et al., 1983; Lopez-Garriga et al., 1986a,b; Deng and Callander, 1987). Although the ND shift is largest in R, it drops by only 2 cm -1 in going from R to B. This difference is anomalously small for an isomerization moving the - C 15=NH- moiety away from a fixed counterion. Any attempt to maintain a strong hydrogen bond between the imine proton and the counterion following a one-bond 11-cis ~ 11-trans photoisomerization will fail to accommodate the observed oscillator strength and spectral shifts. Many proponents of hydrogen bonding between the counterion and the imine proton in R suggest that the imine proton is hydrogen bonded to an uncharged protein residue in B. This model accommodates the bathochromic shift, but does not explain adequately the remarkable similarity Of Vc__nu and VC_NDin R, B and I. It is concluded, based on the force field calculations of Deng and Callender (1987), and the model compound studies ofBaasov et al. (Baasov and Sheves, 1985; Baasov et al., 1987a,b), that VC_NH and vC_ND are rather sensitive to the charge on the hydrogen bonding species. The above model is also incapable of rationalizing a

64

JEFFREYA. STUART and ROBERT R. BIRGE

weaker hydrogen bond to the counterion along with a blue shifted absorption maximum in I. Hydrogen bonding of the imine proton to water provides the best model, because it explains both the magnitude and the similarity in the ND shift observed in R, B and I. We conclude that Figure 6 provides the most realistic model for the binding site in rhodopsin.

II!. THE PRIMARY EVENT IN BACTERIORHODOPSIN The primary event in photoactive proteins is loosely defined as the immediate response by the protein to stimulus by light. In most of the systems studied, this response takes the form of an isomerization or a charge separation. Retinal based proteins, including bacteriorhodopsin and rhodopsin, fall into the former category. Photosynthetic reaction centers constitute an example of the latter. It is now commonly accepted that the primary event in bacteriorhodopsin is a very fast trans to cis isomerization about the C13-C14 double bond of the retinal chromophore (Figure 3) (Honig et al., 1979; Braiman and Mathies, 1980; Tsuda et al., 1980; Kuschmitz and Hess, 1982; Fang et al., 1983; Albeck et al., 1986). Absorption of light initiates a photochemical cycle consisting of at least seven spectrally distinct intermediates, several of which can be trapped at low temperatures (see Figure 2). The initial trans to cis isomerization is the only photochemical transformation in the cycle, the remainder being thermally driven. The isomerization is therefore of special interest, as it acts to "prime" the protein to perform its function, namely the translocation of protons across the cell membrane. Energy that is stored from the absorbed photon is used to drive the subsequent thermally driven steps (the dark reactions) in the photocycle that act as the proton pump. The K intermediate is generally still recognized as the primary photoproduct as it is the first intermediate that can be trapped at low temperatures (77 K). There is at least one known precursor to K, denoted as J, which many authors now consider as the primary photoproduct, although it is not stable even at liquid helium temperatures. It is slightly shifted to the red with respect to K (J ~max~ 610--425 nm). Spectra of bR and K are shown in Figure 7. Evidence for the involvement of a trans to cis isomerization about the C13=C14 double bond of the retinal chromophore came from low temperature resonance Raman studies of the trapped K intermediate (Braiman and Mathies, 1980, 1982). Studies with artificial pigments designed to block isomerization provide the most compelling evidence in support of the C13"-C14 location (Sheves, et al., 1985). Consequent studies of the isomerization have employed time-resolved techniques, including absorbance, Raman, and fluorescence spectroscopies (examined in more depth below). Most of these studies are consistent with the following general reaction scheme for the primary event:

bR~~-~sJ i!~s K

R

440 nm CAg" state)

300 K

•

~350nm

][,} 77K

.

ii

---,',, ,~498 am (505 nm; 77K) ' (f = 0.98; 77K)

[

~'

_~ .~ ~ .~ ~ -

~ ~.. o .....,........... .................................

~

i

543om I,°107!

\

]

i

300K

i; 77K) K)

bR

i

560 nm CBu" state) .... ,vv,,v,~-..<."q.~ ............577 , 568 nm nm (f=0.87) (300 K)

77K

488nm ("Ag" state) ~

%

"

~ ", ~:~ii!:"*: 4:" !

398 nm 375 n m

.

, .

.

.

.

.

.

.

.

.

__.[

i

I

K

77K 440 nm: 4

300

400

.......,

620 nm (f=0.95) ~ : ~ - ~ " 7~-~~*-~-~ ' "":" -" "°*';'

500

600

i

[

I

700

Wavelength (nm) Figure 7. Electronic (one-photon) absorption spectra of rhodopsin (R), bathorhodopsin (B), isorhodopsin (I), light adapted bacteriorhodopsin (bR) and K. Temperatures of the one-photon spectra are shown below the labels. The two-photon thermal lens spectrum of rhodopsin and the two-photon double resonance spectrum of bacteriorhodopsin (both in D20 at room temperature) are shown in the inserts. The two-photon spectra have been shifted to lower absorptivity so that only the maxima are shown. The "Ag" labels indicate two-photon maxima assigned to ,1Ag*-" ~-- So transitions and ~,u ~-- So transitions (see the "Bu" labels indicate two-photon maxima assigned to ,,1 ~,+,, text). Oscillator strengths (f) determined by log-normal fits of the ~.max bands are indicated in parentheses. (One photon spectra are from Birge et al., 1989, Birge et al., 1988 and Crouch et al., 1975 and two-photon spectra are from Birge et al., 1985 and Zhang, 1989). 65

66

JEFFREYA. STUART and ROBERT R. BIRGE

A. The I Intermediate As mentioned earlier, the assignment of J as a distinct cycle intermediate is controversial because it cannot be trapped at low temperatures. Early studies by Applebury et al. (1978) were the first to report a precursor to K on the basis of time-resolved kinetic studies. A similar precursor was reported by Shichida and coworkers in 1983, but an instrument resolution of 30 ps precluded any definitive characterization (Shichida et al., 1983). Structurally, J is generally regarded as being an isomerized ground-state photoproduct (Mathies et al., 1991). The possibility that it might be a mixture of ground and trapped excited state species has been proposed based on theoretical calculations (Birge et al., 1987). Mathies and coworkers disagree with this proposal and suggest from their time-resolved data a fully populated ground state within the formation time of J, and conclude that they directly observe the trans to cis isomerization on the same time-scale (Mathies et al., 1988). However, recent fluorescence lifetime studies of bacteriorhodopsin by Duet al. indicate that as much as 40% of the molecules remain in the excited state after 500 fs, which provide s support for the theoretical model of Birge et al. (Du and Fleming, 1993). Because it is well known that the isomerization to (a possibly distorted) 13-cis-retinal has already occurred by the formation of the K intermediate (Braiman and Mathies, 1982; Polland et al., 1984), it is usually assumed that J also has a 13-cisoid structure as well (due to similarities in the absorption spectra) (Nuss et al., 1985; Polland et al., 1986; Dobler et al., 1988; Mathies et al., 1988; Pollard et al., 1989). In a recent picosecond time-resolved resonance Raman study by Doig et al. (1991) strong hydrogen-out-of-plane intensities (HOOP) were seen at 1,000 and 956 cm -1, which dropped in intensity by the formation of K. The fingerprint region showed a broad band of lines from 1,155 to 1,200 cm -1, and an ethylenic line at 1,518 cm-1. On formation of K, the fingerprint region collapsed to a single strong mode at 1,189 cm-~, while the ethylenic mode remained unchanged (Doig et al., 1991). It was concluded that the chromophore in the J intermediate is a highly twisted and thermally excited 13-cis configuration, and that the J ~ K transition is due to vibrational cooling and conformational relaxation of the chromophore to a more planar 13-cis configuration (as evidenced by decreased HOOP intensities) (Doig et al., 1991). van den Berg and coworkers assigned the band at 1,518 cm-1 to the J intermediate due to its correlation with the 625 nm absorption maximum. Furthermore they report the appearance of a vibration at 1,195 cm-~ (absent in the bR spectrum) that is indicative of isomerization at the C13"--C14 locale (i.e., formation of the 13-cis retinal chromophore) (van den Berg et al., 1990). We conclude that to a first approximation, J is best viewed as vibrationally hot K.

B. The K Intermediate K has traditionally been assigned as the primary photoproduct because it is the first that can be trapped cryogenically. As such, it is much better characterized than

Characterization of the Primary Photochemical Events

67

is J. It is also perhaps the most important of the intermediates with respect to function, as the energy of the photon absorbed by bR is stored in K (to be discussed below). The absorption maximum is about 600 nm at room temperature (625 nm at 77 K). It is formed from J in as little as 3 ps, and has been identified as having the isomerized 13-cis chromophore (Braiman and Mathies, 1982; Brack and Atkinson, 1989; van den Berg et al., 1990; Doig et al., 1991), and has a lifetime of 2 gs (Lohrmann et al., 1991; Lohrmann and Stockburger, 1992). There is some support for the existence of multiple conformers of K at low temperatures (Kalisky and Ottolenghi, 1982; Rothschild et al., 1985; Balashov et al., 199 lc). Balashov et al. (1991 c) observed at least three spectrally distinct species at 77 K, which were apparently formed from differem conformers of trans-bacteriorhodopsin. The existence of parallel photocycles was posited, which may contribute to the room temperature photocycle if the conformers are stable (Balashov et al., 1991c). Although there is some support for this possibility (Hanamoto et al., 1984; Dancshfizy et al., 1988; Baloshov et al., 1991 a), any contribution to the primary event is as of yet unknown. As alluded to earlier, the transition from J to K is likely to be a thermal relaxation (vibrational cooling) of the chromophore. The absence of HOOP intensities in the resonance Raman spectrum implies that the chromophore is more relaxed about its single and double bonds, giving a more planar chromophore than in the J intermediate (Doig et al., 1991). This is in comrast with the low temperature spectra of K reported by Braiman and Mathies (1982), and might indicate that at low temperatures the amino acid residues that interact with the retinal moiety are more rigidly fixed and thereby prevent full chromophore relaxation so that it is held in a distorted 13-cis configuration (Doig et al., 1991). Transient spectra obtained in room temperature picosecond kinetic IR experiments display similarities to low temperature (77 K) spectra ofK (Diller et al., 1991; Rothschild, 1992), perhaps lending support for the thermal transition from J. It should be noted, however, that despite the evidence for a relaxed conformation, the chromophore in K is still not what would be described as classically planar. Strong HOOP modes have been identified in room temperature resonance Raman studies, indicating strong chromophore distortions (Lohrmann et al., 1991; Lohrmann and Stockburger, 1992). FTIR bR -~ K difference spectra also indicate a twisted chromophore confined by steric interactions in the binding site (Siebert and Mantele, 1983; Rothschild et al., 1986). In addition, FTIR has provided evidence for a change in the structure of K at low (77 K) vs. high (13 5 K) temperature, by examination of the C = C and C--N stretching modes, and the HOOP modes. Both frequencies shift to higher energy at the higher temperature, and may indicate a more relaxed, and less restricted, chromophore and protein conformation (Rothschild, 1992). Differences in the FTIR spectra of K and L were attributed to a distortion in the 13-cis retinal chromophore in the vicinity of the protonated Schiff base (Maeda et al., 1991). NMR and resonance Raman experiments have shown the C15=N configuration to be anti (tram) as in bR (Harbison et al., 1984b; Smith et al., 1984), and recent resonance Raman studies

68

JEFFREYA. STUART and ROBERT R. BIRGE

have confirmed previous work (Temer et al., 1979) that the Schiffbase nitrogen is protonated (Lohrmann et al., 1991; Lohrmann and Stockburger, 1992). It also appears that there is a change in the environment at the Schiffbase, due to a 30 cm-] shift in the frequency of the C--N bond with respect to bR (which, as has been discussed above, is a sensitive measure of the local milieu) (Rothschild et al., 1984). This shift is probably due to an increased delocalization of electron density in the polyene chain and/or a shift in the frequency of the N-H in-plane bending mode, the net result of which is a weakening of the electrostatic interactions with the Schiff base (Rothschild, 1992). A plausible structure for K is shown in Figure 3. The assignment of the ground state configuration of the C~4-C15bond remains a matter of some controversy, centering around a model for the primary photochemical event proposed by Tavan and Schulten that involves a concerted isomerization from 13-trans,14-s-trans to 13-cis,14-s-cis (Schulten and Tavan, 1978; Schulten et al., 1984; Tavan and Schulten, 1986; Tavan, 1988):

bR

J,K

15

7

H"

0"Cl

C,ls.,___(14-s-cis)

C131

~

(13-cis)

There are three purported advantages to this model. First, the concerted motion requires only modest mass movement, allowing the isomerization to proceed very quickly. Second, the all-trans to 13,14-cis di-isomerization is stable against thermal back reaction whereas the mono-isomerization to 13-cis is not. Third, an indirect sequential re-isomerization enforces a deprotonation and subsequent reprotonation of the chromophore in attune with the proton transport function of the protein (Tavan and Schulten, 1986). Although FTIR (Gerwert and Hess, 1987) and picosecond (Pollard et al., 1986) experiments have been reported to be consistent with this model, there are three problems with the notion of a concerted 13,14-cis primary event that preclude enthusiasm. First, the kinetics of the primary event as well as the wavelength and temperature independence of the quantum efficiency of the primary event requires a barrier-less excited state surface coupling bR and the

Characterization of the Primary Photochemical Events

69

primary photoproduct (see below). It is unlikely that a concerted photochemical reaction involving the simultaneous rotation about two polyene bonds would produce a barrier-less excited state surface. Second, the detailed resonance Raman studies of Mathies and coworkers are not consistent with a 14-s-cis geometry in K or L (Smith et al., 1986, 1987; Fodor et al., 1988a,b). (However, Tavan and Schulten (1986) have presented MNDO calculations which contradict the Raman assignments.) Third, photostationary state spectroscopic studies are inconsistent with a 13,14-cis primary photoproduct (Birge et al., 1989). The mole fraction ofK (Z~00) in the 77K, 500 nm photostationary state is observed to equal 0.46 + 0.04 (Birge et al., 1989). The calculated absorption spectrum of K at 77K has a maximum absorbance at 620 nm, and a molar absorptivity at ~maxof 63,900 M-1 cm-1. The oscillator strength associated with excitation into the ~max band, fK, is determined to be 0.95 based on log-normal regression analysis (Birge et al., 1989). The corresponding values for bR at 77K are: )gmax "- 577 nm, 8ma x = 66,100 M -1 cm-1, and fbR = 0.87. The observation that fK > fbR is consistent with the displacement of the C15=NH portion of the retinyl chromophore away from a negatively charged counterion as a consequence of the all-trans to 13-cis photoisomerization. It is difficult to reconcile the observation that fK > fbR with the proposal that the primary event involves an all-trans to 13-cis, 14-s-cis photoisomerization, because the latter geometry is predicted to have a significantly lower ~maxband oscillator strength relative to the all-trans precursor, regardless of counterion location. Several recent studies by the progenitors of the 13,14-cis model of the primary event seemingly provide further support. In a series of spectroscopic experiments (FTIR, linear dichroism, photoselection) combined with theoretical analysis, Fahmy et al. (1989) reported that the C14--C15 bond contains a 37 ° twist in bR. This distortion is also reported to be found in the L state of the photocycle (Fahmy et al., 1991). The 13,14-cis model proposed by Tavan and Schulten requires that the chromophore have a 14-s-cis structure in the K and L states (Schulten and Tavan, 1978). Concrete experimental evidence for this assignment in any of the intermediates of the photocycle is lacking, with several studies in support (Orlandi and Schulten, 1979; Gerwert and Siebert, 1986; Tavan and Schulten, 1986; Fahmy et al., 1989, 1991), and several others opposed (Smith et al., 1985a; Fodor et al., 1988b; Mathies et al., 1991). A molecular dynamics study by Nonella et al. (1991) examined both the 13-cis and 13,14-cis models and concluded that the latter is more theoretically appealing. The studies were able to predict a photoisomerization time of about 400 fs which is close to that measured experimentally. In addition, they report that the 13,14-cis model is accompanied by little rearrangement ofLys-216 and the protein backbone, and that it can more easily explain the proton transport mechanism of bacteriorhodopsin (as an acid-base catalyzed system coupled to the deprotonation and protonation of the Schiff base nitrogen) (Nonella et al., 1991). According to their calculations, the 13-cis model, however, is reported to have large steric hindrances (> 10 kT at physiological temperatures), induces significant structural changes in the binding site, and requires longer times for the isomerization to

70

JEFFREYA. STUART and ROBERT R. BIRGE

be complete (~800 fs) (Nonella et al., 1991). Additionally, the 13-cis model supposedly is accompanied by a co-rotation of the C14--C15bond of about 150 °, such that the final post-isomerization geometries of both models are nearly identical (Nonella et al., 1991). It must be reemphasized, however, that experimental evidence for the 13,14-cis model is still lacking.

C. Ultrafast Spectroscopic Studies Subpicosecond studies performed by several groups confirmed the formation of J in ~500 fs with a lifetime 3-4 ps (Nuss et al., 1985; Sharkov et al., 1985; Polland et al., 1986; Dobler et al., 1988; Mathies et al., 1988, 1989, 1991; Doig et al., 1991). No deuterium isotope effect is observed in this transition indicating that proton transfers do not play a roll at this stage of the photocycle (Nuss et al., 1985; Polland et al., 1986). J is generally assumed to be formed from the first singlet excited state S 1 of bR, which has a blue shifted absorption maximum (460 nm) and a reported lifetime of 0.7 ps (Nuss et al., 1985; Sharkov et al., 1985; Polland et al., 1986). On the basis of relative formation and decay times for bR* and J, however, it has recently been suggested that there might be parallel reaction channels for both of these species from the Franck-Condon excited state; J seems to form with a time constant faster than bR*, such that bR* could not therefore act as a common excited state to bR and J (Kandori et al., 1993). The relaxation time from the Franck-Condon excited state has been measured by femtosecond spectroscopy to be 200+70 fs, and represents the motion of the retinal molecule along its excited-state potential surface to the 13-cis conformations (Dobler et al., 1988). This time constant has recently been confirmed by femtosecond time-resolved fluorescence spectroscopy; 200 and 500 fs relaxation components were measured which were interpreted to represent relaxation in the excited state, and a nonradiative process from the excited-state to the ground-state to form reactam and product (Du and Fleming, 1993). The authors conclude that the data is consistent with a model that displays a 100-200 fs relaxation process in the excited state before electronic relaxation to the ground state and branching to photoproduct and reactant geometries (Du and Fleming, 1993). This conclusion, however, should be interpreted somewhat loosely; the fluorescence decays were best fitted as a sum of three exponentials with time constants of 90-240 fs (40-70% ofthe amplitude), 0.6-0.9 ps (20-40%), and 9.0-13.0 ps (5-25%) (Du and Fleming, 1993). These results are not totally consistent with those of transient absorption spectroscopy experiments (described below). The initial time constant was attributed to an excited-state relaxation process, or due to the formation of reactant and product from the excited-state. Assignment of the longer time constants is somewhat more difficult, and it is suggested that they are due to reactive processes from the excited-state to the photoproduct and the bR ground-state: possibilities may include thermal relaxation from J to K, the lifetime of the K intermediate, protein (as opposed to chromophore) relaxation processes, characteristic processes inherent

Characterization of the Primary Photochemical Events (U (J

rlIJ 4J

~ YTZ////2//~

71

Pulse Spectrum i i i |

•

__

_

i

....

998

fs

_ "

443-fs

fD

-=--~222 fs

= o

142 f s

p-

98

i--I

58 f s

or-4

4J

28

~

o

fs

-

,,-,

560

~

, ~

580

-

,~-,-

600

620

,

, -, 640

Wavelength (nm)

,.

~--~-

660

.

.

.

.

~

, ~--Y~680 700

-

2

8

-54

fs

fs

fs

Figure 8. Transient absorption spectra of light-adapted bacteriorhodopsin as a function of delay time in femtoseconds (1 fs = 10 -15 s). The pump pulse at 618 nm (dashed line) had a duration of 60 fs, and the delayed probe pulses were 6 fs in duration. (Reproduced with permission from Mathies et al., 1988).

to barrierless excited-state potential surface, or a distribution of unidentified parameters that might influence the reactive process, such as the existence of protein conformers (Du and Fleming, 1993). As noted above, however, their results are consistent with the model proposed by Birge in which a portion of the excited state trajectory is trapped in an excited state potential minimum (Birge et al., 1987). The existence of protein conformers in bacteriorhodopsin has been reported by kinetic resonance Raman studies (Diller and Stockburger, 1988; Eisfeld et al., 1993) and time-resolved, flash-induced difference spectra (Dancshfizy et al., 1988). The high degree of nonexponential behavior of the fluorescence decay (in contrast to the transient absorption studies) indicates that the excited-state torsional surface is fairly complex. The femtosecond dynamic hole burning experiments by Mathies and coworkers, however, yielded a more facile interpretation: direct observation of the retinal isomerization (Figure 8) (Mathies et al., 1988, 1989, 1991; Pollard et al., 1989). Their results are consistent with a model that displays rapid torsional dynamics along the excited state potential surface (--100 fs), followed by the formation of the isomerized ground state photoproduct after about 200 fs (Pollard et al., 1989). Work by Nuss et al. (1985) determined an overall formation time for J of 430+50 fs, which was later confirmed by Dobler and coworkers with a measured time of 500+_100 fs (Dobler et al., 1988). More recently, femtosecond impulsive excitation experiments at 568 nm observed an approximate 200 fs delay, consistent with the relaxation of the Franck-Condon geometry bR(FC*) to bR(~ 13)*

72

JEFFREY A. STUART and ROBERT R. BIRGE

(where the latter term represents the excited state with the chromophore distorted out of the Franck-Condon region along the C13=C14 torsional coordinate), which was followed by a ground state recovery time of 450 fs for the formation of J (Dexheimer et al., 1992). The same study was able to demonstrate the existence of oscillatory behavior due to coherent vibrational motion of the chromophore at each of the three wavelengths examined (568, 620, and 656 nm). Fourier analysis revealed frequency components which corresponded well to dominant vibrational modes measured by resonance Raman for bactedorhodopsin (Dexheimer et al., 1992; Polland et al., 1986). No contributions were seen from the photoproduct; the detection wavelength of 656 nm is within the absorption band of K, but no frequencies characteristic of the 13-cis-chromophore were evident. The authors suggest that the absence of such oscillatory behavior might be due to damping of the initial impulsively excited coherent vibrational motion in the excited state as the system rapidly moves out of the Franck-Condon region, and as a result, the largest contribution to the oscillatory response is made by the ground state reactant dynamics (Dexheimer et al., 1992). Picosecond time-resolved fluorescence studies by Atkinson and coworkers have measured fluorescence signals that they have attributed to K (Atkinson et al., 1989), as well as to vibrationally excited ground state species BR' and K' (Atkinson et al., 1991; Blanchard et al., 1991). Kinetic models are proposed to accommodate both the picosecond time-resolved absorption and fluorescence data which incorporate these vibrationally excited species. In light of these studies, it is noted that vibrationally excited species may have an important role in models of the primary evem.

D. Quantum Efficiency of the Primary Event Determination of the quantum efficiency of the primary event has proven to be an additional topic of controversy in bacteriorhodopsin research. An accurate assessment of this fundamental value is of importance not only as a measure of the efficiency of the primary event, but also as an aid in characterizing other basic properties of the protein (such as proton pumping efficiency and energy storage). Two efficiencies have proven to be of interest, namely those of the forward reaction, q)a (bR ~ K), and the reverse reaction, q)2 (K ~ bR). After the initial determination by Oesterhelt and Hess (1973) of 0.79 for the forward reaction, subsequent determinations seem to have fallen into two categories: those around 0.3 (Goldschmidt et al., 1976, 1977; Becher and Ebrey, 1977; Hurley and Ebrey, 1978; Dioumaev et al., 1989), and those around 0.6 (Oesterhelt et al., 1985; Polland et al., 1986; Schneider et al., 1989; Govindjee et al., 1990; Tittor and Oesterhelt, 1990; Balashov et al., 199 lb; Rohr et al., 1992). Most of the recent determinations have favored the latter value. The reverse reaction quantum yield, q02, is generally accepted to be around 1 (Balashov et al., 199 lb; Dioumaev et al., 1989; Govindjee et al., 1990; Xie, 1990). Results of some selected literature assignments are presented in Table 2.

Characterization of the Primary Photochemical Events

73

Table 2. Literature Assignments of the Primary Quantum Yields of Light Adapted Bact eri orh od ops i n I n v e s t i •g a t o r s a

O & H (1973)

~l b

0.79 (0

G O K (1976) B&E (1977)

ci)2c

0.30!-0.03

(I)l/(I)2

(I)l + (I)2

TClK

Reaction a

Conditions e

(>0.79)

--

300

bRiM

--

0.40

--

300

bRc:>K

HS/ether aqueous

0.77_!-0.12

0.39+0.15

1.07+0.15

233

bR<=>M

glycerol aqueous

GKRO (1977)

0.25+0.05

0.63+0.20

0.40+0.1

0.88_+0.2

300

bRc=>K

H&E (1978)

0.33+0.05

0.67_+0.04

0.49_+0.10

1.00+0.09

77

bR<==>K

glycerol

OHT (1985)

>0.6

--

(>0.6)

--

300

bR-->M

HS/ether

300

bR--->K

aqueous

77

bRc:>K

glycerol

300

bR-->L

aqueous

P e t al. (1986)

-0.6

B et al. (1989)

<0.49

--

0.45+0.03

0.67

~

(>0.67)

D et al. (1989)

0.31+0.10

0.93+0.14

T&O (1990)

0.64+0.04

SDS (1989)

Xie (1990)

(>0.6) <1.49

"-0.33

,-,1.24

300

bR--->K

aqueous

(>0.64)

--

300

bR--->M

aqueous

< 0.57

<0.96

0.55+0.02

--1.53

110

bRc:>K

glycerol

GBE (1990)

0.64+0.04

0.94+0.06

0.67+0.06

--1.6

300

bR--->K

aqueous

BIGE (1991)

0.66+0.04

0.93+0.07

0.7+0.02

-1.6

110

bRe:>K

glycerol

R et al. (1992)

0.6+0.05

0.6+0.05

1.0

1.2

300

bR--->K

aqueous

Notes:

alnvestigators are defined as follows: O&H(1973) = (Oesterhelt and Hess, 1973), GOK (1976) = (Goldschmidt, Ottolenghi, and Korenstein, 1976), B&E (1977) = (Becher and Ebrey, 1977), GKRO (1977) = (Goldschmidt, Kalisky, Rosenfeld, and Ottolenghi, 1977), H&E (1978) = (Hurley and Ebrey, 1978), OHT (1985) = (Oesterhelt, Hegemann, and Tittor, 1985), Pet al. (1986) = (Polland et al., 1986), Bet al. (1989) = (Birge, Findsen, Lawrence, Masthay, and Zhang, 1989), SDS (1989) = (Schneider, Diller, and Stockburger, 1989), D et al. (1989) = (Dioumaev et al., 1989), T&O (1990) = (Tittor and Oesterhelt, 1990), Xie (1990) = (Xie, 1990), GBE (1990) = (Govindjee, Balashov, and Ebrey, 1990), BIGE (1991) = (Balashov et al., 1991),, R et al. (1992) = (Rohr et al., 1992) bQuantum yield for the formation of the primary photoproduct, K, from bR. Boldface numbers indicate direct measurements. Some investigators assigned this value by measuring the quantum yield of the bR--->M photoreaction and by assuming that the quantum yield for the bR--+M photoreaction is identical to that for the bR--->Kreaction (i.e., no branching back to bR occurs during the dark steps). CQuantum yield for the formation ofbR from the primary photoproduct, K. Boldface numbers indicate idrect measurements. Some investigators assigned this value by measuring the quantum yield of the M---~bR photoreaction and by assuming that the quantum yield for the M-,bR photoreaction is identical to that for the K--+bR reaction. aThe measurement temperature (in Kelvin) and the photoreaction studied. The symbol "<:>" is used to represent a photostationary state measurement. eSolvent conditions used in the experimental measurement. When specific solvent conditions are not provided, "aqueous" is assumed• HS represents high salt, and "glycerol" conditions are typically mixtures of glycerol and water. Individual references should be consulted for more detailed descriptions of the experimental conditions. fValues shown in boldface were measured directly and the error ranges, when reported, are those provided by the investigators. Remaining values in the same row were derived from the data in boldface.

74

JEFFREYA. STUART and ROBERT R. BIRGE

Experimental measurements of the quantum yield have utilized primarily two methods: direct determinations and photoequilibrium measurements. Photoequilibruim methods involve establishing a photostationary state between bR and K. For this method to be applied successfully, the molar extinction coefficient of K must be known accurately, which must be calculated from a spectrum of pure K. Once this information is known, the mole fraction of K can be determined, and the ratio of the forward and reverse quantum yields can be obtained. Unfortunately it is very difficult to assign an absolute absorption spectrum of K and thereby calculate the extinction coefficient. Direct determinations involve quantitation of the amount of bR that has been phototransformed into cycle intermediates, usually by measuring the accumulation of a certain intermediate. Because of the relatively short lifetime of the K state, and the large amount of spectral overlap between bR and K, a subsequent intermediate is usually chosen. M is the logical choice because it can be easily trapped, and because of its relatively well defined spectrum. The L state has also been used. Which ever state is chosen, the extinction coefficient of that state must be well known. Implicit in the use of a post-K intermediate to determine the bR to K quantum yield is the assumption that formation of the intermediate proceeds linearly (i.e., every K state that is formed goes on to form L or M with no branched pathways or back reactions, rather than decaying back into the initial bR state). In addition, the actinic pulse must be sufficiently short (< 3 ps) so that only the forward reaction (bR -~ K) is allowed to take place. The inadvertent inclusion of back reactions might lead to anomalously low values, if care is not taken in implementing the experiment and/or data analysis. Furthermore, the assumption of a linear reaction scheme precludes the existence of multiple forms of bR and K. Several studies have provided evidence in support of multiple forms, especially at the low temperatures needed to trap K or M (Kalisky and Ottolenghi, 1982; Hanamoto et al., 1984; Rothschild et al., 1985; Dancsh~izy et al., 1988; Diller and Stockburger, 1988; Balashov et al., 1991a, c). It has also been pointed out that neglect of filtering effects and ofphotoselection can lead to errors in quantum yield determination (Nagle et al., 1983; Dioumaiv et al., 1989; Xie, 1990). The most recent assignments have obtained values for the forward quantum efficiency centered around 0.6, and the reverse q~2of slightly less than unity. Using resonance Raman scattering to quantitate accumulation of the L intermediate, Schneider et al. (1989) arrived at a value for q~l of 0.67 (although a paper by Xie claimed that this value was too high due to neglect of the photoselection effect (Xie, 1990)). Tittor and Oesterhelt (1990) determined a value of 0.64 + 0.04 by examining the accumulation of M, and found qo1 to be independent of the actinic wavelength, laser intensity, and ionic strength within the regions probed. There was no pH dependence in the range of pH 5-11, but below 5 a decrease was observed. This observation can be explained in light of the fact that no M intermediate exists at these pH values, so that the authors interpreted the pH dependence as a result of different photocycles due to the acidic forms of the protein (Tittor and Oesterhelt, 1990). Govindjee and coworkers determined a forward quantum efficiency of

Characterization of the Primary Photochemical Events

75

0.64+0.04 based on a comparison to that ofrhodopsin (Govindjee et al., 1990). Xie (1990) obtained a somewhat lower value of q~l < 0.57 from photoequilibrium measurements. However, the most recent photoequilibrium measurements by Balashov et al. (1991 b) yielded a higher value of 0.66+0.04 (which was temperature independent). Finally, using femto- and nanosecond optoacoustic spectroscopy, Rohr et al. (1992) arrived at a value of 0.6 for both q~l and q~2.

E. Photoelectric Studies Photoelectric studies have been of limited usefulness for studying the primary event for several reasons. The time resolution of these is fairly low, and interpretation of the results can be ambiguous. The observed photoelectric signals do not always correspond temporally to the observed changes in the absorption spectrum for the formation of the various intermediates. Additionally, the signals cannot differentiate between different types of charge separation, that is, movement of a charged functional group versus photoinduced redistribution of electron density along the polyene chain. However, several conclusions have been drawn and can provide some insight. All of the studies indicate an ultrafast charge separation that is faster than the instrument response function. Early studies showed this charge separation to take place in less than 100 ps (Trissl, 1985; Trissl and G~,rtner, 1987). Improved instrumentation decreased this value to approximately 20 ps (Groma et al., 1988; Trissl et al., 1989), and most recently the photoinduced charge separation has been shown to occur in under 5 ps (Simmeth and Rayfield, 1990). A time-resolved photoelectric study of both the forward and reverse conversions between bR and K was able to observe light-induced charge separations (of opposite polarity) with time constants of 20 and 30 ps, respectively (Trissl et al., 1989). Keszthelyi has attempted to calculate the displacement distance that corresponds to the observed charge separation to be -0.13 nm (Keszthelyi, 1988). Interestingly enough, the initial photoelectric response is in a direction opposite to proton translocation by bacteriorhodopsin, substantiating the results of deuterium studies which show that this does not play a role in the primary event.

F. Molecular Dynamics of the Primary Event We have discussed a number of models for the primary event in previous sections. Based on the ultrafast pico- and femtosecond spectroscopic studies that have been done on bacteriorhodopsin to date, the following reaction scheme provides the most consistent description of the primary sequence of events: hv

bR(all-trans)--+ bR(FC)* ~0-~s bR(~13)* --~-°-°-~Sj(a3-cis) 7 ~ s K(13-cis) where bR(FC)* represents the Franck-Condon excited state, bR(~13)* represents the excited state with the chromophore distorted out of the Franck-Condon region along the C13--C14 torsional coordinate, J is a 13-cis ground state species, and K is

76

JEFFREYA. STUART and ROBERT R. BIRGE

the first "trappable" ground state species (Atkinson et al., 1986, 1989, 1991; Mathies et al., 1988, 1989, 1991; Pollard et al., 1989; Doig et al., 1991; Du and Fleming, 1993). Thus, the photoisomerization is nearly complete in about 500 fs, but equilibration of the chromophore and protein to form a stable ground state species requires an additional --3 ps. The kinetic studies of Du and Fleming, however, indicate that a portion of the photoactivated species remain in the excited state as much as 13 ps after excitation (Du and Fleming, 1993). This observation supports the proposal by Birge et al. (1987) that some of the molecules are trapped in a potential well in the excited state. Birge and coworkers have carried out molecular dynamics simulations of the primary event in bacteriorhodopsin, and the results of one model are shown in Figure 9 (bottom) (Birge et al., 1987, 1989; Birge, 1990b). A number of different models of the binding site were investigated, and one of the key observations was the high sensitivity of the calculated quantum yields to small changes in the binding site geometry. For example, a small change in the chromophore-counterion separation shifted the calculated forward quantum efficiency from 0.27 to 0.74 (Birge et al., 1989). More recent calculations suggest that our neglect of divalent ions within the binding site may have been responsible, in part, for the instability. Nevertheless, the calculated kinetics appear to be fairly stable to changes in the binding site models, and thus we show the results of Figure 9 primarily to emphasize the examine the kinetics of photoisomerization, and to contrast the dynamics with those calculated for rhodopsin (Figure 9, top). The simulations shown in Figure 9 were generated by using all-valence electron molecular orbital theory with extensive single and double configuration interaction to calculate the adiabatic potential surfaces and semiclassical dynamics to simulate the trajectories. The procedures have been reviewed in the literature and the interested reader is referred to Birge and Hubbard (1980), Birge and Hubbard (1981), and Tallent et al. (1992) for the details. The rationale for using the above approach is the inability of direct dynamic procedures to generate fully correlated surfaces, because analytical derivatives involving doubly excited states are not available. Furthermore, the use of S-matrix theory to calculate the quantum efficiencies can be carried out with excellent precision and nonadiabatic effects can be included explicitly (Tallent et al., 1992). Warshel et al. (1991) have examined the dynamics of the primary event of bacteriorhodopsin from a different perspective, and provide an excellent overview of the different isomerization models. No theoretical model of the primary event of bacteriorhodopsin has provided insight into the quantum efficiencies, however. In contrast, the quantum efficiencies of rhodopsin have been predicted with good accuracy (Tallent et al., 1992). This might appear to be anomalous because we know more about the binding site of bacteriorhodopsin than rhodopsin. However, the counterion environment of bacteriorhodopsin is electrostatically complex and the binding site is extremely ionic, if not charged (Harbison et al., 1984a, 1985; Smith et al., 1987, 1989, 1990; Birge and Zhang, 1990a). In contrast, the binding site ofrhodopsin is neutral and the location

Characterization of the Primary Photochemical Events

--•

77

80

70

~ ~

80

~

40

~ ~

ao

~

o

50

o

20

:ll-cis

-30

R60

II-trLlU

30

•0

Cll=Cl2

60

90

I20

150

180

DIHEDRAL ANGLE (DEGREES)

80 70

o

-

,,o

,~

~

!

i

i

i

60

•

,~ o~ . o

i ii ,

50 40 r~

30 20 .

I0 bR'

0

l-tran,~ 120

160

,,

/

J,K

-"-,,.L/ bR 200

13-cis-7 240

280

320

i

360

C13=C14 DIHEDRAL ANGLE (DEGREES)

Figure 9. Potential energy surfaces and molecular dynamics of the primary photochemical transformations of rhodopsin (top) and light adapted bacteriorhodopsin (bottom) based on semiclassical molecular dynamics theory. Each dot represents a time interval of 10 fs (BR) or 50 fs (rhodopsin), and selected time increments relative to the initial excitation are labelled. The two trajectories which are primarily responsible for repopulating the ground state surface are shown. Adapted from Birge et al., 1987; Birge et al., 1989; and Tallent et al., 1992).

of the counterion is fairly well defined (Birge et al., 1985, 1988; Han et al., 1993). We believe the complexity of the counterion environment surrounding the retinyl chromophore in bacteriorhodopsin is primarily responsible for making the theoretical simulation of the dynamics so difficult.

78

JEFFREYA. STUART and ROBERT R. BIRGE

G. Energy Storage in the Primary Event An accurate assignment of the energy storage associated with the primary event of bR is important to an understanding of the molecular mechanism and the stoichiometry of proton pumping. The early cryogenic photocalorimetric studies ofbacteriorhodopsin estimated that ~ 16 kcal mo1-1 is stored in the K photoproduct, an energy sufficient to pump two protons per photocycle (Birge and Cooper, 1983; Cooper et al., 1984; Birge et al., 1989). However, this enthalpy (AH12) was assigned assuming that the forward (O1) and reverse (02) quantum yields associated with the bR ~ - K photoreaction a r e • 1 -- 0.33 and 01)2 = 0.67 (Becher and Ebrey, 1977). As we have noted in Section III-D, these values appear to have been significantly underestimated. The revised experimental values for • 1 and • 2 prompted Birge and coworkers to reanalyze the photocalorimetric data. The results shown in Figure 10 were generated by treating both • 1 and O 2 as independent variables, and carrying out a weighted least-squares regression to assign AH12 (Birge et al., 1991). The best estimates for • 1 and O 2 appropriate to the experimental conditions of the cryogenic studies are • 1 = 0.65+0.05 and 0I) 2 = 0.95+0.05 (Govindjee et al., 1990; Balashov et al., 199 lb). Evaluation of the results shown in Figure 10 indicates that the primary event stores AH12 = 11.6 + 3.4 kcal mo1-1 = 49 + 14 kJ mo1-1 (01 = 0.65; • 2 = 0.95),--4.4 kcal mo1-1 less than the previous assigmnent of~16 kcal mo1-1. A minimum of~6 kcal mo1-1 of energy storage is required to pump a proton under ambient conditions (Birge et al., 1989). When entropic contributions are included, the revised value of AH12 is not sufficient to pump two protons (Lanyi, 1978; Birge et al, 1989; V~ir6 and Lanyi, 1991b,c). This observation contradicts those reports indicating that two protons are pumped per photocycle. However, these results are consistent with those experimental studies that indicate that the number of protons pumped per photocycle is approximately equal to ~0.6/01 (see discussions in Bogomolni et al., 1980; Govindjee et al., 1980; Marinetti and Mauzerall, 1983; Marinetti, 1987a,b). The revised values of AH12 is consistent with the enthalpy levels of the subsequent intermediates predicted by Varo and Lanyi based on their extensive kinetic studies (see for example Figure 4 in Varo and Lanyi, 199 l a). However, a recent time resolved optoacoustic study of bacteriorhodopsin by Rohr et al. (1992) concluded that the K intermediate "stores" 38 kcal mo1-1 (160 kJ mol-1), a value three times larger than that predicted by the cryogenic photocalorimetry experiments. However, as noted by these authors, it is not clear that the two values should be comparable. First, the energy storage value was measured for the protein a few microseconds following excitation based on an optoacoustical signal. Thus, the K intermediate must transfer all of its excess energy through the protein backbone into the solution during the kinetic window assigned to the K state. It is unlikely that this process will be complete on the microsecond time scale, and thus the optoacoustic value ofAH12 = AEI( = 38 kcal mo1-1 (160 kJ mo1-1) should be viewed as an upper limit to the energy of K relative to bR. (It is an upper limit because heat

Characterization of the Primary Photochemical Events i ........i " ,

.......i ...... ~

".

.

.

.

................ i

.

.

...... ..

% 0.750.7

i

--~:.~

I

". . . . . . . . . . . .

,

-

_7

~ ' C

......... :{ .......

79 i .......... ~....

:

I

[

2 5

.............. a......... 3.5

......i....... :

.

,,_

"'[$.............

o 6,

..... -

0.6 ~

0.3

"

'

5.5

0.4

:___:~ ..17 ,

0.5

0.6

0.7

"

0.8

Figure 10. Contour plot of energy storage in K, the primary photoproduct of lightadapted bacteriorhodopsin. The enthalpy, AH12 = H K - HbR, is plotted as a function of the forward (@I) and reverse (@2)quantum yields associated with the photochemical interconversion of bR and K at 77K based on the experimental data from Birge and. Cooper (I 983); Birge et al. (I 989, 1991 ). The black contours indicate the AH12 values (enthalpies above contours on right). The white lines represent the error contours (AH12 standard deviations in grey on contours). The black dot on the left indicates the previously assigned value of energy storage ( A H 1 2 ~ 16 kcal mol-l"@1, = 0.33, C~2 = 0.67). The black dot at upper right indicates the revised value of energy storage (AH12 = 11.6+3.4 kcal mol-l"@1, = 0.65, @2 = 0.95). (Adapted from Birge et al., 1991 ).

stored in the vibrational manifold of the protein backbone and the protein-solution interface should not be included in the ground state energy of K.) Second, the optoacoustic energy storage value was determined assuming forward and reverse quantum yields of 0.6. As is clear from Figure 10, the values of the two quantum yields have a large impact on the assignment of AH~2. If we a s s u m e (I) 1 -'- (I) 2 -- 0.6, we calculate AHI2 = 16.5 kcal mo1-1 (69 kJ mol -l) based on the cryogenic data shown in Figure 10. Alternative, if we assume cI)1 = 0.6 and (I) 2 -- 0 . 9 , the optoacoustic data predict a value ofAHl2 = 15.5 kcal mo1-1 (65 kJ mo1-1) (see Table 1 ofRohr et al. (1992)). Clearly, a portion of the difference between the cryogenic and optoacoustic enthalpies is associated with the difference in the assumed values

80

JEFFREY A. STUART and ROBERT R. BIRGE

of the quantum yields in the two investigations. Rohr et al. (1992) suggest that their data are more consistent with quantum yields of ~1 = (92 = 0.6. Until there is agreement on the quantum yield values, a more accurate assignment of the energy storage is not possible. Nevertheless, we believe the cryogenic value ofAH12 = 11.6 + 3.4 kcal mo1-1 = 49 + 14 kJ mol-~ represents the maximum energy available to pump the proton. What remains to be explained is the origin of the energy storage, that is, the fraction that is associated with charge separation versus the fraction associated with conformational distortion. No accurate assignment is possible, but a comparison to energy storage in rhodopsin suggests that roughly one-half is due to charge separation and the other half is associated with conformational distortion (Birge and Cooper, 1983).

IV. THE PRIMARY EVENT IN RHODOPSIN Like bacteriorhodopsin, the primary event in rhodopsin is an isomerization of the retinal chromophore. Unlike bacteriorhodopsin, this isomerization is cis to trans; 11-cis retinal isomerizes to the all-trans photoproduct (see Figure 6). Also unlike bacteriorhodopsin, in most of the visual rhodopsins absorption of light initiates a bleaching sequence, the ultimate result of which is stimulation of the optic nerve and the consequent expulsion of retinal from the binding site. The exceptions to this are the invertebrate rhodopsins, which typically display a photocyclic response to light absorption. The nature of the primary event was for some time an issue of debate. Most studies indicated that the above-mentioned isomerization was the sole response to light, while other authors argued for a concomitant proton transfer. Although most researchers agree that isomerization plays a role, others argue whether it is mandatory for activation of the protein (see below). Hubbard and Kroft were among the first to suggest that an 11-cis to trans isomerization is central to the primary event (Hubbard and Kropf, 1958). Yoshizawa and Wald supported this view with their models (Yoshizawa and Wald, 1963). Confirmation of the ll-cis to all-trans isomerization comes from many sources, including resonance Raman studies (Palings et al., 1989; Deng et al., 199 l a), chemical extraction data (Hubbard and Kropf, 1958; Wald, 1968; Rosenfeld et al., 1977a), chromophore analog studies (Albeck et al., 1989; Kandori et al., 1989b; Mizukami et al., 1993), and FTIR (Bagley et al., 1985, 1989; DeGrip et al., 1988). The observation that bathorhodopsin is generated in a few picoseconds prompted the suggestion that the primary event involved a proton translocation rather than an isomerization (Peters et al., 1977). This suggestion was based in part on the assumption that an isomerization involving a large change in conformation would take 100-1,000 times longer, but subsequent molecular dynamics calculations predicted picosecond isomerization times (Birge and Hubbard, 1980, 1981). The controversy has been resolved by a series of spectroscopic studies which indicate

Characterization of the Primary Photochemical Events

81

that the chromophore in bathorhodopsin is an all-trans protonated Schiffbase and that the primary photoproducts formed from isorhodopsin (9-cis) and rhodopsin (11-cis) are identical; this observation is perhaps the most compelling of all (Oseroff and Callender, 1974; Callender et al., 1976; Callender and Honig, 1977; Green et al., 1977; Hurley et al., 1977; Rosenfeld et al., 1977b; Honig et al., 1979; Monger et al., 1979; Mao et al., 1980, 1981; Tsuda et al., 1980; Suzuki and Callender, 1981; Mathies, 1982; Klinger et al., 1984; Pande et al., 1984; Bagley et al., 1985; Doukas et al., 1985; Eyring and Mathies, 1979; Birge and Callender, 1987; Deng and Callender, 1987; Palings et al., 1987). In addition, based on an analysis of the results from picosecond spectroscopic studies using Englman-Jortner theory of radiationless transitions, Milder and Kliger demonstrated that proton translocation was not necessary to explain the kinetics of the primary event; their results were consistent with a mechanism that involved a simple chromophore isomerization in the formation of bathorhodopsin (Milder and Kliger, 1986). Thus, the primary photochemical event in vertebrate rhodopsin involves an 11-cis to all-transphotoisomerization, as shown in Figure 6. Recent observations by Foster and coworkers indicate that Chlarnydomonas reinhardtii rhodopsin can be activated even when "non-isomerizable" chromophores are incorporated into the binding site (Foster and Saranak, 1988; Foster et al., 1988, 1989). However, comparisons with vertebrate rhodopsins must be made with care; the binding site of Chlamydornonas rhodopsin is quite different from that of vertebrate rhodopsin (see Birge (1990a) and the references cited therein). Furthermore, it is possible that Chlamydomonas rhodopsin may be activated via photoisomerization at the Schiffbase linkage. This question has yet to be resolved. Finally, the observations of Longstaff and Rando (Longstaff et al., 1986, 1987; Govindjee et al., 1988) indicate that deprotonation of the chromophore is required for activation of rhodopsin. Because deprotonation occurs upon the formation of metarhodopsin II, which is formed only after isomerization of the chromophore, it follows that chromophore isomerization is prerequisite for vertebrate activation. This observation, however, does not preclude the possibility that a chromophore analog might be found that deprotonates upon excitation, and that deprotonation of the chromophore may be sufficient to activate rhodopsin. To date, no such analog for vertebrate rhodopsin has been found, however. In contrast, rhodopsin from Chlamydomonas reinhardtii appears to be activated by virtually all chromophore analogs that form protonated Schiffbases within the binding site (Foster and Saranak, 1988; Foster et al., 1988, 1989). One possibility in this particular case is that the chromophore binding site is sufficiently unstable such that its deprotonation (and perhaps expulsion) upon excitation can occur with, or without, concomitant isomerization. Regardless of the explanation, it is concluded that the binding site and the activation mechanism of Chlarnydomonas reinhardtii rhodopsin is distinctly different from that of vertebrate rhodopsin. Further work on this system is important, however, for it is often through comparison and contrast of the differences between two nominally similar systems that new insights are gained.

82

JEFFREY A. STUART and ROBERT R. BIRGE

The identity of the primary photoproduct is still an area of contention. Three candidates emerge as possibilities: photorhodopsin (~max -~ 570--580 nln), hypsorhodopsin (~max ~ 435 nm) and bathorhodopsin (~max ~ 530 nlTl). It is now generally accepted that hypsorhodopsin is an artifact generated via multiphoton processes, and plays no true physiological role (Yoshizawa et al., 1984; Shichida, 1986; Kandori et al., 1989a; Yoshizawa and Kandori, 1991). It converts to bathorhodopsin above 23 K, and recent FTIR studies indicated a protonated Schiff base, a twisted all-trans chromophore conformation, and chromophore-protein interactions that make it more similar to bathorhodopsin than to rhodopsin or isorhodopsin (Yoshizawa et al., 1984; Sasaki et al., 1992). The spectra ofrhodopsin and bathorhodopsin are shown in Figure 7.

A. Photorhodopsin It is now generally accepted that photorhodopsin is the primary photoproduct of light absorption by rhodopsin, despite the fact that it cannot be trapped atlow (liquid helium) temperatures. It is therefore unclear ifphotorhodopsin should be treated as a distinct intermediate (which implies a potential energy minimum along the reaction surface) or a spectral event characteristic of a slow relaxation process into bathorhodopsin. However, an apparent precursor to bathorhodopsin at 4 K has been reported (Peters et al., 1977), and was denoted as PBATHO (Dinur et al., 1981). Whether or not this precursor is analogous to photorhodopsin is unclear (Ottolenghi and Sheves, 1989). Photorhodopsin was first reported by Shichida et al. as a result of room temperature picosecond photolysis experiments (Shichida, 1986; Shichida et al., 1984). Most of the earlier studies regarded photorhodopsin as an excited state ofrhodopsin, until Honig et al. (1979) described it as a ground state species. Kandori and coworkers reported the absolute absorption spectrum ofphotorhodopsin at room temperature, with a ~max at 570 nm and an oscillator strength smaller than that of rhodopsin and bathorhodopsin (Kandori et al., 1989c). The question naturally arises as to whether or not the actual primary event occurs concurrent to the formation of photorhodopsin, i.e. has the chromophore isomerized. This question has been addressed via several different techniques, including locked chromophore analogues and ultrafast laser spectroscopy (see below). Experiments incorporating 11-cis-locked chromophore analogues into rhodopsin have proven to be very insightful. Nakanishi and coworkers have done a series of experiments fixing the 11-ene retinal with five-, seven-, and eight-membered rings (Kandori et al., 1989b; Mizukami et al., 1993; Hu et al., 1994b). Excitation of the five-membered ring analog (Rh5) resulted in a long-lived excited singlet state with a fluorescence decay time constant of 85 ps; the longevity of the excited state was undoubtedly due to the total inhibition of isomerization (Kandori et al., 1989b). The seven-membered ring analog exhibited properties consistent with hindered isomerization, forming two intermediates; the first transient species absorbed at 580 nm and converted to a species with )~max= 630 on absorption of a second photon (Kandori et al., 1989b).

Characterization of the Primary Photochemical Events

83

The latter species thermally reverted to the original ground state pigment within 44 ps (Kandori et al., 1989b). There are definite similarities between the first transient species of the analog pigment (denoted as Rh7 (580) by the authors) and photorhodopsin, including the absorption maximum and decay time (20--40 ps vs. 45 ps for photorhodopsin) (Kandori et al., 1989b). Because the only difference between Rh5 and Rh7 is the flexibility of the ring locking the 11-ene bond of the chromophore, the photochemical intermediates observed in Rh7 (Rh7(580) and Rh7(630)) must be due to different degrees of isomerization away from the 11-cis configuration. Based on comparison with the results of isomerization in locked model compounds, the authors concluded that the chromophore in Rh7(580) is in a strained trans-like configuration, which eventually relaxes back into the comparatively more relaxed 11-cis form (as opposed to photorhodopsin which is thermally driven to bathorhodopsin) (Kandori et al., 1989b). More recently a pigment was produced utilizing an eight-membered ring to fix the C: :-C12 double bond of retinal (Mizukami et al., 1993; Hu et al., 1994b). These studies further demonstrated the posit that the appearance of early intermediates is a function of the flexibility of the retinal Cll~Cl2 bond. In the first study, excitation of the synthetic pigment (Rh8) resulted in the formation of photorhodopsin-like ()~m~x= 585 nm) and bathorhodopsin-like intermediates (~max = 577 nm), with a consequent relaxation back into the original pre-excited state after 50 ns (Mizukami et al., 1993). If the photorhodopsin-like intermediate produced by the pigment Rh7 has a strained trans-like configuration, it then follows that the Rh8 pigment should have a more relaxed and less distorted all-trans chromophore configuration, allowing the subsequent formation of the batho-like state (Mizukami et al., 1993). In the study that followed, a methylated opsin was used to yield MeRh8 which displayed higher stability than Rh8, while retaining properties very similar to the native pigment (Hu et al., 1994b). This study confirmed the formation of photo-like (within 15 ps) and batho-like (1 ns) intermediates, neither of which could be isolated at low temperatures (as is the case with native photorhodopsin). The inability ofbatho-Rh8 to be isolated at low temperatures (unlike native batho-Rh) is probably a reflection of ring strain which disallows a stable trans configuration. In addition, a series of dihydroretinals was used to form pigments in order to evaluate the role of charge redistribution in the excited state (with respect to the primary event and visual transduction); an increased loss in activity Was seen as the location of the reduced bond was moved closer to the C:I-C:2 double bond. This excited state polarization, also referred to as the sudden polarization model (Salem and Bruckmann, 1975; Albert and Ramasesha, 1990), assumes a sudden redistribution of charge as a function of the isomerization. This polarization of the chromophore (concurrent with isomerization) has often been theorized to be responsible for triggering the process of visual transduction. None of the synthetic pigments examined in these studies were capable of activity, as measured by phosphodiesterase activation and kinase phosphorylation. A charge redistribution in the excited state due to isomerization was seen, however, as demonstrated by the dihydroretinal studies. It is therefore evident

84

JEFFREYA. STUART and ROBERT R. BIRGE

that although there is an excited state polarization as a result of isomerization induced charge redistribution, this alone is insufficient, and it is the complete isomerization that is of prime importance for the initiation of visual transduction (Chen et al., 1992; Hu et al., 1994b). Attempts to incorporate an ll-ene retinal locked with a nine-membered ring were unsuccessful (Hu et al., 1994b). Based on the absolute absorption spectrum, the photorhodopsin oscillator strength is smaller than those ofrhodopsin and bathorhodopsin. This fact also lends support to the assumption of a distorted all-trans configuration for retinal, in that the oscillator strength is sensitive to deviations from planarity in the polyene chain (Sperling, 1972; Kandori et al., 1989c); this is consistent with the results of the cis-locked analog studies discussed above. Accordingly, photorhodopsin must be considered as a ground state species as it is defined on the basis of the state of the chromophore. Kandori and coworkers describe the species as an intermediate state in isomerization of the chromophore (Kandori et al., 1989c). In addition, the lifetime of photorhodopsin is much longer than that published for the rhodopsin excited state, which is reportedly in the subpicosecond range based on measurements of the fluorescence quantum yield (Doukas et al., 1984; Kandori et al., 1989a).

B. Bathorhodopsin Bathorhodopsin (formerly denoted as prelumirhodopsin) was first observed at low temperature in bovine rhodopsin by Yoshizawa and Kito in 1958 (Yoshizawa et al., 1984), and until recently, it was considered to be the primary photoproduct of light absorption by rhodopsin. It remains the first which can be trapped at low temperatures, and as such is still considered to be the primary photoproduct by some. Furthermore, it is in this intermediate that the energy of the absorbed photon is stored, and is 33 kcal/mol higher in energy than rhodopsin (see below). Bathorhodopsin (bovine) forms from rhodopsin (or isorhodopsin) via photorhodopsin in approximately 40-45 ps (Shichida et al., 1984; Kandori et al., 1989a), and as demonstrated by the locked 11-cis retinal analog studies described above (especially those with seven- and eight-membered tings) the cis to trans isomerization is a prerequisite to its formation. Further support for the mandatory cis to trans isomerization comes from circular dichroism (Kakitani et al., 1977; Shichida et al., 1978; Horiuchi et al., 1980), the change in the angle of the transition dipole on absorption of light (of26 °) (Kawamura et al., 1979), and the observation that both 7-cis- and 9-cis-rhodopsin (isorhodopsin) yield the same photoproduct as 11-cisrhodopsin, which can only be accomplished via an isomerization (Kawamura et al., 1980; Kliger et al., 1984). In squid and octopus rhodopsins batho is formed in approximately 200 ps (Ottolenghi and Sheves, 1989). The mechanism of the isomerization, and whether other single bonds are involved will be discussed below. There is some evidence for the existence of multiple spectrally and kinetically distinct forms of bathorhodopsin; these will not be considered herein as they are

Characterization of the Primary Photochemical Events

85

probably artifacts of experimental conditions and as such play no relevant physiological role (Einterz et al., 1987a,b; Sasaki et al., 1980a,b). It has been suggested, however, that clockwise and counterclockwise isomerizations of the chromophore might be responsible for the existence of two interconvertible batho forms (Sasaki et al., 1980b). It is not out of the question that there might even be multiple forms of rhodopsin due to the flexible nature of proteins in general, but again, there is no evidence that these play any relevant physiological role. Because bathorhodopsin can be trapped at low temperatures, much more detailed information is available regarding its chromophore configuration and interactions. FTIR studies by Bagley et al. (1985) concluded that retinal in batho is a distorted all-trans, the Schiff base is protonated, and that the C = N has a trans (anti) conformation. Furthermore, the C--N stretching frequency is invariant between rhodopsin, isorhodopsin, and bathorhodopsin, and the location of the C-10/C-11 stretch in batho indicates a 10-s-trans conformation, providing additional support for an 11-cis to all trans isomerization as the primary event (Bagley et al., 1985). These findings have been confirmed in toto via resonance Raman studies (Deng and Callender, 1987; Palings et al., 1987), and 13C NMR studies (Smith et al., 1991). In addition to the C10-Cll single bond, the C14-C15 single bond was also found to be in an s-trans configuration, such that any model for the primary event that employs a "concerted twist" (as that proposed in Liu and Asato (1985)) is incompatible with the available data (Palings et al., 1987). The invariance of the C = N stretching frequency in rhodopsin, isorhodopsin, and bathorhodopsin not only indicates that there is no C--N isomerization in the primary event, but it also fails to lend support to the idea that there is any significant change in the electrostatic environment of the protonated Schiff base during the primary event, which raises questions with regard to the nature of the associated mechanism for energy storage (to be discussed below). The only perturbation in this vicinity supported by the resonance Raman data is a weakened hydrogen bond to the Schiffbase upon going to from rhodopsin to bathorhodopsin (Palings et al., 1987). These latter observations were confirmed by Deng and Callender (1987) in their resonance Raman studies, thereby showing that energy storage models which rely heavily on charge separation of the protonated Schiff base and its counterion are largely incorrect (Deng and Callender, 1987). This being the case, the energy storage must be accomplished via other protein-chromophore interactions, including ionic perturbations to the polyene chain and/or distortions in the chromophore configuration. Support for such perturbations was provided by Palings et al. (1989) in a study which assigned the HOOP wagging vibrations in the batho chromophore. Based upon comparison to model compounds, the force constants for the C10H, C12H, and C14H wags were found to be lower than expected (Palings et al., 1989). In addition, the C-C stretching force constants in the C10-Cll--C12-C13 vicinity differed from those of the all-trans protonated Schiff base in solution, especially at C12; this is consistent with a protein induced charge perturbation in this region, and may account for the reduced force constants observed for the HOOP wagging vibrations

86

JEFFREYA. STUART and ROBERT R. BIRGE

described above (due to distortions in the ~- and n-bonding systems), as well as the reduced C12-C13 stretching vibration (Palings et al., 1989). Indeed, there is strong evidence for such an interaction in rhodopsin at C12 and C13 (see section above on rhodopsin active site) (Mollevanger et al., 1987; Lugtenburg et al., 1988; Smith et al., 1990). And in an earlier study by Palings et al. (1987) the C-C stretching modes in bathorhodopsin (especially at C8--C 9 and C14--C15) were found to be significantly different from that of all-trans retinal protonated Schiffbase in methanol, in strong contrast to the results from rhodopsin, which showed largely unperturbed C-C stretching frequencies as compared to l l-cis-retinal protonated Schiff base in methanol (Palings et al., 1987). Also in comrast to the results discussed above, this study concluded that there was no charge perturbation at Cl3 in rhodopsin (Palings et al., 1987). The question of the location of the polyene charge perturbation is discussed further below. Resonance Raman studies have also indicated that the protonated Schiff base is strongly hydrogen bonded in both rhodopsin and bathorhodopsin. On deuteration, an unusually large shift of 33 cm-1 is seen for the protonated Schiff base band in bovine bathorhodopsin, which is indicative of a particularly strong hydrogen bond when compared to the -24 crn-1 shift seen for a completely solvated retinal protonated Schiff base in methanol (Deng et al., 199 la). Circular and linear dichroism studies have been performed on both rhodopsin and bathorhodopsin, and support the conclusion of a conformationally strained and distorted chromophore in the latter species. Linear dichroism studies have demonstrated that the chromophore rotates out of the plane of the membrane upon formation of bathorhodopsin, based upon determination of the change in the transition dipole moment (Lewis and Kliger, 1992). Rhodopsin displays a positive circular dichroism (CD) band while that of bathorhodopsin is large and negative (Yoshizawa et al., 1984). Studies on a rhodopsin analog protein incorporating 11-cis-retinal locked with a 5-membered ring showed no CD activity, which indicates that the rhodopsin CD band must be due to actual distortions in the chromophore (Yoshizawa et al., 1984). Therefore, the change in the CD signal due to the primary event must be the result of structural changes in the chromophore. The observation of a large negative signal for bathorhodopsin could be explained by large enough twisting in the chromophore such that the double bonds of the polyene follow the path of a left-handed helix (Lewis and Kliger, 1992). Indeed, Palings et al. (1987) offered this as a tentative explanation to accoum for the results of their Raman studies; a twist of 20 ° around the C10--Cll, C12-C13, and C14--C15 bonds would allow for a minimal spatial difference to result from the isomerization in the primary event, with little change in the Schiff base region. In their model, the positions of the ionone ring and the a-carbon of lysine were fixed during the cis-to-trans isomerization. This model is appealing in that the current evidence seems to support little change in the binding site during the primary event. Several comparative studies between bovine and octopus bathorhodopsins have been made. Both rhodopsins share the 11-cis-to-trans isomerization as the primary

Characterization of the Primary Photochemical Events

87

event, although the active site geometries have many differences (Bagley et al., 1989). Despite this fact, FTIR (Bagley et al., 1989) and Raman studies (Deng et al., 1991a) show many similarities in the fingerprint region (800-1700 cm-1), including a protonated Schiff base and similar configurations for the all-trans chromophores. The Schiff base C--N configuration is trans (anti) (Deng et al., 199 lb), and hydrogen bonding between the protonated Schiffbase and the primary counterion appears to be weaker in octopus bathorhodopsin, as evidenced by a smaller shift (25 crn-1) of the protonated Schiff base Raman band (Eyring and Mathies, 1979; Deng and Callender, 1987; Palings et al., 1987; Deng et al., 1991a). Resonance Raman studies by Deng and coworkers indicated a torsionally distorted chromophore with a twist of 4.7 ° around the Cl1=C12 as determined by an analysis of HOOP mode intensities (Deng et al., 1991a). For the most part, the HOOP modes of octopus bathorhodopsin were found to be very similar to those of the chromophore in solution. Unlike bovine bathorhodopsin, the CllH and Cl2H wags in octopus bathorhodopsin were found to be coupled into a single Cll=Cl2 HOOP Au-like mode, as is found in all-trans-retinal. In bovine bathorhodopsin this mode is decoupled into two separate HOOP bands; a larger twist about the Cll=C~2 bond might be responsible for this observation. Such an effect might be caused by a negative amino acid residue in the vicinity of Cl2, perhaps Glu-113 or- 122 (Deng et al., 1991a). As discussed earlier, Glu- 113 is thought to be the primary counterion to the protonated Schiffbase (see above). Neither of the aforementioned glutamic acid residues are conserved in the case of the octopus pigment, therefore, the absence of the decoupling between the CI~H and C12H wags might well be due to a much weaker, or absent, negative perturbation at C12 (Deng et al., 199 la). A recent 13C NMR study by Smith et al. (1991) has provided more insight into the problem of the extent to which the purported ionic perturbation at C13 interacts with the retinal polyene chain in both bathorhodopsin and rhodopsin. A series of synthetic retinals was produced with specific ~3C labels at positions 8, 10, 11, 12, 13, 14, and 15, which were regenerated with bovine opsin. On comparison to an all-trans protonated Schiffbase chloride salt, the largest chemical shift (6.2 ppm) in bathorhodopsin was found at 13C; as chemical shift is a direct reflection of electron density, this result indicates that there is indeed a charge perturbation in the vicinity of 13C. Chemical shifts at positions 10, 11, and 12 showed slight differences as well. Similar experiments previously done for rhodopsin and isorhodopsin, and the 11-cis protonated Schiffbase chloride salt, yielded nearly identical results, demonstrating that there is little significant change in the location of the ionic perturbation with respect to the chromophore upon isomerization (Smith et al., 1991). If such a change did occur, one would anticipate changes in the corresponding chemical shifts of the ~3C resonances along the polyene chain, especially at positions 13 and 15; only small differences were found (< 2.5 ppm), corresponding to changes of no more than 0.015 electron equivalents (Smith et al., 1991). A change of at least an order of magnitude larger is expected if substantial differences exist along the retinal polyene chain before and after isomerization. In

88

JEFFREYA. STUART and ROBERT R. BIRGE

conjunction with the results of the Raman studies discussed above, it would seem that the hydrogen bonding and ionic environment of the rhodopsin active site is little changed during the primary event (Bagley et al., 1985; Palings et al., 1987, 1989; Deng et al., 1991 a; Smith et al., 1991). In addition, Ganter et al. (1988b) observed little or no change in IR absorptions due to water during the phototransition. It is only on the formation of the lumirhodopsin intermediate that large changes in the interactions between the protein and chromophore are seen (Ganter et al., 1988a).

C. Ultrafast Spectroscopic Studies The time-course dynamics of the primary event has been a topic of considerable interest. As in the study ofbacteriorhodopsin, the advent of femtosecond spectroscopy has yielded a great deal of insight into the nature of the transitions between the initial photoproducts. Early studies of the primary event indicated that bathorhodopsin was formed as the first intermediate in less than 6 ps at room temperature (Busch et al., 1972; Green et al., 1977). Fluorescence quantum yield studies suggested that chromophore isomerization was likely to occur on a subpicosecond time-scale, and must be the result of a barrier-less excited state potential energy curve (Doukas et al., 1984). The results of recent femtosecond studies have shown that the primary photochemical event in vision occurs much more quickly than initially believed possible, and have provided the most complete picture of the primary event to date. As such, only these studies will be considered herein. Schoenlein et al. (1991) were the first to implement femtosecond studies of the rhodopsin photochemistry. Using a 35 fs pump pulse at 500 nm with a 10 fs probe pulse, the primary event was observed between 450 and 580 nm. Transient changes in absorption were monitored at 500, 535, 550, and 570 nm. The initial bleach and subsequent partial recovery of ground state 11-cis-rhodopsin was observed by the 500 nm probe. An absorption at about 500 nm was assigned to the S 1-~S n transition, which interfered with the ground state bleach signal. The changes monitored at 550 and 570 nm elucidated the nature of photoproduct formation. A rapid absorbance developed at 570 nm, reaching its maximum by 200 fs, beyond which little change was observed. Only a slight decay of the signal was found in subsequent measurements out to 6 ps (Schoenlein et al., 1991). Measurements at 535 nm were found to be a bit more complex, exhibiting contributions from the excited state (S 1-~Sn, 0-100 fs) and photoproduct absorption bands appearing between 100 and 200 fs. These results were interpreted as indicating that complete isomerization, and therefore formation of the primary photoproduct, had taken place by 200 fs (Schoenlein et al., 1991). Indeed the absorption maximum of 570 nm is close to that measured for photorhodopsin (~max ~ 570-580 nm). The authors point out that this data confirms earlier suggestions that a rapid non-radiative isomerization is needed to account for the high photosensitivity and quantum yield of rhodopsin (Schoenlein et al., 1991). No stimulated emission was seen from the excited state

Characterization of the Primary Photochemical Events

89

absorption, suggesting that the torsional wavepacket created in the first excited state moves rapidly away from the Franck-Condon region. The observation of oscillatory behavior between 100-200 fs at the four wavelengths probed demonstrates excitation of nonstationary vibrational states; the frequency of these vibrations (--135 cm-1) is characteristic of low-frequency modes observed in rhodopsin and may play a role in modulating the formation of photoproduct (Schoenlein et al., 1991). As discussed above, similar behavior has been observed for bacteriorhodopsin photochemistry (Dexheimer et al., 1992). The study concluded that the primary event in rhodopsin is characteristic of an essentially barrierless transition that proceeds along a nonadiabatic potential surface (see below); such a transition is expected from a vibrationally coherent system and indicates strong coupling between the rhodopsin excited-state and the photoproduct ground-state (photorhodopsin) (Schoenlein et al., 1991). Yan et al. (1991) arrived at different conclusions based upon the results of their femtosecond studies. In their studies, they reported the formation ofbathorhodopsin in 3.0+0.7 ps, preceded by two intermediate species: the first was identified as the vertically excited Franck-Condon state which decayed in--200 fs to a second state, which in turn decayed to bathorhodopsin in 3.0 ps. The second intermediate (denoted as Rh*(90°)) was attributed to an excited-state with a 90 ° twist in the C11-C12 bond of the retinal chromophore (Yen et al., 1991). Subsequent measurements up to 100 ps showed no further changes, and the authors identified the final state as bathorhodopsin, and implied that the Rh*(90 °) state corresponds to photorhodopsin (Yen et al., 1991). The following scheme was proposed:

Rh(ll-cis) -~Rh*(FC) ~=---> 200 fs Rh*(90 °) ~2_~s batho(trans) The rational for attributing the second intermediate (Rh*(90°)) to an excited transient near 90 ° is based on good agreement with molecular dynamic calculations by Birge and coworkers, who predicted the formation of such a state in 375 fs, followed by batho formation in 1.4 ps (Birge and Hubbard, 1980, 1981; Birge, 1990b). In addition, no deuterium effect was found for the primary rhodopsin photochemistry (Yen et al., 1991). In a subsequent study by the same authors (Yan et al., 1993), a wider spectral range was used in order to address the results obtained by Schoenlein and coworkers (Schoenlein et al., 1991). Over the range examined (570-650 nm) the results of the previous study were confirmed, namely that a rapid 200+100 fs decay was followed by a 3.0+0.7 ps decay to a long lived species that remained for at least 100 ps (Yen et al., 1993). The interpretation posited in the scheme above was maintained in this study, and the same rate equations used to fit their data were applied to that obtained by Schoenlein and coworkers, yielding identical rate constants (200 fs and 3 ps) (Yen et al., 1993). The discrepancy in interpretation is ascribed to the 200 fs rise seen at 550 nm. In the model by Yan et al. (1993) 550 nm represents an isosbestic point between Rh*(90 °) and its decay products (approximately 2/3 batho and 1/3

90

JEFFREYA. STUART and ROBERT R. BIRGE

rhodopsin, based upon a quantum yield of 0.67). The transient species Rh*(90 °) absorbs more strongly at wavelengths greater than 570 n m (~max~ 620 nm): a region not examined by Schoenlein et al. (1991) in their initial work. Therefore, the decay of the signal due to Rh*(90 °) would not be totally evident at wavelengths less than 550 nm; only the accumulation ofphotoproducts would be seen (2/3 batho and 1/3 rhodopsin) (Yan et al., 1993). According to Yan et al. (1993) it was this reason, that is, the absence of the signal due to Rh*(90°), that lead Schoenlein and coworkers to their incomplete conclusions (Schoenlein et al., 1991; Yan et al., 1993). In a follow-up study, Peteanu and collaborators reaffirmed the results of Schoenlein's work by examining the spectral range from 490 to 670 nm (Peteanu et al., 1993). Their data is presemed in Figure 11. As the earlier study showed, the completely isomerized photoproduct (photorhodopsin) is formed by 200 fs; spectral features corresponding to excited-state absorption, reactant ground-state bleaching, photoproduct formation and relaxation, and subsequent recovery processes are all clearly discernible. The following events were apparent (Figure 11) (Peteanu et al., 1993): The excited-state absorption appeared centered at 510 nm, and was prevalent at times less than 50 fs. It was immediately followed by the ground-state bleach centered at the same location (510 nm), which reached its largest magnitude by 150 fs. The subsequent formation ofphotoproduct (575 nm) attained its maximum amplitude by 200 fs, which remained wholly unchanged for the next 6 ps, aside from a 10 nm blue shift of the absorbance maximum (to 565 nm); however, the integrated area of the absorption during the blue shift remained the same, implying only one chemical species (Peteanu et al., 1993). The authors also note that they observed a shift in the isosbestic point associated with the decay of the bleach signal, from approximately 535 to 515 nm at 6 ps. Changes that occurred between 200 fs and 6 ps were analyzed in terms of a combination of dynamic ground-state processes, including intramolecular vibrational energy redistribution, vibrational cooling, and conformational relaxation (Peteanu et al., 1993). The isomerization was again described as a vibrationally coherent process which proceeds along a nonadiabatic potential surface that allows for strong coupling between the excited-state of the reactant with the grotmd state of the product (Schoenlein et al., 1991, 1993; Peteanu et al., 1993). In the interpretation of this data, as well as that presented by Yan et al. (1993), of prime importance is the correct identification of the species responsible for the observed spectral and kinetic properties. Similar data is seen by both groups, yet interpretation differs. The 200 fs decay time is attributed to the transition between the Franck-Condon state and Rh*(90 °) by Yan et al. (1993) while Schoenlein and Peteanu ascribe this to the formation time for the primary photoproduct. The interpretation of the latter authors (Peteanu et al., 1993) seems to be the more consistent with their data, and is summarized as follows: upon absorption of a photon, the chromophore is immediately excited into the Franck-Condon region of the potential energy surface, whereby the wavepacket can be further excited into the S1---~Sn transition (510 nm). The wavepacket then rapidly proceeds to evolve

Characterization of the Primary Photochemical Events •

-150

.

1

.A

.4~

............. L-- i-i- --:---:-=

"

•

I

;?i

?~?. 2 ................

I

.

l

~

,~,~ . . . . . . . .................................. i

I - -

91

.,...~

....

I

-A T/T=2% .,

I

__ I

"~

I

..

I

-.

,

.

100 150 v

>,

a

200 300 400

3000 6000

~

~

I

I

I

I

I

I

I

500

550

600

650

Wavelength (nm)

Figure 11. Differential absorption spectra of rhodopsin at various time delays of the probe pulse after excitation with a 35-fs pump pulse at 500 nm. The breaks in the curves near 570 nm indicate where the spectra from blue (490-570 nm) and red (570-670 nm) 10 fs probe pulses were joined. (Reproduced with permission from Peteanu et al., 1993).

along the potential energy surface away from the Franck-Condon region, as evidenced by the disappearance of the S 1--->Sn absorption feature by 100 fs, at which point a strong bleach is found in the same location. Indeed, resonance Raman intensity analysis supports torsional departure from the Franck-Condon region on this time-scale (Loppnow et al., 1992). It should be noted that there was very little absorption in the red spectral region due to the SI-~S n transition. As the bleach evolves, a concomitant photoproduct absorption band appears between 550 and 650 nm reaches its maximum by 200 fs, and remains largely unchanged for subsequent measurements out to 6 ps. The ~maxof this feature is about 570 nm and is consistent with that of what is widely recognized as the primary photoproduct, photorhodopsin. The following changes were seen between 200 fs and 6 ps (Peteanu et al., 1993):

92

JEFFREYA. STUART and ROBERT R. BIRGE

(i) A slight loss in intensity was seen at the red edge of the photoproduct band, between 610 and 630 nm, as well as a shift of the photoproduct maximum from 575-565 nm, and a concomitant increase in absorption at 535 nm. (ii) The isosbestic point shifted from 540 nm at 200 fs to 520 nm at 6 ps, and (iii), after 200 fs the hole due to the bleaching of the reactant had recovered. The gain in absorbance at 535 nm is consistent with the formation of bathorhodopsin. These results contradict those of the Yan and coworkers, and are inconsistent with their interpretation (see above) for the following reasons (according to Peteanu et al. (1993)). The absorption at 620 nm seen in their study (which appeared within the 300 fs time resolution of their experiment) (Yen et al., 1991, 1993) is apparently part of a larger absorption band due to the photoproduct, which is not fully developed until 200 fs; the time resolution of their work did not allow for the observation of the evolution of this signal. Moreover, the signal in the 580--620 nm range purportedly due to an excited state with a 90 ° twist along the polyene chain is expected to be associated with the evolution of a spectral signal, which is not observed between 510-600 nm (Peteanu et al., 1993). The last reason cited by Peteanu and coworkers for the incongruence in interpretation is ascribed to the observation that there was no notable signal evolution seen in their data between 200 fs and 6 ps; one would expect some spectral signature attributable to the 3 ps decay of Rh*(90 °) into the more planar all-trans ground state (Schoenlein et al., 1991; Peteanu et al., 1993). As pointed out in their work, the theoretical predictions of Tallent et al. (1992) indicated observable spectral differences in oscillator strength, bandwidth, and absorption maximum between the twisted excited-state species and the ground-state photoproduct. Photochemical hole-burning studies on both rhodopsin and bacteriorhodopsin by Loppnow and collaborators lend further support for the observed ultrafast isomerization times, based upon the observation of extremely broad homogeneous line widths (1300 cm-1 at 1.5 K) for both proteins; the absence of narrow components in spectra burned and probed at 2 cm-1 resolution is consistent with chromophore isomerization times of approximately 200-500 fs on the excited-state potential surface (Loppnow et al., 1992). Furthermore, the broad and diffuse nature of the holes is consistent with the --25 fs optical relaxation times predicted from resonance Raman intensity analyses (Loppnow et al., 1992). In a recent study on the primary event in octopus rhodopsin, the photoproduct, and therefore the isomerization, was reported to have formed within 400 fs after leaving the excited-state (Taiji et al., 1992). Femtosecond spectroscopy with a range of400-1000 nm revealed an induced absorption between 680-730 nm and640--660 nm that appeared within 100 fs and decayed with time constants of 140+70 fs and 360+180 fs respectively, followed by another induced absorption in the 540--600 nm region that had a rise time of 200-400 fs. The appearance of this signal coincided with the decay of the absorptions in the 640-730 nm region, and was attributed to photoproduct. The initial signal was due to excited-state absorption. After 100 ps, the spectrum matched that of the relaxed all-trans octopus bathorhodopsin (Taiji et al., 1992). The spectral properties of the transient species formed by 400 fs were

Characterization of the Primary Photochemical Events

93

found to be consistent with those ofphotorhodopsin (denoted as "primerhodopsin" by the authors), indicating that the photoisomerization has already taken place (Taiji et al., 1992). Femtosecond studies have also revealed the dynamics of isorhodopsin isomerization, which takes place at the C9~C10bond as opposed to the Cl1~C12bond (Figure 12) (Schoenlein et al., 1993). After excitation with the 500 nm probe pulse, an induced absorption (480-570 nm) appeared within 50 fs that lasted for 100-150 fs, and was attributed to excited state absorption (S~-->Sn). A bleach of the reactant followed, attaining its maximum by 300 fs; recovery of this bleach was found to

O3

"O

-ATFF=1%

-150 fs +33 fs + 100 fs +200 fs +400 fs +600 fs

I

I

+900 fs +3 ps

!

I

1

+6 ps 450

500

550

600

650

700

wavelength (nm)

Figure 12. Differential absorption spectra of isorhodopsin at various time delays of the probe pulse after excitation with a 40-fs pump pulse at 500 nm. Measurements from 450-570 nm were obtained by using a 10 fs probe pulse at 500 nm. Measurements from 570-670 were obtained by using a 10 fs probe pulse at 620 nm. (Reproduced with permission from Schoenlein et al., 1993).

94

JEFFREYA. STUART and ROBERT R. BIRGE

be biphasic with time constants of several hundred femtoseconds and 1.5 ps. The biphasic behavior was attributed to two processes, including the return to the ground state of nonisomerized reactant with a consequential redistribution of vibrational energy on a picosecond time-scale, and dynamics of the photoproduct. The former of the two should account for about 80% of the recovery, due to the low isomerization quantum yield ofisorhodopsin (0.22) (Schoenlein et al., 1993). Photoproduct absorption was seen in the 570-640 nm region. At early times (0-150 fs) an emission process was identified as the S 1-~S 0 transition, the result of the excitedstate wave packet leaving the Franck-Condon region (by 150 fs). As this feature decayed, an absorption signal appeared (150-200 fs) in the same region which exhibited a wavelength-dependent rise time, reaching maximum absorbance at earlier times for longer wavelengths (500 fs at 570 nm, 300 fs at 620 nm). Subsequent photoproduct relaxation (after several hundred femtoseconds) was also found to be wavelength dependent in a similar manner (1.3-0.6 ps); this behavior is consistent with vibrational cooling and conformational relaxation (Schoenlein et al., 1993). From 200 to 600 fs, the photoproduct absorption band narrowed and shifted to the blue, as evidenced by an isosbestic point shift from 525 to 515 nm (see below). By 6 ps, the difference spectrum was consistent with that ofbathorhodopsin (Schoenlein et al., 1993). By comparison of the results from the rhodopsin studies, it is apparent that although both proteins share qualitatively similar behavior, the isorhodopsin isomerization occurs approximately three times more slowly, consistent with its lower quantum yield (0.22). Whereas a distinct stimulated emission was observed for isorhodopsin, none was evident in the rhodopsin dynamics, indicating that the initial velocity of the wave packet out of the Franck-Condon region occurs on a slower time-scale (for isorhodopsin) (Schoenlein et al., 1993). Because the photoproduct band is likely to have had components from vibrationally hot species, including contributions (approximately 80%) from non-isomerized reactant molecules (with--20,000 cm-1 excess vibrational energy from the absorbed photon), it became necessary to model the effects of these species on the absorption spectrum (Schoenlein et al., 1993). After deconvolution of the different contributions, an approximate isomerization time of 600 fs was found. The vibrational coherence of the rhodopsin photoisomerization was not evident in isorhodopsin, again consistem with a slower isomerization. There are several conceivable explanations cited by the authors to account for the observed differences between the proteins. Perhaps the most substantial is the difference in the interactions of the chromophore, namely the lack of an interaction between the 13-methyl group and the 10-hydrogen in isorhodopsin. In rhodopsin this steric interaction is thought to result in a torsionally distorted Cl1~C12bond that may aid in driving the isomerization forward (thus helping to account for the unusually high quantum yield of 0.67). No equivalent steric interaction exists in isorhodopsin; however, this does not rule out the existence of other protein-chromophore interactions in the binding site, which are certainly different from those of rhodopsin. Indeed, such interactions have been

Characterization of the Primary Photochemical Events

95

suggested to account for the temperature and wavelength dependence of the isorhodopsin quantum yield (Hurley et al., 1977; Birge, 1990a; Schoenlein et al., 1993). A small activation barrier of 0.2 kcal/mol has been calculated to be associated with the 9-cis ~-~ 9-trans excited-state torsional surface. This barrier is small enough to be unimportant at ambient temperatures, but is comparable to thermal energy at low temperatures (at 77 K, kT ~ 0.15 kcal/mol vs. 0.6 kcal/mol at ambient), which would therefore have the effect of lowering the quantum yield (see Figure 14) (Birge et al., 1988; Birge, 1990a). It is important to consider the implications of the ultrafast spectroscopic studies upon the nature of the primary intermediate, photorhodopsin, and its transition into the better defined batho-intermediate. As previously mentioned, it is a matter of some controversy as to whether photorhodopsin can be accurately described as a true intermediate due to its inherent inability to be isolated at low (liquid helium) temperatures. The evidence that the primary event, that is, the photoisomerization, has already occurred by the formation ofphotorhodopsin is conclusive, based upon the results of the spectroscopic and chromophore analog studies described above; the photorhodopsin chromophore has an all-trans configuration, albeit highly distorted. Generally speaking, the ultrafast studies indicate a fast femtosecondisomerization, followed by a slight blue-shift on a picosecond time-scale. This shift is most likely the result of thermal cooling of the chromophore, such that photorhodopsin can be described as a thermally "hot" bathorhodopsin. The transition from photo- to bathorhodopsin is then simply a "cooling off'' step, involving the redistribution and dissipation of vibrational energy modes between the chromophore and the protein over several picoseconds. At room temperature, this transition is probably barrierless on a nonadiabatic potential surface (Taiji et al., 1992; Peteanu et al., 1993), and corresponds to a structural relaxation from a highly distorted to a more relaxed planar all-trans configuration. As noted above, bathorhodopsin has a formation time reported to be on a picosecond time-scale, consistent with conformational relaxation times measure by time-resolved vibrational spectroscopy for the bacteriorhodopsin chromophore (van den Berg et al., 1990; Doig et al., 1991). The results cited above by Peteanu et al. (1993) for the temporal period between 200 fs and 6 ps, namely the shift in photoproduct absorption and simultaneous gain at 535 nm, the blue shift in the isosbestic point, and the recovery of the reactant bleach, are all consistent with vibrational cooling and conformational relaxation processes (Peteanu et al., 1993). Observations of these types of behavior were also made in the octopus studies, where the authors described primerhodopsin (syn. photorhodopsin) as "quasi-thermal" bathorhodopsin, which then "thermalizes" with a time constant of 20+10 ps to form the fully relaxed bathorhodopsin (Taiji et al., 1992). Additional support for the vibrationally hot nature ofphotorhodopsin was reported in the studies on isorhodopsin described above. The shift in the isosbestic point concomitant with the blue-shift and narrowing of the photoproduct absorption band, as well as the wavelength-dependent partial relaxation of the same band within several hundred femtoseconds yielding spectral features consis-

96

JEFFREYA. STUART and ROBERT R. BIRGE

tent with bathorhodopsin, both indicate vibrational cooling and conformational relaxation processes of the chromophore to a less distorted all-trans configuration (Peteanu et al., 1993).

D. Quantum Efficiency of the Primary Event The quantum yields for the photoconversions involving rhodopsin, bathorhodopsin and isorhodopsin are shown in Figures 13 and 14 and the salient assignments are given below:

''''''"1'"''''"1'""''"1

0.9 0.8

0.7 0.6

0.5

0.3 1

450

470

490

510

Wavelength (nm)

--.d)2(R-->R~ --.

,,,,,,,,I,,,,,,,,,I,,,,,,,,,I,,,,,,,,,I,,,~

530

450

" .... '1' ...... "1'"'""'1" ....... I'"'_:

0.14 0.12 0.10

¢I)3(B_.__)I )

..... ''"1'"~

~

470

490

510

Wavelength (nm)

' 0.16

--

530 I......

. . . . . .

0.14

*0.08

0.040.06

I

T

T~

I

I

~

T'~

I _

,,,,,,,,I,,,,,,,,,I~,,,,,,,,I,,,,,,,,,I,,,

450

470

490

510

Wavelength (nm)

530

(I)0.12 0.080"10 1

(D41][-)B) "~""~1.

~-.

I= ,,h,,,,,,,,I,,,,,,,,,h,,,,,,,,t,,,,,,,,~h,,,,,,,,h,,,,,,,,h .... ,,,,h,,,

420 440 460 480 500 520 540 560 Wavelength (nm)

Figure 13. Wavelength dependence of the quantum yields for the photoreactions

involving rhodopsin (R), bathorhodopsin (B) and isorhodopsin (I) determined based on photostationary state data collected at 77K (solid circles) and 70K (open circles) (Birge and Callender, 1987; Birge et al., 1988). These graphs support the assumption of wavelength independent values for @1 (R-~B) = 0.67, @2 (B-->R) = 0.49 and @3 (B~I) = 0.076 and the horizontal lines indicate the assigned wavelength independent values (Birge and Callender, 1987). However, the graph for @4 (I-->B) indicates a wavelength dependence that is characteristic of a small barrier in the isorhodopsin --> bathorhodopsin excited state potential surface (Birge et al., 1988). The dotted and solid lines display two models for the quantum yield for the latter photoisomerization, both of which assume a barrier of less than ~70 cm -1. The barrier is so small that the wavelength dependence of the @4 (I-->B) is not observable at ambient temperature [@4 (300 K) = 0.22, (Hurley et al., 1977)].

Characterization of the Primary Photochemical Events •

= 0.67

rhodopsin (R) ~

97

el)_ = 0.073

bathorhodopsin (B) ~

O 2 = 0.49

isorhodopsin (I)

(I)4 (L,T)

where the values cI)1 = 0.67 + 0.02, cI)2 = 0.49 + 0.03, cI)3 = 0.076 + 0.006 are independent with respect to both temperature and excitation wavelength within the error ranges specified (Birge and Callender, 1987). In sharp contrast, (I) 4 is observed to be both temperature and wavelength dependent (Rosenfeld et al., 1977a; Schick et al., 1987; Birge et al., 1988). A summary of the experimental results are shown in Figure 13. This latter observation has interesting implications which have been

-85 %

~15 %

I i

>95%

'

(0.06) "

!

0.49

~, /

~

04) I

~

{(o.6)1"},

•

{(o.2)$}

I

I I

Bathorhodopsin

(0.33)

I I I I

0.073

[0.2251

I I

0.67

~40 %

! ! ! I

-27 kcal/moi -32 kcal/mol

I Isorhodopsin

--5 k c a l / m o l .

.

.

.

.

.

.

.

.

.

.

.

.

. . . . .

. . . . . .

. . . . . . .

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

Rhodopsin Figure 14. Schematic representation of the ground and first excited singlet state surfaces connecting rhodopsin, bathorhodopsin and isorhodopsin using a simplified (linearized) reaction coordinate. The shapes of the ground and excited state surfaces are based on all-valence electron molecular orbital calculations including single and double configuration interaction (Birge et al., 1988). Ground state enthalpies are taken from the experimental measurements of Cooper (1979a,b) and Schick et al. (1987). Absolute quantum yields of photoisomerization are displayed at the tips of the arrows indicating the processes (Birge and Callender, 1987; Birge et al., 1988). Values given in parentheses (o) are predicted by using semiempirical molecular dynamics theory to calculate the reverse/forward yield ratios and multiplying these values by the experimental forward yields (shown without parentheses). Values listed in brackets {o} are ambient temperature quantum yields which display temperature dependence. The arrows indicate the effect that lowering the temperature will have on these values, such as, {0.225} indicates that at lower temperatures, the quantum yield will be lower than 0.22 (Adapted from Birge et al., 1988.)

98

JEFFREY A. STUART and ROBERT R. BIRGE

discussed earlier. At ambient temperature, however, this quantum efficiency is equal to 0.22, and is independent of both temperature and wavelength (Hurley et al., 1977). A slight dependence of the isorhodopsin quantum yield upon laser pulse photon density has also been reported (Kandofi et al., 1988); the ratio relative to rhodopsin was measured to be 0.37 in steady actinic light and 0.39 with a weak picosecond laser pulse. Use of laser pulses of higher energy gave ratios > 0.39, indicating saturation effects (Kandori et al., 1988). The quantum yield of the primary event was originally assigned by Dartnall (1972) and subsequent studies have yielded values in agreement within experimental error (Hurley et al., 1977; Rosenfeld et al., 1977a; Birge and Callender, 1987). The observation that O)1 of rhodopsin is larger by a factor of at least two compared to the photoisomefization quantum efficiency of the 11-cis retinyl protonated Schiffbase (RPSB) in solution (Freedman and Becker, 1986; Freedman et al., 1986) is one indication that the protein has a binding site optimized for 11-cis --~ 11-trans photoisomerization. The fact that the primary event also stores --32 kcal mo1-1 (Cooper, 1979b; Boucher and Leblanc, 1985; Schick et al., 1987), whereas the photoisomerization of the 11-cis RPSB in solution generates a more stable species, provides further evidence that the protein is modifying both the ground and the excited state potential surfaces. The observation that (I)~ + (I)2 add up to a number larger than unity (1.16+0.05) indicates that while a common excited state intermediate may be populated during photochemistry, coupling into the ground state is a trajectory dependent (i.e., dynamic) process (see below). A key conclusion is that the excited state Cl1=C~2 torsional surface is barrierless (Honig et al., 1979; Birge and Hubbard, 1980, 1981; Birge, 1981, 1982; Honig, 1982; Doukas et al., 1985; Birge and Callender, 1987; Birge et al., 1988) (Figures 9 and 14). As discussed above, it appears that the C9=C10 excited state torsional surface is not barrierless (Birge et al., 1988) (Figure 14). A study by G~irtner et al. (1991) recently incorporated insect retinals into bovine opsin and determined the quantum yields. Many insects utilize retinals that are hydroxylated at C 3 or Ca; such additions result in chromophores that are more polar and capable of hydrogen bonding to functional groups of the protein, although it is not well understood as of yet what, if any, physiological role these interactions play in the insect rhodopsins. It was found that the analog rhodopsin formed with 3-hydroxy-retinal had a 20% larger quantum yield than the bovine pigment (G~irtner et al., 1991). It was postulated that the higher quantum yield may have been due to steric forces in the binding site that would favor the photoproduct. Slow regeneration times were indeed observed for bovine opsin with 3-hydroxy-retinal, indicating steric restrictions in the bovine opsin binding site which had to be overcome on formation of the active synthetic pigment (G~irtner et al., 1991).

E. Molecular Dynamics of the Primary Event Tallent et al. (1992) recently carried out molecular dynamic simulations of the primary event in rhodopsin based on the binding site model shown in Figure 6. The

Characterization of the Primary Photochemical Events

99

ground state and excited state surfaces connecting rhodopsin (R) and bathorhodopsin (B) along the ~1~,~2dihedral reaction path were partially adiabatically mapped. The ground state surface was generated by using MNDO/AM 1 procedures and the excited state surface was generated by using INDO-PSDCI procedures including both single and double configuration interaction. The resulting surfaces are shown in Figure 9. The first excited singlet state exhibits a barrierless reaction path for Cl1=C12 dihedral torsion with a local minimum (activated complex) centered at ~11,12 - 90°- Semiempirical molecular dynamics procedures were used to simulate the forward and reverse photochemistry, and the forward dynamics are shown in Figure 9. In the following sections we overview some of the more important observations of this theoretical study as it relates to various experimental observables and recent time-resolved studies of the primary event.

Origin of the Barrierless Excited State Surface We know from experimental studies that the excited state potential surfaces connecting rhodopsin and bathorhodopsin are barrierless (see above). In this section we provide a quantum mechanical perspective on this important characteristic. Many of the comments also apply to bacteriorhodopsin, but the lack of an accurate electrostatic map of the counterion environment in the bR binding site precludes definitive analysis. One of the key aspects of the binding site responsible for the generation of a barrier-less excited state potential surface is associated with the atomic charges on the chromophore in the ground and lowest excited singlet states. The imine nitrogen atom is negatively charged in protonated Schiff bases. The positive charge on the chromophore is highly delocalized, and the most positively charged atoms are C15, C13 and the imine proton (Tallem et al., 1992). Excitation of rhodopsin into the lowest-lying Franck-Condon excited state generates a large redistribution of charge resulting in the transfer of~-0.27 electron units of negative charge into the C13..-N16 portion of the polyene chain. The net charge on the C13--C14--C15--N moiety changes from +0.120 to-0.153 upon excitation (Tallem et al., 1992). This charge reorganization alters the electrostatic interaction with the counterion from a ground state stabilization into an excited state destabilization and forces the chromophore away from the counterion. Torsion about the C l l = C 12bond is the path of minimum energy, and thus the photoisomerization is initiated with a negative barrier. Torsion about the C~l=C12 bond is calculated to mix the second excited ,,1Ag•. state . into . the . lowest-lying . . ~Bu+,, state. In the torsional region 75--105 ° (90 ° = orthogonal), the lowest excited state has considerable ,,lAg,-,, character. This generates a local minimum in the excited state potential surface which is referred to as the "activated complex." The combination of charge reorganization within the chromophore and the partial switch from ionic to covalent character near orthogonality combine to generate a barrierless excited state potential surface for 11-c/s --~ 11-trans photoisomerization.

100

JEFFREY A. STUART and ROBERT R. BIRGE

Kinetics of the Photoisomerization Recent experimental studies of the kinetics of the rhodopsin primary event were discussed in detail in Section IV-C. The femtoseconds experiments of Schoenlein et al. (1993) and Peteanu et al. (1991) predict the formation ofbathorhodopsin in ~200 fs. In contrast, Yan et al. (1991, 1993)predict the formation of an excited state complex in --200 fs with relaxation imo ground state bathorhodopsin in -3 ps. Regardless of which model is used, it is clear that the kinetic simulation of the excited state dynamics is overestimating the time necessary to reach the activated complex ((I)11_12 " 9 0 ° ) . We calculate ~375 fs (Figure 9) but the observed value is 200 fs or less. Thus, we are underestimating the steepness of the potential surface, overestimating the inertial component of the chromophore motion, or both. In addition, the simulations start the molecule from a rest position which eliminates thermal assistance of motion in the excited state. Because the center of mass of the ]3-ionylidene ring was fixed during the dynamics, it is likely that the translation of the center of mass of the chromophore during photoisomerization is overestimated. It is difficult to quantitatively analyze the other potential sources of error. Nevertheless, the above observations provide some insight into the two experimentally based models of the primary event. The theoretical dynamics shown in Figure 9 predict a 1/e time for bathorhodopsin formation of 886 fs (see Table III in Tallent et al., 1992). This number should be compared to the 200 fs formation time ofbathorhodopsin predicted by Peteanu et al. (1993) and Schoenlein et al. (1991). Again, the calculated value is too large, but that observation is consistent with our previous observation that the simulations overestimate the time to reach the activated complex. The excited state spectra that were calculated are, in general, consistent with the interpretations advanced by Schoenlein et al. (1991) and Peteanu et al. (1993) in analyzing the transient spectra shown in Figure 11. A key complication is the theoretical prediction that the principal absorption band in the 400-700 nm region is very similar in intensity and position to the ground state absorption spectrum of bathorhodopsin (see Figure 9 ofTallent et al., 1992). However, the calculated spectra display a doublet structure when the ensemble enters the activated complex, and there is an additional absorption band in the 650-850 nm region (see Figure 9 of Tallent et al., 1992). As the ensemble relaxes in the activated complex, the red band decreases in intensity and blue shifts. If one compares the spectra of Figure 9 and 10 from Tallent et al., 1992 to those shown in Figure 11, one could reasonably interpret the spectral features to indicate the following sequence:

R(ll-cis)

hv

> R(FC)*

~ 150)fs

R(Zll)

--300fs

~ Bhot(ll_transoid ) ~__~s B(ll-trans)

where FC refers to the Franck-Condon vertically excited level, x lx refers to the Cl1~C12 orthogonal activated complex, and the subscript "hot" describes a vibrationally activated ground state. The above kinetic assignments are based on an

Characterization of the Primary Photochemical Events

101

analysis of the weak bands that appear above 600 nm and a comparison of these bands with the calculated S 1---~Sn transitions under the assumption that the experimentally observed bands are indicative of excited state occupation. This assumption is consistent with the observation that the reformation time of rhodopsin lies somewhere between 2 and 6 ps. Our excited state absorption calculations are also consistent with the model of the rhodopsin dynamics proposed by Yan et al. (1991, 1993). However, our dynamics simulations conflict with the ratio of bathorhodopsin formation versus activated complex formation. The model proposed by Yan et al. (1991, 1993) can be described by the following series: hv

R(ll-cis) ~

R(FC)*

"~Fs

,

R0:ll) 2 ~

B(ll-trans)

We make the assumption that the 3 ps time constant for batho formation represents the time leading to formati'on of the relaxed bathorhodopsin product. This assignment yields a ratio of B/R(z11)* ~ 15. The dynamics shown in Figure 9 yield the following times: hv

R(ll-cis) ---> R(FC)* 3]~s R(z11), 8_~s Bhot(ll_transoid ) l:~_~SB(ll_trans ) Thus, the calculated ratio of B/R(Zll)* is 3.6. This is a rather significant difference which indicates an intrinsic flaw in either the calculated dynamics or the model used to interpret the time-resolved data. Yan and coworkers are currently carrying out additional time resolved experiments to help address some of the issues discussed above as well as the significant differences between the two experimental models. These experiments hope to observe some of the red-shifted absorption bands that are predicted to populate the excited state absorption spectra in the 700-900 nm region. The theoretical simulations predict that these features will be highly diagnostic of the excited state torsional angle (see Figures 9 and 10 in Tallent et al., 1992).

Quantum Yield of the Photoisomerization Weiss and Warshel (1979) have proposed that nonadiabatic coupling is the primary contributor to the transfer of rhodopsin from the excited singlet state surface into the ground state manifold. Their analysis, using a pair oforthogonalized LOwdin P7 atomic orbitals, yielded Pna ~-0.5, where Pna is the probability of crossing associated with nonidiabatic coupling as the trajectory passes through the orthogonal region. This value yields excellent agreement with the observed rhodopsin --~ bathorhodopsin quantum yield assuming that this probability is identical on each pass through ~ ) 1 1 , 1 2 " - 90°: 1 tI) 1 =

-

2

-

- Pna

1 --

2 - 0.5

--

0.67

102

JEFFREYA. STUART and ROBERT R. BIRGE

There are two features of the Weiss and Warshel model, however, that deserve fiLrther examination. First, if Pna ~ 0.5 for all trajectory passes, then we may conclude that • 2 v_ O1, and thus the bathorhodopsin ~ rhodopsin quantum yield calculated by using the Warshel-Weiss model will be significantly overestimated unless bathorhodopsin has an unusually large non-dynamic decay mode into the grotmd state. The latter possibility is unlikely given the large excited state gradient that is calculated for bathorhodopsin. Second, extrapolation of integral calculations based on two orthogonalized L6wdin Pz atomic orbitals to a polyene with six double bonds will likely overestimate the <~gl/&g0/90> integrals. However, previous simulations based on purely dynamic (semiclassical S-matrix) coupling consistently underestimated the quantum yields (Birge and Hubbard, 1980, 1981). Tallent et al. (1992) investigated the contributions of nonadiabatic coupling to the quantum efficiencies associated with the dynamic simulations shown in Figure 9. The nonadiabatic probability was explicitly calculated by reference to the phased ground and excited state wavefunctions, and improved the calculated values from OI(R-~B ) = 0.62 and O2'(B--+R ) = 0.48 to • 1 = 0.70 and O 2' -" 0.52, where (I) 2' represents the quantum yield for the B--~R photochemistry when the path to isorhodopsin has been ignored. The experimental values are • 1 = 0.67 +_ 0.02 and • 2' - 0.53 _+0.03. Bathorhodopsin photoconverts to both rhodopsin (0 2 -- 0.49 _+0.03) and isorhodopsin (9-cis-chromophore) (0 3 - 0.076 +_0.006) (see Section IV-D). Thus the adjusted (isorhodopsin pathway excluded) value is: 02 02'-1-0 3

0.49 + 0.02 = 0.53 + 0.03 1-0.076+0.006

When nonadiabatic coupling is included, the forward (O1) quantum yield is overestimated by--4% and the reverse is underestimated by 2%. Given the level of approximation inherent in the theoretical procedures, this represents good agreement with experiment. The nonadiabatic coupling term, when properly phased, changes sign at *11,12 ~ 92 ° . Thus, coupling into the ground state is enhanced preferentially for the forward (R -~ B) trajectory relative to the reverse (B -~ R) trajectory through the (~11,12 -90 ° crossing point. Thus, the contribution of nonadiabatic coupling to the probability of crossing into the ground state is partitioned into trajectory dependent contributions which preferentially enhances • 1 relative to • 2. Nonadiabatic coupling also increases the overall efficiency of coupling into the ground state and decreases the product formation time for both the forward and reverse photochemistry (Tallent et al., 1992). The lower quantum yield of the bathorhodopsin -~ rhodopsin photoisomerization is due to the above partitioning of the nonadiabatic coupling as well as the rapid arrival of the trajectory into the activated complex. The latter precludes equilibration of the excited state prior to arrival at the activated complex and lowers the dynamic coupling term.

Characterization of the Primary Photochemical Events

103

F. Energy Storage in the Primary Event The first photocalorimetric measurement of the energy stored in the primary photochemical event of rhodopsin was carried out by Alan Cooper in 1979, and his measurement of AHRB - 34.7+_2.2 kcal mo1-1(Cooper, 1979b) prompted considerable interest in the mechanistic origins. First, this value indicates that-%0% of the absorbed photon energy is converted into stored energy, an efficiency that seems unrealistically high given the concomitant high quantum efficiency of 0.67 (net system efficiency ~ 40%). Second, models of the primary event published during the same year predicted much lower values: (e.g., 26 kcal mo1-1 (Birge and Hubbard, 1980) and 14-28 kcal mol-l (Honig et al., 1979). A subsequent experimental study using a different technique (pulsed laser photocalorimetry) and a range of excitation wavelengths yielded AHRB= 32.2+_0.9 kcal mo1-1 (Schick et al., 1987) in good agreement with Cooper's measurement. This study also measured the energy stored in the isorhodopsin -~ bathorhodopsin phototransformation and observed AHm = 27.1_+3.2 kcal mo1-1 (Schick et al., 1987). This value of AHIB indicates that the bathorhodopsins formed from rhodopsin and from isorhodopsin are energetically equivalent, because isorhodopsin has an enthalpy -~5 kcal mo1-1 higher than rhodopsin (Cooper, 1979a). This observation, combined with optical spectroscopic studies, confirms the fact that the bathorhodopsins formed from rhodopsin and isorhodopsin are energetically identical though minor spectroscopic differences can be noted (Monger et al., 1979; Mao et al., 1980; Einterz et al., 1987a,b; Hug et al., 1988). A schematic diagram showing the ground and excited state energetics is presented in Figure 14. The molecular origins of energy storage in bathorhodopsin remain a subject of debate. Some of the early models emphasized energy storage due to charge separation (Honig et al., 1979) while other early models emphasized energy storage due to conformational distortion (Birge and Hubbard, 1980, 1981). Virtually all models, both past and present, recognize that both mechanisms contribute, and discussions center on the extent to which one mechanism dominates the other. An experimental and theoretical study of energy storage in bathorhodopsin yielded the following partitioning: charge separation (-~12 kcal mol-1), intrachromophore-lysine conformational distortion (-~10 kcal mol-l), and chromophore-protein conformational distortion (-10 kcal mo1-1) (Birge et al., 1988). Palings et al. (1989) note the contributions of electrostatic interactions in determining the energy storage based on the observation of a perturbed hydrogen out-of-plane wagging and C-C stretching vibrations associated with the C10-C1~=C~2-C13 region of the chromophore.

G. Molecular Origins of Photoreceptor Noise The human visual system can detect faint stars at night and distinguish objects in direct sunlight for an effective operating range of about 10 log units of light

104

JEFFREYA. STUART and ROBERT R. BIRGE

intensity. At high light intensities the key limitation of visual function is the bleaching of photoreceptor pigments. At low light levels the key limiting factor is noise in the photoreceptors. We can reliably detect pulses of light that send roughly 100 photons through the pupil and activate approximately 10-20 rhodopsin molecules in as many rod photoreceptors. We see such dim flashes of light against a background of noise caused by each photoreceptor eliciting false signals ("noise") which are indistinguishable from the signals triggered by single photon absorptions (Baylor et al., 1980). The origin of this photoreceptor noise remains a subject of debate and a number of recent articles have examined this subject from diverse perspectives (Cornsweet, 1970; Baylor et al., 1980; Aho et al., 1988; Barlow, 1988; Barlow, Jr. and Kaplan, 1989; Barlow, Jr. and Silbaugh, 1989; Birge, 1990a, 1993; Fahmy and Sakmar, 1993; Barlow, Jr. et al., 1993; Birge and Barlow, 1994). In this section we examine the origin of photoreceptor noise and demonstrate that the available evidence supports a two-step molecular process inside the protein binding site of rhodopsin as the primary source. We also discuss how nature has optimized the active site of rhodopsin to minimize the thermal reactions that are responsible for generating the false signals.

Rhodopsin is the Source of Photoreceptor Noise Although the origin of photoreceptor thermal noise may remain a subject of debate, investigators agree that thermal and light activated photoreceptor signals are identical with respect to intensity and temporal profile (Baylor et al., 1980, 1984; Aho et al., 1988; Barlow, 1988; Barlow, Jr. and Kaplan, 1989; Barlow, Jr. and Silbaugh, 1989; Barlow, Jr. et al., 1993; Birge and Barlow, 1994). Thus, it is important to understand the nature of the light activation and amplification process. The thermal noise phenomenon must involve either a side reaction or corruption of this process prior to amplification, or a side reaction that undergoes nearly identical amplification. Following the primary event, a series of dark reactions occur which ultimately deprotonate the chromophore and activate the protein. The activated protein initiates a complex biochemical process that hyperpolarizes the plasma membrane of the rod cell in the retina (Liebman et al., 1987; Stryer, 1986, 1987). An analysis of the amplification mechanism coupled with the observation that the intensity and duration of photoreceptor noise signals are identical to light-induced signals leads to the conclusion that if the source of the noise involves a thermal corruption of the above process, rhodopsin must be the source of the noise. That is, activated rhodopsin is being generated by a thermal process, and the resulting R** (i.e., R~hermal)is identical to R* (i.e., Rhv) with respect to the catalytic amplification process described above. There is, however, one other possibility that needs to be considered. The connections among the neural elements in the retina are complex and are responsible for extensive signal processing prior to transferring the photoreceptor signals to the brain. Because the bipolar cells that mediate the photoreceptor signals prior to transfer to the ganglion network have a signal leveling effect, the observation that dark signals have the same shape and intensity

Characterization of the Primary Photochemical Events

105

may indicate that the bipolar cells have mediated the signal. Thus, the origin of photoreceptor noise may be within the neural network rather than the photoreceptor system. There are two observations that argue against this interpretation. First, the nature of the photoreceptor noise in humans is invariably perceived as point source noise (Cornsweet, 1970). This observation supports a model in which thermal activation involves individual photoreceptors rather than mediating neural elements. Perhaps the strongest argument in favor ofrhodopsin is the observation that there is a near linear proportionality between dark noise and the rhodopsin content of photoreceptors in a range of animals (Figure 15). Because the animals analyzed in Figure 15 have significantly different neural architectures, the observed linear relationship suggests that photoreceptors and not neural elements are responsible. The data of Figure 15 coupled with the observed invariance between light-induced and thermal photoreceptor signals points to rhodopsin as the source. There are many possible mechanisms that could be responsible for the thermal activation of rhodopsin. In the following section, we examine and compare the various possibilities with the goal of eliminating those which are inconsistent with thermodynamic, kinetic, or spectroscopic observation.

Possible Mechanisms of Thermal Activation of Rhodopsin Baylor and coworkers have carried out a detailed analysis of electrical dark noise in toad retinal rod outer segments, and assigned the thermodynamic properties of the thermally activated dark processes: (E a = 21.9+1.6 kcal moV l, AG* = 31.9+0.13 kcal mo1-1, AH* = 21.6+1.6 kcal mo1-1, AS* =-35.3+5.6 e.u.) (Baylor et al., 1980). The activation energies measured for the horseshoe crab (Limulus) agree within experimental error (E a = 26.3+7.8 kcal tool-1 (day), 27.9+6.5 kcal roof 1 (night), 26.5+7.5 kcal mo1-1 (in vitro)) (Barlow, Jr. and Kaplan, 1989; Barlow Jr. and Silbaugh, 1989; Barlow, Jr. et al., 1993). We will discuss the Limulus data in more detail below, because this animal is capable of dramatically decreasing photoreceptor noise at night (see Figure 15). A comparison of these data with denaturation activation energies measured by Hubbard for cattle rhodopsin (E a = ---100kcal mol-l), frog rhodopsin (E a = --'45 kcal mo1-1) and squid rhodopsin (E a = -72 kcal moV 1) indicates that protein denaturation is not the origin of the dark signal (Hubbard, 1958). Measurements of thermal isomerization of 11-cis retinal, however, appear to offer a much more compatible set of thermodynamic properties (E a = 22.4 kcal mo1-1, AG* = 29.3 kcal mo1-1, AH* = 21.7 kcal mo1-1, AS* = - 2 1 . 4 e.u.) (1-propanol solution) (Hubbard, 1966). Comparison of the latter measurements on 11-cis retinal with those observed by Baylor on rod segments has prompted some investigators to propose that thermal isomerization of the chromophore is responsible for dark activation of rhodopsin (Baylor et al., 1980; Aho et al., 1988; Barlow, 1988). However, this hypothesis is not consistent with the energetics of ground state isomerization of the protein bound chromophore (Birge, 1990a). The protein bound chromophore is not 11-cis retinal, but the protonated Schiff base of l l-cis retinal. The ground state barrier to

106

JEFFREY A. STUART and ROBERT R. BIRGE

LIMULUS • (day) r.~

r.~

¢

>

0.1

m m mm m mm

o~

m

BULLFROG

Z ell

TOAD

HUMAN

~, 0.01

m m

"

m

m

MONKEY eJ

mu mm m n m

,4,,,I

mm

DOGFISH

m

LIMULUS (night) "

LOCUST

0.001

I

,

m

m

,m,,,l

I

I iIiiim

10 0.1 1 Rhodopsin Content (10 9 molecules cell "1)

Figure 15. Dark noise and rhodopsin content of photoreceptors. Dark noise for these various visual systems is approximately proportional to rhodopsin content (slope of dashed line = 1) supporting the concept that the rate constants for noise are about equal and that rhodopsin is the source of the thermal noise (Adapted from Birge and Barlow, 1994). isomerization of the protein bound chromophore is estimated to be AH$ = 45+3 kcal mo1-1 (Birge, 1990a). We can establish a lower limit ofAH $>_42+3 kcal mo1-1 based on the relative enthalpy ofbathorhodopsin (AHRB = 32.2+0.9 kcal mo1-1) (Schick et al., 1987) plus the activation enthalpy of the bathorhodopsin --->lumirhodopsin dark reaction (AH~ = 10+_2 kcal mo1-1) (Grellmann, et al., 1962) and assuming additive errors. In contrast, the activation energies for thermal activation are all less than 36 kcal/mol, and a majority are less than 30 kcal/mol. Thus, thermal (ground state) isomerization of the native (protonated) chromophore cannot be responsible for thermal activation of the protein.

Characterization of the Primary Photochemical Events

107

Rhodopsin might have sufficient conformational flexibility to undergo a conformational change (R --%R**) that is interpreted (incorrectly) by transducin (T,~C GDP) to represent photochemically activated rhodopsin (R*). Thus the initial step in the amplification process takes place involving this thermally activated rhodopsin (TaI3cGDP + R** ~ R** - T@cGDP + GTP --~ R**-Ta~cGTP + GDP). The key problem with this mechanism, however, involves the observation that the thermal noise signals have intensities identical to the light activated signals. Thus, R** (thermally activated rhodopsin) must have a lifetime nearly identical to R* (photochemically activated rhodopsin). This observation precludes enthusiasm for spontaneous conformational distortion, because the spontaneous process would either generate a very short-lived R**, or would require denaturation of the protein. Another possibility is that there is an equilibrium within the rhodopsin binding site coupling protonated versus unprotonated chromophores. The experiments of Longstaff and Rando have demonstrated that deprotonation of the Schiff base of retinal is obligate for rhodopsin activation (Longstaff et al., 1986; Longstaff and Rando, 1987). This mechanism assumes that deprotonation of the Schiff base is sufficient for activation. More precisely, deprotonation generates a form of rhodopsin that is interpreted by transducin as activated (i.e., R*), and the cascade is initiated. (Deprotonation would break the above mentioned salt bridge.) The observation that the thermal noise signals have intensities identical to the light activated signals requires that the rhodopsin molecule with an unprotonated chromophore have a lifetime identical to R*. This is unlikely due to the instability of the R d species. Recent site directed mutagenesis studies by Fahmy and Sakmar (1993) provide more explicit evidence that deprotonation is not sufficient for activation. These investigators replaced the primary counterion of rhodopsin, the glutamate (E) residue at position 113, with glutamine (Q) and regenerated the mutant opsin with 11-cis retinal. The E 113Q substitution dramatically decreases the pK a of the Schiffbase proton and generates a pigment containing a mixture of protonated and unprotonated retinyl chromophores with absorption maxima at 490 and 380 nm, respectively. The key observation is that the unprotonated (380 nm) species does not activate transducin which indicates that deprotonation is not sufficient to generate R**. The above observations coupled with the theoretical and experimental studies described below led us to conclude that the most likely mechanism for thermal activation ofrhodopsin is a two step process (Birge, 1990a, 1993; Barlow, Jr. et al., 1993). The first step is deprotonation of the 11-cis protonated Schiffbase chromophore. The second step is thermal 11-cis to 11-trans isomerization of the chromophore. This mechanism is schematically diagrammed in Figure 16 and the adiabatic potential surfaces generated by using MNDO/AM 1 and INDO-PSDCI molecular orbital theory are shown in Figure 17. The calculated adiabatic activation energies from Figure 17 are given below:

01%

~ .,

H~,~ ""

-,',<,<,<' .... H

/

A c%

!

Deprotonated Rhodopsin ( ~ ~ ) .......

/\

kl

....

:~__r~,

V

AH

Ckh [

Deprotonated Bathorhodopsin ( ~ ,~ )

~

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

........

:

"~

A

Figure 16. Molecular schematics of the photochemical and thermal pathways of activation of rhodopsin. The membrane spanning helix to which the chromophore is attached via Lys297 is labeled G. The primary counterion, Glu113, is attached to helix C (Zhukovsky and Oprian, 1989; Sakmar et al., 1989). Only one water molecule is shown for simplicity, and the symbol A represents one or more negatively charged amino acids outside of the binding site which ultimately accepts the proton. Note that the rhodopsin (R) binding site is neutral (Birge et al., 1985), and thus the acceptor group must be separated from the chromophore by a larger distance than implied by this figure. Spectroscopic and theoretical studies indicate that the glutamic acid counterion interacts primarily with the chromophore in the region C12-C13=C14C15 region of the chromophore and that at least one water molecule is hydrogen bonded to the imine proton (Birge, 1986, 1990a,b; Deng et al., 1994; Kakitani et al., 1985). (Adapted from Birge and Barlow, 1994). Figure 17. Theoretical analysis of the photochemical and thermal pathways of activation of rhodopsin based on the model shown in Figure 16. Shown are the ground and first excited singlet state adiabatic potential surfaces of the protein bound chromophore of rhodopsin based on MNDO/AM1 and INDO-PSDCI molecular orbital theory. The model of the binding site (minus water) and the molecular orbital procedures are those described in this review. The surfaces are calculated by minimizing the geometry of the chromophore and three water molecules as a function of two coordinates, the Cl1=C12 double bond torsional angle (0°=11 -cis, 180°=11 -trans) and the deprotonation coordinate (0 = covalently bonded; 1 = deprotonated). The key ground state minima and calculated relative energies (kcal/mol) are as follows: R (rhodopsin, AE = 0), B (bathorhodopsin, AE = 31 ), Rd (rhodopsin with a deprotonated 11 -cis chromophore, AE=I 7), Bd (bathorhodopsin with a deprotonated 11-trans chromophore, AE=23). The corresponding excited state potential surface is shifted vertically to facilitate viewing of the ground state surface. (continued) 108

140 120

100 8O 6O 4O

4O 20

Figure 17. (Continued) The symbol RFc designates the Franck Condon excited state of rhodopsin. The photochemical activation of rhodopsin proceeds by excitation into RFc, partial isomerization in the excited state to a Cl1=C12 distorted geometry (-90°), crossing into the ground state and continued rotation about the ~11=12 to form bathorhodopsin (Bd). The thermal activation of rhodopsin involves a two-step process involving deprotonation of the chromophore [rate = kl = A1 exp(-27/RT)] followed by isomerization of deprotonated chromophore [rate = k2 = A2 exp(-23/RT)], where simple Arrhenius rate equations are assumed with prefactors (unknown) of A1 and A2 and activation energies in kcal/mol. Activation to the first excited singlet state of Rd requires absorption of higher energy photons (;Lmax=380nm, 73 kcal mo1-1) than for native rhodopsin, but the ground state barrier for isomerization of 23 kcal mo1-1 is considerably lower than for native rhodopsin making the unprotonated form relatively less stable. (Adapted from Barlow Jr. et al., 1993 and Birge and Barlow, 1994.) 109

110

JEFFREY A. STUART and ROBERT R. BIRGE AE1 ~ 27 kcal/mol AE2 "23 kcal/mol R

j • N AE_I ~ 10 kcal/mol

Rd

f

Bd

The above energetics would produce an apparent (experimental) Arrhenius--like activation energy of approximately 24-25 kcal mo1-1 (see below) which is consistent with the experimental activation energies (see above). We note that the B d species generated via the above mechanism would be virtually identical to the R* generated via the light induced photobleaching sequence, and that both species would decay to form all-trans retinal and opsin. We conclude that the above two step mechanism is viable with respect to observed energetics as well as observed thermal noise photoreceptor signals. If the above mechanism is correct, and the binding site is accessible for titration from the aqueous bulk medium, then changes in extracellular pH should affect the ratio of unprotonated versus protonated chromophores within the binding site. Recent experiments on bovine rhodopsin indicate that the binding site ofrhodopsin is accessible to changes in extracellular pH (Deng et al., 1994). If the pH were decreased, then the equilibrium would favor the protonated species, and the rate of thermal activation processes would decrease. Barlow and coworkers carried out experiments on the Limulus photoreceptor system and observed a pH dependence that supported this.

The Effect of Extracellular pH on Limulus Photoreceptor Noise Photoreceptor noise exhibits a circadian rhythm in the lateral eyes of the horseshoe crab, Limulus. At night, a circadian clock in the brain transmits efferent optic-nerve activity to the lateral eyes and other visual organs of the animal (Kaplan and Barlow, 1980; Barlow et al., 1987). The efferent input reduces the rate of "noise" events generated in the dark by photoreceptors and second-order cell without affecting the events evoked by light. In photoreceptors the noise events take the form of discrete voltage signals called "quantum bumps," and in secondorder cells they are efferent optic-nerve impulses the eye transmits to the brain. Barlow and coworkers found that changes in the temperature of the retina affected both types of noise equally day and night (Barlow, Jr. et al., 1993), that is, the Arrhenius activation energies were the same day and night. During the day, an apparent activation energy of 26.3+7.8 kcal/mol was observed and at night an activation energy of 27.9+6.5 kcal/mol was observed (Barlow, Jr. et al., 1993). Because the Arrhenius activation energies do not change with time of day, one can conclude that the circadian clock does not lower photoreceptor noise at night by increasing the energy required to generate it. As noted below, this observation is consistent with the two-step mechanism of Figure 16, which exhibits an activation energy determined primarily by the second (thermal isomerization) step.

Characterization of the Primary Photochemical Events

111

Barlow and coworkers propose that the clock lowers photoreceptor noise by first lowering retinal pH which in turn reduces the small (<0.01%) population of rhodopsin molecules with unprotonated Schiff-base chromophores (Barlow, Jr. et al., 1993). Because the protonated species have a thermal barrier to isomerization of--45 kcal mo1-1 versus --22 kcal mo1-1, the former do not contribute to an observable thermal activation reaction. Lowering extracellular pH might further push the equilibrium towards the protonated species by protonating the acceptor group (A in Figure 16), which increases the energy of the unprotonated species. The two-step molecular mechanism was tested further by determining whether the clock's efferent input lowers the extracellular pH of the retina. It was noted that activating the efferent input to the retina with current shocks delivered to the optic nerve decreased retinal pH by 0.14 units (Barlow Jr. et al., 1993). Changes in retinal pH were detected in 6 out of 22 experiments. In other experiments it was found that injecting mildly acidic saline also reduces photoreceptor noise (Barlow Jr. et al., 1993). While these experiments do not prove the two-step mechanism shown in Figure 16, they are consistent with that mechanism. None of the other models for photoreceptor noise discussed above are predicted to exhibit a pH dependence.

How Nature Minimizes Intrinsic Photoreceptor Noise It is likely that Limulus has developed a method of controlling photoreceptor noise in part due to the high content of rhodopsin and the necessity to minimize photoreceptor noise during night when vision is important for finding mates. The more highly developed visual systems of land-based animals contain much less rhodopsin in the photoreceptor cells and appear to maintain low levels of dark noise at all times (see Figure 15). Indeed, rhodopsin has a very low probability of thermal activation with a rate of approximately 10-1~thermal events per rhodopsin molecule per second. The question we address in this section is how nature has optimized the binding site ofrhodopsin to minimize photoreceptor noise. We also discuss a kinetic model ofphotoreceptor noise that provides a better understanding of how photoreceptor noise can be reduced significantly without a change in activation energy (Birge, 1993; Birge and Barlow, 1994). Steinberg et al. (1993) recently measured the pK a of the protonated Schiffbase of rhodopsin to be 16 or greater by using a series of model retinal chromophores with electron-withdrawing substituents. This pK a value is significantly larger than corresponding values in model compounds or related proteins (see below). We conclude that the binding site has been designed to minimize noise by maximizing the pK a of the chromophore. The implications are best analyzed in terms of the following kinetic scheme. ka k2 R -----~ Rd --+ B d

k_l

112

JEFFREYA. STUART and ROBERT R. BIRGE

Simple reaction rate theory suggests that the total rate of this process can be described approximately by the following equation (Birge, 1993):

(-hvisom1 klk2~kT t IOAPK / 1 - e x p ~k/~_)T kt°t - k-1 - -if- 1 + 10ApI': E2 exp

(1)

where ApK = pK A - pr,, _,,ass a , pK Arepresents the pK a of the principal proton acceptor group within the protein binding site, pr~ _v-asa a represents the pK a of the protonated Schiff base chromophore, Visom is the frequency of the Cll--C12 ground state torsional mode, E 2 is the activation energy of the isomerization step, h is Planck's constant, k is Boltzmann's constant and T is the temperature. If we assume _TTPSB pr~ a = 16, pK A a =7,Visom= 300 cm-1, E 2 = 22 kcal mo1-1 (see Hubbard, 1966) and T = 310 K (body temperature) we calculate ktot = 1.5x 10-12 s-1. Given the level of approximation inherent in Eq. 1 and the above assignments, the agreement with the observed dark noise rate of--10 -11 events rhodopsin -1 s-l is encouraging. One might anticipate based on the above rate equation that the observed activation energy would be significantly higher than E2, the barrier to thermal isomerization of the unprotonated chromophore. In fact, for reasons outlined below, the observed activation energy should be only slightly larger than E 2. We can write Equation (1) in the form of a simple Arrhenius-like equation as follows:

,ot Aoexp

-E a

(-Ea"~

Ae expL j

where

/e

kT10 ApK xp Ea = E2 +

hvisom-

kT In

//

om _1

(3)

hAo(1 + 10ApK)

and

kT A° =

10ApKexp/-(Ea- EkT hVis°m)](exp(hVis°m ] kT ) - 1 h(1 +

(4)

10ApK)

note, however, that the prefactor, A o, is a function of E a and that the activation energy, Ea, is a function ofA 0. Thus, Equation (1) cannot be formally separated into "pure" prefactor and activation energy terms. By using numerical methods, however, we can solve for E a and A o, and a series of calculations over a broad range of variable assignments indicates that the value of E a never exceeds E 2 + hvisom for

We

Characterization of the Primary Photochemical Events

113

reasonable values of Visom. In fact, to a good approximation, we can equate the activation energy to E 2 + lhvisom. This assumption leads to the following set of equations" z.

1

Ea = E2 + ~hVisom

(5)

and

i Vsi°m 1-1, , jhigk+kkOPexp_hk heft--

(hvisom~

hexp~ kT j ( l + 10ApK) We now have nearly pure prefactor and activation energy terms and this decomposition allows us to examine the effective (Arrhenius-like) activation energy in a straight-forward fashion. We note that in our above example, we assumed, Visom = 300 cm-1. This equates to an energy contribution of 1/2 h Visom = 0.43 kcal mol-~. The value of Visom is poorly defined for a complex molecule such as the chromophore in rhodopsin. In particular, there is no single ground state vibrational mode which is a pure torsional distortion mode involving the C1~--C12 double bond (see for example, vibrational assignments for the chromophore in bacteriorhodopsin and related retinyl polyenes [Smith et al., 1985,a,b, 1987a,b]). Nevertheless, we can state with confidence that Visom must be less than 1,000 cm-~, which will lead to a value of 1/2 h Visom = 1.4 kcal mol-~. Thus, we can conclude that the observed activation energy ofphotoreceptor noise will be less than 1.4 kcal mol-~ larger than the barrier to isomerization of the unprotonated Schiff base chromophore. This observation is fully consistent with the experimental activation energies, which tend to be larger, but not significantly larger, than the barrier for isomerization of 11-cis retinal in solution (E a - 22 kcal mol-~). The effect of pH on the rate of photoreceptor noise is basically observed as an adjustment of the prefactor. We can write ApK approximately as PHex - - P_.,PSB ~a ' where PHex represents the extracellular pH. A decrease of 1 pH unit in extracellular pH will induce a decrease in rate of-0.1 (see Eq. 6). This will have no effect on the observed activation energy, however, which is consistent with the experimental measurements on Limulus (see above). We can offer no definitive explanation for why extracellular pH has a larger impact on photoreceptor noise in Limulus than is predicted by Equations (1) and (6) (Barlow, Jr. et al., 1993). Extracellular buffeting is one possibility, but protonation of an acceptor group (A in Figure 16) would also effectively multiply the effect. Further work will be necessary to fully resolve the apparent enhancement effect. What remains to be explained is how nature has designed the rhodopsin binding site in order to diminish intrinsic photoreceptor noise. For purposes of discussion, we can write the total rate as proportional to two factors:

114

JEFFREYA. STUART and ROBERT R. BIRGE

ktot =

k lk 2

n-1

,-,-PSB k2 lO -p a

(7)

It follows from Equation (7) that a high pK a of the protonated Schiff base chromophore is important to the biological control of photoreceptor noise. Model protonated retinyl Schiffbase chromophores in solution exhibit pK a values around 7 (Druckmann et al., 1982; Sheves et al., 1986; Sandorfy et al., 1987; Steinberg et al., 1993). The retinyl chromophore in bacteriorhodopsin, the light transducing protein in the purple membrane o f H a l o b a c t e r i u m halobium, exhibits a pK a of---13 (Dmckmann et al., 1982). An increase in pK a of six units on incorporation of the retinyl chromophore in bacteriorhodopsin is impressive. The corresponding increase of more than nine units in rhodopsin is extraordinary. Nature rarely explores the limits of physical phenomena without simultaneously improving the comparative advantage of the system. We propose that the adjustment of the pK a in rhodopsin is intimately related to the natural selection of a photoreceptor protein exhibiting minimal dark noise. For comparison, if the chromophore of rhodopsin had the same pK a as that observed in bacteriorhodopsin, the visual pigment would have a dark noise rate three to four orders of magnitude larger than observed. Further experimental and theoretical work will be required before the mechanistic origins of the high pK a of the rhodopsin chromophore can be established. Molecular orbital calculations indicate that the strength of the Schiff base N-H bond increases as the counterion is moved down the chain from the C--N group towards atom C12 (see Figure 16) (Birge, 1993). Thus, the shift in pK a appears to be associated at least in part with the specificity of the chromophore-counterion interaction. Indeed, a recent NMR study by Han et al. (1993) concludes that the major interaction of the counterion with the chromophore is at C~2. We anticipate that other amino acids near the binding site may establish a local field at the retinyl chromophore which further enhances the pK a. This proposal, however, must remain a hypothesis pending the assignment of the tertiary structure of the rhodopsin protein.

H. Parallels between Rhodopsin and Bacteriorhodopsin Photochemistry It is interesting at this point to note the parallels between the nature of the photoproducts of the bacteriorhodopsin and rhodopsin primary events. In both proteins excitation by light causes an isomerization to occur on a femtosecond time scale in what is believed to be a barrierless excited-state process. These isomerizations are very stereospecific, and occur with high quantum yields. Neither of the commonly recognized photoproducts, J or photorhodopsin, respectively, can be isolated at low temperatures, and both are characterized as having distorted chromophore configurations. Additionally, thermal relaxations lead to the formations on a picosecond time-scale of the next well characterized intermediates, K and bathorhodopsin, respectively, both of which serve to store the energy of the

Characterization of the Primary Photochemical Events

115

absorbed photon that fuels the remaining of the thermally driven transformations. Finally, there have been reports of multiple conformers for both K and bathorhodopsin. That nature should converge on almost identical mechanisms and modes of action for proteins that function for totally unrelated purposes remains to this day a mystery of evolutionary biology. This is an interesting paradox that has recently been discussed by Henderson et al. (1990b).

V. RESPONSE OF THE PROTEIN DURING THE PRIMARY EVENT Most of the spectroscopic studies of the primary event have focused on the response of the chromophore, as this is the light absorbing moiety responsible for the subsequent protein reaction. All following changes in the protein are the result of that initial isomerization, which acts to disrupt preexisting protein-chromophore interactions as its effects radiate from the isomerized chromophore to the surrounding protein. Because the isomerization occurs on a femtosecond timescale, it is expected that changes in the protein residues that are in van der Waals contact with the chromophore would occur on this time-scale as well. What is the nature of these early changes in those interactions? Several studies have addressed this question, but unfortunately, most of the studies lack the time resolution needed to address the primary event. None have yet been reported that address the bR--~J transition, although several have examined the overall bR--+K transition. It is probably safe to assume that the former transition is isolated to the chromophore, and it is not until subsequent picosecond processes (i.e., the formation of K) that distinct changes in the protein begin to be seen. Some interactions between the chromophore and opsin are certainly expected. Moreover, formation of K is accompanied by a change in the absorption maximum which is a function of both the protein and the chromophore. Each of these statements is equally valid for rhodopsin and its corresponding photoproducts. Many reports of the protein response after the formation of K (and bathorhodopsin) are available, but will not be treated herein.

A. Bacteriorhodopsin Several studies have addressed the role of Lysine-216 in the primary event. This is an obvious choice, as it is the amino acid to which the chromophore is directly bound. It is known that the chemical shift of the ~-C of Lys-216 is sensitive to the Schiffbase linkage configuration (Farrar et al., 1993); NMR, however is too slow to be able to resolve primary event processes. McMaster and Lewis were among the first to report a Lys-216 perturbation; using isotopically labeled lysine, they found evidence via FTIR difference spectroscopy for conformational changes of Lys-216 in concert with those of the chromophore induced by light absorption (McMaster and Lewis, 1988). It was suggested that the altered lysine vibrations might be associated with the lysine side chain (a four-carbon tether terminated with

116

JEFFREYA. STUART and ROBERT R. BIRGE

a nitrogen), to which the chromophore is bound. This tether is known to be somewhat flexible in the binding site, as evidenced by its ability to accommodate different retinal analogues. It has been suggested to act in such a manner during the primary event that allows the cyclohexene ring to remain fixed (Liu et al., 1991). In a follow-up study to McMaster and Lewis, Gat et al. (1992) demonstrated that a coupling of vibrational modes exists between the Lys-216 alpha-carbon and retinal that was altered upon phototransformation to K. A twist in the C15--N bond may also be involved. Finally, Fahmy and coworkers have reported a dichroic difference band in the amide I region accompanying the formation of K, that was attributed to alterations of the protein backbone induced by the isomerization. This band was due either to changes in the electrostatic environmem or to structural changes transmitted to the protein via the Lys-216 attachment (Fahmy et al., 1989). Takei et al. (1994) have recently examined conformational changes in the Lys-216 backbone during formation of the L and M intermediates. Most of the other studies have concentrated on changes in the protonation states of specific amino acids as a function of the primary event, especially Tyrosine-185. As discussed previously, assignment of tyrosine protonation states is somewhat controversial, and many conflicting results have been published. FTIR and UV difference spectroscopy studies done by Rothschild and coworkers have argued for a deprotonated Tyr-185 (tyrosinate) in bR which is protonated upon the formation of K (Dollinger et al., 1986, 1987; Rothschild et al., 1986, 1989a; Roepe et al., 1987, 1988; Ahl et al., 1988; Dunch et al., 1990). In their model ofthe binding site, the ionized Tyr- 185 is stabilized by interaction with the protonated Schiffbase (see section on the bacteriorhodopsin active site); on chromophore isomerization tyrosinate is destabilized and protonates upon the formation of K (Rothschild et al., 1989a). Contradictory evidence for tyrosinate formation in K is provided by site-directed mutagenesis studies (Jang et al., 1990), solid-state NMR (Herzfeld et al., 1990; McDermott et al., 1991), and ultraviolet resonance Raman (Ames et al., 1990, 1992). The recem UV resonance Raman studies by Ames and coworkers have found no evidence of the existence of tyrosinate in bR (Ames et al., 1990, 1992). The most recent NMR studies also argue against the existence of tyrosinate in any form, in both light- and dark-adapted bR (McDermott et al., 1991). As previous studies have indicated a possible role for Tyr- 185 in stabilizing Asp-212, it has been suggested by several authors that a strong, highly polarizable, hydrogen bond exhibiting a double-well potential minima may be present between the two residues. On formation of K, it is conceivable that the proton might shift in favor ofTyr-185, thus giving the appearance of a protonation (Rothschild et al., 1990; Ames et al., 1992). The diffraction data provided by Henderson et al., (1990a) supports the assertion of an interaction between Try- 185 and Asp-212. At this time this controversy remains unresolved; see Rothschild (1992) for a recent review. Perturbations to tryptophan residues have also been observed for the bR--~K transition, via low temperature UV and FTIR difference spectroscopies (Rothschild et al., 1986; Roepe et al., 1988b). Based upon comparison to model compounds it

Characterization of the Primary Photochemical Events

117

appears that or more tryptophan residues experience an increase in hydrogen bonding to the N-H indole functional group upon formation of K (Roepe et al., 1988b). A later site-directed mutagenesis study identified Trp-86 as the affected residue; it was found to undergo either a change in its immediate environment or in its configuration (Rothschild et al., 1989b). Apossible change in the environment of Asp-115 during the primary event has also been reported (Rothschild, 1992). Tryptophan residues at locations 182 and 189 also seem to interact, as do Tyr-185 and Asp-85, with Trp-86 during the primary event (Rothschild et al., 1989b). The most logical explanation of the observed interactions between the different tryptophan residues is their participation in the structure of the retinal binding pocket (as discussed above). If they are in van der Waals contact with the retinal chromophore, changes due to the primary event would be expected to occur. Mutant proteins W 182F and W 189F have been reported to display shifted absorption maxima (Mogi et al., 1989a,b), implying direct interaction of some sort with the chromophore. Photoelectric measurements have indicated small (1 to 4 A) charge movements toward the cytoplasm during the first 100 ps of the photocycle, which corresponds to the K, and possibly the KL and L intermediates (Keszthelyi and Ormos, 1980; Rayfield, 1983; Groma et al., 1988; Holz et al., 1988). Picosecond charge separations measured photoelectrically are most likely isolated motion of the chromophore alone (Trissl et al., 1989; Simmeth and Rayfield, 1990). Unfortunately, as alluded to earlier, this type of data can be difficult to fully interpret. Doig and coworkers observed changes in Stokes HOOP intensities between 20 and 100 ps that were attributed to isomerization induced protein relaxation processes involved in the decay of the K intermediate to form KL (Doig et al., 1991). Resonance Raman studies by Lohrmann et al. indicated the removal of a negatively charged counterion from the positively charged protonated Schiff base during the bR to K transition, and a subsequent movement of a positive charge from the surrounding protein toward the 13-ionone ring of the chromophore (during the formation of L) (Lohrmann et al., 1991; Lohrmann and Stockburger, 1992). It is not clear, however, if the charge separation was due to movement of the chromophore or the actual counterion. Picosecond infrared spectroscopic studies imply changes in the protein in less than 10 ps, with definite changes seen under 50 ps (Diller et al., 1991, 1992). The first evidence for a subpicosecond response by the protein has just been reported (no details yet available) (Diller et al., 1994).

B. Rhodopsin Much less information has been reported for the protein response during the primary event for rhodopsin, although there have been studies for later intermediates (DeGrip et al., 1988; Shichida, 1992). On the basis ofthe structural information available for rhodopsin and bathorhodopsin, one fact is very intriguing: the C--N stretching frequency remains invariant during the primary event. The implication of this is that there is little, if any, structural change in the environment of the Schiff

118

JEFFREYA. STUART and ROBERT R. BIRGE

base during the isomerization. As alluded to earlier, the nature of the energy storage mechanisms is therefore .limited to those that favor conformational interactions between protein and chromophore, as opposed to charge separation at the protonated Schiffbase (unlike bacteriorhodopsin). There is some evidence from infrared difference spectroscopy for protein related conformational changes immediately after the isomerization, upon the formation of bathorhodopsin. DeGrip et al. (1988) was able to assign differences in the IR spectra between rhodopsin and bathorhodopsin to changes in protein secondary structure, primarily in the backbone conformation. Support was also found for protonation or hydrogen bonding changes in carboxylic residues (Asp or Glu). The authors speculate that the changes might be isolated to the region of the peptide backbone that includes Lys-296, to which the chromophore is covalently attached; this region is bounded on both sides by proline residues, which are known to be helix breakers and would therefore effectively isolate the observed effects (DeGrip et al., 1988). Protonation and/or environmental changes of carboxylic groups were also seen by Bagley and coworkers, and were assigned to Glu-122 and/or Glu-134 (bovine) (Bagley et al., 1989). In addition, evidence was found for an environmental alteration or a protonation change involving a tyrosine residue. Similar types of changes were observed in the octopus rhodopsin phototransformation (Bagley et al., 1989). Other perturbations were idemified as arising from carbonyls in the peptide backbone near the retinal binding pocket that experience environmental changes during the isomerization, and differences in signals due to amide bonds were tentatively attributed to changes in the amide bond structure of Lys-296 (Bagley et al., 1989).

VI.

PROTEIN CONTROL OF ISOMERIZATION

The observation described above that isomerization is driven by steric or other forces in the binding sites of bacteriorhodopsin and rhodopsin has long been a central question in the elucidation of mechanisms of the primary events of these proteins. What forces are responsible for the large stereospecificity of the respective C13=C14 trans-to-cis and Cll--C12 cis to trans photoisomerizations? That the binding sites have been specifically engineered by nature to accommodate their chromophores cannot be disputed. Both proteins have high forward quantum yields, which are associated in part with protein-chromophore imeractions within the active site that allow for the extreme stereospecificity of the isomerizations. Consider the case of the complicated dynamics and the small potemial barrier to isomerization found for the non-naturally occurring visual pigment isorhodopsin (see above), as compared to the barrierless transition in rhodopsin that occurs threefold faster with a significantly larger quantum yield. In addition, visual opsin will not bind the all-trans chromophore, which simply does not fit in the active site. Most of the studies that address this issue directly employ site-directed mutagenesis of amino acid residues in the binding site of bacteriorhodopsin. Considerable

Characterization of the Primary Photochemical Events

119

insight into this problem has been inferred from these studies, but it is beyond the scope of this paper to review them here. More than 90 bacteriorhodopsin mutant proteins have been reported, of which under a third displayed significantly different rates of chromophore regeneration, proton pumping rates, differences in absorption maxima, and/or different compositions of retinal isomer ratios. These mutated residues are therefore generally assumed to be part of the retinal binding site. Several recent studies, however, are worth mentioning. In a study by Greenhalgh et al. (1993) a number of hydrophobic amino acid site-directed mutations of the binding site were examined, several of which succeeded in affecting the ratio of 13-c/s- to all-trans-retinal isomers in the dark- and light-adapted states, as well as a formation of nonnative 11-c/s- and 9-c/s-retinal isomers. This is direct evidence for site-specific control of isomerization by the protein. Indeed, part of the retinal binding site consists of a hydrophobic pocket formed by tryptophan and tyrosine residues (Mogi et al., 1989a,b; Rothschild et al., 1989a) that are thought to play a valuable role in preserving the stereospecific nature of the isomerization. Replacement of a number of the residues involved in this binding pocket (especially tryptophans) has been reported to lead to the formation of new chromophore isomers, implying that they provide an active site structure that prevents photoisomerization at any other location than the C13--C14 bond (Ahl et al., 1987, 1988; Mogi et al., 1989a,b; Rothschild et al., 1989b). Song et al. (1993)examined the role of protein catalysis in the bacteriorhodopsin primary event by employing mutants exhibiting different ionic environments in the active site. Charged residues (Asp-85, -212, Arg-82) were found to be the only ones to catalyze photoisomerization; substitution of neutral residues for Tyr-185 and Asp-115 in the wild type protein was found to have little effect, whereas replacement of Asp-85, -212, or Arg-82 with neutral residues resulted in a drastic decrease in the photoisomerization rate. The Asp-212 mutant also exhibited an increased retinal excited-state lifetime, perhaps due to its inability to participate in the hydrogen bonding network thought to exist in the vicinity ofthis residue. Arole in stabilization of Asp-85 was postulated for Arg-82. It seems that in the excited state, the negatively charged Asp residues serve to stabilize the positive charge of C13, as well as that of the protonated Schiff base in the ground state. Excited state ionic stabilization of C13 is expected to decrease the double bond character of the C13=C14 bond, and therefore the barrier to photoisomerization at this location. In addition, the two Asp residues might also serve to reduce the rates of thermal isomerization in the dark (Song et al., 1993). The results indicated that the three charged amino acids mentioned above have a strong catalytic function in the retinal photoisomerization via charge stabilization processes in the ground and excited states; the location of negative charges, as well as variations in the local dielectric constant, may act to exert a specific and selective control over the rate ofphotoisomerization in bacteriorhodopsin (Song et al., 1993). Retinal extraction has also been a useful method of determining the control of the apoprotein upon isomerization. Isomerization in bacteriorhodopsin always occurs at the C13=C14 bond; no other isomers have ever been identified. However,

120

JEFFREY A. STUART and ROBERT R. BIRGE

the protonated retinal Schiff base in solution, seldom, if ever, undergoes the all-trans to 13-cis isomerization (Koyama et al., 1993). Photostationary state experiments typically yield compositions that are predominantly all-trans and 11-cis (in about the same amounts) with small amounts of 9-cis and 13-cis: a situation very different from light adapted bR (all-trans with a trace of 13-cis) (Koyama et al., 1993). This lead Koyama et al. (1993) to examine the question of whether the 11-cis isomer was ever found in bR, either as a result of different environmental or excitation conditions, or if it is simply not allowed sterically in the binding site (Koyama et al., 1993). Only in extreme conditions of high pH or photon density were the 11-cis (and 9-cis in lesser concentrations) isomers seen in the extracts; at physiological conditions, the 11-cis configuration was sterically forbidden by the apoprotein, indicating its strong influence on the chromophore (Koyama et al., 1993). Finally, it must also be pointed out that not only is the active site of bacteriorhodopsin specifically engineered to facilitate the initial photoisomerization, but it must also facilitate the re-isomerization back to the all-trans chromophore geometry during the subsequent thermal transformations.

ACKNOWLEDGMENTS The authors thank Drs. G.H. Atkinson, R.B. Barlow, R.A. Bogomolni, R.H. Callender, A. Cooper, T.G. Dewey, T.G. Ebrey, M.A. E1-Sayed, K.W. Foster, J. Herzfeld, B. Honig, H.G. Khorana, D.S. Kliger, B. Knox, R.S.H. Liu, T. Marinetti, R.A. Mathies, D. Oesterhelt, M. Ottolenghi, R.R. Rando, K.J. Rothschild, T.P. Sakmar, M. Sheves, Y. Shichida, W. Stoeckenius, and A. Watts for interesting and helpful discussions and for providing manuscripts prior to publication. Acknowledgment does not represent endorsement by the above scientists of the analyses presented in this review. Research in our laboratory was funded in part by grants to R.R.B. from the National Institutes of Health, USAF Rome Laboratory, New York State Center for Advanced Technology in Computer Applications and Software Engineering and the W.M. Keck Foundation.

VIII.

REFERENCES

Ahl, P. L., Stem, L. J., Hackett,N. R., Rothschild, K. J., & Khorana, H. G. (1987). The effects ofaromatic amino acid substitutions in the putative helix F on the bacteriorhodopsin photocycle. Biophys. J. 51,416 (abstr.). Ahl, P. L., Stem, L. J., DOting, D., Mogi, T., Khorana, H. G., & Rothschild, K. J. (1988). Effects of amino acid substitutions in the F helix ofbacteriorhodopsin. Low temperature ultraviolet/visible difference spectroscopy. J. Biol. Chem. 263, 13594-13601. Aho, A. C., Donner, K., Hyden, C., Larsen, L. O., & Reuter, T. (1988). Low retinal noise in animals with low body temperature allows high visual sensitivity. Nature 334, 348-350. Albeck, A., Friedman, N., Sheves, M., & Ottolenghi, M. (1986). Role of retinal isomerizations and rotations in the photocycle of bacteriorhodopsin. J. Am. Chem. Soc. 108, 4614--4618. Albeck, A., Friedman, N., Ottolenghi, M., Sheves, M., Einterz, C. M., Hug, S. J., Lewis, J. W., & Kliger, D. S. (1989). Photolysis intermediates of the artificial visual pigment cis-5, 6-dihydro-isorhodopsin. Biophys. J. 55, 233-241.

Characterization of the Primary Photochemical Events

121

Albeck, A., Livnah, N., Gottlieb, H., & Sheves, M. (1992). 13C NMR studies of model compounds for bacteriorhodopsin: factors affecting the retinal chromophore chemical shifts and absorption maximum. J. Am. Chem. 114, 2400-2411. Albert, I. D. L., & Ramasesha, S. (1990). Sudden polarization in push-pull polyenes; a model exact study. J. Phys. Chem. 94, 6540-6543. Ames, J. B., Fodor, S. E A., Gebhard, R., Raap, J., van den Berg, E. M. M., Lugtenburg, J., & Mathies, R. A. (1989). Bacteriorhodopsin's M412 intermediate contains a 13-cis, 14-s-trans, 15-anti-retinal Schiff base chromophore. Biochemistry 28, 3681-3687. Ames, J. B., Bolton, S. R., Netto, M. M., & Mathies, R. A. (1990). Ultraviolet resonance raman spectroscopy of bacteriorhodopsin: evidence against tyrosinate in the photocycle. J. Am. Chem. Soc. 112, 9007-9009. Ames, J. B., Ros, M., Raap, J., Lugtenburg, J., & Mathies, R. A. (1992). Time-resolved ultraviolet resonance Raman studies of protein structure: applications to bacteriorhodopsin. Biochemistry 31, 5328-5334. Applebury, M. L., Peters, K. S., & Rentzepis, R. (1978). Primary intermediates in the photochemical cycle ofbacteriorhodopsin. Biophys. J. 23,375-382. Atkinson, G. H., Blanchard, D., & Brack, T. L. (1991). Fluorescence from the primary products of the bacteriorhodopsin photocycle. J. Lumin. 48, 49, 410-414. Atkinson, G. H., Blanchard, D., Lemaire, H., Brach, T. L., & Harashi, H. (1989). Picosecond time-resolved fluorescence spectroscopy of K-590 in the bacteriorhodopsin photocycle. Biophys. J. 55, 263-274. Atkinson, G. H., Brack, T. L., Grieger, I., Rumbles, G., Blanchard, D., & Siemenkowski, L. (1986). Picosecond time-resolved resonance Raman spectroscopy of bacteriorhodopsin intermediates. Proc. SPIE-Int. Soc. Opt. Eng. 620, 82-88. Aton, B., Doukas, A. G., Callender, R. H., Becher, B., & Ebrey, T. G. (1977). Resonance raman studies of the purple membrane. Biochemistry 16, 2995-2999. Baasov, T., & Sheves, M. (1985). On the absorption maxima of protonated retinal Schiff bases. An interaction with external charges. Israel J. Chem. 25, 53-55. Baasov, T., Friedman, N., & Sheves, M. (1987a). Factors affecting the C=N stretching in protonated retinal schiff base: a model study for bacteriorhodopsin and visual pigments. Biochemistry 26, 3210-3217. Baasov, T., Friedman, N., & Sheves, M. (1987b). On the C=N stretching frequency ofprotonated retinal schiffbases. In T. G. Ebrey, H. Frauenfelder, B. Honig, & K. Nakanishi (Eds.), Biophysical studies of retinal proteins (pp. 252-261). University of Illinois Press. Bagley, K., Dollinger, G., Eisenstein, L., Singh, A. K., & Zimanyi, L. (1982). Fourier-transform infrared difference spectroscopy of bacteriorhodopsin and its photoproducts. Proc. Natl. Acad. Sci. USA 79, 4972-4976. Bagley, K., Balogh-Nair, V., Croteau, A. A., Dollinger, G., Ebrey, T. G., Eisenstein, L., Hong, M. K., Nakanishi, K., & Vittitow, J. (1985). Fourier-transform infrared difference spectroscopy of rhodopsin and its photoproducts at low temperature. Biochemistry 24, 6055-6071. Bagley, K. A., Eisenstein, L., & Ebrey, T. G. (1989). A comparative study of the infrared difference spectra for octopus and bovine rhodopsins and their bathorhodopsin photointermediates. Biochemistry 28, 3366-3373. Balashov, S. P., Govindjee, R., & Ebrey, T. G. (199 l a). Red shift of the purple membrane absorption band and the deprotonation of tyrosine residues at high pH. Biophys. J. 60, 475-490. Balashov, S. P., Govindgee, R., Imasheva, E., Misra, S., Ebrey, T. G., Feng, Y., Crouch, R. K., & Menick, D. R. (1994). Substitution of Arg82 with lysine in bacteriorhodopsin: effects on thermal isomerization, photocycle, and proton release. Biophys. J. 66, A45. Balashov, S. P., Imasheva, E. S., Govindjee, R., & Ebrey, T. G. (1991b). Quantum yield ratio of the forward and back light reactions of bacteriorhodopsin at low temperature and photosteady-state concentration of the bathoproduct K. Photochem. Photobiol. 54, 955-961.

122

JEFFREY A. STUART and ROBERT R. BIRGE

Balashov, S. P., Kameyeva, N. V., Litvin, F. E, & Ebrey, T. G. (1991 c). Bathoproducts and conformers of all-trans- and 13-cis-bacteriorhodopsin at 90 K. Photochem. Photobiol. 54, 949-953. Barlow, H. B. (1988). The thermal limit to seeing. Nature 334, 296-350. Barlow, Jr., R. B., & Kaplan, E. (1989). What is the origin ofphotoreceptor noise. Biol. Bull. 177, 323. Barlow, Jr., R. B., & Silbaugh, T. H. (1989). Is photoreceptor noise caused by thermal isomerization of rhodopsin. Invest. Ophthalmol. Vis. Sci. Suppl. 30, 61. Barlow, Jr., R. B., Kaplan, E., Renninger, G. H., & Saito, T. (1987). Cicadian rhythms in Limulus photoreceptors I. Intracellular studies. J. Gen Physiol. 89, 353-378. Barlow, Jr., R. B., Birge, R. R., Kaplan, E., & Tallent, J. R. (1993). On the molecular origin of photoreceptor noise. Nature 366, 64--66. Bashford, D., & Gerwert, K. (1992). Electrostatic calculations of the pK a values ofionizable groups in bacteriorhodopsin. J. Mol. Biol. 224, 473-486. Baylor, D. A., Matthews, G., & Yau, K. W. (1980). Two components of electrical dark noise in toad retinal rod outer segments. J. Physiol. 309, 591-621. Baylor, D. A., Nunn, B. J., & Schnapf, J. L. (1984). J. Physiol. 357, 575-607. Becher, B., & Ebrey, T. G. (1977). The quantum efficiency for the photochemical conversion of the purple membrane protein. Biophys. J. 17, 185-191. Bibikov, S. I., Grishanin, R. N., Marwan, W., Oesterhelt, D., & Skulachev, V. P. (1991). The proton pump bacteriorhodopsin is a photoreceptor for signal transduction in Halobacterium halobium. FEBS Lett. 295,223-226. Birge, R. R. (1981). Photophysics of light transduction in rhodopsin and bacteriorhodopsin. Ann. Rev. Biophys. Bioeng. 10, 315-354. Birge, R. R. (1982). Simulation of the primary event in rhodopsin photochemistry using semiempirical molecular dynamics theory. In R. R. Alfano (Ed.), Biological events probed by ultrafast laser spectroscopy (pp. 299-317). New York: Academic Press. Birge, R. R. (1986). Two-photon spectroscopy of protein-bound chromophores. Ace. Chem. Res. 19, 138-146. Birge, R. R. (1990a). Nature of the primary photochemical events in rhodopsin and bacteriorhodopsin. Biochim. Biophys. Acta 1016, 293-327. Birge, R. R. (1990b). Photophysics and molecular electronic applications of the rhodopsins-. Annu. Rev. Phys. Chem. 41,683-733. Birge, R. R. (1993). The unusual pK a of the rhodopsin chromophore. Is this how nature minimizes photoreceptor noise? Biophys. J. 64, 1371-1372. Birge, R. R., & Barlow, R. B. (1994). On the molecular origins of thermal noise in vertebrate and invertebrate photoreceptors. Biophys. Chem. (in press). Birge, R. R., Bocian, D. E, & Hubbard, L. M. (1982). Origins of inhomogeneous broadening in the vibronic spectra of visual chromophores and visual pigments. J. Am. Chem. Soc. 104, 1196-1207. Birge, R. R., & Callender, R. H. (1987). Quantum efficiencies of primary photochemical processes in vertebrate rhodopsin. In T. Ebrey, H. Frauenfelder, B. Honig, & K. Nakanishi (Eds.), Biophysical studies of retinal proteins (pp. 270-281). University of Illinois Press. Birge, R. R., & Cooper, 1". M. (1983). Energy storage in the primary step of the photocycle of bacteriorhodopsin. Biophys. J. 42, 61-69. Birge, R. R., Cooper, T. M., Lawrence, A. E, Masthay, M. B., Vasilakis, C., Zhang, C. E, & Zidovetzki, R. (1989a). A spectroscopic, photocalorimetric and theoretical investigation of the quantum efficiency of the primary event in bacteriorhodopsin. J. Am. Chem. Soc. 111, 4063-4074. Birge, R. R., Cooper, T. M., Lawrence, A. E, Masthay, M. B., Zhang, C. E, & Zidovetzki, R. (1991). Revised assignment of energy storage in the primary photochemical event in bacteriorhodopsin. J. Am. Chem. Soc. 113, 4327-4328. Birge, R. R., Einterz, C. M., Knapp, H. M., & Murray, L. P. (1988). The nature of the primary photochemical events in rhodopsin and isorhodopsin. Biophys. J. 53, 367-385.

Characterization of the Primary Photochemical Events

123

Birge, R. R., Findsen, L. A., Lawrence, A. F., Masthay, M. B., & Zhang, C. F. (1989b). Origins of the quantum efficiency duality in the primary photochemical event of bacteriorhodopsin. In R. R. Birge & H. H. Mantsch (Eds.), Biomolecular spectroscopy (pp. 103-112). Bellingham, Washington: The International Society for Optical Engineering. Birge, R. R., Findsen, L. A., & Pierce, B. M. (1987). Molecular dynamics of the primary photochemical event in bacteriorhodopsin. Theoretical evidence for an excited singlet state assignment for the J intermediate. J. Am. Chem. Soc. 109, 5041-5043. Birge, R. R., & Hubbard, L. M. (1980). Molecular dynamics of cis-trans isomerization in rhodopsin. J. Am. Chem. Soc. 102, 2195-2204. Birge, R. R., & Hubbard, L. M. (1981). Molecular dynamics of trans-cis isomerization in bathorhodopsin. Biophys. J. 34, 517-534. Birge, R. R., Murray, L. P., Pierce, B. M., Akita, H., Balogh-Nair, V., Findsen, L. A., & Nakanishi, K. (1985). Two-photon spectroscopy of locked-11-cis rhodopsin: Evidence for a protonated Schiff base in a neutral protein binding site. Proc. Natl. Acad. Sci. USA 82, 4117-4121. Birge, R. R., Murray, L. P., Zidovetzki, R., & Knapp, H. M. (1987). Two-photon, 13C and two-dimensional 1H NMR spectroscopic studies of retinal schiff bases, protonated schiff bases, and schiff base salts: Evidence for a protonation induced ~x excited state level ordering reversal. J. Am. Chem. Soc. 109, 2090-2101. Birge, R. R., & Pierce, B. M. (1983). The nature of the primary photochemical events in bacteriorhodopsin and rhodopsin. In A. H. Zewail (Ed.), Photochemistry and Photobiology (pp. 841-855). Harwood Academic. Birge, R. R., & Zhang, C. F. (1990a). Two-photon double resonance spectroscopy ofbacteriorhodopsin. ~ and d'po ~ lar pr operties of the low-lying 1 Ag, --like and 1 Bu*+ -like x, Ass'gnme nt of th e e 1ectr on'c states. J. Chem. Phys. 92, 7178-7195. Birge, R. R., & Zhang, C. F. (1990b). Two-photon spectroscopy of light adapted bacteriorhodopsin. J. Chem. Phys. 92, 7178-7195. Blanchard, D., Gilmore, D. A., Brack, T. L., Lemaire, H., Hughes, D., & Atkinson, G. H. (1991). Picosecond time-resolved absorption and fluorescence in the bacteriorhodopsin photocycle: Vibrationally-excited species. Chem. Phys. 154, 155-170. Blatz, P. E., & Mohler, J. H. (1972). Effect of selected hydrogen bonding solvents on the absorption maxima of n-retinylidene-n-butylammonium salts. Biochemistry 11, 3240--3242. Boehm, M. F., Gawinowicz, M. A., Foucault, A., Derguini, F., & Nakanishi, K. (1990). Photoaffinity . . . . . 3 • labehng studies ofbacterlorhodopsm with [ 15- H]-3-dlazo-4-keto-all-trans-retmal. J. Am. Chem. Soc. 112, 7779-7782. Bogomolni, R. A., Hwang, S.-B., Tseng, T.-W., King, G. I., & Stoeckenius, W. (1977). Orientation of the bacteriorhodopsin transition dipole. Biophys. J. 17, 98a (abstr.). Bogomolni, R. A., Baker, R. A., Lozier, R. H., & Stoeckenius, W. (1980). Action spectrum and quantum efficiency for proton pumping in Halobacterium Halobium. Biochemistry 19, 2152-2159. Boucher, F., & Leblanc, R. M. (1985). Energy storage in the primary photoreaction of bovine rhodopsin. A photoacoustic study. Photochem. Photobiol. 41,459-465. Brack, T. L., & Atkinson, G. H. (1989). Picosecond time-resolved resonance raman spectrum of the K-590 intermediate in the room temperature bacteriorhodopsin photocycle. J. Molec. Struct. 214, 289-303. Braiman, M., & Mathies, R. (1980). Resonance raman evidence for an all-trans to 13-cis isomerization in the proton-pumping cycle of bacteriorhodopsin. Biochemistry 19, 5421-5428. Braiman, M., & Mathies, R. (1982). Resonance raman spectra of bacteriorhodopsin's primary photoproduct for a distorted 13-cis retinal chromophore. Proc. Natl. Acad. Sci. USA 79, 403-407. Braiman, M. S., Mogi, T., Marti, T., Stem, L. J., Khorana, H. G., & Rothschild, K. J. (1988). Vibrational spectroscopy of bacteriorhodopsin mutants: Light-driven proton transport involves protonation changes ofaspartic acid residues 85, 96, and 212. Biochemistry 27, 8516-8520.

124

JEFFREYA. STUART and ROBERT R. BIRGE

Busch, G. E., Applebury, M. L., Lamola, A. A., & Rentzepis, P. M. (1972). Formation and decay of prelumirhodopsin at room temperature. Proc. Natl. Acad. Sci. USA 69, 2802-2806. Callender, R. H., & Honig, B. (1977). Resonance Raman studies of visual pigments. Ann. Rev. Biophys. Bioeng. 6, 33-55. Callender, R. H., Doukas, A., Crouch, R., & Nakanishi, K. (1976). Molecular flow resonance Raman effect from retinal and rhodopsin. Biochemistry 15, 1621-1629. Cao, Y., V~6, G., Chang, M., Ni, B., Needleman, R., & Lanyi, J. K. ( 1991). Water is required for proton transfer from aspartate-96 to the bacteriorhodopsin schiffbase. Biochemistry 30, 10972-10979. Ceska, T. A., Henderson, R., Baldwin, J. M., Zemlin, E, Beckmann, E., & Downing, K. (1992). An atomic model for the structure of bacteriorhodopsin, a seven-helix membrane protein. Acta. Physiol. Scand. 146, 31-40. Chabre, M. (1975). X-ray diffraction studies of retinal rods I structure of the disc membrane, effect of illumination. Biochim. Biophys. Acta 382, 322-335. Chabre, M., & Worcester, D. L. (1982). X-ray and neutron diffraction of retinal rod outer segments. In L. Packer (Ed.), Methods in Enzymology (pp. 593-604). New York: Academic Press. Chabre, M., Breton, J., Michel-Villaz, M., & Saibil, H. (1982). Linear dichroism studies in the visible, UV, and IR on oriented rod suspensions. In L. Packer (Ed.), Methods in Enzymology (pp. 605-616). New York: Academic Press. Chen, A.-H., Derguini, E, Franklin, P., Hu, S., Nakanishi, K., Silvo, B. R., & Wang, J. (1992). The triggering process of visual transduction. In J.-L. Rigaud (Ed.), Colloques INSERM (Structures and functions of retinal proteins), 221 (pp. 243-246). Dourdan, France: John Libbey and Company Ltd. Clark, N. A., Rothschild, K. J., Luippold, D. A., & Simon, B. A. (1980). Surface-induced lamellar orientation of multilayer membrane arrays: Theoretical analysis and a new method with applications to purple membrane fragments. Biophys. J. 3 l, 65-96. Cooper, A. (1979a). Energies of rhodopsin and isorhodopsin. FEBS Lett. 100, 382-384. Cooper, A. (1979b). Energy uptake in the first step of visual excitation. Nature 282, 531-533. Cooper, T. M., Schmidt, H. H., Murray, L. P., & Birge, R. R. (1984). Theory and design of a tunable pulsed dye laser cryogenic photocalorimeter. Rev. Sci. Instrm. 55, 896-904. Copi6, V., McDermott, A. E., Beshah, K., Williams, J. C., Spijker-Assink, M., Gebhard, R., Lugtenburg, J. L., Herzfeld, J., & Griffin, R. G. (1994). Deuterium solid-state nuclear magnetic resonance studies of methyl group dynamics in bacteriorhodopsin and retinal model compounds: evidence for a 6-s-trans chromophore in the protein. Biochemistry 33, 3280-3286. Comsweet, T. N. (1970). Visual Perception. New York: Academic. Crouch, R., Purvin, V., Nakanishi, K., & Ebrey, T. (1975). Isorhodopsin II: Artificial photosensitive pigment formed from 9, 13-dicis retinal. Proc. Natl. Acad. Sci. USA 72, 1538-1542. Cruezet, F., McDermott, A., Gebhard, R., van der Hoef, K., Spijker-Assink, M. B., Herzfeld, J., Lugtenburg, J., Levitt, M. H., & Griffin, R. G. (1991). Determination of membrane protein structure by rotational resonance NMR: Bacteriorhodopsin. Sci. 251,783-786. DancstY,izy, Z., Govindjee, R., & Ebrey, T. G. (1988). Independent photocycles of the spectrally distinct forms ofbacteriorhodopsin. Proc. Natl. Acad. Sci. USA 85, 6358-6361. Dartnall, H. J. A. (1972). Photosensitivity. In H. J. A. Dartnall (Ed.), Handbook of sensory physiology (pp. 122-145). Springer-Verlag, Heidelberg. Das, E K., Kogan, G., & Becker, R. S. (1979). Spectroscopy of polyenes III. Absorption and emission spectral investigation of polyene Schiff bases and protonated Schiff bases related to visual pigments. Photochem. Photobiol. 30, 689-695. de Groot, H. J. M., Harbison, G. S., Herzfeld, J., & Griffin, R. G. (1989). Nuclear magnetic resonance study of the Schiff base in bacteriorhodopsin: counterion effects on the 15N shift anisotropy. Biochemistry 28, 3346-3353. DeGrip, W. J., Gray, D., Gillespie, J., Bovee, P. H. M., Van den Berg, E. M. M., Lugtenburg, J., & Rothschild, K. J. (1988). Photoexcitation ofrhodopsin: conformation changes in the chromophore,

Characterization of the Primary Photochemical Events

125

protein and associated lipids as determined by FTIR difference spectroscopy. Photochem. Photobiol. 48, 497-504. Deng, H., & Callender, R. H. (1987). A study of the Schiff base mode in bovine rhodopsin and bathorhodopsin. Biochemistry 26, 7418--7426. Deng, H., Manor, D., Weng, G., Rath, P., Koutalos, Y., Ebrey, T., Gebhard, R., Lugtenburg, J., Tsuda, M., & Callender, R. H. (1991a). Resonance raman studies of the HOOP modes in octopus bathorhodopsin with deuterium-labeled retinal chromophores. Biochemistry 30, 4495-4502. Deng, H., Manor, D., Weng, G., Rath, P., Koutalos, Y., Ebrey, T., Gebhard, R., Lugtenburg, J., Tsuda, M., & Callender, R. H. (1991b). A resonance Raman study of octopus bathorhodopsin with deuterium labeled retinal chromophores. Photochem. Photobiol. 54, 1001-1007. Deng, H., Haung, L., Callender, R., & Ebrey, T. (1994). Evidence for a bound water molecule next to the retinal Schiffbase in bacteriorhodopsin and rhodopsin: a resonance Raman study of the Schiff base hydrogen/deuterium exchange. Biophys. J. 66, 1129-1136. Dexheimer, S. L., Wang, Q., Peteanu, L. A., Pollard, W. T., Mathies, R. A., & Shank, C. V. (1992). Femptosecond impulsive excitation of nonstationary vibrational states in bacteriorhodopsin. Chem. Phys. Lett. 188, 61-66. Diller, R., & Stockburger, M. (1988). Kinetic resonance Raman studies reveal different conformational states ofbacteriorhodospin. Biochemistry 27, 7641-7651. Diller, R., Iannone, M., Bogomolni, R., & Hochstrasser, R. M. (1991). Ultrafast infrared spectroscopy ofbacteriorhodopsin. Biophys. J. 60, 286-289. Diller, R., Iannone, M., Cowen, B. R., Maiti, S., Bogomolni, R. A., & Hochstrasser, R. M. (1992). Picosecond dymanics of bacteriorhodopsin, probed by time-resolved infrared spectroscopy. Biochemistry 31, 5567-5572. Diller, R., Walker, G. C., Maiti, k. S., Cowen, B. R., Pippenger, R. S., Bogomolni, R., & Hochstrasser, R. M. (1994). The J-K transition in bacteriorhodopsin, studied by femtosecond time-resolved infrared-spectroscopy. Biophys. J. 66, A 113. Dinur, U., Honig, B., & Ottolenghi, M. (1981). Analysis of primary photochemical processes in bacteriorhodopsin. Photochem. Photobiol. 33,523-527. Dioumaev, A. K., Savroansky, V. V., Tkachemko, N. V., & Chukharev, V. I. (1989). Quantum yield and extinction measurements in strongly overlapping reactant and photoproduct absorption bands II. Bathointermediate formation in bacteriorhodopsin photocycle at room temperature. J. Photochem. Photobiol. B. 3, 397-410. Dobler, J., Zinth, W., Kaiser, W., & Oesterhelt, D. (1988). Excited-state reaction dynamics ofbacteriorhodopsin studied by femtosecond spectroscopy. Chem. Phys. Lett. 144, 215-220. Doig, S. J., Reid, P. J., & Mathies, R. A. (1991). Picosecond time-resolved resonance raman spectroscopy of bacteriorhodopsin's J, K, and KL intermediates. J. Phys. Chem. 95, 6372-6379. Dollinger, G., Eisenstein, L., Lin, S.-L., Nakanishi, K., & Termini, J. (1986). Fourier transform infrared difference spectroscopy of bacteriorhodopsin and its photoproducts regenerated with deuterated tyrosine. Biochemistry 25, 6524--6533. Dollinger, G., Eisenstein, L., Lin, S. L., Nakanishi, K., & Termini, J. (1987). The role of tyrosine in the proton pump of bacteriorhodopsin. In T. G. Ebrey, H. Frauenfelder, B. Honig, & K. Nakanishi (Eds.), Biophysical studies of retinal proteins (pp. 120-125). University of Illinois Press. Doukas, A. G., Junnarkar, M. R., Alfano, R. R., Callender, R. H., Kokitani, T., & Honig, B. (1984). Fluorescence quantum yield of visual pigments: Evidence for subpicosecond isomerization rates. Proc. Natl. Acad. Sci. USA 81, 4790-4794. Doukas, A. G., Junnarkar, M. R., Alfano, R. R., Callender, R. H., & Balogh-Nair, V. (1985). The primary event in vision investigated by time resolved fluorescence spectroscopy. Biophys. J. 47, 795-798. Druckmann, S., Ottolenghi, M., Pande, A., Pande, J., & Callender, R. H. (1982). Acid-base equilibrium of the Schiff base in bacteriorhodopsin. Biochemistry 21, 4953-4959.

126

JEFFREYA. STUART and ROBERT R. BIRGE

Du, M., & Fleming, G. R. (1993). Femtosecond time-resolved fluorescence spectroscopy of bacteriorhodopsin: direct observation of the excited state dynamics in the primary step of the proton pump cycle. Biophys. Chem. 48, 101-111. Dufiach, M., Berkowitz, S., Marti, T., He, Y.-W., Subramanian, S., Khorana, H. G., & Rothschild, K. J. (1990). Ultraviolet-visible transient spectroscopy of bacteriorhdopsin mutants. Evidence for two forms oftyrosine-185 ~ phenylalanine. J. Biol. Chem. 265 (28), 16978-16984. Earnest, T. N., Roepe, P., Braiman, M. S., Gillespie, J., & Rothschild, K. J. (1986). Orientation of the bacteriorhodopsin chromophore probed by polarized fourier transform infrared difference spectroscopy. Biochemistry 25, 7793-7798. Ebrey, T. G. (1993). Light energy transduction in bacteriorhodopsin. In M. B. Jackson (Ed.), Thermodynamics of membrane receptors and channels (pp. 353-387). Boca Raton: CRC Press. Einterz, C. M., Lewis, J. W., & Kliger, D. S. (1987a). Spectral and kinetic evidence for the existence of two forms ofbathorhodopsin. Proc. Natl. Acad. Sci. USA 84, 3699-3703. Einterz, C. M., Lewis, J. W., & Kliger, D. S. (1987b). Two forms of bathorhodopsin: Laser power dependence. In T. G. Ebrey, H. Frauenfelder, B. Honig, & K. Nakanishi (Eds.), Biophysical studies of retinal proteins (pp. 282-286). University of Illinois Press. Eisfeld, W., Pusch, C., Diller, R., Lohrmann, R., & Stockburger, M. (1993). Resonance Raman and optical transient studies on the light-induced proton pump of bacteriorhodopsin reveal parallel photocycles. Biochemistry 32, 7196-7215. Eyring, G., & Mathies, R. (1979). Resonance Raman studies of bathorhodopsin: Evidence for a protonated Schiffbase linkage. Proc. Natl. Acad. Sci. USA 76, 33-37. Fahmy, K., & Sakmar, T. P. (1993). Light-dependent transducin activation by an ultraviolet-absorbing rhodopsin mutant. Biochemistry 32, 9165--9171. Fahmy, K., Siebert, F., Grol3jean, M. F., & Tavan, P. (1989). Photoisomerization in bacteriorhodopsin studied by FTIR, linear dichroism and photoselection experiments combined with quantum chemical theoretical analysis. J. Molec. Struct. 214, 257-288. Fahmy, K., Siebert, F., & Tavan, P. (1991). Structural investigation ofbacteriorhodopsin and some of its photoproducts by polarized fourier transform infrared spectroscopic methods-difference spectroscopy and photoselection. Biophys. J. 60, 989-1001. Fang, J. M., Carriker, J. D., Balogh-Nair, V., & Nakanishi, K. (1983). Evidence for the necessity of double-bond (13-Ene) isomerization in the proton pumping of bacteriorhodopsin. J. Am. Chem. Soc. 105, 5162-5164. Farrar, M. R., Lakshmi, K. V., Smith, S. O., Brown, R. S., Raap, J., Lugtenburg, J., Griffin, R. G., & Herzfeld, J. (1993). Solid state NMR study of [e-13C]Lys-bacteriorhodopsin: Schiffbase photoisomerization. Biophys. J. 65, 310-315. Fodor, S. P. A., Ames, J. A., Gebhard, R., van den Ber, E. M. M., Stoechenius, W., Lugtenburg, J., & Mathies, R. A. (1988a). Chromophore structure in bacteriorhodopsin's N intermediate: Implications for the proton-pumping mechanism. Biochemistry 27, 7097-7101. Fodor, S. P. A., Pollard, W. T., Gebhard, R., van den Berg, E. M. M., Lugtenburg, J., & Mathies, R. A. (1988b). Bacteriorhdopsin's L550 intermediate contains a C 14--C 15 s-trans-retinal chromophore. Proc. Natl. Acad. Sci. USA 85, 2156-2160. Foster, K. W., & Saranak, J. (1988). Rhodopsin activation: A novel view suggested by in vivo Chlamydomonas experiments. J. Am. Chem. Soc. 110, 6588--6589. Foster, K. W., Saranak, J., & Zarrilli, G. (1988). Autoregulation of rhodopsin synthesis in Chlamydomonas reinhardtii. Proc. Natl. Acad. Sci. USA 85, 6379--6383. Foster, K. W., Saranak, J., Derguini, F., Zarrilli, G. R., Johnson, R., Okabe, M., & Nakanishi, K. (1989). Activation ofChlamydomonas rhodopsin in vivo does not require isomerization retinal. Biochemistry 28, 819-824. Freedman, K., & Becker, R. S. (1986). Comprehensive investigation of the photoisomerization of the protonated and unprotonated Schiff bases of 9-cis, 11-cis, 13-cis, and all-trans retinals. J. Am. Chem. Soc. 108, 1245-1251.

Characterization of the Primary Photochemical Events

127

Freedman, K., Becker, R. S., Hannak, D., & Bayer, E. (1986). Investigation into the spectroscopy and photoisomerization of a series of poly(ethyleneglycol) peptide Schiff bases of 11-cis retinal. Photochem. Photobiol. 43, 291-295. Fukuda, Y., Okano, T., Shichida, Y., Yoshizawa, T., Trehan, A., Mead, D., Denny, M., Asato, A., & Liu, R. S. H. (1990). Comparative study on the chromophore binding sites of rod and red-sensitive cone visual pigments by use of synthetic retinal isomers and analogues. Biochemistry 29, 3133-3140. Ganter, U. M., Gartner, W., & Siebert, F. (1988a). Rhodopsin-lumirhodopsin phototransition of bovine rhodopsin investigated by Fourier transform infrared difference spectroscopy. Biochemistry 27, 7480-7488. Ganter, U. M., Schmid, E. D., & Siebert, F. (1988b). The photoreaction of vacuum-dried rhodopsin at low temperature: evidence for charge stabilization by water. J. Photochem. Photobiol. B. 2, 417-426. Garcia-Quintana, D., Garriga, P., & Manyosa, J. (1993). Quantitative characterization of the structure ofrhodopsin in disc membrane by means of fourier transform infrared spectroscopy. J. Biol. Chem. 268, 2403-2409. G/irmer, W., Ullrich, D., & Vogt, K. (1991). Quantum yield of chapso-solubilized rhodopsin and 3-hydroxy retinal containing bovine opsin. Photochem. Photobiol. 54, 1047-1055. Gat, Y., & Sheves, M. (1993). A mechanism for controlling the pK a of the retinal protonated Schiffbase in retinal proteins. A study with model compounds. J. Am. Chem. Soc. 115, 3772-3773. Gat, Y., Grossjean, M., Pinevsky, I., Takei, H., Rothman, Z., Sigrist, H., & Lewis, A. (1992). Participation of bacteriorhodopsin active-site lysine backbone in vibrations associated with retinal photochemistry. Proc. Natl. Acad. Sci. USA 89, 2434--2438. Gerwert, K., & Hess, B. (1987). Initiation of light-induced proton transfers in bacteriorhodopsin monitored by FTIR difference spectroscopy. In T. G. Ebrey, H. Frauenfelder, B. Honig, & K. Nakanishi (Eds.), Biophysical studies of retinal proteins (pp. 144-148). University of Illinois Press. Gerwert, K., & Siebert, F. (1986). Evidence for light-induced 13-cis, 14-s-cis isomerization in bacteriorhodopsin obtained by FTIR difference spectroscopy using isotopically labelled retinals. EMBO 5,805-811. Gibson, N. J., & Cassim, J. Y. (1987). Retinal can regulate global conformational changes of the bacteriorhodopsins in the light-mediated sodium-borohydride-reduced purple membrane. In T. G. Ebrey, H. Frauenfelder, B. Honig, & K. Nakanishi (Eds.), Biophysical studies of retinal proteins (pp. 71-79). University of Illinois Press. Gilson, H. S. R., Honig, B. H., Croteau, A., Zarrilli, G., & Nakanishi, K. (1988). Analysis of the factors that influence the C=N stretching frequency of polyene schiffbases. Biophys. J. 53, 261-269. Goldschmidt, C. R., Ottolenghi, M., & Korenstein, R. (1976). On the primary quantum yield in the bacteriorhodopsin photocyle. Biophys. J. 16, 839-843. Goldschmidt, C. R., Kalisky, O., Rosenfeld, T., & Ottolenghi, M. (1977). The quantum efficiency of the bacteriorhodopsin photocycle. Biophys. J. 17, 179-183. Govindjee, R., Ebrey, T. G., & Crofts, A. R. (1980). The quantum efficiency of proton pumping by the purple membrane of Halobacterium halobium. Biophys. J. 30, 231-240. Govindjee, R., Dancshazy, Z., Ebrey, T. G., Longstaff, C., & Rando, R. R. (1988). Photochemistry of methylated rhodopsin. Photochem. Photobiol. 48, 493-496. Govindjee, R., Balashov, S. P., & Ebrey, T. G. (1990). Quantum efficiency of the photochemical cycle of bacteriorhodopsin. Biophys. J. 58, 597-608. Green, B., Monger, T., Alfano, R., Anton, B., & Callender, R. H. (1977). Cis-trans isomerization of rhodopsin in picoseconds. Nature 269, 179-180. Greenhalgh, D. A., Farrens, D. L., Subramaniam, S., & Khorana, H. G. (1993). Hydrophobic amino acids in the retinal-binding pocket of bacteriorhodopsin. J. Biophys. Chem. 268, 20305-20311.

128

JEFFREYA. STUART and ROBERT R. BIRGE

Greenhalgh, D. A., Subramaniam, S., Alexiev, U., Otto, H., Heyn, M. E, & Khorana, H. G. (1992). Effect of introducing different carboxylate-containing side chains at position 85 on chromophore formation and proton transport in bacteriorhodopsin. J. Biol. Chem. 267, 25734-25738. Grellmann, K. H., Livingston, R., & Pratt, D. (1962). A Flash-photolytic investigation of rhodopsin at low temperatures. Nature 193, 1258-1260. Groma, G. I., R~iksi, F., Szabr, G., & V~irr, G. (1988). Picosecond and nanosecond components in bacteriorhodopsin light-induced electric response signal. Biophys. J. 54, 77--80. Hackett, N. R., Stem, L. J., Chao, B. H., Kronis, K. A., & Khorana, H. G. (1987). Structure-function studies on bacteriorhodopsin: V. effects of amino acid'substitutions in the putative helix F. J. Biol. Chem. 269, 9277-9284. Han, M., DeDecker, B. S., & Smith, S. O. (1993). Localization of the retinal protonated Schiff base counterion in rhodopsin. Biophys. J. 65, 899-906. Hanamoto, J. H., Dupuis, P., & E1-Sayed, M. A. (1984). On the protein (tyrosine)-chromophore (protonated Schiff base) coupling in bacteriorhodopsin. Proc. Natl. Acad. Sci. USA 81, 7083-7087. Harada, I., Yamagishi, T., Uchida, K., & Takeuchi, H. (1990). Ultraviolet resonance raman spectra of bacteriorhodopsin in the light-adapted and dark-adapted states. J. Am. Chem. Soc. 112, 24432445. Harbison, G. S., Herzfeld, J., & Griffin, R. G. (1983). Solid-state nitrogen-15 nuclear magnetic resonance study of the schiffbase in bacteriorhodopsin. Biochemistry 22, 1-5. Harbison, G. S., Mulder, P. P. J., Pardoen, H., Lugtenburg, J., Herzfeld, J., & Griffin, R. G. (1985). High-resolution carbon-13 NMR of retinal derivatives in the solid state. J. Am. Chem. Soc. 107, 4809-4816. Harbison, G. S., Roberts, J. E., Herzfeld, J., & Griffin, R. G. (1988). Solid-state NMR detection ofproton exchange between the bacteriorhodopsin Schiff base and bulk water. J. Am. Chem. Soc. 110, 7221-7223. Harbison, G. S., Smith, S. O., Pardoen, J. A., Courtin, J. M. L., Lugtenburg, J., Herzfeld, J., Mathies, R. A., & Griffin, R. G. (1985). Solid-state 13C NMR detection of a perturbed 6-s-trans chromophore in bacteriorhodopsin. Biochemistry 24, 6955-6962. Harbison, G. S., Smith, S. O., Pardoen, J. A., Mulder, P. P. J., Lugtenburg, J., Herzfeld, J., Mathies, R., & Griffin, R. G. (1984a). Solid-state 13C NMR studies of retinal in bacteriorhodopsin. Biochemistry 23, 2662-2667. Harbison, G. S., Smith, S. O., Pardoen, J. A., Winkle, C., Lugtenburg, J., Herzfeld, J., Mathies, R., & Griffin, R. G. (1984b). Dark-adapted bacteriorhodopsin contains 13-cis, 15-syn and all-trans, 15-anti retinal schiffbases. Proc. Natl. Acad. Sci. USA 81, 1706-1709. Hauss, T., Grzesiek, S., Otto, H., Westerhausen, J., & Heyn, M. P. (1990). Transmembrane location of retinal in bacteriorhodopsin by neutron diffraction. Biochemistry 29, 4904-4913. Henderson, R., Baldwin, J. M., & Ceska, T. A. (1990a). Model for the structure of bacteriorhodopsin based on high-resolution electron cryo-microscopy. J. Mol. Biol. 213, 899-929. Henderson, R., Schertler, F. R. S., & Schertler, G. R. X. (1990b). The structure ofbacteriorhodopsin and its relevance to the visual opsins and other seven-helix G-protein coupled receptors. Phil. Trans. R. Soc. Lond., B. 326, 379-389. Herzfeld, J., Das Gupta, S. K., Farrar, M. R., Harbison, G. S., McDermott, A. E., Pelletier, S. L., Raleigh, D. P., Smith, S. O., Winkel, C., Lugtenburg, J., & Griffin, R. G. (1990). Solid-state 13C NMR study of tyrosine protonation in dark-adapted bacteriorhodopsin. Biochemistry 29, 5567-5574. Heyn, M. P., Cherry, R. J., & Mialler, U. (1977). Transient and linear dichroism studies on bacteriorhodopsin: Determination of the orientation of the 568 nm all-trans retinal chromophore. J. Mol. Biol. 117, 607-620. Hildebrandt, P., & Stockburger, M. (1984). Role of water in bacteriorhodopsin's chromophore: resonance raman study. Biochemistry 23, 5539-5548.

Characterization of the Primary Photochemical Events

129

Holz, M., Lindau, M., & Heyn, M. P. (1988). Distributed kinetics of the charge movements in bacteriorhodopsin: evidence for conformational substrates. Biophys. J. 53, 623-633. Honig, B. (1982). Theoretical aspects of photoisomerization in visual pigments and bacteriorhodopsin. In R. R. Alfano (Ed.), Biological events probed by ultrafast laser spectroscopy (pp. 281-296). New York: Academic Press Inc. Honig, B., Ebrey, T., Callender, R. H., Dinur, U., & Ottolenghi, M. (1979). Photoisomerization, energy storage, and charge separation: A model for light energy transduction in visual pigments and bacteriorhodopsin. Proc. Natl. Acad. Sci. USA 76, 2503-2507. Honig, B., Greenberg, A. D., Dinur, U., & Ebrey, T. G. (1976). Visual-pigment spectra: Implications of the protonation of the retinal schiffbase. Biochemistry 15, 4593-4599. Honig, B., Hudson, B., Sykes, B. D., & Karplus, M. (1971). NMR and theory predict 6-s-cis 40-70 deg PSB in solution. Proc. Natl. Acad. Sci. USA 68, 1289-1293. Honig, B., Warshel, A., & Karplus, M. (1975). Theoretical studies of the visual chromophores. Acct. Chem. Res. 8, 92-100. Horiuchi, S., Tokunaga, F., & Yoshizawa, T. (1980). Circular dichrosim of cattle rhodopsin and bathorhodopsin at liquid nitrogen temperatures. Biochim. Biophys. Acta 591,445-457. Hu, J., Griffin, R. G., & Herzfeld, J. (1994a). Synergy in the spectral tuning of retinal pigments--complete accounting of the opsin shift in bacteriorhodopsin. Biophys. J. 66, A45. Hu, S., Franklin, P. J., Wang, J., Chen, A.-H., Silva, B. E. R., Derguini, F., & Nakanishi, K. (1994b). Unbleachable rhodopsin with an 11-cis-locked eight-membered ring retinal: The visual transduction process. Biochemistry 33,408-416. Huang, J. Y., & Lewis, A. (1989). Determination of the absolute orientation of the retinylidene chromophore in purple membrane by a second-harmonic interference technique. Biophys. J. 55, 835-842. Huang, K.-S., Radhakrishnan, R., Bayley, H., & Khorana, H. G. (1982). Orientation of retinal in bacteriorhodopsin as studied by cross-linking using a photosensitive analog of retinal. J. Biol. Chem. 257, 13616-13623. Hubbard, R. (1966). The stereoisomerization of 11-cis-retinal. J. Biol. Chem. 241, 1814-1818. Hubbard, R., & Kropf, A. (1958). Proc. Natl. Acad. Sci. USA 44, 130-139. Hug, S. J., Lewis, J. W., & Kliger, D. S. (1988). Evidence for a common batho intermediate ofrhodopsin and isorhodopsin. J. Am. Chem. Soc. 110, 1998-1999. Hurley, J. B., & Ebrey, T. G. (1978). Energy transfer in the purple membrane ofHalobacterium halobium. Biophys. J. 22, 49-66. Hurley, J. B., Ebrey, T. G., Honig, B., & Ottolenghi, M. (1977). Temperature and wavelength effects on the photochemistry of rhodopsin, isorhodopsin, bacteriorhodopsin, and their photoproducts. Nature 270, 540-542. Ito, M., Katsuta, Y., Imamoto, Y., Shichida, Y., & Yoshizawa, T. (1992). Conformational analysis of the rhodopsin chromophore using bycyclic retinal analogues. Photochem. Photobiol. 56, 915-919. Iwasa, T., Ito, M., & Tokunaga, F. (1992). A novel bacteriorhodopsin analogue with conformationally 6-s-cis fixed retinals. Photochem. Photobiol. 56, 921-927. Jang, D.-J., E1-Sayed, M. A., Stem, L. J., Mogi, T., & Khorana, H. G. (1990). Effect of genetic modification of tyrosine- 185 on the proton pump and the blue-to-purple transition in bacteriorhodopsin. Proc. Natl. Acad. Sci. USA 87, 4103-4107. Jonas, R., & Ebrey, T. G. (1991). Binding of a single divalent cation directly correlates with the blue-to-purple transition in bacteriorhodopsin. Proc. Natl. Acad. Sci. USA 88, 149. Kakitani, T., Kakitani, H., & Yomosa, S. (1977). Molecular mechanism for the initial process of visual excitation. II. Theoretical analysis of optical activity in rhodopsin and bathorhodopsin. J. Phys. Soc. Japan 42, 996-1004. Kakitani, H., Kakitani, T., Rodman, H., & Honig, B. (1985). On the mechanism of wavelength regulation in visual pigments. Photochem. Photobiol. 41, 471-479.

130

JEFFREYA. STUART and ROBERT R. BIRGE

Kakitani, H., Kakitani, T., Rodman, H., Honig, B., & Callender, R. J. (1983). Correlation of vibrational frequencies with absorption maxima in polyenes, rhodopsin, bacteriorhodopsin and retinal analogues. J. Phys. Chem. 87, 3620-3628. Kalisky, O., & Ottolenghi, M. (1982). Branching pathways in the photocycle of bacteriorhodopsin. Photochem. Photobiol. 35, 109-115. Kandori, H., Matuoka, H., Nagai, H., Shichida, Y., & Yoshizawa, T. (1988). Dependency of apparent relative quantum yield of isorhodopsin to rhodopsin on the photon density of picosecond laser pulse. Photochem. Photobiol. 48, 93-97. Kandori, H., Matuoka, S., Shichida, Y., & Yoshizawa, T. (1989a). Dependency of photon density on primary process of cattle rhodopsin. Photochem. Photobiol. 49, 181-184. Kandori, H., Matuoka, S., Shichida, Y., Yoshizawa, T., Ito, M., Tsukida, K., Balogh-Nair, V., & Nakanishi, K. (1989b). Mechanism of isomerization of rhodopsin studied by use of 11-cis-locked rhodopsin analogues excited with a picosecond laser pulse. Biochemistry 28, 6460-6467. Kandori, H., Shichida, Y., & Yoshizawa, T. (1989c). Absolute absorption spectra of batho- and photorhodopsins at room temperature. Biophys. J. 56, 453-457. Kandori, H., Yoshihara, K., Tomioka, H., Sasabe, H., & Shichida, Y. (1993). Comparative studies of primary photochemical events of two retinal proteins, bacteriorhodopsin and halorhodopsin, by use of subpicosecond time-resolved spectroscopy. Chem. Phys. Lett. 211,385-391. Kaplan, E., & Barlow, R. B., Jr. (1980). Circadian clock in Limulus brain increases response and decreases noise of retinal photoreceptors. Nature 286, 393-395. Kawamura, S., Miyatani, S., Matsumoto, H., Yoshizawa, T., & Liu, R. S. H. (1980). Photochemical studies of 7-cis-rhodopsin at low temperatures. Nature and properties of the bathointermediate. Biochemistry 19, 1549-1553. Kawamura, S., Tokunaga, F., Yoshizawa, T., Sarai, A., & Kakitani, T. (1979). Orientational changes of the transition dipole moment of retinal chromophore of the disk membrane due to the conversion of rhodopsin to bathorhodopsin and to isorhodopsin. Vision Res. 19, 879-884. Keszthelyi, L. (1988). Primary charge motions and light-energy transduction in bacteriorhodopsin. Biophys. Chem. 29, 127-136. Keszthelyi, L., & Ormos, P. (1980). Electric signals associated with the photocycle ofbacteriorhodopsin. FEBS Lett. 109, 189-193. Kliger, D. S., Horowitz, J. W., Lewis, J. W., & Einterz, C. M. (1984). Evidence for a common batho-intermediate in the bleaching of rhodopsin and isorhodopsin. Vision Res., 24, 1465-1470. Korenstein, R., & Hess, B. (1978). Immobilization ofbacteriorhodopsin and orientation of its transition moment in purple membrane. FEBS Lett. 89, 15-20. Koutalos, Y., Ebrey, T. G., Tsuda, M., Odashima, K., Lien, T., Park, M. H., Shimizu, N., Derguini, F., Nakanishi, K., Gilson, H. R., & Honig, B. (1989). Regeneration of bovine and octopus opsins in situ with natural and artificial retinals. Biochemistry 28, 2732-2739. Koyama, Y., Natasu, H., Mukai, Y., & Tokunaga, F. (1993). Isomerization of the retinylidene chromophore ofbacteriorhodopsin in light adaptation: Intrinsic isomerization of the chromophore and its control by the apo-protein. Photochem. Photobiol. 57, 732-738. Kropf, A., & Hubbard, R. (1958). The colors of the visual pigments. Ann. NY Acad. Sci., 74, 266-280. Kuschmitz, D., & Hess, B. (1982). Trans-cis isomerization of the retinal chromophore ofbacteriorhodopsin during the photocycle. FEBS Lett. 138, 137-140. Lanyi, J. K. (1978). Light energy conversion in Halobacterium halobium. Microbiol. Rev. 42, 682-712. Leder, R. O., Helgerson, S. L., & Thomas, D. D. (1989). The transverse location of the retinal chromophore in the purple membrane by diffusion-enhanced energy transfer. J. Mol. Biol. 209, 683-701. Lewis, A., Spoonhower, J., Bogomolni, R. A., Lozier, R. H., & Stoeckenius, W. (1974). Tunable laser resonance raman spectroscopy ofbacteriorhodopsin. Proc. Nat. Acad. Sci. USA 71, 4462--4466. Lewis, J. W., & Kliger, D. S. (1992). Photointermediates of Visual Pigments. J. Bioenerg. Biomembr. 24, 201-210.

Characterization of the Primary Photochemical Events

131

Liebman, E A., Parker, K. R., & Dratz, E. A. (1987). The molecular mechanism of visual excitation and its relation to the structure and composition of the rod outer segment. Ann. Rev. Physiol. 49, 965-991. Lin, S. W., & Mathies, R. A. (1989). Orientation of the protonated retinal Schiff base group in bacteriorhodopsin from absorption linear dichroism. Biophys. J. 56, 653-660. Lin, S. W., Sakmar, T. P., Franke, R. R., Khorana, H. G., & Mathies, R. A. (1992). Resonance Raman microprobe spectroscopy ofrhodopsin mutants: effects of substitutions in the third transmembrane helix. Biochemistry 31, 5105-5111. Liu, R. S. H., & Asato, A. E. (1985). The primary process of vision and the structure ofbathorhodopsin, a mechanism of photoisomerization of polyenes. Proc. Natl. Acad. Sci. USA 82, 259-263. Liu, R. S., & Mirzadegan, T. (1988). The shape of a three-dimensional binding site ofrhodopsin based on molecular modeling analyses of isomeric and other visual pigment analogues. Bioorganic studies of visual pigments. 11. J. Am. Chem. Soc. 110, 8617-8623. Liu, R. S. H., Liu, C. W., X.-Y., L., & Asato, A. E. (1991). Butyl conformational reorganization as a possible explanation for the longitudinal flexability of the binding site ofbacteriorhodopsin. The azulene and C-22 retinoid analogs. Photochem. Photobiol. 54, 625-631. Livnah, N., & Sheves, M. (1993). The Schiffbase bond configuration in bacteriorhodopsin and in model compounds. Biochemistry 32, 7223-7228. Lohmaann, R., & Stockburger, M. (1992). Time-resolved resonance Raman studies ofbacteriorhodopsin and its intermediates K590 and L550: biological implications. J. Raman Spect. 23,575-583. Lohrmann, R., Grieger, I., & Stockburger, M. (1991). Resonance Raman studies on the intermediate K-590 in the photocycle ofbacteriorhodopsin. J. Phys. Chem. 95, 1993-2001. Longstaff, C., Calhoon, R. D., & Rando, R. R. (1986). Deprotonation of the Schiffbase of rhodopsin is obligate in the activation of the G protein. Proc. Natl. Acad. Sci. USA 83, 4209-4213. Longstaff, C., & Rando, R. R. (1987). Deprotonation of the Schiffbase ofbacteriorhodopsin is obligate in light-induced proton pumping. Biochemistry 26, 6107-6113. Longstaff, C., Seckler, B., Calhoon, R. D., & Rando, R. R. (1987). Active-site lysine methylation and the mechanisms of action ofrhodopsin and bacteriorhodopsin. In T. G. Ebrey, H. Frauenfelder, B. Honig, & K. Nakanishi (Eds.), Biophysical studies of retinal proteins (pp. 64-70). University of Illinois Press. Lopez-Garriga, J. J., Babcock, G. T., & Harrison, J. E (1986a). Factors influencing the C=N stretching frequency in neutral and protonated Schiff's bases. J. Am. Chem. Soc. 108, 7241-7251. Lopez-Garriga, J. J., Babcock, G. T., & Harrison, J. E (1986b). Increase in the C=N stretching frequency upon complexation of trans-retinylidene-n-butylamine with general Lewis acids. J. Am. Chem. Soc. 108, 7131-7133. Loppnow, G. R., Barry, B. A., & Mathies, R. A. (1989). Why are blue visual pigments blue? A resonance raman microprobe study. Proc. Natl. Acad. Sci. USA 86, 1515-1518. Loppnow, G. R., Mathies, R. A., Middendorf, T. R., Gottfried, D. S., & Boxer, S. G. (1992). Photochemical hole-buming spectroscopy of bovine rhodopsin and bacteriorhodopsin. J. Phys. Chem. 96, 737-745. Loppnow, G. R., Miley, M. E., Mathies, R. A., Liu, R. S., Kandori, H., Shichida, Y., Fukada, Y., & Yoshizawa, T. (1990). Structure of retinal chromophore in 7, 9-dicis-rhodopsin. Biochemistry 29, 8985-8991. Lugtenburg, J., Mathies, R. A., Griffin, R. G., & Herzfeld, J. (1988). Structure and function ofrhodopsins from solid state NMR and resonance Raman spectroscopy of isotopic retinal derivatives. Trends Biochem. Sci., 13,388-393. Lugtenburg, J., Muradin-Szweykowska, M., Heeremans, C., & Pardoen, J. A. (1986). Mechanism for the opsin shift of retinal's absorption in bacteriorhodopsin. J. Am. Chem. Soc. 108, 3104-3105. Maeda, A., Saski, J., Pfefferl~, J.-M., Shichida, Y., & Yoshizawa, T. (1991). Fourier transform infrared spectral studies on the Schiffbase mode ofall-trans bacteriorhodopsin and its photointermediates, K and L. Photochem. Photobiol. 54, 911-921.

132

JEFFREY A. STUART and ROBERT R. BIRGE

Maeda, A., Sasaki, J., Shichida, Y., & Yoshizawa, T. (1992). Water structural changes in the bacteriorhodopsin photocycle: analysis by fourier transform infrared spectroscopy. Biochemistry 31,462467. Maeda, A., Sasaki, J., Tamazake, T., Needleman, R., & Lanyi, J. K. (1994). Interaction of aspartate-85 with a water molecule and the protonated Schiffbase in the L intermediate ofbacteriorhodopsin: a fourier-transform infrared spectroscopic study. Biochemistry 33, 1713-1717. Mao, B., Ebrey, T. G., & Crouch, R. (1980). Bathoproducts of rhodopsin, isorhodopsin I, and isorhodopsin II. Biophys. J. 29, 247-256. Mao, B., Tsuda, M., Ebrey, T., Akita, H., Balogh-Nair, V., & Nakanishi, K. (1981). Flash photolysis and low temperature photochemistry of bovine rhodopsin with a fixed 11-ene. Biophys. J. 35,543-546. Marinetti, T. (1987a). Abrupt onset of large scale nonproton ion release in purple membranes caused by increasing pH or ionic strength. Biophys. J. 51,875-881. Marinetti, T. (1987b). Large scale nonproton ion release and bacteriorhodopsin's state of aggregation in lipid vesicles. Biophys. J. 52, 115-121. Marinetti, T., & Mauzerall, D. (1983). Absolute quantum yields and proof of proton and non-proton transient release and uptake in photoexcited bacteriorhodopsin. Proc. Natl. Acad. Sci.USA 80, 178-180.

Marti, T., Rrsselet, S. J., H., O., Heyn, M. P., & Khorana, H. G. (1991). The retinylidene schiff base counterion in bacteriorhodopsin. J. Biol. Chem. 266, 18674-18683. Mathies, R. (1982). Resonance Raman spectroscopy of rhodopsin and bacteriorhodopsin isotopic analogs. In L. Packer (Ed.), Methods enzymol (pp. 633-643). New York: Academic Press. Mathies, R. A., Brito Cruz, C. H., Pollard, W. T., & Shank, C. V. (1988). Direct observation of the femtosecond excited-state cis-trans isomerization in bacteriorhodopsin. Science 240, 777-779. Mathies, R. A., Lin, S. W., Ames, J. B., & Pollard, W. T. (1991). From femtoseconds to biology: Mechanism ofbacteriorhdopsin's light-driven proton pump. Annu. Rev. Biophys. Biophys. Chem. 20, 491-518. Mathies, R. A., Lugtenburg, J., & Shank, C. V. (1989). From femtoseconds to biology: mechanism of the light-driven proton pump in bacteriorhodopsin. SPIE (Biomolecular Spectroscopy), 1057, 138-145. McDermott, A. E., Thompson, L. K., Winkel, C., Farrar, M. R., Pelletier, S., Lugtenburg, J., Herzfeld, J., & Griffin, R. G. (1991). Mechanism of proton pumping in bacteriorhodopsin by solid-state NMR: the protonation state of tyrosine in the light-adapted and M states. Biochemistry 30, 8366-8371. McMaster, E., & Lewis, A. (1988). Evidence for light-induced lysine conformational changes during the primary event of the bacteriorhodopsin photocycle. Biochem. Biophys. Res. Comm. 156, 86-91. Metz, G., Siebert, F., & Engelhard, M. (1992). High-resolution solid state 13C NMR ofbacteriorhodopsin: characterization of [4-13C]asp resonances. Biochemistry 31,455-462. Milder, S. J., & Kliger, D. S. (1986). A new approach to understanding the initial step in visual transduction. Biophys. J. 49, 567-570. Mizukami, T., Kandori, H., Shichida, Y., Chen, A.-H., Derguini, F., Caldwell, C. G., Bigge, C. F., Nakanishi, K., & Yoshizawa, T. (1993). Photoisomerization mechanism of the rhodopsin chromophore: Picosecond photolysis of pigment containing 11-cis-locked eight-membered ring retinal. Proc. Natl. Acad. Sci. USA 90, 4072-4076. Mogi, T., Marti, T., & Khorana, H. G. (1989a). Structure-function studies on bacteriorhodopsin. J. Biol. Chem. 264, 14197-14201. Mogi, T., Stem, L. J., Chao, B., & Khorana, H. G. (1989b). Structure-function studies on bacteriorhodopsin. J. Biol. Chem. 264, 14192-14196. Mogi, T., Stem, L. J., Marti, T., Chao, B. H., & Khorana, H. G. (1988). Aspartic acid substitutions affect proton translocation by bacteriorhodopsin. Proc. Natl. Acad. Sci. USA 85, 4148-4152.

Characterization of the Primary Photochemical Events

133

Mollevanger, L. C. R J., Kentgens, A. E M., Pardoen, J. A., Courtin, J. M. L., Veeman, W. S., Lugtenburg, J., & de Grip, W. J. (1987). High-resolution solid-state laC-NMR study of carbons C-5 and C-12 of the chromophore of bovine rhodopsin. Eur. J. Biochem. 163, 9-14. Monger, T., Alfano, R. R., & Callender, R. H. (1979). Photochemistry of rhodopsin and isorhodopsin investigated on a picosecond time scale. Biophys. J. 27, 105-115. Nagle, J. F., Bhattacharjee, S. M., Parodi, L. A., & Lozier, R. H. (1983). Effect ofphotoselection upon saturation and the dichroic ratio in flash experiments upon effectively immobilized systems. Photochem. Photobiol. 38, 331-339. Nakanishi, K., Balogh-Nair, V., Arnaboldi, M., Tsujimoto, K., & Honig, R. (1980). An external point-charge model for bacteriorhodopsin to account for its purple color. J. Am. Chem. Soc. 102, 7945-7947. Nakayama, T. A., & Khorana, H. G. (1990). Orientation of retinal of bovine rhodopsin determined by cross-linking using a photoactivatable analog of 11-cis-retinal. J. Biol. Chem. 265, 15762-15769. Nathans, J. (1990a). Determinants of visual pigment absorbance: identification of the retinylidene schiff's base counterion.in bovine rhodopsin. Biochemistry 29, 9746-9752. Nathans, J. (1990b). Determinants of visual pigment absorbance: Role of charged amino acids in the putative transmembrane segments. Biochemistry 29, 937-942. Nathans, J. (1992). Rhodopsin: structure, function, and genetics. Biochemistry 3 l, 4923-4929. Nonella, M., Windemuth, A., & Schulten, K. (1991). Structure of bacteriorhodopsin and in situ isomerization of retinal: A molecular dynamics study. Photochem. Photobiol. 54, 937-948. Nuss, M. C., Zinth, W, Kaiser, W., Kolling, E., & Oesterhelt, D. (1985). Femtosecond spectroscopy of the first events of the photochemical cycle in bacteriorhodopsin. Chem. Phys. Lett. 117, 1-7. Oesterhelt, D., & Hess, B. (1973). Reversible photolysis of the purple complex in the purple membrane of Halobacterium halobium. Eur. J. Biochem. 37, 31 6-326. Oesterhelt, D., & Tittor, J. (1989). Two pumps, one principle: light-driven ion transport in Halobacteria. Trends Biochem. Sci. 14, 57-61. Oesterhelt, D., Hegemann, P., & Tittor, J. (1985). The photocycle of the chloride pump halorhodopsin. II. Quantum yields and a kinetic model. EMBO J. 4, 2351-2356. Ohtani, H., Kobayashi, T., Tsuda, M., & Ebrey, T. G. (1988). Primary processes in photolysis of octopus rhodopsin. Biophysical J. 53, 17-24. Orlandi, G., & Schulten, K. (1979). Coupling of stereochemistry and proton donor-acceptor properties of a schiff base. A model of a light-driven proton pump. Chem. Phys. Lett. 64, 370-374. Oseroff, A. R., & Callender, R. H. (1974). Resonance Raman spectroscopy of rhodopsin in retinal disk membranes. Biochemistry 13, 2443--4348. Otomo, J., Tomioka, A., Kinoshita, K., Miyata, H., Takenaka, Y. T. K., & Ikegami, A. (1988). Chromophore of bacteriorhodopsin is closer to the cytoplasmic surface of purple membrane. Biophys. J. 54, 57-64. Ottolenghi, M., & Sheves, M. (1989). Synthetic retinals as probes for the binding site and photoreactions in rhodopsins. J. Membrane Biol. 112, 193-212. Palings, I., Pardoen, J. A., van den Berg, E., Winkle, C., Lugtenburg, J., & Mathies, R. A. (1987). Assignment of fingerprint vibrations in the resonance raman spectra of rhodopsin, isorhodopsin, and bathorhodopsin: Implications for chromophore structure and environment. Biochemistry 26, 2544--2556. Palings, I., van den Berg, E. M. M., Lugtenburg, J., & Mathies, R. A. (1989). Complete assignment of the hydrogen out-of-plane wagging vibrations of bathorhodopsin: chromophore structure and energy storage in the primary photoproduct of vision. Biochemistry 28, 1498-1507. Pande, A. J., Callender, R. H., Ebrey, T. G., & Tsuda, M. (1984). Resonance Raman study ofthe primary photochemistry of visual pigments. Hypsorhodopsin. Biophys. J. 45, 573-576. Pande, C., Deng, H., Rath, P., Callender, H., & Schwemer, J. (1987). Resonance Raman spectroscopy of an ultraviolet-sensitive insect rhodopsin. Biochemistry 26, 7426-7430.

134

JEFFREYA. STUART and ROBERT R. BIRGE

Papadopoulos, G., Dencher, N. A., Zaccai, G., & B~ildt, G. (1990). Water molecules and exchangeable hydrogen ions at the active centre ofbacteriorhodopsin localized by neutron diffraction. Elements of the proton pathway? J. Mol. Biol. 214, 15--19. Park, M. H., Yamamoto, T., & Nakanishi, K. (1989). Probes which reflect the distance between the retinal chromophore and membrane surface in bacteriorhodopsin (bR). Direction of retinal 9-methyl in bR. J. Am. Chem. Soc. 111, 4997-4998. Peteanu, L. A., Schoenlein, R. W., Wang, Q., Mathies, R. A., & Shank, C. V. (1993). The first step in vision occurs in femtoseconds: Complete blue and red spectral studies. Proc. Natl. Acad. Sci. USA 90, 11762-11766. Peters, K. S., Applebury, M. L., & Rentzepis, P. M. (1977). Primary photochemical event in vision: proton translocation. Proc. Natl. Acad. Sci. USA 74, 3119-3123. Polland, H. J., Franz, M. A., Zinth, W., Kaiser, W., Kolling, E., & Oesterhelt, D. (1984). Optical picosecond studies of bacteriorhodopsin containing a sterically fixed retinal. Biochim. Biophys. Acta 767, 635-639. Polland, H. J., Franz, M. A., Zinth, W., Kaiser, W., Kolling, E., & Oesterhelt, D. (1986). Early picosecond events in the photocycle of bacteriorhodopsin. Biophys. J. 49, 651-662. Pollard, W. T., Cruz, C. H. B., Shank, C. V., & Mathies, R. A. (1989). Direct observation of the excited-state cis-trans photoisomerization ofbacteriorhodopsin: Multilevel line shape theory for femtosecond dynamic hole burning and its application. J. Chem. Phys. 90, 199-208. Racker, E., & Stoeckenius, W. (1974). Reconstitution of purple membrane vesicles catalyzing lightdriven proton uptake and adenosine triphosphate formation. J. Biol. Chem. 249, 662-663. Rafferty, C. N., & Shichi, H. (1981). The involvement of water at the retinal binding site in rhodopsin and early light-induced intramolecular proton transfer. Photochem. Photobiol. 33,229-234. Rayfield, G. W. (1983). Events in proton pumping by bacteriorhodopsin. Biophys. J. 41,109-117. Renk, G. E., Or, Y. S., & Crouch, R. K. (1987). A rhodopsin pigment containing a spin-labeled retinal. J. Am. Chem. Soc. 109, 6163-6168. Roepe, P., Ahl, P. L., DasGupta, S. K., Herzfeld, J., & Rothschild, K. J. (1987). Tyrosine and carboxyl protonation changes in the bacteriorhodopsin photocycle. 1. M412 and L550 intermediates. Biochemistry 26, 6696--6707. Roepe, P., Ahl, P. L., Herzfeld, J., Lugtenburg, J., & Rothschild, K. J. (1988a). Tyrosine protonation changes in bacteriorhodopsin. J. Biol. Chem. 263, 5110-5117. Roepe, P., Gray, D., Lugtenburg, J., van den Berg, E. M. M., Herzfeld, J., & Rothschild, K. J. (1988b). FTIR evidence for tryptophan perturbations during the bacteriorhodopsin, photocycle. J. Am. Chem. Soc. 110, 7223-7224. Roepe, P., Scherrer, P., Ahl, P. L., Das Grupta, S. K., Bogomolni, R. A., Herzfeld, J., & Rothschild, K. J. (1987). Tyrosine and carboxyl protonation changes in the bacteriorhodopsin photocycle. 2. Tyrosines-26 and -64. Biochemistry 26, 6708-6717. Roepe, P. D., Ahl, P. L., Herzfeld, J., Lugtenburg, J., & Rothschild, K. J. (1988). Tyrosine protonation changes in bacteriorhodopsin. A fourier transform infrared study of br548 and its primary photoproduct. J. Biol. Chem. 263, 5110-5117. Rohr, M., G/irtner, W., Schweitzer, G., Holzwarth, A. R., & Braslavsky, S. E. (1992). Quantum yields of the photochromic equilibrium between bacteriorhodopsin and its bathointermediate K. Femptoand nanosecond optoacoustic spectroscopy. J. Phys. Chem. 96, 6055-6061. Rosenfeld, T., Honig, B., & Ottolenghi, M. (1977a). Cis-trans isomerization in the photochemistry of vision. Pure Appl. Chem. 49, 341-351. Rosenfeld, T., Honig, B., Ottolenghi, M., Hurley, J. B., & Ebrey, T. G. (1977b). On the role ofthe protein in the photoisomerization of the visual pigment chromophore. Pure Appl. Chem 49, 341-351. Rothschild, K. J. (1992). FTIR difference spectroscopy of bacteriorhodopsin: Toward a molecular model. J. Bioenerg. Biomembr. 24, 147-167.

Characterization of the Primary Photochemical Events

135

Rothschild, K. J., Argade, E V., Eamest, T. N., Huang, K. S., London, E., Liao, M.-J., Bayley, H., Khorana, H. G., & Herzfeld, J. (1982). The site of the attachment of retinal in bacteriorhodopsin. J. Biol. Chem. 257, 8592-8595. Rothschild, K. J., Braiman, M. S., He, Y.-W., Marti, T., & Khorana, H. G. (1990). Vibrational spectroscopy of bacteriorhodopsin mutants. Evidence for the interaction of aspartic acid 212 with tyrosine 185 and possible role in the proton pump mechanism. J. Biol. Chem. 265(28), 1698316991. Rothschild, K. J., Braiman, M. S., Mogi, T., Stem, L. J., & Khorana, H. G. (1989a). Conserved amino acids in F-helix of bacteriorhodopsin form part of a retinal binding pocket. FEBS Lett. 250, 448--452. Rothschild, K. J., Gray, D., Mogi, T., Marti, T., Braiman, M. S., Stem, L. J., & Khorana, H. G. (1989b). Vibrational spectroscopy of bacteriorhodopsin mutants: chromophore isomerization perturbs tryptophan-86. Biochemistry 28, 7052-7059. Rothschild, K. J., Roepe, E, Ahl, E L., Earnest, T. N., Bogomolni, R. A., Das Gupta, S. K., Mulliken, C. M., & Herzfeld, J. (1986). Evidence for a tyrosine protonation change during the primary phototransformation of bacteriorhodopsin at low temperature. Proc. Natl. Acad. Sci. USA 83, 347-351. Rothschild, K. J., Roepe,. P., & Gillespie, J. (1985). Fourier transform infrared spectroscopic evidence for the existence of two conformations of the bacteriorhodopsin primary photoproduct at low temperature. Biochim. Biophys. Acta 808, 140-148. Rothschild, K. J., Roepe, P., Lugtenburg, J., & Pardoen, J. A. (1984). Fourier transform infrared evidence for Schiff base alteration in the first step of the bacteriorhodopsin photocycle. Biochemistry 23, 6103-6109. Sakmar, T. P., Franke, R. R., & Khorana, H. G. (1989). Glutamic acid 113 serves as the retinylidene Schiff base counterion in bovine rhodopsin. Proc. Natl. Acad. Sci. USA 86, 8309-8313. Sakmar, T. P., Franke, R. R., & Khorana, H. G. (1991). The role ofthe retinylidene Schiffbase counterion in rhodopsin in determining wavelength absorbance and Schiff base pica. Proc. Natl. Acad. Sci. USA 88, 3079-3083. Salem, S., & Bruckrnann, P. (1975). Conversion of a photon to an electrical signal by sudden polarization in the n-retinylidene visual chromophore. Nature 258, 526-528. Sampogna, R. V., & Honig, B. (1994). Environmental effects on the protonation states of active site residues in bacteriorhodopsin. Biophys. J. 66, 1341-1352. Sandorfy, C., Lussier, L. S., Thanh, H. L., & Vocelle, D. (1987). Fourier-transform infrared study of the protonation ofretinylidene Schiffbase. In T. G. Ebrey, H. Frauenfelder, B. Honig, & K. Nakanishi (Eds.), Biophysical studies of retinal proteins (pp. 247-251). University of Illinois Press. Sasaki, J., Maeda, A., Shichida, Y., Groesbeek, M., Lugtenburg, J., & Yoshizawa, T. (1992). Structure of hypsorhodopsin: analysis by fourier transform infrared spectroscopy at 10 K. Photochem. Photobiol. 56, 1063-1071. Sasaki, N., Tokunaga, F., & Yoshizawa, T. (1980a). Existence of two forms ofbathorhodopsins. FEBS Lett. 114, 1-3. Sasaki, N., Tokunaga, F., & Yoshizawa, T. (1980b). The formation of two forms ofbathorhodopsin and their optical properties. Photochem. Photobio. 32, 433-441. Schertler, G. F. X., Lozier, R., Michel, H., & Oesterhelt, D. (1991). Chromophore motion during the bacteriorhodopsin photocycle: polarized absorption spectroscopy of bacteriorhodopsin and its M-state in bacteriorhodopsin crystals. EMBO J. 10, 2253-2261. Schertler, G. F. X., Villa, C., & Henderson, R. (1993). Projection structure of rhodopsin. Nature 362, 770-772. Schick, G. A., Cooper, T. M., Holloway, R. A., Murray, L. P., & Birge, R. R. (1987). Energy storage in the primary photochemical events of rhodopsin and isorhodopsin. Biochemistry 26, 2556-2562. Schneider, G., Diller, R., & Stockburger, M. (1989). Photochemical quantum yield ofbacteriorhodopsin from resonance Raman scattering as a probe for photolysis. Chem. Phys. 131, 17-29.

136

JEFFREYA. STUART and ROBERT R. BIRGE

Schoenlein, R. W., Peteanu, L. A., Mathies, R. A., & Shank, C. V. (1991). The first step in vision: femptosecond isomerization of rhodopsin. Sci. 254, 412-415. Schoenlein, R. W., Peteanu, L. A., Wang, Q., Mathies, R. A., & Shank, C. V. (1993). Femtosecond dynamics of cis-trans isomerization in a visual pigment analog: isorhodopsin. J. Phys. Chem. 97, 12087-12092. Schulten, K., & Tavan, P. (1978). A mechanism for the light-driven proton pump of Halobacterium halobium. Nature 272, 85-86. Schulten, K., Schulten, Z., & Tavan, P. (1984). An isomerization model for the pump cycle of bacteriorhodopsin. In L. Bolis, E. J. M. Helmrich, & H. Passow (Eds.), Information and energy . transduction in biological membranes (pp. 113-131). New York: Alan R. Liss. Sharkov, A. V., Pakulev, A. V., Chekalin, S. V., & Matveetz, Y. A. (1985). Primary events in bacteriorhodopsin probed by subpicosecond spectroscopy. Biochim. Biophys. Acta 808, 94-102. Sheves, M., Albeck, A., Ottolenghi, M., Bovee-Geurts, P. H. M., De Grip, W. J., Einterz, C. M., Lewis, J. W., Schaechter, L. E., & Kliger, D. S. (1986). An artificial visual pigrnent with restricted C9-C11 motion forms normal photolysis intermediates. J. Am. Chem. Soc. 108, 6440-6441. Sheves, M., Friedman, N., & Albeck, A. (1985). Primary photochemical event in bacteriorhodopsin: study with artificial pigments. Biochemistry 24, 1260-1265. Shichida, Y. (1986). Primary intermediates of photobleaching of rhodopsin. Photobiochem. Photobiophys. 13, 287-307. Shichida, Y. (1992). Changes in the chromophore-opsin interaction in the photobleaching processes of visual pigments. In J.-L. Rigaud (Ed.), Colloques INSERM (Structures and functions of retinal proteins), (pp. 291-294). Dourdan, France: John Libbey and Company Ltd. Shichida, Y., Matuoka, S., Hidaka, Y., & Yoshizawa, T. (1983). Absorption spectra of intermediates of bacteriorhodopsin measured by laser photolysis at room temperature. Biochim. Biophys. Acta. 723, 240-246. Shichida, Y., Matuoka, S., & Yoshizawa, T. (1984). Formation of photorhodopsin, a precursor of bathorhodopsin, detected by a picosecond laser photolysis at room temperature. Photobiochem. Photobiophys. 7, 221-228. Shichida, Y., Tokunaga, F., & Yoshizawa, T. (1978). Circular dichroism of squid rhodopsin and its intermediates. Biochim. Biophys. Acta 504, 413-430. Siebert, F., & Mantele, W. (1983). Investigation of the primary photochemistry ofbacteriorhodopsin by low temperature Fourier-transform infrared spectroscopy. Eur. J. Biochem. 130, 565--573. Simmeth, R., & Rayfield, G. W. (1990). Evidence that the photoelectric response ofbacteriorhodopsin occurs in less than 5 picoseconds. Biophys. J. 57, 1099-1101. Simmons, C. J., Liu, R. S. M., Denny, M., & Serf, K. (1981). Acta Crystallogr., Sect. B: Struct. Crystallogr. Cryst. Chem. B37, 2197-2205. Skulachev, V. P. (1992). Rhodopsins: from ion pumps to specialized photoreceptors. In J.-L. Rigaud (Ed.), Colloques INSERM (Structures and functions of retinal proteins), (pp. 229-232). Dourdan, France: John Libbey and Company Ltd. Smith, S. O., Braiman, M. S., Myers, A. B., Pardoen, J. A., Courtin, J. M. L., Winkel, C., Lugtenburg, J., & Mathies, R. A. (1987c). Vibrational analysis of the all-trans-retinal chromophore in light-adapted bacteriorhodopsin. J. Am. Chem. Soc. 109, 3108-3125. Smith, S. O., Courtin, J., de Groot, H., Gebhard, R., & Lugtenburg, J. (1991). 13C magic-angle spinning NMR studies of bathorhodopsin, the primary photoproduct of rhodopsin. Biochemistry 30, 7409-7415. Smith, S. O., de Groot, H. J. M., Gebhard, R., Courtin, J. M. L., Lugtenburg, J., Herzfeld, J., & Griffin, R. G. (1989). Structure and protein environment of the retinal chromophore in light- and dark-adapted bacteriorhodopsin studied by solid-state NMR. Biochemistry 28, 8897-8904. Smith, S. O., Harbison, G. S., Raleigh, D. P., Griffin, R. G., Winkel, C., Pardoen, J. A., Lugtenburg, J., & Herzfeld, J. (1987d). Solid-state NMR studies of bacteriorhodopsin: Recent results on the structure of the Schiffbase in the dark-adapted and M412 states. In T. G. Ebrey, H. Frauenfelder,

Characterization of the Primary Photochemical Events

137

B. Honig, & K. Nakanishi (Eds.), Biophysical studies of retinal proteins (pp. 219-225). University of Illinois Press. Smith, S. O., Hornung, I., van der Steen, R., Pardoen, J. A., Braiman, M. S., Lugtenburg, J., & Mathies, R. A. (1986). Are C 14--C15 single bond isomerizations of the retinal chromophore involved in the proton-pumping mechanism ofbacteriorhodopsin. Proc. Natl. Acad. Sci. USA 83, 967-971. Smith, S. O., Lugtenburg, J., & Mathies, R. A. (1985a). Determination of retinal chromophore structure in bacteriorhodopsin with resonance raman spectroscopy. J. Membrane Biol. 85, 95-109. Smith, S. O., Myers, A. B., Mathies, R. A., Pardoen, J. A., Winkel, C., Van Den Berg, E. M. M., & Lugtenburg, J. (1985b). Vibrational analysis of the all-trans retinal protonated Schiff base. Biophys. J. 47, 653-664. Smith, S. O., Myers, A. B., Pardoen, J. A., Winkel, C., Mulder, P. P. J., Lugtenburg, J., & Mathies, R. (1984). Determination of retinal Schiffbase configuration in bacteriorhodopsin. Proc. Natl. Acad. Sci. USA 81, 2055-2059. Smith, S. O., Palings, I., Miley, M. E., Courtin, J., de Groot, H., Lugtenburg, J., Mathies, R. A., & Griffin, R. G. (1990). Solid-state NMR studies of the mechanism of the opsin shift in the visual pigment rhodopsin. Biochemistry 29, 8158-8164. Smith, S. O., Palings, J., Copie, V., Raleigh, D. R., Courtin, J., Pardoen, J. A., Lugtenburg, J., Mathies, R. A., & Griffin, R. G. (1987a). Low temperature solid state 13C NMR studieS of the retinal chromophore in rhodopsin. Biochemistry 26, 1606-1611. Smith, S. O., Pardoen, J. A., Lugtenburg, J., & Mathies, R. A. (1987b). Vibrational analysis of the 13-cis-retinal chromophore in dark-adapted bacteriorhodopsin. J. Phys. Chem. 91,804-819. Song, L., E1-Sayed, M. A., & Lanyi, J. K. (1993). Protein catalysis of the retinal subpicosecond photoisomerization in the primary process of bacteriorhodopsin photosynthesis. Science 261, 891-894. Sperling, W. (1972). Conformations of 11-cis retinal. In H. Langer (Ed.), Biochemistry and physiology of visual pigments (pp. 19-28). New York: Springer-Verlag Inc. Spudich, J. L., McCain, D. A., Nakanishi, K., Okabe, M., Shimizu, N., Rodman, H., Honig, B., & Bogomolni, R. A. (1986). Chromophore/protein interaction in bacterial sensory rhodopsin and bacteriorhodopsin. Biophys. J. 49. 479-483. Steinberg, G., Ottolenghi, M., & Sheves, M. (1993). pica of the protonated Schiff base of bovine rhodopsin. A study with artificial pigments. Biophys. J. 64, 1499-1502. Stem, L. J., & Khorana, H. G. (1989). Structure-function studies on bacteriorhodopsin. J. Biol. Chem. 264, 14202-14208. Stryer, L. (1986). Cycle GMP cascade in vision. Ann. Rev. Neurosci. 9, 87-119. Stryer, L. (1987). The molecules of visual excitation. Sci. Am. 257, 42-50. Subramaniam, S., Greenhalgh, D. A., & Khorana, H. G. (1992). Aspartic acid 85 in bacteriorhodopsin functions both as proton acceptor and negative counterion to the Schiff base. J. Biol. Chem. 267, 25730-25733. Suzuki, T., & Callender, R. H. (1981). Primary photochemistry and photoisomerization of retinal at 77K in cattle and squid rhodopsins. Biophys. J. 34, 261-265. 3+ 3+ Sweetman, L. L., & E1-Sayed, M. A. (1991). The binding ofthe strongly bound Eu in Eu -regenerated bacteriorhodopsin. FEBS Lett. 282, 436. Taiji, M., Bryl, K., Nakagawa, M., Tsuda, M., & Kobayashi, T. (1992). Femtoseeond studies of primary photoprocesses in octopus rhodopsin. Photochem. Photobiol. 56, 1003-1011. Takei, H., Gat, Y., Rothman, Z., Lewis, A., & Sheves, M. (1994). Active site lysine backbone undergoes conformational changes in the bacteriorhodopsin photocycle. J. Biol. Chem. 269, 7387-7389. Tallent, J. R., Hyde, E. Q., Findsen, L. A., Fox, G. C., & Birge, R. R. (1992). Molecular dynamics of the primary photochemical event in rhodopsin. J. Am. Chem. Soc. 114, 1581-1592. Tavan, P. (1988). Stereodynamic coupling of light energy and ion transport in the retinal proteins bacteriorhodopsin and halorhodopsin. Phys. Chem. 92, 1040-1045.

138

JEFFREYA. STUART and ROBERT R. BIRGE

Tavan, P., & Schulten, K. (1986). Evidence for a 13, 14-cis cycle in bacteriorhodopsin. Biophys. J. 50, 81-89. Terner, J., Hsieh, C., Bums, A. R., & E1-Sayed, M. A. (1979). Time-resolved resonance Raman spectroscopy of intermediates of bacteriorhodopsin: The bK590 intermediate. Proc. Natl. Acad. Sci. USA 76, 3046-3050. Thomas, D. D., & Stryer, L. (1982). Transverse location of the retinal chromophore of rhodopsin in rod outer disc membranes. J. Mol. Biol. 154, 145-157. Thompson, L. K., McDermott, A. E., Raap, J., van der Wielen, C. M., Lugtenburg, J., Herzfeld, J., & Griffen, R. G. (1992). Rotational resonance NMR study of the active site structure in bacteriorhodopsin: conformation of the schiffbase linkage. Biochemistry 31,7931-7938. Tindall, B. J. (1992). In A. Balows, H. G. TrOper, M. Dworkin, W. Harder, & K.-H. Schleifer (Eds.), The Prokaryotes: A Handbook on the Biology of Bacteria (pp. 769--808). Berlin: Springer. Tittor, J., & Oesterhelt, D. (1990). The quantum yield ofbacteriorhodopsin. FEBS Lett. 263,269-273. Trissl, H.-W. (1985). I. Primary electrogenic processed in bacteriorhodopsin probed by photoelectric measurements with capacitative metal electrodes. Biochim. Biophys. Acta 806, 124-135. Trissl, H.-W., & G/irmer, W. (1987). Rapid charge separation and bathochromic absorption shift of flash-excited bacteriorhodopsin containing 13-cis or all-trans forms of substituted retinals. Biochemistry 26, 751-758. Trissl, H.-W., G/irtner, W., & Leibl, W. (1989). Reversed picosecond charge displacement from the photoproduct K of bacteriorhodopsin demonstrated photoelectrically. Chem. Phys. Lett. 158, 515-518. Tsuda, M., Glaccum, M., Nelson, B., & Ebrey, T. G. (1980). Light isomerizes the chromophore of bacteriorhodopsin. Nature 287, 351-353. Ulrich, A. S., Heyn, M. E, & Watts, A. (1992). Structure determination ofthe cyclohexene ring of retinal in bacteriorhodopsin by solid-state deuterium NMR. Biochemistry 31, 10390-10399. Urabe, H., Otomo, J., & Ikegami, A. (1989). Orientation of retinal in purple membrane determined by polarized raman spectroscopy. Biophys. J. 56, 1225-1228. van den Berg, R., Jang, D.-J., Bitting, H. C., & E1-Sayed, M. A. (1990). Subpicosecond resonance Raman spectra of the early intermediates in the photocycle ofbacteriorhodopsin. Biophys. J. 58, 135-141. van der Steen, R., Biesheuvel, P. L., Mathies, R. A., & Lugtenburg, J. (1986). Retinal analogues with locked 6-7 conformations show that baeteriorhodopsin requires the 6-s-trans conformationof the chromophore. J. Am. Chem. Soc. 108, 6410--6411. Varo, G., & Lanyi, J. K. (199 la). Effects of the crystalline structure of purple membrane on the kinetics and energetics of the bacteriorhodopsin photocyele. Biochem. 30, 7165-7171. Var6, G., & Lanyi, J. K. (1991b). Kinetic and spectroscopic evidence for an irreversible step between deprotonation and reprotonation of the schiffbase on the bacteriorhodopsin photocycle. Biochemistry 30, 5008-5015. V~ir6, G., & Lanyi, J. K. (1991c). Thermodynamics and energy coupling in the bacteriorhodopsin photocycle. Biochemistry 30, 5016-5022. Wada, M., Sakurai, M., Inoue, Y., Tamura, Y., & Watanabe, Y. (1994). Ab Initio study of 13C NMR chemical shifts for the chromophores of rhodopsin and bacteriorhodopsin. 1. Theoretical estimation of their ring-chain conformations. J. Am. Chem. Soc. 116, 1537-1545. Waddell, W. H., Schaffer, A. M., & Becker, R. S. (1973). Visual pigments. III. Determination and interpretation of the fluorescence quantum yields of retinals, Sehiffbases and protonated Sehiff bases. J. Am. Chem. Soc. 95, 8223-8227. Wald, G. (1968). The molecular basis of visual excitation. Nature 219, 800-807. Warshel, A., Chu, Z. T., & Hwang, J.-K. (1991). The dynamics of the primary event in rhodopsins revisited. Chem. Phys. 158, 303-314. Weiss, R. M., & Warshel, A. (1979). A new view of the dynamics ofsinglet cis-trans photoisomerization. J. Am. Chem. Soc. 101, 6131-6133.

Characterization of the Primary Photochemical Events

139

Xie, A. (1990). Quantum efficiencies of bacteriorhodopsin photochemical reactions. Biophys. J. 58, 1127-1132. Yan, M., Manor, D., Alfano, R. R., Callender, R. H., & Rothberg, L. J. (1993). Initial photochemistry of the visual pigment rhodopsin. Proc. Int. Conf. Lasers 1992, 422. Yan, M., Manor, D., Weng, G., Chao, H., Rothberg, L., Jedju, T. M., Alfano, R. R., & Callender, R. H. (1991). Ultrafast spectroscopy of the visual pigment rhodopsin. Proc. Natl. Acad. Sci. USA 88, 9809-9812. Yoshizawa, T., & Kandori, H. (1991). Primary photochemical events in the rhodopsin molecule. Yoshizawa, T., & Wald, G. (1963). Pre-lumirhodopsin and the bleaching of visual pigments. Nature 197, 1279-1286. Yoshizawa, T., Shichida, Y., & Matuoka, S. (1984). Primary intermediates of rhodopsin studied by low temperature spectrophotometry and laser photolysis. Vision. Res. 24, 1455-1463. Zhang, C.-F. (1989). Two-Photon Spectroscopic and Molecular Orbital Studies of Light-Adapted Bacteriorhodopsin. Ph.D., Carnegie-Mellon University. Zhang, Y. N., Sweetman, L. L., Awad, E. S., & E1-Sayed, M. A. (1992). Nature of the individual Ca2+ binding sites in CaZ+-regenerated bacteriorhodopsin. Biophys. J. 61,1201-1206. Zhang, Y. N., E1-Sayed, M. A., Bonet, M. L., Lanyi, J. K., Chang, M., Ni, B., & Needleman, R. (1993). Effects of genetic replacements of charged and H-bonding residues in the retinal pocket on Ca 2+ binding to deionized bacteriorhodopsin. Proc. Natl. Acad. Sci. USA 90, 1445-1449. Zhukovsky, E. A., & Oprian, D. D. (1989). Effect of carboxylic acid side chains on the absorption maximum of visual pigments. Science 246, 928--930. Zhukovsky, E. A., Robinson, P. R., & Oprian, D. D. (1992). Changing the location of the schiff base counterion in rhodopsin. Biochemistry 31, 10400-10405.

This Page Intentionally Left Blank

LIGHT-INDUCED PROTEIN-PROTEIN INTERACTIONS ON THE ROD PHOTORECEPTOR DISC MEMBRANE

Klaus Peter Hofmann and Martin Heck

Io II.

III.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C o m p o n e n t s o f the Visual Cascade . . . . . . . . . . . . . . . . . . . . . . . A. R h o d o p s i n . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

142 143 146

B.

Transducin

C.

Phosphodiesterase

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

D.

R h o d o p s i n Kinase

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

151

E. F.

Arrestin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Other C o m p o n e n t s . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

152 152

G.

The Disc M e m b r a n e

154

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

148 150

Interactions o f R h o d o p s i n . . . . . . . . . . . . . . . . . . . . . . . . . . . .

156

A.

156

Signaling States o f R h o d o p s i n

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

B.

Interaction with Transducin

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

158

C.

M e c h a n i s m o f Gt Activation

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

162

Biomembranes Volume 2A, pages 141-198 Copyright © 1996 by JAI Press Inc. All rights of reproduction in any form reserved. ISBN: 1-55938-659-2

141

142

IV.

V.

VI.

VII.

KLAUS PETER HOFMANN and MARTIN HECK

D. Formation of a Signaling State without Light . . . . . . . . . . . . . . . . E. Interaction with Rhodopsin K_inase . . . . . . . . . . . . . . . . . . . . . F. Interaction with Arrestin . . . . . . . . . . . . . . . . . . . . . . . . . . . Interactions of the Phosphodiesterase . . . . . . . . . . . . . . . . . . . . . . A. Activated Gt Relieves an Inhibitory Constraint on PDE . . . . . . . . . . B. Mechanism of Interaction and Activation . . . . . . . . . . . . . . . . . . C. Interaction Domains . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Timing and Deactivation of the Cascade . . . . . . . . . . . . . . . . . . . . . A. The Disc Amplifier . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Deactivation of Rhodopsin . . . . . . . . . . . . . . . . . . . . . . . . . C. Deactivation of Transducin and PDE . . . . . . . . . . . . . . . . . . . . Appendix: Posttranslational Modifications . . . . . . . . . . . . . . . . . . . A. Acylation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Isoprenylation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Methylation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D. Phosphorylation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Recent Reviews Related to the Topic . . . . . . . . . . . . . . . . . . . . . . Abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

165 165 168 170 170 170 174 176 176 177 178 180 180 180 182 182 182 183 183 184

I. I N T R O D U C T I O N Membrane-bound protein-protein interactions are key processes in biological signal transduction. The biomembrane reduces the dimensionality of the interactions and provides a barrier for the soluble components involved. With the vertebrate rod outer segment, nature has provided a model system for the study of such interactions. It contains the relevant proteins in abundance, and is accessible to biophysical as well as biochemical techniques. The analysis is simpler than in most related systems, since the rod photoreceptive apparatus is basically specialized to just one purpose, the monochromatic detection of low light levels. Besides the high quantum yield of the receptor protein, the fidelity of signal transduction via protein-protein interactions is most important for the high sensitivity achieved. Absorption of light initiates a cascade of biochemical reactions in the rod outer segment, which finally leads to the hydrolysis of intracellular cyclic GMP, hyperpolarization of the cell and synaptic signaling. Many biochemical reactions can now be assigned to specific steps of the physiological signal transduction pathway. Interactions of the receptor protein, rhodopsin, of the G-protein, transducin, and of the effector, cyclic GMP phosphodiesterase, have been analyzed in detail. In this review, aspects of signal transduction and regulation will be discussed with emphasis on the role of these membrane bound interactions. Other recent reviews dealt with the visual cascade from different perspectives; they are listed in a separate section at the end of the text.

Light-Induced Protein-Protein Interactions

143

II. COMPONENTS OF THE VISUAL CASCADE As shown in Figure 1A, the rod photoreceptor cell consists of an outer and an inner segment, with a thin connecting cilium. The inner segment bears all the organelles for cellular metabolism, such as the nucleus, mitochondria, and the endoplasmic reticulum. All components needed for transduction of the light signal are contained

A

B c

Plasma membrane Cytoplasm

E !._

5'GMP cGMP

0

Na ÷

Ca 2÷

Light

Figure 1. Schematic representation of a rod cell (A) and of the main components of the visual cascade (B). A: The rod consists (from below) of the synaptic body, the nucleus, the inner segment (with the metabolic machinery) and, connected by a cilium, the outer segment, which contains the stack of discs. A bovine disc is approximately 1.5 l~m in diameter. B: The picture shows a section through the rim region of two discs, in apposition to the plasma membrane, and the components of the visual cascade, namely, the visual pigment, rhodopsin (R), the G-protein transducin (G, Gt), and the cGMP phosphodiesterase (PDE). Cytoplasmic components are the nucleotides GTP and GDP and the second messenger cGMP. In the plasma membrane, only the cGMP dependent channel (Ch) is shown. Rhodopsin is an integral protein of the disc membrane, G and PDE are bound to its cytoplasmic surface; the density of R in the disc membrane is 30,000 !~-2; the molar ratio of R:G:PDE is approximately 100:10:1.

144

KLAUS PETERHOFMANN and MARTIN HECK

in the outer segment. Most proteins are associated with the plasma membrane or with the discs, intracellular fiat membrane vesicles forming a long stack in the outer segment. A Na~-ATPase maintains a steady dark current that flows into the outer segment through ion channels in the plasma membrane. Two soluble cytoplasmic transmitter substances have been identified in the outer segment; Ca 2+ and cyclic GMP, which are responsible for the transduction of the signal between discs and plasma membrane. The cellular response, a hyperpolarization of the plasma membrane, is caused by the closure of cyclic GMP dependent ion channels (Ch, Figure 1B). The "negative" transmitter function of cyclic GMP arises from its hydrolysis through a cyclic GMP phosphodiesterase (PDE), which is activated by coupling to the G-protein transducin (G, Gt). The G-protein in turn is catalytically converted into the active, GTP-bound form by interaction with light-activated rhodopsin (R). Rhodopsin, Gt and PDE constitute an archetype of a receptor/G-protein/effector system, with cyclic GMP as the second messenger (see Chabre and Deterre, 1989, and references therein). The closure of hundreds of channels within 100 ms after photon absorption is solely due to the rapid successive activation of these components. Absorption by rhodopsin is the only input for the light signal. On the time scale of seconds, other proteins come into play, which serveto regulate the activity of the cascade and to reset it for graded and repeated activation. Each of the components involved can be represented as part of a reaction cycle, as shown in Figure 2A. The fast transduction chain from light absorption to closure of the ion channels is marked by a background shadow. The light-activated form of rhodopsin, R*, is related to the Meta intermediates of the rhodopsin reaction sequence (see below). Only as the holoprotein, and with GDP bound to the a-subunit, can the G-protein collisionally interact with R* (Figure 2B). Upon binding, the affinity of Gt for GDP becomes very low and the nucleotide dissociates (Bennett and Dupont, 1985). An "emptysite" complex ofG t with R* (R*-Ge) is formed, which remains stable in the absence of GTP. Binding of GTP to the Gt a-subunit within the R*-Gecomplex enables a conformational change that eventually leads to dissociation of the GTP bound Ga-subunit (GaGTP) from R* and GI3~. The GaGTP-subunit can interact with the PDE and activate it. The high rate of receptor-catalyzed nucleotide exchange, typical for the rod system, leads to the rapid accumulation of Gt in its GTP-bound, active form. Non-catalytic binding of

Figure 2. Reactions cycles (A) and kinetics of subunit interactions (B) in the visual cascade. A: Four reaction cycles constitute the cascade: 1. The rhodopsin cycle; it comprises rhodopsin (R), phosphorylated R (Rp), opsin (0), and the proteins that regulate rhodopsin's activity, namely, rhodopsin kinase (RK), arrestin (A), retinal dehydrogenase, (RDH).

A Light (

NAD

1-cis

~J retinal

,~

R ~

RDH/~ V '". p, NADP~ I Op all.traqs.__~,~ A

retinal "~

R~A

A

Phd

I

I

Hyper-

R* ~ RK I

R*Ge

GTP

R*RK

G(~TPPDE~

GGTP~

k-ATP

polarlsatton

PDE

GGDP

PDE

)GC CBP ~

i

:

'

Ca~+ "

........... ]"......... . . . ~ 9

R~Kp R~,RK - V ~ ATP

t

Cao+

+

~---Na

l Disc Membrane

Cytoplasm

Plasma Membrane

,~.~

B 0.3 m/

Light

:','.:,:,®:~',, --"

.,.,:""N'

<~ 1 s ~

5'GMP

'!,~~i1~Ni";", i,, ;i,i"~,i~'~ilil~,'4i,"@ I ~",~,:,;G~,,

cGMP

(~

100

ms

(~

< 200

4 Na+

.~

< 5 ms

(?;

(?) 0.3 ms ~ ,

~

~

"

.........

c~,,P

s (?)

Figure 2. (Continued) 2. The transducin cycle, with photoactivated rhodopsin (R*) as a catalyst of GDP/GTP exchange in the G-protein, and the activation of the PDE. Phosducin (Phd) interacts with the Gt 13y-subunit. 3. The guanine nucleotide cycle, containing the enzymes for resynthesis of cGMP (only the guanylate cyclase, GC, is shown). 4. The calcium cycle; Ca2+ flows through the Ca channel (Ch) and the Na/Ca exchanger (E) and interacts with Ca-binding proteins (CBP), which in turn bind to the GC and to other proteins (?). Section of A, showing in more detail the subunits involved, and a tentative assignment of the reaction times (see text). 145

146

KLAUS PETER HOFMANN and MARTIN HECK

active GaGTP to the PDE, in a stoichiometry of 2:1, keeps the enzyme active, and hydrolysis of cyclic GMP leads to the closure of the channels. After a short photoexcitation of the dark-adapted rod, the membrane potential returns to the dark level within 1-2 seconds (Baylor et al., 1984). The necessary restoration of the cyclic GMP dark level is satisfyingly explained by a feedback loop that activates GMP cyclase via calcium-binding protein(s) (CBP). CBP's are sensors of the drop of intracellular calcium arising from channel closure and persistent extrusion of Ca2+ by the Na/Ca exchanger (E). A complete reset of the transduction system also implies the deactivation of the cascade on all its stages, namely, receptor, G-protein and effector. Gt and PDE activities are thought to be deactivated simultaneously by the intrinsic GTPase activity of the G-protein (see below), closing the Gt/PDE cycle. GaGDP recombines with G~v. Shut off of rhodopsin's activity toward the G-protein involves phosphorylation by rhodopsin kinase and subsequent binding of arrestin (Wilden et al., 1986; Palczewski et al., 1992b). Slow regeneration of rhodopsin closes the rhodopsin cycle; it requires arrestin to be removed first (Hofmann et al., 1992), before dephosphorylation and regeneration with 11-cis retinal can occur. See the section on the timing of the cascade for a more detailed discussion of Figure 2B.

A. Rhodopsin Vertebrate rhodopsins are intrinsic membrane proteins of approximately 40 kDa molecular weight (Applebury and Hargrave, 1986). They share their seven transmembrane helix structure with both the G-protein coupled receptors and the retinal proteins (for recent reviews, see Findlay, 1986; Hargrave, 1991; Hargrave and McDowell, 1992; Hargrave et al., 1993). Bovine rhodopsin, to which this chapter predominantly refers, is 348 amino acid residues in length, identical to human rhodopsin at all but 23 positions (for a general review on visual pigments see Sakmar, 1994). The disposition of the helical transmembrane stretches became recently visible in maps of rhodopsin's projection structure, as obtained from two-dimensional crystals of bovine rhodopsin (Schertler et al., 1993); based on this structural data and previous information, the probable arrangement of the helices could now be determined (Baldwin, 1993). About one half of rhodopsin's protein mass is buffed in the lipid bilayer and one half is exposed in equal portions to the cytoplasmic and intradiscal surfaces. As Figure 3A shows, part of the carboxyl terminus is anchored to the bilayer by palmitoyl-cysteines, so that four "loop" regions and the carboxyl terminus face the rod cytoplasm (discussed in more detail by Hargrave and McDowell, 1992). Rhodopsin has many characteristics in common with the large family of G-protein coupled receptor proteins. Besides the anchoring palmitoylation sites (Ovchinnikov et al., 1988), they include an amino terminal glycosylation site and multiple carboxy terminal phosphorylation sites for a receptor kinase (see Palczewski and Benovic, 1991). It applies for all G-protein coupled receptors that a specific kinase recognizes

Cytoplasm

A

j O O H

Dis~l~~

'

memb~lle:~

I

,,,l~fl,,l,,,: ,

II~~"'~)'iI'

R

MII'Ge

D C B

B

"I~I~'~',~!,,,,

':I! ,,,,;,,,,,,,

R

RK

c R

g

MI

Gt

MII a

MII b

A

MII-P ii-phosphorylated

Loop conformation d e v e l o p e d Schiff base proton translocated Steric trigger operated ( •

-> •

)

C h r o m o p h o r e isomerized

Figure 3. Signaling states of rhodopsin. A: Seven-helix structureof rhodopsin and its disposition in the membrane, simplified representation of the four cytoplasmic loops, and of the 3-loop interaction mode of the G-protein in its empty site complex (MII-Ge). B: Hydrophobic core of the protein, with the protonated retinal Schiff base bond to Lysin-296, the Glu-113 counterion, and some functionally important charged residues. The righthand picture symbolizes rhodopsin's ground state, with the chromophore in its twisted 11-cis conformation, the location of the Schiff base proton, and the steric situation (filled circle). C: Sequence of rhodopsin photoproducts and signaling states for interacting proteins. Light-induced retinal cis/trans-isomerization yields the batho form (B); the subsequent "steric trigger" (limited conformational change by steric interaction between chromophore and apoprotein)is necessary for the formation of Meta I (MI); translocation of the Schiff base proton and altered lipid-protein interaction is linked to Mlla, while MIIb requires the additional uptake of H+ from the aqueous phase. The G-protein (Gt) is bound and activated by MIIb, while the rhodopsin kinase (RK) interacts with forms Mi and MII; arrestin (A) binds to phosphorylated MII. 147

148

KLAUS PETER HOFMANN and MARTIN HECK

the activated form and that an arrestin-like protein is bound with high affinity to the phosphorylated receptor. These interactions occur in activated or "signaling" state(s), in which a bound agonist holds the protein in a state compatible with the interacting proteins. In these terms, the photochemical effect of light absorption (the isomerization of the chromophore retinal, see below), may be understood as a light-induced transformation of the chromophore from an antagonist to an agonist. The ligand-binding pockets of different receptors show surprising similarities (Oprian, 1992 a,b). The chromophore, 11-cis retinal is bound as a protonated Schiffbase to Lys-296, in the center of helix G. The pK a of the Schiffbase in the environment of the retinal binding site is extremely high (apparent pK a above 16; Steinberg et al., 1993). Negative charges are buried in the hydrophobic part of the protein, located along the helix C, at carboxylic acids of residues Glu-113, Glu-122, and Glu-134, and at helix B (Asp-83). The residue Glu-113, at the luminal border of the third helix, provides the counterion for the protonated retinal Schiffbase (Sakmar et al. 1989; Nathans, 1990; Zhukovsky and Oprian, 1989). The loop regions are loci of interaction with Gt, rhodopsin kinase and arrestin (Figure 3); three cytoplasmic loops have been identified as parts of the interaction domain with the G-protein (see below). We wish to know how the signaling states arise from the initial light-induced chromophore isomerization. It is known that these states are only reached after a multistep thermal transformation of the protein, subsequent to the initial photochemical events. In this sequence oftransforrrations, protonation changes play a major part, which establishes a profound relationship to other retinal proteins, including some with quite different biological function. B. Transducin

Rod transducin belongs to the extended family of the heterotrimeric G-proteins, which form a sub-group ofthe GTPases (Gilman, 1987; Bourne et al., 1990, 1991). G-proteins generally mediate between sensory, hormonal and neurotransmitter receptor proteins and effect of proteins such as cyclic GMP phosphodiesterase (rod and cone transducins, Gtr and Gtc), adenylyl cyclase (Gs and Gi) , phospholipase C (Gq) and various channels (Gs, Gi, Go). The heterotrimers are composed of the nucleotide-binding t~-subunit with a molecular weight of 39-45 kDa, a fl-subunit of 35-36 kDa and a y-subunit of 5 kDa. For all heterotrimeric G-proteins, the activation cycle is basically as shown in Figure 2. In contrast to most heterotrimeric G-proteins, which are anchored in their host membrane (for a review, see Gilman, 1987), Gt is only peripherally bound to the disc membrane. The Gt holoprotein in its inactive, GDP-binding form can be easily extracted from the membrane at low ionic strength (KOhn, 1980); under these conditions, Ga leaves the undissociated GI3~ dimer. The transition from solution to the membrane bound state occurs with a considerable activation energy (44 kJmol-1), and may therefore be accompanied by a change of conformation and/or

Light-Induced Protein-Protein Interactions

149

interaction in the subunit polypeptides or hydrocarbon modifications (Schleicher and Hofmann, 1987; see appendix). It has been reported that the affinity of Ga for G~v is much higher when the Gy-subunit is famesylated (Fukada et al., 1990; Ohguro et al., 1991; Perez-Sala et al., 1991), and that the famesylation is indispensable for GTP-binding (Fukada et al., 1990). The high degree of conservation of the nucleotide-binding sites on ot-subunits indicates structural similarity also to ras p21 and EF-Tu, whose three-dimensional structures have been determined (for a review, see Hamm, 1991). However, differences even in the highly conserved nucleotide binding region are indicated by the finding that caged GTE which binds avidly to ras proteins, promotes G t activation with more than 10 times lower efficiency than GTP (Heck and Hofmann, unpublished observations). Structural models for the G t ot-subunit have been proposed, based on the ras p21 and EF-Tu crystal structures (reviewed by Hamm, 1991). The tertiary structure of the GTP~S-bound G t ot-subunit has been determined recently (Noel et al., 1993; see below). Each of the three subunits of the rod transducin heterotrimer is highly homologous to the respective subunits in other G-proteins (see Hamm, 1991). Rod and cone cells contain different 13subunits, 131 in rods and [33 in cones (reviewed in Yau et al., 1992). The 137-subunit must be present in the interaction ofG t with rhodopsin, to enable the Meta II stabilizing empty site interaction; neither the c~-subunit nor the ]37-subunit alone can stabilize the active Meta II conformation suggesting that either binding sites on both subunits or a [3v-induced change of the conformation of the ot-subunit is required for catalytic interaction with rhodopsin. Evidence has been presented that 137 alone can interact with rhodopsin (Kelleher and Johnson, 1988), via interaction with the C-terminus (Phillips and Cerione, 1992); it remains open whether this is a rapid light-induced interaction. The 137-subunit of other G-proteins can play an active role and can stimulate effectors (Camps et al., 1992; Hamm, 1991; Tang and Gilman, 1991). An analogy for the rod system is not known, but [3T interacts with rod phosphoproteins (Lee et al., 1987; Suh and Hamm, 1991). Considerable heterogeneity of transducin has been found in frog rods (Umbarger et al., 1992). Functions of both the ~x and 13T subunits require different interaction sites with other proteins. This is obvious for the et-subunit, which must have specialized surface regions for its [3Vpartner, for rhodopsin, and for the PDE (see below). In a variety of G proteins, myristoylation at the amino terminal glycine appears to be important for high affinity interaction with 137and with the membrane (Mumby et al., 1990; Jones et al., 1990). Transducin shows considerable heterogeneity of its fatty acyl moiety at this position, with a mixture of C12, C14:0 (myristate) and C 14:1 modifications (Neubert et al., 1992). The amino-terminal 17 amino acids of G~ are involved in the interaction with 13T (Navon and Fung, 1987), comparable to the residues 2-29 of the G s ot-subunit (Journot et al., 1991). This may explain why peptides from the N-terminal region of Ga relatively destabilize Meta II (Hamm et al., 1988), since stabilization requires the Gt-holoprotein (K6nig, unpublished).

150

KLAUS PETER HOFMANN and MARTIN HECK

C. Phosphodiesterase The PDE hotoenzyme is heterotrimeric (Baehr et at., 1979), composed of two undissociabte, similar PDE~ (88 kDa) and PDEI3 (84 kDa) subunits and two tightly associated, identical PDEv subunits (11-13 kDa) (Deterre et at., 1988; Whalen and Bitensky, 1989; Fung et at., 1990). Cross-linking studies suggest that each of the larger subunit binds one •-subunit (Fung et at., 1990; Cterc et at., 1992, Artemyev and Hamm, 1992). The finding that artificial (maximal)activation can be achieved by tryptic proteolysis of the ~/-subunit (Hurley and Stryer, 1982) which is reversed on addition of purified PDEv (Hurley and Stryer, 1982; Wensel and Stryer, 1986) shows that in PDEal3resides the catalytic (hydrolytic) activity, which is blocked in the latent muttimeric form by the two inhibitory PDEcsubunits. Sequence comparisons with other cyclic nucteotide-binding proteins predict a catalytic cGMPbinding site located .in the C-terminal region of PDE~ and PDEI3, respectively (Ovchinnikov et at., 1987; Lipkin et at., 1990; Li et at., 1990). In addition, the studies on bovine PDE presume up to two non-catalytic cGMP binding sites on each catalytic subunit (Charbonneau et al., 1990; Lipkin et at., 1990; Li et at., 1990), which is further confirmed by quantitative analysis of tightly bound [endogenous/copurified] cGMP [ 1.8+0.3 mot cGMP per mot PDE, (Gittespie and Beavo, 1989]) and by labeling studies (one binding site on PDE~, Thompson and Khorana, 1990). Related studies on frog PDE have been recently reviewed by Yamazaki (1992). Despite the high degree of homology among PDE isoenzymes in different animal species, some functional differences were reported, including the affinity and role of the non-catalytic cGMP-binding sites on the one hand and the intermediate states of the activation process on the other (Whalen and Bitensky, 1989; Whalen et al., 1990; reviewed in Yamazaki, 1992 and Pfister et al., 1993). Unless otherwise noted, we refer in the following to bovine PDE. Native PDE is peripherally bound to the disc membrane (Baehr et al., 1979); a soluble isoenzyme with low affinity for transducin has been reported to account for 20-30% of the PDE activity in bovine ROS (Giltespie et at., 1989). The PDE can be reversibly detached from the membranes either by isotonic dilution of the membranes (Catty et at., 1992) or by extraction with low ionic strength buffers (Miki et al., 1975; Baehr et al., 1979). Furthermore, PDE is irreversibly released from the membranes by limited treatment with trypsin (Wenset and Stryer, 1986), which cleaves a short peptide from the C-terminus of both, PDEa and PDEI3, prior to the functional degradation of the PDEv subunits (Ong et at., 1989; Catty and Deterre, 1991). The membrane association is therefore thought to be mediated by the C-terminus of PDE~ and PDE~, respectively. Both subunits possess a C-terminal CAAX-motif (C = Cys, A = mostly aliphatic residue, X = any residue), which appears to signal posttranstationat prenytation of the cysteine, proteotytic removal

Light-Induced Protein-Protein Interactions

151

of AAX, and carboxymethylation of the terminal cysteine (see Marshall, 1993). Based on the pattern of substrate specificity of prenyl-transferases, a geranyl-geranylation (C20) of PDEf~ and a famesylation (C 15) of PDE~ has been predicted (see Clarke, 1992); this is indeed found in vivo in rat retinal PDE (Anant et al., 1992) and in vitro for mouse PDE, expressed in bacteria (Qin et al., 1992). In addition, a small fraction of both termini is found to be carboxymethylated in vitro (Ong et al., 1989; Catty and Deterre, 1991). The geranylgeranyl group seems to be responsible for most of the membrane attachment of PDE (Catty and Deterre, 1991; Anant et al., 1992), while the famesyl group may be involved in functions other than membrane binding (possibly including protein-protein interactions), as proposed for the Gcsubunit (Fukada et al., 1990).

D. Rhodopsin Kinase Rhodopsin kinase is a member of the family of cyclic nucleotide-independent receptor kinases (for reviews see Palczewski and Benovic, 1991, and Lefkowitz et al., 1992). Similarity to rhodopsin kinase in the primary structure was demonstrated for the [3-adrenergic kinase (Lorenz et al., 1991). Rhodopsin kinase is a single polypeptide chain, with a molecular weight of 67,000 daltons. It is highly specific towards its protein and nucleotide substrate. ATP is the preferred nucleotide, and only light-activated rhodopsin can be a substrate of the kinase. Although rhodopsin kinase shares with the other receptor kinases the principal task of receptor phosphorylation and inactivation, interesting differences exist to the 13-adrenergic kinase and, presumably, other members of the family (Lefkowitz et al., 1992; Haga and Haga, 1989). For example, rhodopsin kinase, but not ]3-adrenergic kinase, is effectively inhibited by sangivamycin. Importantly, rhodopsinkinase is farnesylated, and this modification is indispensable for its interaction with photoactivated rhodopsin (Inglese et al., 1992; see Appendix). The ]3-adrenergic kinase, however, is not prenylated; it is thought to bind to the membrane via the prenylated [3T-subunit of the G-protein (after G subunit dissociation upon interaction with the receptor). A very significant feature of rhodopsin kinase is its ability to autophosphorylate (Palczewski et al., 1992a); this light-induced event is probably involved in the dissociation of the enzyme from the receptor (see below). Rhodopsin kinase is localized in the outer segments of the rods and cones (for human and bovine retinae; Palczewski et al., 1993); its subcellular localization is unknown. Since its purification requires detergent, it appears to be at least loosely membrane bound in situ. From the yield of kinase preparations, a 1% rnol fraction in bovine rod outer segments, relative to rhodopsin, is estimated.

152

KLAUS PETER HOFMANN and MARTIN HECK

E. Arrestin Like rhodopsin kinase, the arrestin of the bovine rod cell belongs to a protein family with considerable sequence homology, including the well-studied 13-arrestin (Lohse et al., 1990). Arrestin is identical to S-antigen (Pfister et al., 1985). A recent study has suggested that autoantibodies reactive with 13-arrestin and arrestin are present in sera from multiple sclerosis patients, and may be related to the course of progression of the disease (Ohguro et al., 1993). The retinal protein was originally described as a 48 kDa protein that redistributed from solution to rod outer segment membranes upon light-activation (Ktihn, 1978); it was termed arrestin, based on its proposed role as an arresting transmitter from photoactivated rhodopsin to the effector phosphodiesterase (Zuckerman and Cheasty, 1986, 1988). cDNA analysis has revealed that bovine arrestin has a molecular weight of 45.3 kDa (404 amino acids in a single polypeptide chain; Shinohara et al., 1987). In toad rod outer segments, the relative amount of arrestin increases with illumination, by migration from the inner segment (Mangini and Pepperberg, 1988). Retina preparations yield arrestin in a mol fraction of 10%, relative to rhodopsin. Insect photoreceptors contain two differentially regulated arresting (LeVine et al., 1991). The 49 kDa form (Arr2) binds to photoactivated insect rhodopsin and quenches light-induced GTPase activity (Byk et al., 1993). For the Arr2 homolog of the blowfly, binding of the purified protein to isolated photoreceptor membranes could now be studied. The results suggest a regulatory mechanism, in which the arrestin blocks the visual pigment in its unphosphorylated form and stimulates phosphorylation as a secondary inactivation step (Bentrop et al., 1993). Assuming fast binding of bovine arrestin to (phosphorylated) bovine rhodopsin (Schleicher et al., 1989), the mechanism could explain the rapid inactivation of invertebrate pigments (100 ms in Limulus; Richard and Lisman, 1992).

F. Other Components The interactions of the proteins presented so far will be discussed below in detail. In the following, some other proteins will be briefly reviewed, whic h are probably involved in light-induced interactions.

Phosphataseand Retinol Dehydrogenase Only on reduction ofall-trans-retinal to all-trans-retinol, is arrestin released from phosphorylated opsin (Figure 2; Hofmann et al., 1992); this unlocks the regeneration of opsin with 11-cis-retinal (Hofmann et al., 1992) and its dephosphorylation by a phosphatase 2A (Palczewski et al., 1989). Retinol dehydrogenase was purified from bovine rod outer segments. The enzyme is membrane-bound and labile when solubilized by detergents; it can be stabilized by NADP (Ishiguro et al., 1991; for assays of retinoid dehydrogenases, see Saari et al., 1993).

Light-Induced Protein-Protein Interactions

153

Guanylate Cyclase, Calcium-Binding Proteins and Phosducin When the cGMP dependent channels in the plasma membrane of a rod outer segment close, the balance between Ca 2÷ influx through the channels and the effiux through the Na/Ca exchanger is disturbed, and the cytosolic calcium concentration decreases rapidly (with a half mean time of 0.5 s). It is widely accepted that this drop of calcium represents a signal for restoration of the depleted cGMP pool, that is, for reset of the photoresponse after a short stimulus and for the adaptation to background illumination (for reviews from the electrophysiological and biochemical point of view see Yau, 1991, and Koch, 1992, 1994, respectively). Although this input alone cannot explain the full complexity of the calcium-dependence of light adaptation (Detwiler and Gray-Keller, 1992; Rispoli and Detwiler, 1992), several loci have been identified, where the drop in calcium is sensed. They include the cGMP-dependent channel, a particulate form ofguanylate cyclase, and the rhodopsin/transducin amplifier. Most, if not all the sensors appear to involve a calciumbinding protein (CBP, Figure 2A; see Koch, 1994, for a recent review). By interaction with target enzymes in the visual cascade, they translate changes in Ca 2÷ concentrations into a change of enzymatic activity or of biochemical gain. Guanylate cyclase (GC) from rat photoreceptors is sensitive to variation in Ca 2÷ concentration (Lolley and Racz, 1982). The effect occurs within the physiologically relevant range and is mediated by a soluble factor isolated from bovine rod outer segments (Koch and Stryer, 1988). This factor was identified as a 26 kDa protein, termed recoverin (Dizhoor et al., 1991; Lambrecht and Koch, 1991a). However, according to recent evidence, recoverin may not be the (or at least not the only) calcium-sensitive factor involved in GC regulation (Hurley et al., 1993; Gray-Keller et al., 1993; Koutalos and Yau, 1993; Koch, 1994). Based on a novel assay of guanylate cyclase activity (Gorczyka et al., 1994a), a new protein with a molecular weight of 20 kDa has been found, which exhibits GC regulating capacity (Gorczyka et al., 1994b). A different role for S-modulin, the frog protein that is closely related to recoverin (Kawamura et al., 1993) is suggested by evidence from biochemical and electrophysiological work. Dependent on calcium, the protein can influence the activation of PDE and the phosphorylation of rhodopsin (Kawamura, 1993). Using recoverin bound to a gel matrix, a direct interaction between recoverin and rhodopsin kinase could now be demonstrated (Hurley, J.B, personal communication). Recoverin affects not only the falling phase but also the size and the rising phase of photoresponses from truncated rod outer segments (Lagnado and Baylor, 1994). These results have led to the notion that recoverin may affect the lifetime of active rhodopsin and/or its efficiency in the catalysis of nucleotide exchange in G c Here, it should be noted that a direct measure of R* lifetime is provided by a characteristic delay of the photoresponse (Pepperberg et al., 1992; Corson et al., 1994); measurements under background light on intact salamander rods do not show the delay

154

KLAUS PETER HOFMANN and MARTIN HECK

expected for a calcium-mediated effect of background light on R* lifetime (Pepperberg et al., 1994). A fundamentally different, calcium-independent feedback loop may involve phosducin, a protein which binds tightly to the 13,{-subunit of transducin (Lee et al., 1987). The notion that this may be a key interaction in the regulation of the visual cascade (Lee et al., 1992), gains support from findings that protein kinase A phosphorylates phosducin on its Ser-73 residue (Lee et al., 1990), and that the inhibition of GTPase of the G-protein G s is regulated in the related 13-adrenergic system by phosphorylation of a phosducin analogue via protein kinase A (Bauer et al., 1992). The guanylate cyclase (Shyian et al., 1992), recoverin (Ray et al., 1992) and phosducin (Bauer et al., 1992) have been cloned and sequenced, and the proteins are all available in purifed form (for a review of the biochemical properties, see Koch, 1994). However, studies on the mechanism of interaction between these and other proteins are not yet available.

Plasma Membrane Channel and Exchanger Both the cGMP-dependent channel and the Na/Ca exchanger are well-studied; the proteins are exclusively located in the plasma membrane and may be associated with one another (Bauer and Drechsler, 1992). The native channel configuration comprises several subunits, and interaction with calmodulin (Molday et al., 1990; Hsu and Molday, 1993; Chen et al., 1993). For these plasma membrane proteins, interaction studies are not yet available; we refer to reviews on the exchanger (Schnetkamp, 1989) and on the channel (Kaupp, 1991; Kaupp and Koch, 1992; Koutalos and Yau 1993). Other Proteins

Besides the calcium-sensitive particulate guanylate cyclase, a soluble, nitric oxide sensitive form has been described (Schmidt et al., 1992; Sitaramayya and Margulis, 1992). Protein kinase C can be involved in the phosphorylation of rhodopsin (Newton and Williams, 1991). Of special potential interest for the topic of this review, but still not assigned functionally, are the reversible methylation reactions of the isoprenylated cysteines observed with a number of rod outer segment proteins (Perez-Sala et al., 1991; Tan and Rando, 1992). Interaction of small G-proteins with photoactivated rhodopsin has been reported (Wieland et al., 1990). Finally, we note the activation of phospholipase A 2 by light and its modulation by GTP-binding proteins (Jelsema, 1987). G. The Disc Membrane The disc membrane is remarkable for several reasons, including its very high fluidity, the high unsaturation of the lipid phase, and the enormous concentration

Light-Induced Protein-Protein Interactions

155

of intrinsic membrane protein it hosts. Many of these properties can now be assigned to details of its composition and molecular architecture.

Lipid

Composition

The major phospholipids in the bovine disc membrane (expressed in moles/mole rhodopsin) are: 32, phosphatidylethanolamine (PE); 30, phosphatidylcholine (PC); 11, phosphatidylserine (PS); values for cholesterol are only 8, and for free fatty acids 6 moles/mol rhodopsin (Miljanich et al., 1981). The phospholipids are polyunsaturated to a high percentage and mainly 20-22 carbons in length (Miljanich et al., 1981; Wang and Anderson, 1992). In recombinant PC membranes with short chain length (C 14 or shorter), the active rhodopsin intermediate Meta II does not form. In PC model membranes, the cholesterol content (Mitchell et al., 1990), temperature-dependent phase transitions (Ryba et al., 1993) and the influence of chain length (Ryba and Marsh, 1992) have been studied with respect to Meta II formation. Furthermore, the identity of the lipid headgroups was varied, with the lipid acyl chain composition held constant at that of egg PC (Gibson and Brown, 1993). These studies agree in the conclusion that any alteration of the lipid host changes the apparent pK a of the MI/MII transition (see the section on Meta II below). Plasma membranes prepared from bovine ROS (Molday and Molday, 1987) have a much lower PDE activating capability than disc membranes, although the rhodopsin content is similar (Boesze-Battaglia and Albert, 1990); the effect can be assigned to the higher cholesterol content (28% of total lipid) of the plasma membrane. The lower activity is parallel to the lower yield ofMeta II photoproduct seen in model membranes with high cholesterol content (Mitchell et al., 1990).

Asymmetry The disc membrane is asymmetric with respect to the unsaturation (the outer monolayer is more unsaturated) and to lipid distribution (Miljanich et al., 1981). Using a novel electron-electron double resonance method, Wu and Hubbell (1993) could show that spin-labeled derivatives of PC, PE and PS redistribute by transmembrane flip-flop with a half time of less than 10 minutes between the outer and inner monolayer of native disc membrane vesicles. At equilibrium, this method predicts that 63% of the PE, 47% of the PC, and 18% of the PS reside in the inner layer (which is the intradiscal layer of the native, stacked discs). The presence at high concentration of the well-oriented and bipolar rhodopsin molecule can explain most of the asymmetry of the membrane (Hubbell, 1990); in natively stacked discs, with their close apposition of the intradiscal surfaces and their high luminal calcium content, the asymmetry may be different (Wu and Hubbell, 1993; Schnetkamp, 1985).

156

KLAUS PETER HOFMANN and MARTIN HECK

Lipid-protein Interaction The induction of phospholipid asymmetry by rhodopsin does not invoke any specific interactions with associated lipid (Hubbell, 1990). This confirms the conclusion drawn from the linear dependence of the free energy of activation for the MI/MII conversion on added lipid to solubilized rhodopsin (Litman et al., 1981). Later analyses came to basically the same conclusion, stating that the bulk lipid environment of rhodopsin is the most critical determinant of rhodopsin function (Ryba et at., 1993). Recent work specifies the interactions in terms of the concept of surface free energy of the protein (Baldwin and Hubbell, 1985) by the suggestion of lateral and/or curvature stresses, which involve both the protein/lipid and the lipid/water interfaces, in concert with functional transitions of the proteins (for the well-studied example of rhodopsin, see Gibson and Brown, 1993, and the discussion on the Meta II state below).

Ion

Permeability

The disc membrane exhibits in the dark no significant permeabilities to charged species other than K÷; the proton permeability is very low (Uhl et al., 1980), and depends on light (Smith and Fager, 1991). The well-defined permeability to ions is highly dependent on the preparation procedure (Uhl et at., 1980); this was also seen in studies of light-induced interfaciat potentials (Cafiso and Hubbell, 1980). In this context, we note that the rate of Gt activation is faster in functioning retinae, as compared to membrane preparations (Kahlert and Hofmann, 1991).

III.

INTERACTIONS OF R H O D O P S I N A. SignalingStates of Rhodopsin

Retinal Isomerization and Steric Trigger Raman spectroscopy has shown that, in the ground state ofrhodopsin, the residue Glu-113 is negatively charged and stabilizes the positively charged Schiff base group by electrostatic interaction (Figure 3; Lin et at., 1992). This work and, even more specifically, 13C-magic angle spinning NMR (Han et al., 1993) have confirmed a model of the chromophore binding site (Birge et at., 1988; Birge, 1990), in which the carboxylic acid group of Glu-113 interacts with a carbon atom (C12) on the retinal hydrocarbon chain. Absorption of a photon results in photochemistry, which leads (with a quantum efficiency of 0.7) from an 11-cis to an 11-trans conformation of the chromophore. In this first step, two thirds of the photon energy are fixed in the protein (see Birge, 1990, for a recent review). From this state, rhodopsin relaxes through intermediates, each representing a specific state of chromophore-protein interaction reflected in its absorption spectrum. These classical intermediates include bathorhodopsin (B), lumirhodopsin, and metarhodopsin I and II (MI and MII; for a review see, for example, Hofmann,

Light-induced Protein-Protein Interactions

157

1986). Additional intermediates have been recently disclosed (for a detailed review see Lewis and Kliger, 1992). The proposed kinetic schemes are complex, including the slow and fast formation of two different Meta II species in branched pathways (Straume et al., 1990) and early Meta II-like forms (Thorgeirsson et al., 1992, 1993). Functionally important forms MII a and MII b could recently be discriminated by a combined kinetic and chemical analysis. Fig. 3C shows these forms and the phosphoryiated MII-P form; they cannot be distinguished by their absorption spectrum, but only by their chemical (H ÷ uptake) or biochemical (e.g. binding of arrestin) function. Evidence for the assignment of the different states to the binding of rhodopsin kinase (RK), G t and arrestin (A) will be discussed below. Rhodopsin shares with bacterial sensory rhodopsins a "steric trigger" mechanism required for its activation (Ganter et al., 1989; Yan et al., 1991); in rhodopsin, it involves the 9-methyl group of the retinal. It precedes any of the transformations linked to Schiffbase deprotonation (discussed below), and is thought of as one of the key events that mediate between retinal isomerization and the signaling conformation. The steric trigger is operated during the transition from Lumirhodopsin to Meta I (Figure 3C).

Deprotonation of the Retinal Schiff Base The photoproduct metarhodopsin II (as defined by its 380 nm absorption maximum, i.e. MII a or MII b in Figure 3C) is formed in milliseconds. It represents the first "bleached" form of the pigment; in molecular terms, it is the state of the protein where the Schiff base bond is still intact, but deprotonated. The deprotonation is directly seen in the change of the maximal absorption to 380 nm absorption of the photoproduct (Doukas et al., 1978). Proton release into the aqueous phase is not observed in the time domain of Meta II formation (Amis and Hofmann, 1993), providing evidence that the Schiff base proton is translocated to another group in the hydrophobic core of the protein. Alterations that occur only in Meta II, and not in the preceding intermediates, include: (see Hofmann, 1986, for a review) (i) an altered orientation of the chromophore, (ii) electrical effects occurring at the time scale of MII formation, (iii) the access of small solutes to the chromophore, (iv) a net volume change of the protein, and (v) the uptake of protons from the aqueous phase. Molecular spectroscopy has indicated that considerable changes occur in the apoprotein moiety during the transition to Meta II (reviewed by DeGrip, 1988; Siebert, 1992). The change in the thermodynamic variables that accompanies the MI/MII conversion, is consistent with partial unfolding of the polypeptide chain (Gibson and Brown, 1993). Meta II remains in equilibrium with its predecessor, Meta I, and its successor Meta III, which are both Schiffbased protonated forms (Kibelbek et al., 1991). The Meta II equilibria depend on temperature, pH, and ionic strength, as well as on properties of the hydrophobic host (lipid bilayer or detergent micelle). Partially digested preparations (Kt~hn and Hargrave, 1981) have been used to investigate the specific contribution of the cytoplasmic loop regions (Ganter et al., 1992). This region has

158

KLAUS PETER HOFMANN and MARTIN HECK

become accessible to direct investigation by spin label techniques (Resek et al., 1993). Site-directed mutagenesis, for example replacements of the Schiff base counterion (Sakmar et al., 1989), has specified parts of the sequence involved in the formation of Meta II. We shall discuss these microscopic assignments after dissecting the different forms of Meta II, MII a and MII b, and their role in the catalysis of nucleotide exchange in the G-protein. Current kinetic schemes (Baumann and Reinheimer, 1973) suggest that both species, Meta II and Meta III, decay in parallel in slow reactions to the end products, the free apoprotein opsin and all-trans-retinal, by hydrolysis of the Schiffbase bond (see Hofmann, 1986). B. Interaction with Transducin

Active Photoproduct It has been shown t.hat the chromophore-protein interaction in the Meta II state is a prerequisite for rhodopsin to adopt a conformation recognized by the G-protein. The first evidence arose from a stabilizing effect, which the G-protein exerts specifically on the Meta II photoproduct (Emeis and Hofmann, 1981; see Hofmann, 1986, 1993). At temperatures near 0 °C and pH > 7, a flash of light yields a low amount of Meta II, in equilibrium with its precursor Meta I. When transducin (but no GTP/GDP) is added, extra Meta II is formed, because stable binding to Meta II prevents its transition back to Meta I (reviewed in Hofmann, 1986; 1993). Inversely, active site methylated rhodopsin, which cannot form Meta II, is inactive toward G t (Longstaff et al., 1986). When the Schiff base nitrogen is methylated instead of protonated, the activity toward transducin is low (Longstaff et al. 1986) and the modified rhodopsin does not form Meta II but decays directly from Meta I to a Meta III-like species (Ganter et al. 1991). Substitution of the active site lysine by glycine (mutant K296G) leads to a form ofrhodopsin that does not bind retinal covalently; however, this mutant can non-covalently bind the protonated retinyl Schiff base 11-cis-retinyl-n-propyl amine and activates transducin with light (Zhukovsky et al. 1991). Not all the proteins that interact with rhodopsin upon light, need the Meta II form to recognize photoactivated rhodopsin (see below). The Meta II stabilization can serve as a quite general criterion, whether a given protein specifically interacts with Meta II.

Cytoplasmic Interaction Domain Peptide competition (K6nig et al., 1989b) and protein engineering (Franke et al., 1990, 1992) have led to the proposal that three cytoplasmic loops of rhodopsin are involved in forming rhodopsin's signaling state. Each of the second, third, and fourth cytoplasmic loops (connecting helices C and D, E and F, and G with the palmitoyl anchor, respectively, see Figure 3A) must contain at least one binding site. For the fourth loop, evidence has been provided that the fixation to the membrane by palmitoylation is not relevant for G t activation (Morrison et al., 1991;

Light-Induced Protein-Protein Interactions

159

Kamik et al., 1993). The 3-loop interactive state is the centerpiece of a sequence of interactions during nucleotide exchange. During transient states, both prior to and after the 3-loop interaction, G t binds to photoactivated rhodopsin but does not stabilize the Meta II form. Loop mutants have shown that the replacement of part of the second loop by an inactive sequence or the excision of part of the third loop impairs the activation of nucleotide exchange catalysis in G t (Franke et al., 1990, 1992).

Proton Uptake and Active Conformation Matthews et al. (1963) have established two major characteristics of the MI/MII conversion: the 480/380 nm absorption change, related to deprotonation of the Schiffbase, and the pH-dependence, favoring Meta II above Meta I by acidic pH. The negative correlation with the concentration of hydrogen ions is in apparent contradiction to the release of the Schiffbase proton during the conversion. Parkes and Liebman (1984) reinvestigated the temperature- and pH-dependence of the reaction kinetics and interpreted them in terms of true forward and reverse rate constants. The decrease of the on-rate in the alkaline branch of the bell-shaped pH-rate curve was assigned to the uptake of protons, while the acidic branch was attributed to base catalysis. The mechanism substantiates the notion of a conformation change enabled and/or stabilized by translocation of the Schiffbase proton and the uptake of proton(s) from the aqueous phase (Emrich and Reich, 1974). By a kinetic study on detergent-solubilized protein, it was recently found that the deprotonation of the Schiffbase and proton uptake do not occur in one and the same concerted transformation of the protein (Amis and Hofmann, 1993). The chemical reactions belong to two different forms of the protein, MII a and MIIb: MI

+nil ÷ MII a MII b

The MIIa/MII b transition has a considerable activation energy (60 kJmol-1). The transformation of the protein determines the apparent pK a (6.75). Both transformations in the coupled equilibria will affect the overall yield ofMeta II photoproduct (i.e. the sum ofMII a and MIIb). Thus the apparent pK a measured for the dependence of the apparent MI/MII equilibrium will depend on all parameters that influence either MII a or MII b formation. The sensitivity of rhodopsin in the Meta II state to the hydrophobic environment in membranes (see above) or detergent (Litman et al., 1981; Schleicher et al., 1989; K6nig et al., 1989a) is already present in MII a. In membranes, MII a may depend on the lipid fatty acid composition and the cholesterol content, related to the volume available for molecular motion within the core of the bilayer (Mitchell et al., 1990) and to a lipid reorganization during the MI/MII conversion (Ryba et al., 1993). Properties of the hydrophilic milieu, including pH and ionic strength, influence the

160

KLAUS PETER HOFMANN and MARTIN HECK

protein only in MII b. It may represent the species that gives Meta II its sensitivity to the lipid polar headgroups (Gibson and Brown, 1993). From the pH dependence of G t activation, it was concluded that both MII a and MII b must be successively formed to generate full activity towards the G-protein. The MIIa/MII b transition may reflect the development of the helix-loop conformation related to the 3-loop interaction pattern of MII-G interaction (K6nig et al., 1989b; Franke et al., 1990). In general terms, rhodopsin appears to consist of two domains: while the chromophore reaction domain remains in the Schiff base deprotonated state, proton uptake and release can shift the cytoplasmic domain between its active or inactive states. The light-induced H+-uptake suggests to understand rhodopsin as an incomplete proton pump. Like in bacteriorhodopsin, translocation of the Schiffbase proton is a necessary precursor of the aqueous proton movement. Beyond this fundamental analogy there is kinetic evidence for two distinct M states in bacteriorhodopsin (Lanyi, 1992).

Residues Involved in Gt Interaction The retinal Schiff base is stabilized in the dark by its counterion, Glu-113 (Lin et al., 1992). Replacement of Glu-113 by Gln (mutant Ell3Q) reduces the Schiff base pK a, so that on regeneration with 11-cis-retinal, a pH-dependem equilibrium between a 380 nm and a 490 nm form of the mutant rhodopsin is observed (Sakmar et al., 1989). Illumination of the 490 nm form with green light yields a 380 nm active species. When all-trans-retinal is added to the mutant opsin, a 380 nm Meta II-like species is formed that is similarly active as the form resulting from illumination of the 490 nm cis-form. Interestingly, the 380 nm-absorbing, deprotonated 11-cis form of the mutant becomes active with UV-light as well (Fahmy and Sakmar, 1993a). This has shown that the transformations of the protein related to retinal isomerization and to the neutralization of E 113 must not necessarily occur sequentially. The mutant protein is flexible enough to exist in an 11-cis form, in which the neutralization of the 113 residue (see below) during the normal light-induced activation process is already anticipated by protein engineering; light absorption just adds the necessary chromophore isomerization. The fact that the protonation state of the Schiff base does not influence light-induced catalytic activity enables the interpretation of pH-activity profiles (see below). Robinson et al. (1992) have demonstrated that the mutant E113Q is not only light-activatable after regeneration with the cis-chromophore, but also (with a shifted pH-dependence) constitutively active, with or without light, and without any binding of retinal. The same applies, when the Lys-296 residue is neutralized (mutant K296G). These results have shown that the protonated Lys-296 and the negatively charged group at position 113 must be simultaneously present to inactivate the receptor towards transducin. Cohen et al. (1992) have recorded pH-rate profiles for the rhodopsin-catalysed GTP uptake in G t, for mutants in which Glu-113 and/or Lys-296 were replaced. The results could be explained by a model

Light-Induced Protein-Protein Interactions

161

that involves breaking of a salt bridge between Lys-296 and Glu-113, deprotonation of Lys-296 and the net uptake of a proton from the aqueous phase. Mutant-dependent shifts in the alkaline branch of the bell-shaped pH-profiles were explained by changes in the pK a of only one protonatable group; the group could not be assigned to His-211, although this residue is among the ones which contribute to the pH-dependence of Meta II (Weitz and Nathans, 1992; Weitz and Nathans, 1993). We can infer that the group(s), which generate the pK a seen in the activity profiles, should belong to a domain that operates by proton uptake, when the MII b state is formed. The mechanism of proton translocation could be investigated by combining site directed mutagenesis (using monkey kidney cells, Oprian et al., 1987, or recombinant baculovirus, Jansen et al., 1991) with Fourier transform infrared difference spectroscopy. Certain bands in the 1730-1770 cm-1 spectral range had previously been interpreted as changes ofprotonated carboxyl groups within the hydrophobic interior of the protein (Siebert et al., 1983; DeGrip et al., 1985; reviewed by Siebert, 1992). Replacements of membrane-embedded residues by neutral amino acids were studied. Rathet al. (1993) found that a band assigned to the C=O stretching mode ofcarboxylic acid groups was absent in the D83N mutant. They concluded that this residue remains protonated in rhodopsin and in all its intermediates (but changes its hydrogen bonding). Replacements of Asp-83 and/or Glu-122 removed two well-defined difference bands, but did not influence a third photoproduct band (1712 cm-1), assigned therefore to protonation of Glu-113 (Fahmy et al., 1993). It was concluded that Asp-83 and Glu- 122 remain protonated upon Meta II formation. This interpretation would confirm the simple notion that, in Meta II, the counterion of the retinal Schiff base becomes eventually neutralized by translocation of the Schiff base proton (Hofmann and Kahlert, 1992). This is further supported by the finding that, in the mutant E113Q, a net protonation change is not observed, probably because proton uptake is compensated by release of the Schiffbase proton in the absence of its Glu- 113 acceptor (Amis, S., Sakmar, T.E, and Hofmann, K.E, unpublished). It remains open, whether the 113 position becomes protonated by direct translocation of the Schiff base proton during formation of MII a, or by a proton relay mechanism involving earlier photoproducts (Ganter et al., 1989). Double and triple mutants have been constructed, to study the effect of a displacement of the Schiff base counterion (Zhukovsky et al., 1991, 1992; Zvyaga et al., 1993). In the double replacement mutant E113A/All7E (in the publication by mistake identified as a triple mutant, Sakmar, T., personal communication), a reversed pH dependence of the spectrophotometry is observed (Zvyaga et al., 1993). At acidic pH, Gt activation is maximal, similar to the wild type and in agreement with the normal MII a ~ MIIb scheme (Figure 3C), although a MI/MIIIlike product dominates in this range. The product decays slower in the presence of Gt, indicating interaction of the G-protein with the "wrong" spectrophotometric species (Sakmar and Fahmy, personal communication). This result shows that the

KLAUS PETER HOFMANN and MARTIN HECK

162

active state of rhodopsin can arise from different types of chromophoreprotein interaction. Different approaches emphasize the importance of the cytoplasmic termination of rhodopsin's third transmembrane helical segment for the interaction with G r Time-resolved electron paramagnetic signals from rhodopsin, spin-labeled at Cys140, has revealed a light-triggered conformation change in this region, the kinetics of which are correlated with the 380 nm absorption change (Farahbakhsh et al., 1993). The 134-136 tripeptide (Glu or Asp/Arg/Tyr) is conserved among most G-protein coupled receptors (see Hargrave, 1991). Mutations of this motif affect G-protein activation, in rhodopsin as well as in other receptors, in peculiar way. Reversal of the Glu/Arg charge pair in the sequence of bovine rhodopsin abolishes binding and activation of the G-protein (Franke et al., 1990). Any of the mutations tested ofArg- 135 severely impairs Gt activation, while replacement of Glu- 134 can result in higher or lower activity, depending on whether a neutral or negative residuum is inserted (Sakmar et al., 1989; Franke et al., 1992). Among five charged, membrane-embedded residues (besides Glu-113 and Lys-296), tested for the effect of replacement by neutral amino acids, Glu-134 was the only to give the mutant opsin constitutive activity (Cohen et, 1993). With respect to light-induced activity, pH-rate profiles show a complex pattern on replacements of Glu- 134 (Cohen et al., 1992; Fahmy and Sakmar, 1993b). For E 134Q, a pronounced shoulder is superimposed on the alkaline branch of the bell-shaped pH-rate curve. However, replacement of Glu- 134 by Asp (mutant E 134D) diminishes the overall activity without a significant shift or distortion of the pH dependence (Fahmy and Sakmar, 1993b). This shows that the activity of the receptor does not depend exclusively on the protonation state of Glu- 134. C. Mechanism of Gt Activation

A "Mesoscopic" Switch Scheme Independent of the assignment of the different states in microscopic detail, a reaction scheme (Figure 4) can be given for the interaction of photoactivated rhodopsin with transducin, summarizing the available "mesoscopic" information: (i) the active Meta II conformation (MIIIb) arises from proton uptake by the deprotonated Schiff base form (MII a, Figure 3); (ii) only G t with an empty nucleotide site (MII-G e complex) can stabilize Meta II (Kohl and Hofmann, 1987; Kahlert et al., 1990); and (iii) GDP or GTP can bind to the MII-G e complex, to form transient "switch" states, with protein-protein interactions other than the 3-loop empty site interaction (Kahlert et al., 1990; Kahlert and Hofmann, 1991). In the transient states, Meta II rhodopsin replaces GDP, or GTP replaces Meta II, respectively. A detailed discussion of the switch scheme has been given elsewhere (Kahlert et al., 1990; Hofmann and Kahlert, 1991; Hofmann, 1992). The activation enthalpies for rhodopsin and transducin are from published work (AHa~ and AH~, see Arnis and Hofmann, 1993; AHgDp,Kahlert and Hofmann, 1991; AHgTp, Kahlert

Light-Induced Protein-Protein Interactions

163 inactive

o

E

.--)

Jr

o

#

AH a

transient

AH #b

Light

/

+GoPII empty-site

R

MI

MII a

~G~ O

ecy, m tos, e Disc membrane

MII b

~G~ _

-

pH - -

,.5

6.75 6.0

\ transient #

AH GTP

active

Figure 4. Reaction model and activation parameters for rhodopsin and transducin. The scheme summarizes the development of the active MIIb conformation of rhodopsin (Figure 3), and the transitions of the G-protein. After the initial photochemistry, all conversions are thermal and in principle reversible. The final, quasi-irreversible transition into the active state of the G-protein (which involves subunit dissociation, Figure 2B) drives the flow of information. The catalytic process involves a sequence of interactions between rhodopsin and the G-holoprotein. Membrane binding of G (top) and the transient states of interaction with Mlib involve less than the 3-loop binding domain in the central empty site complex (see text). MIIb can enter or leave the complex with Gt only via the transient "switch" states. The activation enthalpies (AH #) for the transitions between the Meta forms of rhodopsin and for the association and dissociation with the G-protein are shown on one and the same scale. The activation free energies (AG#) for forming Mlla and MIIb depend on the hydrophobic host (detergent) and pH, respectively.

and Hofmann, 1991, and Kohl and Hofmann, 1987). The values for the standard free energies (AG~# and AGb#) are from Arnis and Hofmann (1993). For the corresponding sites on G t of the three interacting loops on Meta II, we assume three major binding sites (Figure 4); two of these sites, both near to the C-terminus and within stretches 340-350 and 311-329, were assigned to the G t Gt-subunit by synthetic peptide competition (Hamm et al, 1988; for a review see Hamm, 1991). A third synthetic peptide from G a (near the N-terminus) competed against Gt-induced Meta II stabilization; it may affect the interactions of the G t subunits.

164

KLAUS PETER HOFMANN and MARTIN HECK

Complexity of the Interaction Domain The analysis of G t activation by rhodopsin loop mutants has shown that interaction of the second or third loop, together with the fourth loop, can lead into a stable complex; however, such two-loop interaction was not competent to perform the catalytic process and to dissociate active GTP-binding G t (Franke et al., 1990). The finding is in apparent contradiction to the "simple structure" replacement by peptides reported for the IGF-II/mannose 6-phosphate receptor (Okamoto et al., 1990). Such activation by peptides does exist for transducin, but it is much slower; for mastoparan, the rate of activation is five orders of magnitude lower than for intact photoactivated rhodopsin (Heck and Hofmann, unpublished results). By contrast, the PDE effector is effectively activated by peptide from the G t sequence (Rarick et al., 1992). A related problem arises from the observation that high concentrations (1 mM) of synthetic peptides from the G t a-sequence can force the MI/MII-equilibrium to the Meta II side (Hamm et al., 1988). The question is whether this effect is only due to the binding of a given peptide to its own binding site on rhodopsin (where it binds as part of the G a sequence). Most relevant is this problem for the interpretation of Transferred Nuclear Overhauser Effect from G,~ peptides. Dratz et al. (1993) have used this method to study the three-dimensional structure of the C-terminal peptide of the Gtot-subunit , which is involved in the interaction with photoactivated rhodopsin (Hamm et al., 1988). The analysis presented yields a unique, well-defined conformation of the underlying rhodopsin loop sequence, for the ground state as well as for the light-activated state (MII b in our terminology). On the other hand, the synergism in the competition of G~ peptides (Hamm et al., 1988) or rhodopsin loop peptides (K6nig et al., 1989b) was interpreted by the hypothesis that a peptide acts by binding to not only its own principal site but also, with lower affinity, to other sites ("degeneracy"; see K6nig et al., 1989b, and Hofmann and Kahlert, 1992). Future approaches to this unsolved problem may consider the fact that the total conformational energy of bound peptide must replace the energy of the protonations, to shift the equilibria in Figures 3,4 to the right. Willardson et al. (1993) have recently suggested cooperative association of two G t per rhodopsin, or vice versa, in the catalytic interaction. The models were based on the binding of radioactively labeled G t to ROS membranes, which displays a non-linearity similar to the one sometimes observed for Gt-PDE interaction (see below). We note that binding from solution to the membrane, rather than from the membrane to active rhodopsin, can limit the overall effect of G t seen in centrifugation or GTPase assays (see Hofmann, 1993, for a detailed discussion); this may cause the non-linearity, consistent with the reported oligomers of Gt-holoprotein in solution (Hingorani et al., 1988).

Light-induced Protein-Protein Interactions

165

Conclusions from the Structure of GTPSyS-binding Ga The crystal structure of the activated transducin ot-subunit (Noel et al., 1993) shows the nucleotide deeply occluded in a cleft between two domains of the protein. One of the domains is highly homologous to the small GTPases, the other, unique to heterotrimeric G-proteins, is basically helical. The ribose hydroxyls and the N3 ring nitrogen of the nucleotide are hydrogen bonded to the (x-helical domain. Noel et al. propose a mechanism in which the helical domain moves upon binding to activated rhodopsin, thereby allowing the release and access of the nucleotide.

D. Formation of a Signaling State without Light In view of the analogies between rhodopsin and other G-coupled receptors, it is interesting that purified opsin forms with all-trans-retinal a receptor-agonist complex (380 nm "pseudo-photoproduct"; Hofmann et al., 1992). Similar to normal, light-induced MI/MII, a 470 nm species is formed, in equilibrium with a 380 nm form. Surprisingly, G t does not measurably influence this product, while arrestin does. On the other hand, the 380 nm product must interact with Gt, because it catalyses nucleotide exchange in G t (Cohen et al., 1992). Most recent evidence indicates that G t activation via this reversible pathway involves all-trans-retinal non-covalently bound, presumably in the retinal pocket; all-trans retinal forms protonated Schiff bases with the amino groups of lysine residues other than the active site Lys-296 and arrestin interferes with the retinal bound to these groups (J~iger, S., Palczewski, K., Hofmann, K.P., unpublished observations). It is of interest how the activity of the pseudo forms compares to light-induced Meta II. Unfortunately, all available assays tend to underestimate the activity of native Meta II (see Hofmann, 1993). In terms of visual physiology, it is important to know what role, if any, the pseudo-products play in bleaching desensitization. Kahlert et al. (1990) have shown that this phenomenon is not due to a change in the biochemical gain of the visual cascade, but rather to the pre-activating effect of some form of"bleached rhodopsin," which prevents activation by "real" light. For the identity of this ~brm, free opsin with an empty retinal pocket (Jin et al., 1993) or opsin with reversibly bound retinal "agonist" (Lamb, 1990) have been proposed; this problem awaits further investigation.

E. Interaction with Rhodopsin Kinase

The Role of Rhodopsin Phosphorylation It is a remarkable fact that the shut-offreaction ofrhodopsin as an active receptor is not based on the spontaneous decay of the active conformation. Activation of the visual transduction cascade is shut off after 1.5-2 seconds (Pepperberg et al., 1988; Pepperberg et al., 1992), long before the spontaneous decay ofMeta II, which takes minutes under native conditions (reviewed by Hofmann, 1986). Liebman and Pugh (1980) were the first to propose that the shut-off of rhodopsin as a light receptor is

166

KLAUS PETER HOFMANN and MARTIN HECK

an active, energy-consuming process; Wilden et al. (1986) then provided the evidence for a two-step shut-off process, which implied that (i) the binding of rhodopsin kinase leads to phosphorylation of active rhodopsin at several sites on its C-terminal extension, and (ii) phosphorylation provides the switch from the Gt binding to the arrestin binding state, in which any interaction ofrhodopsin with Gt is effectively blocked. Even in the absence of arrestin, phosphorylated rhodopsin is reduced in its efficiency to stimulate PDE activation (Miller et al., 1986). Palczewski et al., (1992b) employed whole-cell preparations and specific inhibitors of rhodopsin kinase or arrestin to study the rapid shut-off of light responses of functioning rods. This work has now confirmed the conclusion from experiments on a reconstituted system that phosphorylation of rhodopsin on multiple sites can in itself act as a turnoff mechanism, but arrestin accelerates the recovery by quenching partially phosphorylated rhodopsin (Bennett and Sitaramayya, 1988). Rhodopsin Acts as an Activator and Substrate

Phosphorylation by rhodopsin kinase occurs at Ser and Thr residues (Thompson and Findlay, 1984), located exclusively in the C-terminal region of rhodopsin (Palczewski and Benovic, 1991). When the C-terminal stretch of the rhodopsin sequence is presented as a synthetic peptide to purified kinase, it is only very poorly phosphorylated. Palczewski et al. (199 l a) concluded that the receptor in its activated form must act as both the activator and the protein substrate of the kinase. Two basically different mechanisms were discussed: either R* induces a change in the kinase conformation which renders it active or the binding of the kinase induces the exposition of phosphorylation sites. The finding that photoactivated rhodopsin stimulates the kinase to phosphorylate synthetic C-terminal peptide has then shown that the kinase is indeed activated by active rhodopsin (Palczewski et al., 1991a). The N-terminal region of the kinase is involved in the interaction with rhodopsin (Palczewski et al., 1993). In a recent study on the kinetics of multiphosphorylation of rhodopsin, it was suggested that the rate of incorporation of phosphate was the slowest for the first phosphate and increased for more highly phosphorylated species (Adamus et al., 1993). Employing both techniques ofproteolysis, combined with sequence analysis of the phosphopeptides or mass spectrometry, several groups have now identified the consecutive phosphorylation sites on photoexcited rhodopsin. Phosphorylation is sequential, with Ser-338 and Ser-343 as the major initial phosphorylation sites (McDowell et al., 1993; Papac et al., 1993; Ohguro et al., 1993); the conditions under which the phosphorylation is performed, appear to influence the sites which become phosphorylated first (Hargrave and Palczewski, personal communications, 1994). Phosphorylation at more than three sites is inhibited by the binding of arrestin or by the reduction of the photolyzed chromophore all-trans-retinal to all-trans-retinol (Ohguro et al., 1993).

Light-induced Protein-Protein Interactions

167

Analysis of the Interaction Based on a recently developed high yield kinase preparation (Palczewski et al., 1992a), the interaction ofrhodopsin kinase with rhodopsin could be separated from its activation as an enzyme (Pulvermtiller et al., 1993). The kinase was held in solution by small amounts of the detergent Tween; binding from this solubilized state to rhodopsin could be measured by centrifugation and, with kinetic resolution, by light scattering binding signals. These measurements allowed the determination of the rate and dissociation constant of the binding reaction between activated rhodopsin and the kinase. A K D of 0.5 ~tM and a bimolecular rate constant of 1 ~tM-ls-1 was found. The kinase binds specifically to R*, as is seen both by the stoichiometry of the LS binding signal and by its competition against Gt-induced stabilization of Meta II. However, the kinase itself does not stabilize the Meta II. Its affinity can therefore not be specific for this intermediate, neither in its MII a nor its MII b, G t binding form. Binding of kinase appears to be less "demanding" than the one of transducin or arrestin. This finding is in apparent contradiction to the conclusion from active site methylation studies that deprotonation of the retinal Schiffbase (i.e., Meta II) is mandatory for rhodopsin phosphorylation (Seckler and Rando, 1989). A possible solution would be that the binding of the kinase can occur to both the conformations Meta I or Meta II, while phosphorylation requires rhodopsin in the Meta II conformation. There is, however, evidence form native preparations that both Meta I and Meta II can be a substrate of the kinase (Paulsen, and Bentrop, 1983). The binding conformation of rhodopsin for the kinase is already expressed in Meta I, and, in vitro, it persists over minutes. In the absence of ATP, the interaction with the kinase remains at a stable equilibrium. From the measured on-rate and the K D (0.5 ~tM), one estimates that a given copy of the complex, once formed, would remain bound for an average time of 4 seconds, before it dissociates in the on/off equilibrium. However, rapid binding of ATP to the complex, once formed, lowers the off rate; the ternary complex between receptor, kinase and the donor substrate is the state of the highest affinity (PulvermOller et al., 1993). Spontaneous dissociation of the complex within the time frame set by the lifetime of active rhodopsin in situ (Pepperberg et al., 1988; Pepperberg et al.~ 1992) would be a rather rare event. According to the hypothesis of Buczytko et al. (1991), the postponed sequential phosphorylation of rhodopsin and autophosphorylation of rhodopsin kinase will terminate the complex. A major problem of all experimental investigations on rhodopsin and the kinase (in its purified form) is that the latter is not yet available in its native but only in the Tween-solubilized form. This preparation has opened the possibility to use the light scattering binding assay but the kinetic parameters measured with this system do not necessarily apply to the system in situ.

168

KLAUS PETER HOFMANN and MARTIN HECK

F. Interaction with Arrestin

Active and Inactive Conformations Binding ofarrestin (in its purified form) to rhodopsin requires several conditions, including: (i) rhodopsin must be in the Meta II conformation (Schleicher et al., 1989), (ii) it must be phosphorylated (Schleicher et al., 1989; Palczewski et al., 1991b), and (iii) arrestin must be in a binding conformation (Palczewski et al., 1991b). The binding conformation ofarrestin, Ab, arises from the predominantly inactive conformation, Ai, by imeraction with Meta II rhodopsin. Under conditions of low yield of Meta II in its equilibrium with Meta I, the interaction between the proteins after a flash of light stabilizes both conformations, at the expense of the inactive conformations (MI and Ai, respectively): MI + A i ~ MII + A i ~ MII-A b By contrast to Ab, which is only stable in the complex with Meta II, subspecies of arrestin appear in isoelectric focussing (Weyand and K~ihn, 1990). The physiological implications of the different forms of arrestin are unknown.

Mechanism of Binding The conformation A b, as compared to the inactive Ai, was characterized by its reaction with a light-dependent M-site and a light-independem P-site (phosphorylation site) on rhodopsin. The M-site may be identical in part with the 3-loop pattern for Gt; preliminary peptide competition data indicate competition of synthetic peptides from the second and the fourth (but not from the third) cytoplasmic loop (PulvermOller, A., Hargrave, and Hofmann, unpublished, 1994). Inversely, the interaction domain of arrestin for the Meta II conformation may overlap with the respective sites of G c The similarity of the C-terminal sequences of G t and arrestin is consistent with this idea (Wistow et al., 1986). Recognition of the phosphorylated C-terminus of active, phosphorylated rhodopsin and of heparin has been proposed to involve a highly basic region ofarrestin, which begins with Lys-163 in the bovine protein (Shinohara et al., 1987) and is similarly found in other species (reviewed in Palczewski et al., 199 l b). Using an in vitro expression system (Gurevich and Benovic, 1992) for bovine arrestin, Gurevich and Benovic could now analyze the whole interaction domain of arrestin. In a mechanistic model, they have dissected the interaction with Meta II and the conformation changes thereby induced in a sequence of reaction steps, based on the concerted action of different binding and activation sites. Sites for "phosphorylation recognition," "activation recognition," and hydrophobic and regulatory domains are distinguished. The concept is related to the "induced fit" model for the interaction and activation of the G-protein (Figure 4); it remains to be demonstrated, whether the binding state ofarrestin, Ab, can only exist in contact with Meta II, in analogy to the empty site state of G c So far, the

Light-Induced Protein-Protein Interactions

169

simpler equilibrium shift in the scheme above appears to describe the available data satisfyingly well.

The Phosphorylation Switch When arrestin is bound to Meta II, the binding and activation of transducin and the excitation cascade of phototransduction is effectively blocked (Wilder et al., 1986). Phosphorylation alone, without binding of arrestin, uncouples transducin to a degree that depends on the number of phosphorylated sites (Miller et al., 1986); the effect cannot be explained by an alteration of the relative yield or lifetime of Meta II but is due to an alteration of Meta II -G t interaction (Mitchell et al., 1992). Meta II rhodopsin with 2-3 phosphate groups still stimulates PDE activation with 50% of the rate of unphosphorylated Meta II (Miller et al., 1986) and forms extra MII with comparable efficiency (Pulvermt~ller and Hofmann, unpublished, 1993); the interaction with arrestin, however, which yields unmeasurably low stabilization for unphosphorylated Meta II, increases dramatically on phosphorylation (KD = 50 nM; Schleicher et al., 1989). Thus the function of the kinase is basically to operate the phosphorylation switch to arrestin binding.

Possible Cofactors Higher inositol phosphates compete against the interaction between arrestin and Meta II, while neither G t nor rhodopsin kinase are inhibited by these compounds (Palczewski et al., 1992c). By this effect, high bleaching levels could potentially influence the concentration of higher inositol phosphates in the cell. The relation to the arrestin-dependent activation of phospholipase C (Ghalayini and Anderson, 1992) is unknown. Meta II stabilization by arrestin does not depend on calcium or ATP in the range of concentration, where these compounds were reported to bind arrestin (Huppertz et al., 1992; Glitscher and Rappel, 1991, 1992); recent studies show that there is no ATPase activity ofarrestin and probably no nucleotide binding either (R/ippel, H., personal communication, 1993). The binding of calcium or ATP could not be confirmed by Palczewski and Hargrave (1991). Wagner et al. (1988) have reported that, in intact stacks of bovine rod outer segments, arrestin requires additional soluble proteins (besides rhodopsin kinase) for rapid deactivation of rhodopsin.

Release of Arrestin The question arises what event triggers the release of arrestin and allows the regeneration of rhodopsin to the light-activatable form. In a study on post-Meta II decay and regeneration of photolyzed rhodopsin (Hofmann et al., 1992), it was found that a key step for the decay reaction of photoactivated rhodopsin is the reduction of the photolyzed chromophore all-trans-retinal into all-trans-retinol. As long as all-trans-retinal is present, it reacts with opsin to form pseudo-photoproducts that bind arrestin and rhodopsin kinase (see the section on signaling states without light).

170

KLAUS PETER HOFMANN and MARTIN HECK

IV.

I N T E R A C T I O N S OF THE P H O S P H O D I E S T E R A S E A. Activated Gt Relieves an Inhibitory Constraint on PDE

It is generally assumed that the PDE is in vivo activated by collisonal coupling with the GTP-bound form of transducin (Lamb and Pugh, 1992; Pugh and Lamb, 1993). By contrast with the catalytic R*-G interaction, the activation of PDE results from a stoichiometric binding of active transducin, which relieves the inhibitory constraint imposed by PDEv on the catalytic PDE~I3core (Hurley and Stryer, 1982; Yamazaki et al., 1982). At least in preparations in vitro, the persistently activated, isolated G~GTPvS- or GaGMPP(NH)P-subunit alone is competent to activate the PDE (Fung et al., 1981; Hurley and Stryer, 1982). Under these conditions, one GaGTPvS binds one molecule of PDEv (Deterre et al., 1986; Fung and GriswoldPrenner, 1989), and the resulting stable complex can be chromatographically separated from PDEal3 (Deterre et al., 1986). Two molecules GaGTP are required to fully activate one molecule PDEal3v~ (Deterre et al., 1988; Fung et al., 1990). Given the high affinity of PDEv to PDEal3 (KD = 10-11-10-13 M; Wensel and Stryer, 1986; Bennett and Clerc, 1989), the observed rapid activation can not result from spontaneous dissociation of PDEv and PDEal3 (t 1/2 = 500-800s; Wensel and Stryer, 1990; Brown, 1992) with subsequent formation of the G~GTP-PDEv complex (Hurley and Stryer, 1982). Activation of PDE by transducin therefore must involve, at least transiently, the interaction of GaGTP with the holoenzyme PDEal3w (symbolized by the transient states in Figure 5), which is experimentally confirmed by the finding that the incorporation of labeled PDEv into the PDE holoenzyme is greatly accelerated in the presence of G~GTP (Wensel and Stryer, 1990). B. Mechanism of Interaction and Activation The initial interaction between GaGTP and PDE induces a change of the PDE holoprotein which lowers the affinity of PDEv to PDE~ and enables the replacement of PDEa~ on PDEv by GaGTP. The mechanism resembles the first step in the activation of Gt, where MII replaces bound GDP, leading to the MII-empty site interaction (Figure 4). The mutual displacement process is common to most mechanisms that have been considered in the literature. For the transition into the active state, two mechanisms were found: either the inhibitory subunit is physically removed from the catalytic core by GaGTP, or it remains bound in an undissociated complex of GaGTP with the PDE holoenzyme (Figure 5).

Removal of the Inhibitory Subunit in Solution In the dissociation model (Figure 5, right column), G~GTP binds to the PDE holoenzyme and subsequently removes one PDEv subunit, to form a partially active PDEa~v complex. Another copy of GaGTP can then carry away the second PDEv subunit, leaving an active PDEa~ (Deterre et al., 1986; Fung and Griswold-Prenner, 1989; Wensel and Stryer 1990). In both consecutive transient states, the interaction

membrane

solution

inactive

transient

GMP

GMP

cGMpj

partially active

transient

GMP

GMP

GMP

GMP

,..,...,.

Figure 5. Reaction scheme for the activation of bovine PDE by transducin, in solution (left column), and on the disc membrane (right column). The scheme summarizes the sequence of events and gives a tentative model of the subunit interactions involved. Inactive PDE (first row) consists of two catalytic subunits (PDEc~I3),each inhibited by one PDE~ subunit (I, dark shaded). Activation of PDE results from the successive displacement of the two PDE~ subunits from their inhibitory sites on PDEczl3by two GTP-bound c~-subunits of transducin (G~GTP). in both cases (solution or membrane), G~GTP must initially interact with the PDE holoenzyme to lower the high affinity of PDE~ to PDEczl3(transient states I), leading to the formation of a PDE~-G~GTP complex. In the absence of membranous binding sites (left column) PDE~-G~GTP dissociates from the catalytic core, leaving the partially active PDE~I3~. A second G~GTP can then interact with PDEc~I3~(second transient state) and in turn remove the second PDE~ subunit from the fully active PDE~I3dimer. When the PDE is associated to membranes (shaded bar in the right column), the interaction between G~GTP and PDE is stabilized by phospholipid-protein interactions and the PDE~-G~GTP complex remains bound to PDE~I3. Both the partially active PDEc~I3~(PDE~G~GTP) and the fully active PDEc~I3 (PDE~ G~ GTP)2 complex display a greatly enhanced membrane binding, as compared to the inactive PDE. Note that the inhibitory sites on active PDE~I3 remain accessible for interaction with additional PDE~. 171

172

KLAUS PETER HOFMANN and MARTIN HECK

of active Gt with the PDE lowers the high affinity of PDEv to PDE~. The experimental results suggest that the PDE~ subunit, once dissociated from the G~-PDEv complex, reestablishes its high affinity to PDEv, which is about 1000 times higher than the affinity of G~GTP to PDEv (Wensel and Stryer 1986). In the resulting coupled equilibria, the amount of active PDE is determined by the ratio of the dissociation constants of PDEv-PDEc~~ and PDEcGa-GTP (KD < 10- 1 0 M, Otto-Bruc et al., 1993) and the concentrations of the components. Since the relation between the KD'S highly favors inhibited PDE, only a large excess of active Gt can fully activate the PDE; this is indeed found experimentally. Generally, the dissociation model applies well to the proteins in their soluble form.

Membrane Binding Stabilizes Gt-PDE Interaction The presence of membranes potentiates the Gt-stimulated PDE activity in vitro (Fung and Nash, 1983; Tyminski and O'Brian, 1984). This indicates that the interaction occurs in situ, like other G-protein-effector interactions, on the cytoplasmatic surface of the membranes. Although the influence of the membranes may be partially due to an increase ofVmax (Gillespie and Beavo, 1988), the strikingly reduced amount of active Gt required to stimulate maximal activity (Bennett and Clerc, 1989; Phillips et al., 1989; Malinski and Wensel, 1992) implies that the activation mechanism is different from that in solution. IfG,~GTP-PDEv dissociates from PDEa~, the membranes could shift the equilibrium in favor of active PDE either by interacting with PDEa~ to lower the affinity for PDEv and/or by stabilization ofthe G~GTP-PDEv complex (Fung and Griswold-Prenner, 1989). On the other hand, several lines of evidence strongly suggest that active bovine PDE is an undissociated, membrane bound PDEa~ (PDEvG~GTP)E-complex: 1. A comparative kinetic study shows that PDEa~ (generated by trypsin) is not equivalent to light-activated PDE with respect to the K m for cGMP and K i for PDEv (Sitaramayya et al., 1986). 2. In cross linking studies, complexes of PDEa and PDE~ with one and two molecules of G,~GTPvS/GaGMPP(NH)P are observed (Hingorani et al., 1988; Clerc et al., 1992). 3. The mainly soluble GaGTPcS-subunit becomes membrane associated in the presence of inhibited PDE, whereas no binding occurs upon addition of purifed PDEv (Clerc and Bennett, 1992). 4. GaGTP activated PDE displays greatly enhanced membrane binding, as compared to the isolated compounds PDEa~vv (Malinski and Wensel, 1992; Catty et al., 1992), PDEa~, PDEa~v and GaGTP-PDEv (Catty et al., 1992). 5. Addition of recombinant PDEv to PDE, which had been persistently activated by GaGTPvS, releases GaGTPvS complexed to native PDEv (Otto-Bruc et al., 1993). Thus dissociation of GaGTP-PDEv from PDEa~ would most likely ensue only in the absence, or at low concentrations (<0.1 mg Rh/ml; Catty et al., 1992) of

Light-Induced Protein-Protein Interactions

173

membranes. In the presence of sufficient high overall concentrations of membranes and proteins, the interaction of G~GTP with the PDE holoenzyme is stabilized by protein-phospholipid interactions (Malinski and Wensel, 1992), leading to the active PDEal3(PDEv Ga GTP)z-complex (Figure 5, left column). IfGaGTP remains in contact with PDEa~w, the amount of active PDE is only determined bythe affinity of GaGTP to membrane bound PDE~ w, which is high enough (<0.25 nM, Malinski and Wensel, 1992; <2.5 nM, Heck and Hofmann, 1993) to ensure efficient activation of PDE. It remains an open question, whether factors other than the membranes can influence the mechanism of PDE activation by transducin (dissociation of G,~GTPPDEv from PDEa~ vs. formation of a PDE~I3(PDEv G~ GTP)2_ complex) in bovine ROS. For frog ROS, it has been recently suggested, that cGMP regulates the interaction between PDEv and PDE~: if non-catalytic binding sites on PDE,~ are occupied by cGMP, G,~GTP remains bound to PDEa~v~; when the sites are empty, G~GTP-PDEv dissociates from PDE~ (Arshavsky et al., 1992). The much slower dissociation rate of the bound cGMP of bovine PDE in vitro (Gillespie and Beavo, 1989) could explain why the PDE~ (PDEv G~ GTP)z-complex exists even in the absence of externally added cGME Intermediate States

The existence of two inhibitory subunits in the PDE holoenzyme poses the question of the relative activity of the partially inhibited P D E ~ or P D E ~ (PDE~ Ga GTP) complex and the affinity of G~GTP and PDE~ to the two binding sites on PDEa~. Differential activities of PDE with different occupation by PDE~ (or alternatively G~GTP) may play a role in the physiological process of activation and deactivation (Deterre et al., 1988; Bennett and Clerc, 1989, Whalen and Bitensky, 1989). A single class of binding sites is found when PDE is activated by G~GTPyS in solution (Phillips et al., 1989). For samples containing membranes or phopholipids, however, different results have been reported: no indication of cooperativity in the activation of PDE by G~GTP or G~GTP~S can be seen in the data of Tyminski and O'Brian (1984), and Malinski and Wensel (1992). On the other hand, a nonlinear dependence on active transducin has been reported by Phillips et al. (1989), Bennett and Clerc (1989), Whalen and Bitensky (1989), Whalen et al. (1990), Clerc and Bennett (1992) and Heck and Hofmann (1993). This sigmoidal dependence on active transducin could be explained by a mechanism, in which the first G~GTP interacts with PDE with a higher affinity than the second, and the relative activity of the intermediate state (PDE~ or P D E ~ (PDE~GaGTP)) is less than 50% of the fully active enzyme. Several studies used the complementary approach, namely, the inhibition of trypsin activated PDE (tPDE) by purified PDE~. Inhibition of the catalytic activity of tPDE by a fluorescein-labeled PDE~ subunit was found to be strictly linear and parallels occupation of PDEcbinding sites, suggesting that t P D E ~ has 50% of the

174

KLAUS PETER HOFMANN and MARTIN HECK

activity oftPDEa~ (Wensel and Stryer, 1990). AHill plot for the inhibition oftPDE with a recombinant PDEv subunit showed no cooperativity (Hamilton et al., 1993); however, a nonlinear dependence of PDE activity on PDEv concentration has been reported by Whalen and Bitensky (1989) and Clerc and Bennett (1992). These inconsistencies may be explained by a heterogeneous transducin population containing only a small fraction oftransducin competent to bind tightly to PDE (Wensel and Stryer, 1986; Malinski and Wensel, 1992) or a loss of activity at low protein concentrations due to surface adsorption and/or proteolysis (Malinski and Wensel, 1992). In any case, it should be emphasized that in most studies the membrane and PDE concentrations used were not high enough to quantitatively bind PDEal3n to the membranes. The nonlinear dependence of the PDE activity on GaGTP could therefore reflect the different efficiency by which soluble and membrane associated PDE is activated by GaGTP, in combination with the strikingly enhanced membrane binding of G~GTP-complexed PDE (Catty et al. 1992, see above). The resulting shift of the equilibrium between the various components in the course of G~GTP titration could produce the observed nonlinearity. C. Interaction Domains

Interaction between PDE~and PDEa,# The PDEr subunit plays a key role in the activation of PDE by GaGTP and must contain binding sites for both the PDE~I3and G~ subunits. Insight into the molecular details of these interactions was recently obtained by employing site-directed mutagenesis and synthetic peptide or antibody competition. PDEr is 87 amino acids long (Ovchinnikov et al., 1986); the overall sequence comprises three different regions, which include the N-terminus, a polycationic central region and the C-terminal extension. Truncated PDE~ subunits described in the literature include the stretches 1-74 (Brown and Stryer, 1989), 1-77 (Takemoto et al., 1992), 1-80 (Lipkin et al., 1988; Cunnick et al., 1990), and 1-82 (Brown, 1992), which all fail to inhibit the hydrolytic activity of tPDE, although they bind tightly to it. This suggests that the C-terminal 82-87 region is primarily responsible for inhibition of the enzymatic activity of PDEal3 and it is consistent with the finding that peptide 46-87-PDEr binds with high affinity to tPDE and effectively inhibits its enzymatic activity (Artemyev and Hamm, 1992). The central 24-46 region is less important for the inhibitory role of PDEr: mutations of Lys-41, -44 and -45 do not alter PDEr inhibitory properties (Brown, 1992) and peptides from this region can only partially inhibit tPDE activity, even at high concentrations (Morrison et al., 1987; Artemyev and Hamm, 1992). On the other hand, the 24--46-PDEr peptide binds tightly to tPDE (Artemyev and Hamm, 1992), showing that this region participates in the high-affinity binding of PDEr to PDEa~.

Light-Induced Protein-Protein Interactions

175

The N-terminal region, 1-12-PDEv, does most likely not interact with PDEal3, since a peptide corresponding to this sequence has no effect on tPDE activity (Cunnick et al., 1990) and a N-terminal truncated PDEv subunit (12-87-PDEv) is as effective as native PDEv in tPDE inhibition (Artemyev and Hamm, 1992).

Interaction Between PDE~ and Transducin At least one binding site for G~ can be localized within residues 24--45 of PDE~, since peptides encompassing this sequence must interact with Ga, based on GTPase (Morrison et al., 1987; Morrison et al., 1989) and competition or binding assays (Artemyev et al., 1992; Takemoto et al., 1992). A monoclonal antibody, which blocks the 24M5-PDEv region, prevents photoactivation of PDE (Lipkin et al., 1988). Furthermore, the G~GTPvS induced activation of PDE, reconstituted with a Lys-41, -44 and -45 PDEv mutant, was found to be highly impaired (Brown, 1992). A second region is involved in the binding of GaGTP to PDEv: a PDEc46-87 peptide interacts with G~ within residues 293-315 (Artemyev et al., 1992). Regions on Ga, identified as loops protruding from the structure (Noel et al., 1993), provide interaction sites with the PDE; a synthetic peptide derived from the sequence 293-314 of Ga mimics GaGTP and activates PDE by reversing the inhibitory effect of PDEv (Rarick et al., 1992). Replacement of the tryptophane at position 70 of PDEv diminishes its affinity for G~GTP and GaGDP 100-fold; PDEal3reconstituted with this mutant is fully inhibited but could no longer be activated by GaGTP (Otto-Bruc et al., 1993). Taken together, PDEa~ and GaGTP appear to interact with sites within the same pair of regions on PDEv suggesting that the binding domains of PDE~ on PDEv come close to, or even overlap the ones of GvGTP. Activation of PDE by G~GTP may therefore involve an (at least partial) displacement of PDE~ from its binding site on PDEa~, relieving the inhibitory effect of the C-terminus of PDEv. The notion is also consistent with the finding that the PDEv binding site on PDEa~ is accessible for added PDEv under conditions, even in the undissociated PDEal3 (PDEvGa GTP)2-complex (Otto-Bruc et al., 1993), and not only under the assumption of dissociation (Wensel and Stryer, 1990).

Interaction Between Transducin and PDEa~ Besides the interactions between G~GTP and PDEv, an additional direct link between G~ and catalytic subunits is suggested by several findings: (i) cross-linked products between G~GTP and PDEa~ (Hingorani et al., 1988; Clerc and Bennett, 1992), (ii) enhanced membrane binding of the purified PDEa~ subunit upon addition of G~GTP (Catty et al., 1992) and (iii) inhibition oftPDE by G~GDP (Kroll et al., 1989; Phillips et al., 1989) and by two synthetic peptides corresponding to the sequence 53-65-G~ and 201-215-G~ (Rarick et al., 1992).

176

KLAUS PETERHOFMANN and MARTIN HECK

Role of the Gpr Subunit As stated above, the isolated G~GTP subunit alone is competent to activate PDE. A direct interaction of the GI3r subunit with the PDE has not been demonstrated; its proximity to the PDE during the activation process is suggested by cross-linking studies (Hingorani et al., 1988). One possibility is that GI3~helps to deliver GaGTP to the PDE. Such a mechanism would require that the G t subunit dissociation is postponed to the interaction with the effector in vivo. Although G~GTP becomes instantaneously soluble under most in vitro conditions, the presence of membrane associated PDE may slow its release (Heck and Hofmann, 1993). A continuously membrane bound pathway from receptor to effector for active G t or its subunits, has been discussed (Liebman et al., 1987; Uhl et al., 1990; Heck and Hofmann, 1993). However, as soon as active G t is rapidly, and effectively captured (Berg and Purcell, 1977) by membrane associated PDE, the measured 5 ms interval between G-protein activation and formation of the Gt-PDE complex (Heck and Hofmann, 1993) may be explained by either membrane-bound or transiently soluble G t ("micro-soluble G", Chabre, M., personal communication, 1993).

V.

T I M I N G A N D D E A C T I V A T I O N OF THE CASCADE

A. The Disc Amplifier The first kinetic analysis ofG t activation arose from the work of Paul Liebman's laboratory (see Liebman et al., 1987, for a review). Ktihn et al. (1981) applied time resolved infrared light scattering to determine the kinetics and stoichiometry of G t activation. The most recent analysis by Bruckert et al. (1992) specifies the time of association of G t to active rhodopsin with 0.25 ms and the overall time of catalysis with 1.2 ms (at room temperature). Fast diffusion of G t (in the order of rhodopsin diffusion, with a lateral diffusion coefficient of 0.5 ~tm2s-1) is consistent with the results (Bruckert et al., 1992; Kahlert and Hofmann, 1991). An even faster diffusion was still in agreement with the data of Bruckert et al. (1992), but could not be unequivocally proven. GDP, which does not compete much at low concentrations (Kahlert et al., 1990) does so very effectively at the high nucleotide concentrations that presumably prevail in the intact cell (Bruckert et al., 1992). Due to the weak binding of GTP to the empty site complex (Figure 3), the overall rate of activation can accelerate even into the physiologically relevant mM range (Kahlert and Hofmann, 1991; Hofmann and Kahlert, 1992). Note that GDP does not simply compete for binding but can induce the back reaction and dissociation from the receptor (Figure 3; Kahlert et al., 1990). Kahlert and Hofmann (1991) recorded light scattering signals from functioning rods of the bovine retina, between 0 °C and 40 °C; they found that the kinetic parameters of the G t activation process,

Light-Induced Protein-Protein Interactions

177

determined from data in vitro (Figure 3), fitted the light scattering signals from the intact preparation reasonably well. The Arrhenius plots were consistent with rate limitation by the MII/GTP protein transformation at low and by protein diffusion at high temperatures, with a transition in the range 18-22 °C. Throughout the temperature range tested, the release or entry of nucleotide to its binding site does not limit per se the reaction; the nucleotide equilibrates with the binding site before it induces the back or forward reaction (Figure 3; Hofmann and Kahlert, 1992). The rates of G t activation were 30, 800, and 4,000 G t per photoactivated rhodopsin molecule at 5, 20, 37 °C, respectively. More recently, Heck, and Hofmann (1993) could give an upper limit of 5 ms (at room temperature) for the time interval between formation of active G t and its interaction with PDE; this time fits well into the time flame determined in situ by electrophysiology (Cobbs and Pugh, 1987). Generally, the whole body of data on the kinetics of the cascade fits into recent analyses, which have integrated the biophysical and biochemical results to predict the electrical responses of photoreceptor cells (Lamb and Pugh, 1992; Pugh and Lamb, 1993). This consistency does not necessarily mean that the tumover numbers are realized in situ. Kahlert and Hofmann (1991) have noted the intrinsic (for data in situ probably untestable) assumption behind all light scattering studies that the maximal signal corresponds to the activation of the entire G t pool; subpopulations ofG t may exist (Wensel and Stryer, 1986). On the other hand, binding studies, which generally yield lower numbers (see Pugh and Lamb, 1993, for a discussion), may severely underestimate the turnover because of the large contribution of the slow soluble G t pool (Schleicher and Hofmann, 1987). We can state that light scattering signals from membrane preparations, which can be calibrated based on their known G t stoichiometry, yield turnover numbers of several hundred (in the single photon limit, and at room temperature; Heck and Hofmann, 1993). A more critical problem is the assumption made so far by all investigators, namely, that deactivation reactions do not influence the activation rate. Pulvermtiller et al. (1993) have extrapolated their kinase binding data to cellular concentrations and found a time for formation of the rhodopsin-kinase complex in the 10-100 ms range. Based on the experimentally determined competition between kinase and Gt, simple calculations yield an inhibition of G t activation after a few hundred cycles. The mechanism would down-regulate the active G t pool, preceding phosphorylation and arrestin binding. Whether such a mechanism occurs physiologically, depends on the extent to which it is balanced by (calcium dependent) feedback regulation (see the section about rhodopsin kinase).

B. Deactivation of Rhodopsin Shut-off of rhodopsin as the activator of the visual cascade takes 1.5-2 s in situ; this result was obtained from two-flash light-scattering experiments (Pepperberg et al., 1988), and independently, from the onset of the falling phase of the electrical

178

KLAUS PETER HOFMANN and MARTIN HECK

response (Pepperberg et al., 1992). The kinetic parameters in vitro are known for rhodopsin kinase and arrestin. Bimolecular rate constants in the order of 1.0 and 0.2 gM-:s-:, respectively, were obtained (Pulvermaller et al., 1993; Schleicher et al., 1989). Extrapolating the rates of association to cellular concentrations yields estimated association times below 200 ms for both proteins. This means that the actual binding reactions would not be rate limiting for the overall rate of quenching.

C. Deactivation of transducin and PDE The deactivation of the effector in an amplified G-protein mediated activation pathway is of considerable complexity. Each of the three components involved in the activation of the effector (receptor, G-protein and the effector unit itself) have to be deactivated and reset to the activatable state. Individual reactions that must occur to reach a complete reset of the cascade, include: 1. deactivation ofrhodopsin, as discussed above; 2. deactivation of the G-protein by its intrinsic GTPase and recombination of the subunits to the receptor-interactive holoenzyme; 3. deactivation of the PDE by re-association of the displaced PDEr subunit to the inhibitory site on the catalytic core. GTP

Hydrolysis

The first suggestion of a subsecond GTPase reaction time was based on the rapid recovery of G t activation light scattering signals (Wagner et al., 1988); it was more directly derived from measurements of time it takes to reach the steady state of heat production (Vuong and Chabre, 1991). Such a fast reaction fits into the time frame for the physiological response of the rod (Cobbs, 1991). For the isolated GaGTP subunit, a much slower GTPase is measured (reaction time of 21 s; Antonny et al., 1993). For native frog ROS, two different rates of GTP hydrolysis were found that could be assigned to two portions of the total G t pool. The relative weight of the fast rate was 20%, equivalent to the amount of G t which participates in PDE activation; the kinetic difference was abolished in the presence of cGMP (Arshavsky et al., 1991). A GTPase accelerating protein (GAP) function was found for both PDE holoprotein and recombinant PDEv from toad ROS, with partial suppresSion of the holoprotein effect by cGMP (Arshavsky and Bownds, 1992). No influence of PDEv was found by Antonny et al. (1993), who studied the GTPase of active G~ eluted from bovine ROS membranes in the presence or absence of recombinant PDEv. The suggestion from this work, that factors other than PDEv are additionally required for the GAP function, was substantiated by Pag6s et al. (1992) who have found that formation of membrane associated PDEal3 (PDEvGaGTP)2 complex correlates with enhanced GTPase rate.

Light-Induced Protein-Protein Interactions

179

PDE Deactivation According to the standard model (Figure 2), PDE deactivation necessarily requires the GTPase reaction to occur before. In the simplest model, newly formed G~GDP would loose its affinity for both PDEv and PDEa~, leading to the release of GaGDP and reinhibition of PDE. Several studies, however, suggest that the GTPase reaction may not be sufficient to deactivate PDE. In frog ROS, hydrolysis of bound GTP after separation of the GaPDEv complex from the membranes does not relieve the interaction; excess G~v is required to displace the PDEv subunit (Yamazaki et al., 1990). Otto-Bruc et al. (1993) have found a high affinity (3nM) interaction between the isolated components Ga GDP and PDEv. It remains open, to which degree this interaction occurs in the PDE holoprotein complex and delays PDE deactivation in situ (see below).

The Convolution Problem The lifetime of the active effector must be short enough to explain the measured short responses of retinal rods (Baylor et al., 1984). However, it must be stressed that even very short individual lifetimes of both active G t and PDE do not necessarily lead to a short period of overall activity of the effector pool (which is the only physiologically relevant output of the cascade). This is evident from Figure 2: even if the lifetime of active G t and PDE were infinitely short, the reactivation in the Gt/PDE cycle would keep the PDE pool active over the full R* lifetime (1.5-2 s; Pepperberg et al., 1992). Computer simulations show that an assumed deactivation of G t and PDE in the order of 1 second leads, by the convolution with the R* lifetime, to an effective activation of the PDE pool extending over several seconds (for an example, see Pulvermt~ller et al., 1993). Since both the R* lifetime and the GTPase reaction time are experimentally confirmed (see above), the simple recycling mechanism cannot apply. Consequently, either G t must be hindered to recycle and/or the PDE must be uncoupled from repeated G t activation. A refractory period, in which G t does neither bind to R nor to the PDE, could delay the recycling of G t and uncouple it from reactivation; such models indeed describe the time course of PDE activation adequately (Pepperberg and Hofmann, unpublished results, 1993). The reported interaction between phosducin and the G~v subunit (Lee et al., 1987; see above) opens a way to realize a refractory state, if phosducin intervenes after a first round of G t activation. For such a mechanism, however, phosducin must also exist in a second form, in which it does not compete with G,~ and allows formation of G t holoprotein and hence activation by rhodopsin. According to results obtained with other systems, such a form is likely to exist (Bauer et al., 1992) but could not be identified yet unambiguously for the visual system. Excess PDEv could downregulate PDE activity (Erickson et al., 1992); however, to allow activation, it must not disturb the light-induced Gt-PDE interaction step (Otto-Bruc et al., 1993).

180

KLAUS PETER HOFMANN and MARTIN HECK

Altematively, the reported inhibition of trypsin activated PDE by GDP bound Ga (Kroll et al., 1989) opens an analogous refractory state for the PDE: G~GDP still bound to the PDE catalytic core after GTP hydrolysis would block the access of recycled active G t to the effector, thereby preventing prolonged activation.

Vi. APPENDIX: POSTTRANSLATIONAL MODIFICATIONS Most of the proteins involved in the visual cascade are posttranslationally modified (Table 1). Modifications can be reversible or irreversible. The latter include: glycosylation (on extracellular side of integral membrane proteins: rhodopsin, channel), acylation (e.g. palmitoylation on rhodopsin C-terminus), isoprenylation with removal of C-terminal residues (e.g., PDEal3). Reversible modifications are involved in regulation and include phosphorylation and carboxymethylation. In the following, we discuss some modifications of potential significance.

A. Acylation For a review see Gordon et al. (1991). Covalent attachment of a lipid to a protein can stabilize subunit interaction (Linder et al., 1991; Kokame et al., 1992), the interaction with the membrane (Buss et al., 1987) or interaction with other proteins. Numerous proteins are N-acylated on the N-terminus, e.g. G~ subunits (Gi, Go, Gt, G z but not Gs; Mumby et al., 1990). Most acylations are with myristic acid (C 14:0)); G~ and recoverin in rods are an exception; they are heterogeneously acylated (C12:0, C14:0, C14:1, C14:2; Neubert et al., 1992; Kokame et al., 1992; Dizhoo'r et al., 1992). Binding of Ca 2÷ can promote the extrusion of the N-acyl group of recoverin, leading to enhanced membrane binding and/or enhanced interaction with other proteins (Ca2+-myristoyl switch; Zozulya and Stryer, 1992). Interestingly, Gsa is not myristoylated but at least as tightly membrane bound as Go,~ or Gia; in the case of Gta , the myristoylation is not sufficient for tight, ionic strength independent membrane association (Yang and Wensel, 1992). Proteolytic removal of N-terminus of Ga (including the myristoyl-Gly) abolishes its ability to activate PDE efficiently on the membrane, but leaves its ability to activate PDE weakly in solution (Fung and Nash, 1983).

B. Isoprenylation For reviews see Maltese (1990) and Clarke (1992). Isoprenylation (prenylation, polyisoprenylation) is a posttranslational modification that involves the formation of thioether bonds between cysteine and isoprenyl groups (famesyl (C15) or geranylgeranyl (C20)). The modified cysteine is always located in the fourth position from the C-terminus (CXXX- or CAAX-motif, A = aliphatic amino acid; X = C-terminal amino acid). The isoprenylation is followed by removal of the last three amino acids (peptidase cleavage at Cys) and the newly exposed Cys can be carboxyl-methylated.

Light-Induced Protein-Protein Interactions

181

Table 1. Posttranslational Modifications of Some ROS Proteins Protein

Modification

Rhodopsin

G t ct

Small G-proteins PDEGt

Y Recoverin

Kinase Source:

glycosylation on N-terminus I; palmitoylation on Cys-322 and Cys-3232; phosphorylation on Ser-338, Ser-343, Thr-336 by rhodopsin kinase (see text) CAAX-motif of C-term.3; acylation on N-term. Gly4"5'6: C12:0, C14:0, C14:1, C14:24'6; removal of N-term. Met 6,7 CAAX-motif of C-term.8; farnesylation 7'9'1°'11 (indispensable for activation 7,11,12); removal of C-term. AAX and N-term. Met8; carboxylmethylation of C-term. Cys 7'1°'11 (reversiblel°). prenylation9; reversible carboxyl-methylation.l° CAAX-motif of C-term. 13; farnesylation of C-term. Cysl4; removal of C-term. AAXIS; carboxyl-methylation of C-term. Cys14'15'16; acetylation of N-term. Gly. 13 CAXX-motif of C-term.17; geranylgeranylation of C-term. Cysl4; removal of Cterm. AAX18; carboxyl-methylation of C-term. Cys (less than PDEa)14'18; acetylation of N-term. Ser. 17 (rev ersi b le ) pho sphorylati on. 19 acylation on N-term. Gly2°'21'22: C12:0, C14:0, C14:1, C14:22°; "calciummyristoyl switch" (calcium-dependent access of the acyl-group); 21 phosphorylation. 23 CAXX-motif of C-term.24; famesylation 25'26 (indispensable for interaction with R* ).26

1. See Findlay (1986). 2. Ovchinnikov et al. (1988). 3. See Lochrie and Simon (1988). 4. Neubert et al. (1992). 5. Yang and Wensel (1992). 6. Kokame et al. (1992). 7. Fukada et al. (1990). 8. Ovchinnikov et al. (1985). 9. Lai et al. (1990). I0. P6rez-Sala et al. (1991). 11. Fukada et al. (1989). 12. Ohguro et al. (1991). 13. Ovchinnikov et al. (1987). 14. Anant et al. (1992). 15. Ong et al. (1989). 16. Swanson and Appleburry (1983). 17. Lipkin et al. (1990). 18. Catty and Deterre (1991). 19. Hayashi et al. (1991). 20. Dizhoor et al. (1992). 21. Zozula and Stryer (1992). 22. See Koch (1994). 23. Lambrecht and Koch (1991). 24. Lorenz et al. (1991). 25. Anant and Fung (1992). 26. Inglese et al. (1992).

182

KLAUS PETER HOFMANN and MARTIN HECK

Physiologically, the isoprenylation either enhances (similar to the acylation) membrane binding or protein-protein (or subtmit) interaction. Examples: (i) isoprenylation of PDEa~ is mandatory for its membrane association (Ong et al., 1989; Catty and Deterre, 1991), hence important for its activation by active Gt; (ii) farnesylation of rhodopsin-kinase is indispensable for its interaction with R* (Inglese et al., 1992); (iii) G~ has much higher affinity for G~v when the Gcsubunit is famesylated (Ohguro et al., 1991) and the farnesylation is indispensable for GTP-binding (Fukada et al., 1990). Geranylgeranylated proteins are tightly anchored to the membranes, while farnesylated proteins can shuttle between membrane-bound and cytosolic states (Anant et al., 1992); even soluble proteins can be isoprenylated.

C. Methylation Reversible carboxyl-methylation of the isoprenylated C-terminal cystein was shown to be mediated by a methyl transferase associated with the ROS-discmembrane (Perez-Sala et al., 1991). Methylations occur on Gv (Perez-Sala et al., 1991; Fukada et al., 1990) and PDE~ (Ong et al., 1989; Anant et al., 1992). The functional significance might be the reversible conversion of the charged carboxylate anion to a neutral ester moiety, which enhances the hydrophobicity of the isoprenylated C-terminus.

D. Phosphorylation Besides the well-studied phosphorylations of rhodopsin and rhodopsin kinase, phosphorylations of unknown function occur on PDEv (Hayashi et al., 1991), phosducin (Lee et al., 1987) and recoverin (Lambrecht and Koch, 1991).

Vii.

RECENT REVIEWS RELATED TO THE TOPIC

In the following, we provide a (noncomprehensive) list of recent reviews. Topics that were dealt with include: • general (Liebman et al., 1987; Lamb, 1990; Detwiler and Gray-Keller, 1992; Lamb and Pugh, 1992; Pugh and Lamb, 1993; Koutalos and Yau, 1993; Hargrave and Hamm, 1994); • rhodopsin and the seven-helix family (Applebury and Hargrave, 1986; Findlay, 1986; Hofmann, 1986; DeGrip, 1988; Birge, 1990; Hargrave, 1991; Hargrave and McDowell, 1992; Khorana, 1992; Lanyi, 1992; Lewis and Kliger, 1992; Nathans, 1992; Siebert, 1992); • transducin and G-proteins (Gilman, 1987; Lochrie and Simon, 1988; Bourne et al., 1990; Bourne et al., 1991; Kaziro et al., 1991); • rhodopsin-transducin interaction (Hamm, 1991; Hofmann and Kahlert, 1992; Oprian, 1992a,b; Hargrave et al., 1993; Hofmann, 1993);

Light-Induced Protein-Protein Interactions

183

PDE-transducin interaction (Yamazaki, 1992; Pfister et al., 1993); • arrestin and rhodopsin kinase (Palczewski and Benovic, 1991; Glitscher and Riippel, 1992; Lefkowitz et al., 1992); • channel/exchanger (Schnetkamp, 1989; Kaupp, 1991; Kaupp and Koch, 1992); • Ca2+-regulation (Yau, 1991; Kaupp and Koch, 1992; Koch, 1992; Koch, 1994); • techniques (Uhl et al., 1990; Hofmann, 1993); • posttranslational modifications (Maltese, 1990; Gordon et al., 1991; Clarke, 1992; Marshall, 1993).

•

ABBREVIATIONS CBP Ch E CAAX GC Gt GaGTP~,S cGMP GTP~S GMPP(NH)P KD Ki

calcium binding protein(s) channel Na/Ca exchanger C = Cys A = mostly aliphatic residue X = any residue guanylate cyclase rod G-protein, transducin, Gary GTPvS-bound ot-subunit of transducin guanosine 3',5'-cyclic monophosphate guanosine 5'-O-(3-thiotriphosphate) guanyl-5'-yl imidodiphosphate dissociation constant concentration for half maximal inhibition Km Michaelis Menten constant MI Meta I, metarhodopsin I MII Meta II, metarhodopsin II Mill Meta III, metarhodopsin III PDE cGMP phosphodiesterase tPDE trypsin activated PDE pKa hydrogen ion dissociation constant R rhodopsin R* light activated rhodopsin RDH retinol dehydrogenase RK rhodopsin kinase Tween polyoxymethylene-sorbitan-monooleate Vmax maximal enzymatic activity

ACKNOWLEDGMENTS We wish to thank the Deutsche Forschungsgemeinschaft (SFB 60, H-5, and SFB 325, H-5) and the Human Frontier Science Program for financial support of our laboratory studies.

184

KLAUS PETER HOFMANN and MARTIN HECK

REFERENCES Adamus, G., Arendt, A., Hargrave, P. A., Heyduk, T., & Palczewski, K. (1993). The kinetics of multiphosphorylation of rhodopsin. Arch. Biochem. Biophys. 304, 443-447. Anant, J. S., Ong, O. C., Xie, H., Clarke, S., O'Brian, P. J., & Fung, B. K-.K. (1992). In vivo di .fferential prenylation of retinal cyclic GMP phosphodiesterase catalytic subunits. J. Biol. Chem. 267, 687--690. Antonny, .B., Otto-Bruc, A., Chabre, M., & Vuong, T. M. (1993). GTP hydrolysis by purifed ct-subunit of transducin and its complex with the cyclic GMP phosphodiesterase inhibitor. Biochemistry 32, 8646-8653. Applebury, M. L., & Hargrave, P. A. (1986). Molecular biology of the visual pigments. Vision Res. 26, 1881-1885. Amis, S., & Hofmann, K. P. (1993). Two forms ofmetarhodopsin II: Schiffbase deprotonation precedes proton uptake and signaling state. Proc. Natl. Acad. Sci. USA 90, 7849-7853. Arshavsky, V. Y., & Bownds, M. D. (1992). Regulation of deactivation of photoreceptor G protein by its target enzyme and cGMP. Nature 357, 416-417. Arshavsky, V. Y., Dumke, C. L., & Bownds, M. D. (1992). Noncatalytic cGMP-binding sites of amphibian rod cGMP phosphodiesterase control interaction with its inhibitory ~,-subunits. J. Biol. Chem. 267, 24501-24507. Arshavsky, V. Y., Gray-Keller, M. P., & Bownds, M. D. (1991). cGMP suppresses GTPase activity of a portion of transducin equimolar to phosphodiesterase in frog rod outer segments. J. Biol. Chem. 266, 18530-18537. Artemyev, N. O., & Hamm, H. E. (1992). Two-site high,affinity interaction between inhibitory and catalytic subunits of rod cyclic GMP phosphodiesterase. Biochem. J. 283, 273-279. Artemyev, N. O., Rarick, H. M., Mills, J. S., Skiba, N. P., & Hamm, H. E. (1992). Sites of interaction between rod G-protein at-subunit and cGMP-phosphodiesterase y-subunit. J. Biol. Chem. 267, 25067-25072. Baehr, W., Devlin, M., & Applebury, M. L. (1979). Isolation and characterization of cGMP phosphodiesterase from bovine rod outer segments. J. Biol. Chem. 254, 11669-11677. Baldwin, J. M. (1993). The probable arrangement of the helices in G protein-coupled receptors. EMBO J. 12. 1693-1703. Baldwin, E A., & Hubbell, W. L. (1985). Effects of lipid environment on the light-induced conformational changes of rhodopsin. 2. Roles of lipid chain length, unsaturation, and phase state. Biochemistry 24, 2633-2639. Bauer, E J., & Drechsler, M. (1992). Association of cyclic GMP-gated channels and Na +- Ca 2+ -K+ exchangers in bovine retinal rod outer segment plasma membranes. J. Physiol. 451, 109-131. Bauer, P. H., M011er, S., Puzicha, M., Pippig, S., Obermaier, B., Helmreich, E. J. M., & Lohse, M. J. (1992). Phosducin is a protein kinase A-regulated G-protein regulator. Nature 358, 73-76. Baumann, C., & Reinheimer, R. (1973). Temperature dependence of slow thermal reactions during the bleaching of rhodopsin in the frog retina. In: Biochemistry and physiology of visual pigments (Langer, H. ed.), pp. 89-99, Springer Verlag, Berlin, Heidelberg, New York. Baylor, D. A., Nunn, B. J., & Schnapf, J. L. (1984). The photocurrent, noise and spectral sensitivity of rods of the monkey macacafascicularis. J. Physiol. 357, 575-607. Bennett, N., & Dupont, Y. (1985). The G-protein of retinal rod outer segments (transducin) - mechanism of interaction with rhodopsin and nucleotides. J. Biol. Chem. 260, 4156--4168. Bennett, N., & Clerc, A. (1989). Activation of cGMP phosphodiesterase in retinal rods: mechanism of interaction with the GTP-binding protein (transducin). Biochemistry 28, 7418-7424. Bennett, N., & Sitaramayya, A. (1988). Inactivation ofphotoexcited rhodopsin in retinal rods: the roles of rhodopsin kinase and 48-kDa protein (arrestin). Biochemistry 27, 1710-1715.

Light-Induced Protein-Protein Interactions

185

Bentrop, J., Plangger, A., & Paulsen, R. (1993). An arrestin homolog of blowfly photoreceptors stimulates visual-pigment phosphorylation by activating a membrane-associated protein kinase. Eur. J. Biochem. 216, 67-73. Berg, H. C., & Purcell, E. M. (1977). Physics of chemoreception. Biophys. J. 20, 193-219. Birge, R. R. (1990). Nature of the primary photochemical events in rhodopsin and bacteriorhodopsin. Biochim. Biophys. Acta 1016, 293-327. Birge, R. R., Einterz, C. M., Knapp, H. M., & Murray, L. P. (1988). The nature of the primary photochemical events in rhodopsin and isorhodopsin. Biophys. J. 53,367-385. Boesze-Battaglia, K., & Albert, A. (1990). Cholesterol modulation of photoreceptor function in bovine retinal rod outer segments. J. Biol. Chem. 34, 20727-20730. Bourne, H. R., Sanders, D. A., & McCormick, F. (1990). The GTPase superfamily: A conserved switch for diverse cell functions. Nature 348, 125-132. Bourne, H. R., Sanders, D. A., & McCormick, F. (1991 ). The GTPase superfamily: conserved structure and molecular mechanism. Nature 349, 117-127. Brown, R. L. (1992). Functional regions of the inhibitory subunit of retinal rod cGMP phosphodiesterase identified by site-specific mutagenesis and fluorescence spectroscopy. Biochemistry 31, 59185925. Brown, R. L., & Stryer, L. (1989). Expression in bacteria of functional inhibitory subunit of retinal rod cGMP phosphodiesterase. Proc. Natl. Acad. Sci. USA 86, 4922-4926. Bruckert, F., Chabre, M., & Vuong, T. M. (1992). Kinetic analysis of the activation of transducin by photoexcited rhodopsirr---influence of the lateral diffusion of transducin and competition of guanosine diphosphate and guanosine triphosphate for the nucleotide site. Biophys. J. 63, 616-629. Buczytko, J., Gutmann, C., & Palczewski, K. (1991). Regulation of rhodopsin kinase by autophosphorylation. Proc. Natl. Acad. Sci. USA 88, 2568-2572. Byk, T., Bar-Yakov, M., Doza, Y. N., Minke, B., & Selinger, Z. (1993). Regulatory arrestin cycle secures the fidelity and maintenance of the fly photoreceptor cell. Proc. Natl. Acad. Sci. USA 90, 1907-1911. Cafiso, D.S., & Hubbell, W. L. (1980). Light-induced interfacial potentials in photoreceptor membranes. Biophys. J. 30, 243-263. Camps, M., Carozzi, A., Schnable, P., Scheer, A., Parker, P. J., & Gierschik, P. (1992). Isozyme-selective stimulation of PLC-132 by G-protein 133,-subunits. Nature 360, 684-686. Catty, P., & Deterre, P. (1991). Activation and solubilization of the retinal cGMP-specific phosphodiesterase by limited proteolysis. Eur. J. Biochem. 199, 263-369. Catty, P., Pfister, C., Bruckert, F., & Deterre, P. (1992). The cGMP phosphodiesterase-transducin complex of retinal rods. J. Biol. Chem. 267, 19489-19493. Chabre, M., & Deterre, P. (1989). Molecular mechanism of visual transduction. Eur. J. Biochem. 179, 255-266. Chen, T.-W., Peng, Y.-W., Dhallan, R. S., Ahamed, B., Reed, R. R., & Yau, K.-W. (1993). A new subunit of the cyclic nucleotide-gated cation channel in retinal rods. Nature 362, 764-767. Charbonneau, H., Prusti, R. K., LeTrong, H., Sonnenburg, W. K., Mullaney, P. J., Walsh, K. A., & Beavo, J. A. (1990). Identification of a noncatalytic cGMP-binding domain conserved in both the cGMP-stimulated and photoreceptor cyclic nucleotide phosphodiesterases. Proc. Natl. Acad. Sci. USA 87, 288-292. Clarke, S. (1992). Protein isoprenylation and methylation at carboxyl-terminal cysteine residues. Annu. Rev. Biochem. 61,355-386. Clerc, A., & Bennett, N. (1992). Activated cGMP phosphodiesterase of retinal rods. J. Biol. Chem. 267, 6620-6627. Clerc, A., Catty, P., & Bennett, N. (1992). Interaction between cGMP-phosphodiesterase and transducin a-subunit in retinal rods. J. Biol. Chem 267, 19948-19953. Cobbs, W. H. (1991). Light and dark active phosphodiesterase regulation in salamander rods. J. Gen. Physiol. 98, 575--614.

186

KLAUS PETER HOFMANN and MARTIN HECK

Cobbs, W. H., & Pugh, E. N., Jr. (1987). Kinetics and components of the flash photocurrent of isolated retinal rods of the larval salamander, amystome tigrinum. J. Physiol. 394, 529-572. Cohen, G. B., Oprian, D. D., & Robinson, P. R. (1992). Mechanism of activation and inactivation of opsin: role of Glu- 113 and Lys-296. Biochemistry 31, 12592-12601. Cohen, G. B., Yang, T., Robinson, P. R., & Oprian, D. D. (1993). Constitutive activation of opsin: influence of charge at position 134 and size at position 296. Biochemistry 32, 6111-6115. Corson, D. W., Cornwall, M. C., & Pepperberg, D. R. (1994). Evidence for the prolonged photoactivated lifetime of an analogue visual pigment containing 11-cis 9-desmethylretinal. Vis. Neuroscience 11, 91-98. Cunnick, J. M., Hurt, D., Oppert, B., Sakamoto, K., & Takemoto, D. J. (1990). Binding of the ~/-subunit of retinal rod-outer-segment phosphodiesterase with both transducin and the catalytic subunits of phosphodiesterase. Biochem. J. 271, 721-727. DeGrip, W. J., Gillespie, J., & Rothschild, K. J. (1985). Carboxyl group involvemem in the meta I and meta II stages in rhodopsin bleaching. A Fourier transform infrared spectroscopic study. Biochem. Biophys. Acta 809, 97-106. DeGrip, W. J. (1988). Recent chemical studies related to vision. Photochem. Photobiol. 48, 799-810. Deterre, P., Bigay, J., Forquet, F., Robert, M., & Chabre, M. (1988). cGMP phosphodiesterase of retinal rods is regulated by two inhibitory subunits. Proc. Natl. Acad. Sci. USA 85, 2424-2428. Deterre, P., Bigay, J., Robert, M., Pfister, C., Ktihn, H., & Chabre, M. (1986). Activation of retinal rod cyclic GMP-phosphodiesterase by transducin: characterization of the complex formed by phosphodiesterase inhibitor and transducin ct-subunit. Proteins 1,188-193. Detwiler, P. B., & Gray-Keller, M. P. (1992). Some unresolved issues in the physiology and biochemistry of phototransduction. Curr. Opin. Neurobiol. 2, 433-438. Dizhoor, A. M., Ray, S., Kumar, S., Nieme, G., Spencer, M., Brolley, D., Walsh, K. A., Philipov, P. P., Hurley, J. B., & Stryer, L. (1991). Recoverin: a calcium sensitive activator of retinal rod guanylate cyclase. Science 251, 915-918. Dizhoor, A. M., Ericsson, L. H., Johnson, R. S., Kumar, S., Olshevskaya, E., Zozulya, S., Neubert, T. A., Stryer, L., Hurley, J. B., & Walsh, K. A. (1992). The NH 2 terminus of retinal recoverin is acylated by a small family of fatty acids. J. Biol. Chem. 267, 16033-16036. Doukas, A. G., Aton, B., Callender, R. H., & Ebrey, T. G. (1978). Resonance Raman studies of bovine metarhodopsin I and metarhodopsin II. Biochemistry 17, 2430-2435. Dratz, E. A., Furstenau, J. E., Lambert, C. G., Thireault, D. L., Rarick, H., Schepers, T., Pakhlevaniants, S., & Hamm, H. E. (1993). NMR structure of a receptor-bound G-protein peptide. Nature 363, 276-281. Emeis, D., & Hofmann, K. P. (1981). Shift in the relation between flash-induced metarhodopsin I and metarhodopsin II within the first 10% rhodopsin bleaching in bovine disc membranes. FEBS Lett. 136, 201-207. Emrich, H. M., & Reich, R. (1974). Uber Prim~irreaktionen beim Sehvorgang. Thermodynamischer und kinetischer Einflul3 des pH-Wertes auf die Metarhodopsin I-II-Umwandlung. Protonenverbrauch als Auswirkung einer Konformations~inderung. Z. Naturforsch. 29c, 577-591. Erickson, M. A., Robinson, P., & Lisman, J. (1992). Deactivation of visual transduction without guanosine triphosphate hydrolysis by G protein. Science 257, 1255-1258. Farahbakhsh, Z. T., Hideg, K., & Hubbell, W. L. (1993). Photoactivated conformational changes in rhodopsin: a time-resolved spin label study. Science 262, 1416-1419. Fahmy, K., & Sakmar, T. P. (1993a). Light-dependent transducin activation by an ultraviolet-absorbing rhodopsin mutant. Biochemistry 32, 9165-9171. Fahmy, K., & Sakmar, T. P. (1993b). Regulation of rhodopsin-transducin interaction by a highly conserved carboxylic acid group. Biochemistry 32, 7229-7236. Fahmy, K., J/iger, F., Beck, M., Zvyaga, T. A., & Sakmar, T. P. (1993). Protonation states of membraneembedded carboxylic acid groups in rhodopsin and metarhodopsin II: a Fourier-transform infrared spectroscopic study of site-directed mutants. Proc. Natl. Acad. Sci. USA 10206-10210.

Light-Induced Protein-Protein Interactions

187

Findlay, J. B. C. (1986). The structure of rhodopsin. Photobiochem. Photobiophys. 13, 213-228. Franke, R. R., K6nig, B., Sakmar, T. P., Khorana, H. G., & Hofmann, K. P. (1990). Rhodopsin mutants that bind but fail to activate transducin. Science 250, 123-125. Franke, R. R., Sakmar, T. P., Graham, R. M., & Khorana, H. G. (1992). Structure and function in rhodopsin: studies of the interaction between the rhodopsin cytoplasmic domain and transducin. J. Biol. Chem. 267, 14767-14774. Fukada, Y., Ohguro, H., Saito, T., Yoshizawa, T., & Akino, T. (1989). 13T-subunit of bovine transducin composed of two components with distinctive T-subunits. J. Biol. Chem. 264, 5937-5943. Fukada, Y., Takao, T., Ohguro, H., Yoshizawa, T., Akino, T., & Shimonishi, Y. (1990). Famesylated T-subunit of photoreceptor G-protein indispensable for GTP-binding. Nature 364, 658-660. Fung, B. K.-K., & Nash, C. R. (1983). Characterization of transducin from bovine outer segments. II Evidence for distinct binding sites and conformational changes revealed by limited proteolysis with trypsin. J. Biol. Chem. 258, 10503-10510. Fung, B. K.-K., Young, J. H., Yamane, H. K., & Griswold-Prenner, I. (1990). Subunit stoichiometry of retinal rod cGMP phosphodiesterase. Biochemistry 29, 2657-2664. Fung, B. K.-K., Hurley, J. B., & Stryer, L. (1981). Flow of information in the light-triggered cyclic nucleotide cascade of vision. Proc. Natl. Acad. Sci. USA 78, 152-156. Fung, B. K.-K., & Griswold-Prenner, I. (1989). G protein-effector coupling: binding of rod phosphodiesterase inhibitory subunit to transducin. Biochemistry 28, 3133-3137. Ganter, U. M., Longstaff, C., Pajares, M. A., Rando, R. R., & Siebert, F. (1991). Fourier-transform infrared studies of active-site methylated rhodopsin---implications for chromophore-protein interaction, transducin activation, and the reaction pathway. Biophys. J. 59, 640--644. Ganter, U. M., Charitopoulos, T., Virmaux, N., & Siebert, E (1992). Conformational changes ofcytosolic loops of bovine rhodopsin during the transition to metarhodopsin II: an investigation by Fourier transform infrared difference spectroscopy. Photochem. Photobiol. 56, 57-62. Ganter, U. M., Schmid, E. D., Perez-Sala, D., Rando, R. R., & Siebert, F. (1989); Removal ofthe 9-metlyl group of retinal inhibits signal transduction in the visual processma Fourier transform infrared and biochemical investigation. Biochemistry 28, 5954-5962. Ghalayini, A. J., & Anderson, R. E. (1992). Activation of bovine rod outer segment phospholipase C by arrestin. J. Biol. Chem. 267, 17977-17982. Gibson, N. J., & Brown, M. F. (1993). Lipid headgroup and acyl chain composition modulate.the MI-MII equilibrium of rhodopsin in recombinant membranes. Biochemistry 32, 2438-2454. Gillespie, P. G., Prusti, R. K., Apel, E. D., & Beavo, J. A. (1989). A soluble form of bovine rod photoreceptor phosphodiesterase has a novel 15-kDa subunit. J. Biol. Chem. 264, 12187-12193. Gillespie, P. G., & Beavo, J. A. (1988). Characterization of a cone photoreceptor phosphodiesterase purified by cyclic GMP-sepharose chromatography. J. Biol. Chem. 263, 8133--8141. Gillespie, P. G., & Beavo, J. A. (1989). cGMP is tightly bound to bovine retinal rod phosphodiesterase. Proc. Natl. Acad. Sci. USA 86, 4311-4315. Gilman, A. G. (1987). G-proteins: Transducers of receptor-generated signals. Ann. Rev. Biochem. 56, 615--649. Glitscher, W., & Riippel, H. (1991). Evidence for ATP-ase activity of arrestin from bovine photoreceptors. FEBS Lett. 282, 431-435. Glitscher, W., & Rtippel, H. (1992). Arrestirr---A member of a new class of regulating proteins in transduction processes. Biomed. Lett. 47, 315-324. Gorczyka, W. A., Van Hooser, J. P., & Palczewski, K. (1994a) Nucleotide inhibitors and activators of retinal guanylyl cyclase. Biochemistry 33, 3217-3222. Gorczyka, W., Gray-Keller, M. P., Detwiler, P. B., & Palczewski, K. (1994b). Purification and physiological evaluation of a guanylyl cyclase activating protein from retinal rods. Proc. Natl. Acad. Sci., in press. Gordon, J. I., Duronio, R. J., Rudnick, D. A., Adams, S. P., & Gokel, G. W. (1991). Protein N-myristoylation. J. Biol. Chem. 266, 8647-8650.

188

KLAUS PETER HOFMANN and MARTIN HECK

Gurevich, V. V., & Benovic, J. L. (1993). Visual arrestin interaction with rhodopsirr-sequential multisite binding ensures strict selectivity toward light-activated phosphorylated rhodopsin. J. Biol. Chem. 268, 11628-11638. Gray-Keller, M. P., Polans, A. S., Palczewski, K., & Detwiler, P. B. (1993). The effect ofrecoverin-like calcium binding proteins on the photoresponse of retinal rods. Neuron 10, 1-20. Haga, K., & Ham, T. (1989). Agonist-dependent phosphorylation of cerebral and atrial muscarinic receptors: Blockade of the phosphorylation by GTP-binding regulatory proteins and its reversal by guanine nucleotides. Biomed. Res. 10, 293-299. Hamilton, S. E., Prusti, R. K., Bentley, J. K., Beavo, J. A., & Hurley, J. B. (1993). Affinities of bovine photoreceptor cGMP phosphodiesterases for rod and cone inhibitory subunits. FEBS Lett. 318, 157-161. Hamm, H. E. (1991). Molecular interactions between the photoreceptor G protein and rhodopsin. Cell. Mol. Neurobiol. 11,563-578. Hamm, H. E., Deretic, D., Arendt, A., Hargrave, P. A., K6nig, B., & Hofmann, K. P. (1988). Site of G Protein binding to rhodopsin mapped with synthetic peptides from the ct subunit. Science 241, 832--835. Han, M., DeDecker, B. S., & Smith, S. O. (1993). Localization of the retinal protonated Schiff base counterion in rhodopsin. Biophys. J. 65, 899-906. Hargrave, P. A. (1991). Seven-helix receptors. Curr. Opin. Struct. Biol. 1,575-581. Hargrave, P. A., & McDowell, J. M. (1992). Rhodopsin and phototransduction: A model system for G-protein linked receptors. FASEB J. 6, 2323-2331. Hargrave, P. A., Hamm, H. E., & Hofmann, K. P. (1993). Interaction of rhodopsin with the G-protein, transducin. BioEssays 15, 43-50. Hargrave, P. A. & Hamm, H. E. (1994). Regulation of visual transduction. In: Regulation of Cellular Signal Transduction Pathways by Desensitization and Amplification (Sibley, D. R., & Houslay, M. D. eds.), pp. 25-67. John Wiley & Sons Ltd., Chichester. Hayashi, F., Lin, G. Y., Matsumoto, H., & Yamazaki, A. (1991). Phosphatidylinositol-stimulated phosphorylation of an inhibitory subunit of cGMP phosphodiesterase in vertebrate rod photoreceptors. Proc. Natl. Acad. Sci. USA 88, 4333-4337. Heck, M., & Hofmann, K. P. (1993). G-protein-effector coupling: a real-time light-scattering assay for transducin-phosphodiesterase interaction. Biochemistry 32, 8220-8227. Hingorani, V. N., Tobias, D. T., Henderson, J. T., & Ho, Y.-K. (1988). Chemical cross-linking of bovine retinal transducin and cGMP phosphodiesterase. J. Biol. Chem. 263, 6916-6926. Hofmann, K. P. (1986). Photoproducts of rhodopsin in the disc membrane. Photobiochem. Photobiophys. 13, 309-338. Hofmann, K. P. (1993). Rhodopsin/G-protein interaction. In: GTPases in BiologymHandbook of Experimental Pharmacology (Dickey, B., & Bimbaumer, L., eds.), pp. 267-290. Springer. Hofmann, K. P., Emeis, D., & Schnetkamp, P. P. M. (1983). Interplay between hydroxylamine, metarhodopsin II and GTP-binding protein in bovine photoreceptor membranes. Biochim. Biophys. Acta 725, 60-70. Hofmann, K. P., & Kahlert, M. (1992). The activation of transducin: studies on its mechanism and modulation. In: Signal Transduction in Photoreceptor Cells (Hargrave, P. H., Hofmann, K. P., & Kaupp, U. B., eds.), Springer 1992. Hofmann, K. P., PulvermOller, A., Buczylko, J., Hooser, P. V., & Palczewski, K. (1992). The role of arrestin and retinoids in the regeneration pathway ofrhodopsin. J. Biol. Chem. 267, 15701-15706. Hsu, Y. T., & Molday, R. S. (1993). Modulation of the cGMP-gated channel of rod photoreceptor cells by calmodulin. Nature 361, 76-79. Hubbell, W. L. (1990). Transbilayer coupling mechanism for the formation of lipid asymmetry in biological membranes---Application to the photoreceptor disc membrane. Biophys. J. 57, 99-108. Huppertz, B., Weyand, I., & Bauer, P. (1990). Ca 2÷ binding capacity of cytoplasmic proteins from rod outer segments is mainly due to arrestin. J. Biol. Chem. 265, 9470-9475.

Light-Induced Protein-Protein Interactions

189

Hurley, J. B., & Stryer, L., (1982). Purification and characterization of the y regulatory subunit of the cyclic GMP phosphodiesterase from retinal rod outer segments. J. Biol. Chem. 257, 11094-11099. Hurley, J. B., Dizhoor, A. M., Ray, S., & Stryer, L. (1993). Recoverin's role: conclusion withdrawn (letter). Science 260, 740. Hsu, Y.-T., Wong S. Y. C., Connell, G. J., & Molday, R. S. (1993). Structural and functional properties of rhodopsin from rod outer segment disk and plasma membrane. Biochim. Biophys. Acta 1145, 85-92. Inglese, J., Koch, W. J., Caron, M. G., & Lefkowitz, R. J. (1992). Isoprenylation in regulation of signal transduction by G-protein-coupled receptor kinases. Nature 359, 147-150. Ishiguro, S., Suzuki, Y., Tamai, M. & Mizuno, K. (1991). Purification of retinol dehydrogenase from bovine retinal rod segments. J. Biol. Chem. 266, 15520-15524. Jansen, J. J. M., Mulder, W. R., DeCaluwe, G. L. J., Vlak, J. M., & DeGrip, W. J. (1991). In vitro expression of bovine opsin using recombinant baculovirus: the role ofglutamic acid (134) in opsm biosynthesis and glycosylation. Biochim. Biophys. Acta 1089, 68-76. Jelsema, C. L. (1987). Light activation of phospholipase A 2 in rod outer segments of bovine retina and its modulation by GTP-binding proteins. J. Biol. Chem. 262, 163--168. Jin, J., Crouch, R. K., Corson, D. W., Katz, B. M., MacNicol, E. F., & Cornwall, M. C. (1993). Noncovalent occupancy of the retinal-binding pocket of opsin diminishes bleaching adaptation of retinal cones. Neuron 11, 513-522. Jones, T. L. Z., Simonds, W. F., Merendino, J. J., Jr., Brann, M. R., & Spiegel, A. M. (1990). Myristoylation of an inhibitory GTP-binding protein ot subunit is essential for its membrane attachment. Proc. Natl Acad. Sci. USA 87, 568-572. Journot, L., Pantaloni, C., Bockaert, J., & Audigier, Y. (1991). Deletion within the amino-terminal region of Gset impairs its ability to interact with 137 subunits and to activate adenylate cyclase. J. Biol. Chem. 266, 9009-9015. Kahlert, M., Pepperberg, D. R., & Hofmann, K. P. (1990). Effect of bleached rhodopsin on signal amplification in rod visual receptors. Nature 345,537-539. Kahlert, M., & Hofmann, K. P. (1991). Reaction rate and collisional efficiency of the rhodopsintransducin system in intact retinal rods. Biophys. J. 59, 375-386. Karnik, S. S., Rodge, K. D., Bhattacharya, S., & Khorana, H. G. (1993). Palmitoylation of bovine opsin and its cysteine mutants in COS cells. Proc. Natl. Acad. Sci. USA 90, 40-44. Kaupp, U. B. (1991). The cyclic nucleotide-gated channels of vertebrate photoreceptors and olfactory epithelium. Trends Neurosci. 14, 150-157. Kaupp, U. B., & Koch, K.-W. (1992). Role of cGMP and Ca2+ in vertebrate photoreceptor excitation and adaptation. Ann. Rev. Physiol. 54, 153-175. Kawamura, S., Hisatomi, O., Kayada, S., Tokunaga, F., & Kuo, C.-H. (1993). Recoverin has S-modulin activity in frog rods. J. Biol. Chem. 268, 14579-14582. Kawamura, S. (1993). Rhodopsin phosphorylation as a mechanism of cyclic GMP phosphodiesterase regulation by S-modulin. Nature 362, 855-857. Kelleher, D. J., & Johnson, G. L. (1988). Transducin inhibition of light-dependent rhodopsin phosphorylation: Evidence for 137subunit interaction with rhodopsin. Mol. Pharmacol. 34, 452-460. Khorana, H. G. (1992). Rhodopsin, photoreceptor of the rod cell. An emerging pattern for structure and function. J. Biol. Chem. 267, 1-4. Kibelbek, J., Mitchell, D. C., Beach, J.M., & Litman, B.J. (1991). Functional equivalence of metarhodopsin II and the Gt-activating form of photolyzed bovine rhodopsin. Biochemistry 30, 67616768. Koch, K.-W. (1992). Biochemical mechanism of light adaptation in vertebrate photoreceptors. Trends Biochem. Sci. 17, 307-311. Koch, K.-W. (1994). Calcium as modulator of phototransduction in vertebrate photoreceptor cells. Rev. Physiol. Biochem. Pharmacol. 125, 149-192.

190

KLAUS PETER HOFMANN and MARTIN HECK

Koch, K.-W., & Stryer, L. (1988). Highly cooperative feedback control of retinal rod guanylate cyclase by calcium ions. Nature 334, 64--66. Kohl, B., & Hofmann, K. P. (1987). Temperature dependence of G-protein activation in photoreceptor membranes - transient extra metarhodopsin II on bovine disc membranes. Biophys. J. 52, 271-277. Krnig, B., Welte, W., & Hofmann, K.P. (1989a). Photoactivation of rhodopsin and interaction with transducin in detergent micelles. FEBS Lett. 257, 163-166. Krnig, B., Arendt, A., McDowell, J. H., Kahlert, M., Hargrave, P. A., & Hofmann, K. P. (1989b). Three cytoplasmic loops of rhodopsin interact with transducin. Proc. Natl. Acad. Sci. USA 86, 68786882. Kokame, K., Fukada, Y., Yoshizawa, T., Takao, T., & Shimonishi, Y. (1992). Lipid modification at the N terminus of photoreceptor G-protein ct-subunit. Nature 359, 749-752. Koutalos, Y., & Yau, K.-W. (1993). A rich complexity emerges in phototransduction. Curr. Opin. Neurobiol. 3, 51 3-519. Kroll, S., Phillips, W. J., & Cerione, R. A. (1989). The regulation of the cyclic GMP phosphodiesterase by the GDP-bound form of the ot subunit of transducin. J. Biol. Chem. 264, 4490--4487. KOhn, H. (1978). Light-regulated binding of rhodopsin kinase and other proteins to cattle photoreceptor membranes. Biochemistry 17, 4389-4395. KOhn, H. (1980). Light- and GTP-regulated interaction of GTPase and other proteins with bovine photoreceptor membranes. Nature 283, 587-589. KOhn, H., & Hargrave, P. A. (1981). Light-induced binding of GTPase to bovine photoreceptor membranes: Effects of limited proteolysis of the membranes. Biochemistry 20, 2410-2417. Kiitm, H., Bennett, N., Michel-Villaz, M., & Chabre, M. (1981). Interactions between photoexcited rhodopsin and GTP-binding protein: Kinetic and stoichiometric analysis from light scattering changes. Proc. Natl. Acad. Sci. USA 78, 6873--6877. Lagnado, L., & Baylor, D. A. (1994). Calcium controls light-triggered formation of catalytically active rhodopsin. Nature 367, 273-277. Lai, R. K., Perez-Sala, D., Cafiada, F. J., & Rando, R. R. (1990). The T subunit of transducin is farnesylated. Proc. Natl. Acad. Sci. USA 87, 7673-7677. Lamb, T. D., & Pugh, E. N. Jr. (1992). A quantitative account of the activation steps involved in phototransduction in amphibian photoreceptors. J. Physiol. 449, 719-758. Lamb, T. D. (1990). Dark adaptation: a re-examination. In: Night vision, Basic, Clinical and applied aspects. (Hess, R. F., Sharpe, L. T., Nordby, K. eds.), pp 177-222, Cambridge University Press. Lambrecht, H.-G., & Koch, K.-W. (1991a). A 26 kDa calcium binding protein from bovine rod outer segment as modulator of photoreceptor guanylate cyclase. EMBO J. 10, 793-798. Lambrecht, H.-G., & Koch, K.-W. (1991b). Phosphorylation of recoverin, the calcium-sensitive activator of photoreceptor guanylyl cyclase. FEBS Lett. 294, 207-209. Lanyi, J. K. (1992). Proton transfer and energy coupling in the bacteriorhodopsin photocycle. J. Bioenerg. Biomemb. 24, 169-179. Lee, R. H., Lieberman, B. S., Yamane, H. K., Bok, D., & Fung, B. K.-K. (1992). A third form of the G protein 13subunit- 1. Immunochemical identification and localization to cone photoreceptors. J. Biol. Chem. 267, 24776-24781. Lee, R. H., Lieberman, B. S., & Lolley, R. N. (1987). A novel complex from bovine visual cells of a 33,000-Dalton phosphoprotein with 13- and T-Transducin: Purification and subunit structure. Biochemistry 28, 3983-3990. Lee, R. H., Fowler, A., McGinnis, J. F., Lolley, R. N., & Craft, C. M. (1990). Amino acid and cDNA sequence of bovine phosducin, a soluble phosphoprotein from photoreceptor cells. J. Biol. Chem. 265, 15867-15873. Lee, R. H., Ting, T. D., Lieberman, B. S., Tobias, D. E., Lolley, R. N., & Ho, Y.-K. (1992). Regulation of retinal cGMP cascade by phosducin in bovine rod photoreceptor cells - interaction ofphosducin and transducin. J. Biol. Chem. 267, 25104-25112.

Light-Induced Protein-Protein Interactions

191

Lefkowitz, R. J., Inglese, J., Koch, W. J., Pitcher, J., Attramadal, H., & Caron, M. G. (1992). G-protein-coupled receptors: regulatory role of receptor kinases and arrestin proteins. Cold Spring Harb. Symp. Qant. Biol. LVII, 127-133. LeVine III, H., Smith, D. P., Whitney, M., Malicki, D. M., Dolph, P. J., Smith, G. F. H., Burkhart, W., & Zuker, C. S. (1991). Isolation of a novel visual-system-specific arrestin: an in vivo substrate for light-dependent phosphorylation. Mech. Develop. 33, 19-26. Lewis, J. L., & Kliger, D. S. (1992). Photointermediates of visual pigments. J. Bioenerg. Biomemb. 24, 201-210. Li, T., Volpp, K., & Applebury, M. L. (1990). Bovine cone photoreceptor cGMP phosphodiesterase structure deduced from cDNA clone. Proc. Natl. Acad. Sci. USA 87, 293-297. Liebman, P. A., & Pugh, E. N. Jr. (1980). ATP mediates rapid reversal of cyclic GMP phosphodiesterase activation in visual receptor membranes. Nature 287, 734-736. Liebman, P. A., Parker, K. R., & Dratz, E. A. (1987). The molecular mechanism of visual excitation and its relation to the structure and composition of the rod outer segment. Ann. Rev. Physiol. 49, 765-791. Lin, S. W., Sakmar, T. P., Franke, R. R., Khorana, H. G., & Mathies, R. A. (1992). Resonance Raman microprobe spectroscopy of rhodopsin mutants: effect of substitutions in the third transmembrane helix. Biochemistry 31, 5105-5111. Linder, M. E., Pang, I.-H., Duronio, R. J., Gordon, J. I., Sternweis, P. C., & Gilman, A. G. (1991). Lipid modifications of G protein subunits. Myristoylation of Goa increases its affinity for 13T.J. Biol. Chem. 266, 4654-4659. Lipkin, V. M., Dumler, I. L., Muradov, K. G., Artemyev, N. O., & Etingof, R. N. (1988). Active sites of the cyclic GMP phosphodiesterase T-subunit of retinal rod outer segments. FEBS Lett. 234, 287-290. Lipkin, V. M., Khramtsov, N. V., Vasilevskaya, I. A., Atabekova, N. V., Muradov, K. G., Gubanov, V. V., Li, T., Johnston, J. P., Volpp, K. J., Applebury, M. L. (1990). 13-subunit of bovine rod photoreceptor cGMP phosphodiesterase. J.Biol. Chem. 265, 12955-12959. Litman, B. J., Kalisky, O., & Ottolenghi, M. (1981). Rhodopsin-phospholipid interactions: dependence of rate of the Meta I to Meta II transition on the level of associated disk phospholipid. Biochemistry 20, 631-634. Lochrie, M. A., & Simon, M. I. (1988). G protein multiplicity in eukariotic signal transduction systems. Biochemistry 27, 4957-4965. Lohse, M. J., Benovic, J. L., Codina, J., Caron, M. G., & Lefkowitz, R. J. (1990). 13-arrestin: a protein that regulates 13-adrenergic receptor function. Science 248, 1547-1550. Lolley, R. N., & Racz, E. (1982). Calcium modulation of cyclic GMP synthesis in rat visual cells. Vision Res. 22, 1481-1486. Longstaff, C., Calhoon, R. D., & Rando, R. R. (1986). Deprotonation of the Schiffbase ofrhodopsin is obligate in the activation of the G protein. Proc. Natl Acad. Sci. USA 83, 4209--4213. Lorenz, W., Inglese, J., Palczewski, K., Onorato, J. J., Caron, M. G., & Lefkowitz, R. J. (1991). The receptor kinase family: primary structure of rhodopsin kinase reveals similarities to the 13-adrenergic receptor kinase. Proc. Natl. Acad. Sci. 88, 8715-8719. Malinski, J. A., & Wensel, T. G. (1992). Membrane stimulation of cGMP phosphodiesterase activation by transducin: Comparison of phospholipid bilayers to rod outer segment membranes. Biochemistry 31, 9502-9512. Maltese, W. A. (1990). Posttranslational modification of proteins by isoprenoids in mammalian cells. FASEB J. 4, 3319-3328. Mangini, N. J., & Pepperberg, D. R. (1988). Immunolocalization of 48K in rod photoreceptors - light and ATP increase OS labeling. Invest. Ophthal. Vis. Sci. 29, 1221-1234. Marshall, C. L. (1993). Protein prenylation: a mediator of protein-protein interactions. Science 259, 1865-1866.

192

KLAUS PETER HOFMANN and MARTIN HECK

Matthews, R. G., Hubbard, R., Brown, P. K., & Wald, G. (1963). Tautomeric forms of metarhodopsin. J. Gen. Physiol. 47, 215-240. MeDowell, J. E, Nawroell, J. P., & Hargrave, P. A. (1993). Phosphorylation sites in bovine rhodopsin. Biochemistry 32, 4968-4974. Mill, N., Baraban, J. M., Keims, J. J., Boyee, J. J., & Bitensky, M. W. (1975). Purification and prol~erties of the light-activated cyclic nueleotide phosphodiesterase of rod outer segments. J. Biol. Chem. 250, 6320-6327. Miljanieh, G. P., Nemes, P. E, White, D. L., & Dratz, E. A. (1981). The asymmetric transmembrane distribution ofphophatidylethanolamine, phosphatidylserine, and fatty acids of the bovine retinal rod outer segment disc membrane. J. Memb. Biol. 60, 249--255. Miller, J. L., Fox, D. A., & Litman, B. J. (1986). Amplification ofphosphodiesterase activation is greatly reduced by rhodopsin phosphorylation. Biochemistry 25, 4983-4988. Mitchell, D. C., Straume, M., Miller, J. L., & Litman, B. J. (1990). Modulation of metarhodopsin formation by cholesterol-induced ordering of bilayer lipids. Biochemistry 29, 9143-9149. Mitchell, D. C., Kibelbek, J., & Litman, B. J. (1992). Effect of phosphorylation on receptor conformation: The metarhodopsin I metarhodopsin II equilibrium in multiply phosphorylated rhodopsin. Biochemistry 31, 8107-8111. Molday, L. L., Cook, N. J., Kaupp, U. B., & Molday, R. S. (1990). The eGMP-gatecl channel of bovine rod photoreeeptor cells is associated with a 240-kDa protein exhibiting immunoehemieal crossreactivity with spectrin. J. Biol. Chem. 265, 18690-18695. Morrison, D. F., Rider, M. A., & Takemoto, D. J. (1987). Modulation of retinal transduein and phosphodiesterase activities by synthetic peptides of the phosphodiesterase T-subunit. FEBS Lett. 222, 266--270. Morrison, D. F., Cunniek, J. M., Oppert, B., & Takemoto, D. J. (1989). Interaction of the y-subunit of retinal rod outer segment phosphodiesterase with transducin. J. Biol. Chem 264, 11671-11681. Morrison, D. E, O'Brien, P. J., & Pepperberg, D. R. (1991). Depalmitoylation with hydroxylamine alters the functional properties of rhodopsin. J. Biol. Chem. 266, 20118-20123. Mumby, S. M., Heukeroth, R. O., Gordon, J. I., & Gilman, A. G (1990). G-protein a-subunit expression, myristoylation, and membrane association in COS cells. Proc. Natl. Aead. Sci. USA 87, 728-732. Nathans, J. (1990). Determinants of visual pigment absorbanee: identification of retinylidene Schiff's base eounterion in bovine rhodopsin. Biochemistry 29, 9746-9752. Nathans, J. (1992). Rhodopsin: structure, functions, and genetics. Biochemistry 31, 4923-4931. Navon, S. E., & Fung, B. K.-K. (1987). Characterization of transduein from bovine retinal rod outer segments. J. Biol. Chem. 262, 15746-15751. Neubert, T. A., Johnson, R. S., Hurley, J. B., & Walsh, K. A. (1992). The rod transducin alpha subunit amino terminus is heterogeneously fatty acylated. J. Biol. Chem. 267, 18274-18277. Newton, A. C., & Williams, D. S. (1991). Involvement of protein kinase C in the phosphorylation of rhodopsin. J. Biol. Chem. 266, 17725-17728. Noel, J. E, Hamm, H. E., & Sigler, E B. (1993). The 2.2 ,~, crystal structure oftransducin-a complexed with GTPTS. Nature 366, 654-663. Ohguro, H., Fukada, Y., Takao, T., Shimonishi, Y., Yoshizawa, T., & Akino, T. (1991). Carboxyl methylation and famesylation of transduein T-subunit synergistically enhance its coupling with metarhodopsin II. EMBO J. 10, 3669--3674. Ohguro, H., Johnson, R. S., Ericsson, L. H., Walsh, K. A., & Palezewsld, K. (1993). Control ofrhodopsin multiple phosphorylation. Biochemistry, in press. Ohguro, H., Chiba, S., Igarashi, Y., Matsumoto, H., Akino, T., & Palezewski, K. (1993). 13-arrestin and arrestin are recognized by autoantibodies in sera from multiple sclerosis patients. Proe. Natl. Acad. Sci. USA 90, 3241-3245. Okamoto, T., Katada, T., Murayama, Y., Ui, M., Ogata, E., Nishimoto, I. (1990). A simple structure encodes G protein-activating function of the IGF-II/Mannose &phosphate receptor. Cell 62, 709-717.

Light-induced Protein-Protein Interactions

193

Ong, O.C., Ota, I. M., Clarke, S., & Fung, B. K.-K. (1989). The membrane binding domain of rod cGMP phosphodiesterase is posttranslationally modified by methyl esterification at the C-terminal cysteine. Proc. Natl. Acad. Sci. USA 86, 9238-9242. Oppert, B., Cunnick, J. M., Hurt, D., & Takemoto, D. J. (1991). Identification of the retinal cyclic GMP phosphodiesterase inhibitory ~/-subunit interaction sites on the catalytic c~-subunit. J. Biol. Chem. 266, 16607-16613. Oprian, D. D., Molday, R. S., Kaufman, R. J., & Khorana, H. G. (1987). Expression of a synthetic bovine rhodopsin gene in monkey kidney cells. Proc. Natl. Acad. Sci. USA 84, 8874-8878. Oprian, D. D. (1992a). Molecular determinants of spectral properties and signal transduction in the visual pigments. Curr. Opinion Neurobiol. 2, 428-432. Oprian, D. D. (1992b). The ligand-binding domain of rhodopsin and other G protein-linked receptors. J. Bioenerg. Biomemb. 24, 211-217. Otto-Bruc, A., Antonny, B., Vuong, Y. M., Chardin, P., & Chabre, M. (1993). Interaction between the retinal cyclic GMP phosphodiesterase inhibitor and transducin. Kinetics and affinity studies. Biochemistry 32, 8636-8645. Ovchinnikov, Y. A., Lipkin, V. M., Shuvaeva, T. M., Bogachuk, A. P., & Shemiakin, V. V. (1985). Complete amino acid sequence of gamma-subunit of the GTP-binding protein from cattle retina. FEBS Lett. 179, 107-110. Ovchinnikov, Y. A., Lipkin, V. M., Kumarev, V. E, Gubanov, V. V., Khramtsov, N. V., Akhmedov, N. B., Zagranichny, V. E., & Muradov, K. G. (1986). Cyclic GMP phosphodiesterase from cattle retina. FEBS Lett. 204, 288-292. Ovchinnikov, Y. A., Gubanov, V. V., Khramtsov, N. V., Ischenko, K. A., Zagranichny, V. E., Muradov, K. G., Shuvaeva, T. M., & Lipkin, V. M. (1987). Cyclic GMP phosphodiesterase from bovine retina. FEBS Lett. 223, 169-173. Ovchinnikov, Y. A., Abdulaev, N. G., & Bogachuk, A. S. (1988). Two adjacent cysteine residues in the C-terminal cytoplasmic fragment of bovine rhodopsin are palmitylated. FEBS Lett. 230, 1-5. Pages, F., Deterre, P., & Pfister, C. (1992). Enhanced GTPase activity of transducin when bound to cGMP phosphodiesterase in bovine retinal rods. J. Biol. Chem. 267, 22018-22021. Palczewski, K., McDowell, J.H., Jakes, S., Ingebritsen, T.S., & Hargrave, EA. (1989). Regulation of rhodopsin dephosphorylation by arrestin. J. Biol. Chem. 264, 15770-15773. Palczewski, K., & Hargrave, P. A. (1991). Studies of ligand binding to arrestin. J. Biol. Chem. 266, 4201-4206. Palczewski, K., Buczytko, J., Kaplan, M. W., Polans, A. S., & Crabb, J. W. (1991a). Mechanism of rhodopsin kinase activation. J. Biol. Chem. 266, 12949-12955. Palczewski, K., Pulvermtiller, A., Buczyiko, J., & Hofmann, K. P. (1991b). Phosphorylated rhodopsin and heparin induce similar conformational changes in arrestin. J. Biol. Chem. 266, 18649-18654. Palczewski, K., & Benovic, J. L. (1991). G-protein-coupled receptor kinases. Trends Biochem. Sci. 16, 387-391. Palczewski, K., Buczytko, J., Van Hooser, P., Carr, S. A., Huddleston, M. J., & Crabb, J. W. (1992a). Identification of the autophosphorylation sites in rhodopsin kinase. J. Biol. Chem. 267, 1899118998. Palczewski, K., Rispoli, G., & Detwiler, P. B. (1992b). The influence of arrestin (48K protein) and rhodopsin kinase on visual transduction. Neuron 8, 11%126. Palczewski, K., Pulvermtiller, A., Buczylko, J., Gutmann, C., & Hofmann, K. P. (1992c). Binding of inositol phosphates to arrestin. FEBS Lett. 295, 195-199. Palczewski, K., Buczytko, J., Lebioda, L., Crabb, J. W., & Polans, A. S. (1993). Identification of the N-terminal region in rhodopsin kinase involved in its interaction with rhodopsin. J. Biol. Chem. 268, 6004-6013. Papac, D. I., Oatis, J. E., Crouch, R. K., & Knapp, D. R. (1993). Mass spectrometric identification of phosphorylation sites in bleached bovine rhodopsin. Biochemistry 32, 5930-5934.

194

KLAUS PETER HOFMANN and MARTIN HECK

Parkes, J. H., & Liebman, P. A. (1984). Temperature and pH dependence of the metarhodopsin I-metarhodopsin II kinetics and equilibria in bovine rod disk membrane suspensions. Biochemistry 23, 5054-5061. Paulsen, R., & Bentrop, J. (1983). Activation of rhodopsin phosphorylation is triggered by the lumirhodopsin-metarhodopsin I transition. Nature 302, 417-419. Pepperberg, D. R., Kahlert, M., Krause, A., & Hofmann, K. E (1988). Photic modulation of a highly sensitive, near-infrared light-scattering signal recorded from intact retinal photoreceptors. Proc. Natl. Acad. Sci. USA 85, 5531-5535. Pepperberg, D. R., Cornwall, M. C., Kahlert, M., Hofmann, K. P., Jin, J., Jones, G. J., & Ripps, H. (1992). Light-dependent delay in the falling phase of the retinal rod photoresponse. Vis. Neuroscience 8, 9-18. Pepperberg, D. R., Jin, J., & Jones, G. J. (1994). Modulation of transduction gain in light adaptation of retinal rods. Vis. Neuroscience 11, 53-62. Perez-Sala, D., Tan, E.W., Cafiada, E J., & Rando, R. R. (1991). Methylation and demethylation reactions of guanine nucleotide-binding proteins of retinal rod outer segments. Proc. Natl. Acad. Sci. USA 88, 3043-3046. Pfister, C., Chabre, M., Plouet, J., Tuyen, V. V., De Kozak, Y., Faure, J. P., & KOhn, H. (1985). Retinal S antigen identified as the 48K protein regulating light-dependent phosphodiesterase in rods. Science 228, 891--893. Pfister, C., Bennett, N., Bruckert, E, Catty, E, Clerc, A., Pag6s, E, & Deterre, E (1993). Interactions of a G-protein with its effector: transducin and cGMP phosphodiesterase in retinal rods. Cellular Signalling 5,235-251. Phillips, W. J., Trukawinski, S., & Cerione, R. A. (1989). An antibody-induced enhancement of the transducin-stimulated cyclic GMP phosphodiesterase activity. J. Biol. Chem. 264, 16679-16688. Phillips, W. J., & Cerione, R. A. (1992). Rhodopsin/transducin interactions. J. Biol. Chem. 267, 17032-17039. Pugh, E.N.Jr., & Lamb, T. D. (1993). Amplification and kinetics of the activation steps in phototransduction. Biochim. Biophys. Acta 1141, 111-149. PulvermOller, A., Palczewski, K., & Hofmann, K. E (1993). Interaction between photoactivated rhodopsin and its kinase: stability and kinetics of complex formation. Biochemistry 32, 1408214088. Qin, N., Pittler, S. J., & Baehr, W. (1992). In vitro isoprenylation and membrane association of mouse rod photoreceptor cGMP phosphodiesterase Gtand 13subunits expressed in bacteria. J. Biol. Chem. 267, 8458-8463. Rarick, H. M., Artemyev, N. O., & Hamm, H. E. (1992). A site on rod G protein ot subunit that mediates effector activation. Science 256, 1031-1033. Rath, P., DeCaluv6, L. L. J., Bovee-Geurts, P. H. M., DeGrip, W. J., & Rothschild, K. J. (1993). Fourier transform infrared difference spectroscopy of rhodopsin mutants: light activation of rhodopsin causes hydrogen-bonding change in residue aspartic acid-83 during meta II formation. Biochemistry 32, 10277-10282. Ray, S., Zozulya, S., Niemi, G. A., Flaherty, K. M., Brolley, D., Dizhoor, A. M., McKay, D. B., Hurley, J., & Stryer, L. (1992). Cloning, expression and crystallization of recoverin, a calcium sensor in vision. Proc. Natl. Acad. Sci. USA 89, 5705-5709. Resek, J., Farahbakhsh, Z., Hubbell, W., & Khorana, H. G. (1993). Formation of the meta II photointermediate is accompanied by conformational changes in the cytoplasmic surface of rhodopsin. Biochemistry 32, 12025-12032. Richard, E. A., & Lisman, J. E. (1992). Rhodopsin inactivation is a modulated process in Limulus photoreceptors. Nature 356, 3336--3338. Rispoli, G., & Detwiler, P. B. (1992). Interaction between calcium and cGMP in dialyzed detached retinal rod outer segments. In: Signal transduction in photoreceptor cells. (Hargrave, P.A.,

Light-lnduced Protein-Protein Interactions

195

Hofmann, K.R~ & Kaupp. U.B.. eds.), pp. 217-235, Springer-Verlag, Berlin, Heidelberg, New York. Robinson, P. R., Cohen, G. B., Zhukovsky, E. A., & Oprian, D. D. (1992). Constitutively active mutants of rhodopsin. Neuron 9, 719-725. Ryba, N. J. P., & Marsh, D. (1992). Protein rotational diffusion and lipid/protein interactions in recombinants of bovine rhodopsin with saturated diacylphosphatidylcholines of different chain length studied by conventional and saturation-transfer electron spin resonance. Biochemistry 31, 7511-7518. Ryba, N. J. P., Marsh, D., & Uhl, R. (1993). The kinetics and thermodynamics of bleaching ofrhodopsin in dimyristoylphosphatidylcholine - identification of meta-I, meta-II, and meta-III intermediates. Biophys. J. 64, 1801-1812. Saari, J. C., Bredberg, D. L., Garwin, G. G., Buczylko, J., Wheeler, T., & Palczewski, K. (1993). Assays of retinoid dehydrogenases by phase partition. Anal. Biochem. 213, 128-132. Sakmar, T. P., Franke, R. R., & Khorana, H. G. (1989). Glutamic acid 113 serves as the retinylidine schiffbase counterion in bovine rhodopsin. Proc. Natl. Acad. Sci. USA 86, 8309-8313. Sakmar, T.P. (1994). Opsins. In: Handbook of receptors and channels; G-protein coupled receptors. Peroutka, S. J., ed.. CRC press, Boca Raton, Ann Arbor, London, Tokyo. Schertler, G. F. X., Villa, C., & Henderson, R. (1993). Projection structure of rhodopsin. Nature 362, 770-772. Schleicher, A., Franke, R., Hofmann, K. P., Finkelmann, H., & Welte, W. (1987). Deoxylysolecithin and a new biphenyl detergent as solubilizing agents for bovine rhodopsin. Functional test by formation of metarhodopsin II and binding of G-protein. Biochemistry 26, 5908-5916. Schleicher, A., & Hofmann, K. P. (1987). Kinetic study on the equilibrium between membrane-bound and free photoreceptor g-protein. J. Neuro. Biol. 95,271-281. Schleicher, A., K~hn, H., & Hofmann, K. P. (1989). Kinetics, binding constant and activation energy of the 48 kDa protein-rhodopsin complex by extra-metarhodopsin II. Biochemistry 28, 1770-1775 Schnetkamp, P. P. M. (1985). Calcium buffer sites in intact bovine rod outer segments: introduction to a new optical probe to measure ionic permeabilities in suspensions of small particles. J. Membr. Biol. 88, 249-262. Schnetkamp, E P. M. (1989). Na-Ca or Na-Ca-K exchange in rod photoreceptors. Prog. Biophys. Molec. Biol. 54, 1-29. Schmidt, K.-F-, N611, G. N., & Yamamoto, Y. (1992). Sodium nitroprusside alters dark voltage and light responses in isolated retinal rods during whole cell recording. Vis. Neurosci. 9, 205-209. Seckler, B., & Rando, R. R. (1989). Schiff-base deprotonation is mandatory for light-dependent rhodopsin phosphorylation. Biochem. J. 264, 489-493. Shinohara, T., Dietzschold, B., Craft, C. M., Wistow, G., Early, J. J., Donoso, L. A., Horwitz, J., & Tao, R. (1987). Primary and secondary structure of bovine retinal S antigen (48-kDA protein). Proc. Natl. Acad. Sci. USA 84, 6975-6979. Shyjan, A. W., de Sauvage, F. J., Gillett, N. A., Goeddel, D. V., & Lowe, D. G. (1992). Molecular cloning of a retina-specific membrane guanylyl cyclase. Neuron 9, 727-737. Siebert, F., M~ntele, W., & Gerwert, K. (1983). Fourier-transform infrared spectroscopy applied to rhodopsin; the problem of the protonation state of the retinylidene Schiffbase reinvestigated. Eur. J. Biochem. 136, 119--127. Siebert, F. (1992). Infrared spectroscopic investigation of retinal proteins. In: Biomolecular Spectroscopy. Advances in Spectroscopy (Clark, R.J.H., & Hester, R.E., eds.), pp. 1-54, J. Wiley & Sons, Chichester, New York. Sitaramayya, A., Harkness, J., Parkes, J. H., Gonzalez-Oliva, C., & Liebman, P. A. (1986). Kinetic studies suggest that light-activated cyclic GMP phosphodiesterase is a complex with G-protein subunits. Biochemistry 25, 651-656.

196

KLAUS PETER HOFMANN and MARTIN HECK

Sitaramayya, A., & Margulis, A. (1992). Effects of lithium on basal and modulated activities of the particulate and soluble guanylate cyclases in retinal rod outer segments. Biochemistry 31, 10652-10656. Smith, H. G., & Fager, R. S. (1991). Light-induced proton permeability changes in retinal rod photoreceptor disk membranes. Biophys. J. 59, 427-432. Steinberg, G., Ottolenghi, M., & Sheves, M. (1993) pKa of the protonated Schiff base of bovine rhodopsin. A study with artificial pigments. Biophys. J. 64, 1499-1502. Straume, M., Mitchell, D. C., Miller, J. L., & Litman, B. J. (1990). Interconversion of metarhodopsins I and II: a branched photointermediate decay model. Biochemistry 29, 9135-9142. Suh, K.-H., & Hamm, H. E. (1991). Cross-linking studies of complex formation between phosphorylated components I and II and G protein in frog rod outer segments. Invest. Ophthalmol. & Vis. Sci. 32, 1005. Swanson, R. J., & Applebury, M. L. (1983). Methylation of proteins in photoreceptor rod outer segments. J. Biol. Chem. 258, 10599-10605. Takemoto, D. J., Hurt, D., Oppert, B., & Cunnick, J. (1992). Domain mapping of the retinal cyclic GMP phosphodiesterase ),-subunit. Biochem. J. 281,637-643. Tan, E. W., & Rando, R. R. (1992). Identification of an isoprenylated cysteine methyl ester hydrolase activity in bovine rod outer segment membranes. Biochemistry 31, 5572-5578. Tang, W.-J., & Gilman, A. G. (1991). Type-specific regulation of adenylyl cyclase by G protein 13~/ subunits. Science 256, 1031-1033. Thompson, D. A., & Khorana, H. G. (1990). Guanosine 3',5'-cyclic nucleotide binding proteins of bovine retina identified by photoaftinity labeling. Proc. Natl. Acad. Sci. USA 87, 2201-2205. Thompson, P., & Findlay, J. B. C. (1984). Phosphorylation of ovine rhodopsin Identification of the phosphorylated sites. Biochem. J. 220, 773-780. Thorgeirsson, T. E., Lewis, J. W., Wallace-Williams, S. E., & Kliger, D. S. (1992). Photolysis of rhodopsin results in deprotonation of its retinal Schiff base prior to formation of metarhodopsin II. Photochem. Photobiol. 56, 1135-1144. Thorgeirsson, T. E., Lewis, J. W., Wallace-Williams S. E., & Kliger, D. S. (1993). Effects of temperature on rhodopsin photointermediates from lumirhodopsin to metarhodopsin II. Biochemistry 32, 13861-13872. Tyminski, P. N., & O'Brian, D. F. (1984). Rod outer segment phosphodiesterase binding and activation in reconstituted membranes. Biochemistry 23, 3986-3993. Uhl, R., Kuras, P. V., Anderson, K, & Abrahamson, E. W. (1980). A light scattering study on the ion permeabilities of dark-adapted bovine rod outer segment disk membranes. Biochim. Biophys. Acta 601,462-477. Uhl, R., Wagner, R., & Ryba, N. (1990). Watching G proteins at work. Trends Neurosci. 13, 64-70. Umbarger, K. O., Yamazaki, M., Hutson, L. D., Hayashi, F., & Yamazaki, A. (1992). Heterogeneity of the retinal G-protein transducin from frog rod photoreceptors---biochemical identification and characterization of new subunits. J. Biol. Chem. 267, 19494-19502. Vuong, T. M., & Chabre, M. (1991). Deactivation kinetics of the transduction cascade of vision. Proc. Natl. Acad. Sci. USA 88, 981 3-9817. Wagner, R., Ryba, N., & Uhl, R. (1988). Rapid transducin deactivation in intact stacks of bovine rod outer segment disks as studied by light scattering techniques. FEBS Lett. 235, 103-108. Wang, N., & Anderson, R. E. (1992). Enrichment of polyunsaturated fatty acids from rat retinal pigment epithelium to rod outer segments. Curr. Eye Res. 8, 783-791. Weitz, C. J., & Nathans, J. (1992). Histidine residues regulate the transition of photoexcited rhodopsin to its active conformation, metarhodopsin II. Neuron 8, 465-472 Weitz, C. J., & Nathans, J. (1993). Rhodopsin activation: Effects on the metarhodopsin I - metarhodopsin II equilibrium of neutralization or introduction of charged amino acids within putative transmembrane segments. Biochemistry 32, 14176-14182.

Light-Induced Protein-Protein Interactions

197

Wensel, T. G., & Stryer, L. (1990). Activation mechanism of retinal rod cyclic GMP phosphodiesterase probed by fluorescein-labeled inhibitory subunit. Biochemistry 29, 2155-2161. Wensel, T. G., & Stryer, L. (1986). Reciprocal control of retinal rod cyclic GMP phosphodiesterase by its T subunit and transducin. Proteins 1, 90-99. Weyand, I., & Kahn, H. (1990). Subspecies of arrestin from bovine retina. Equal functional binding to photoexcited rhodopsin but various isoelectric focusing phenotypes in individuals. Eur. J. Biochem. 193, 459-467. Whalen, M. M., Bitensky, M. W., & Takemoto, D. J. (1990). The effect of the T-subunit of the cyclic GMP phosphodiesterase of bovine and frog (Rana catesbiana) retinal rod outer segments on the kinetic parameters of the enzyme. Biochem. J. 265, 655--658. Whalen, M. M., & Bitensky, M. W. (1989). Comparison of the phosphodiesterase inhibitory subunit interactions of frog and bovine rod outer segments. Biochem. J. 259, 13-19. Wieland, T., Ulibarri, I., Aktories, K., Gierschik, P., & Jakobs, K. H. (1990). Interaction of small G-proteins with photoexcited rhodopsin. FEBS Lett. 263, 195-198. Wilden, U., Hall, S. W., & Ktihn, H. (1986). Phosphodiesterase activation by photoexcited rhodopsin is quenched when rhodopsin is phosphorylated and binds the intrinsic 84-kDa protein of rod outer segments. Proc. Natl. Acad. Sci. USA 83, 1174-1178. Willardson, B. M., Pou, B., Yoshida, T., & Bitensky, M. W. (1993). Cooperative binding of the retinal rod G-protein, transducin, to light-activated rhodopsin. J. Biol. Chem. 268, 6371-6382. Wistow, G. J., Katial, A., Craft, C., & Shinohara, T. (1986). Sequence analysis ofbovine retinal S-antigen Relationships with c~-transducin and G-proteins. FEBS Lett. 196, No. 1, 23-28. Wohlfart, P., Haase, W., Molday, R. S., & Cook, N. J. (1992). Antibodies against synthetic peptides used to determine the topology and site of glycosylation of the cGMP-gated channel from bovine rod photoreceptor. J. Biol. Chem. 267, 644--648. Wu, G., & Hubbell, W. L. (1993). Phospholipid asymmetry and transmembrane diffusion in photoreceptor disc membranes. Biochemistry 32, 879-888. Yamazaki, A. (1992). The GTP-binding protein-dependent activation and deactivation of cyclic GMP phosphodiesterase in rod photoreceptors. In: Advances in Second Messenger and Phosphoprotein Research, 25 (Strada, S.J., & Hidaka, H., eds), pp. 135-146, Raven Press, New York. Yamazaki, A., Bartucca, F., Ting, A., & Bitensky, M. W. (1982). Reciprocal effects of an inhibitory factor on catalytic activity and noncatalytic cGMP binding sites of rod phosphodiesterase. Proc. Natl. Acad. Sci. USA 79, 3702-3706. Yan, B., Nakanishi, K., & Spudich, J. L. (1991). Mechanism of activation of sensory rhodopsin I: evidence for a steric trigger. Proc. Natl. Acad. Sci. USA 88, 9412-9416. Yang, Z., & Wensel, T. G. (1992). N-myristoylation of the rod outer segment G protein, transducin, in cultured retinas. J. Biol. Chem. 267, 23197-23201. Yau, K.-W., Peng, Y. W., & Robishaw, J. D. (1992). Retinal rods and cones have different G-protein T-subunits. Invest. Ophthalmol. & Vis. Sci. 33, 1004. Yau, K.-W. (1991). Calcium and light adaptation in retinal photoreceptors. Curr. Opin. Neurobiol. 1, 252-257. Zhokovsky, E. A., Robinson, P. R., & Oprian, D. D. (1991). Transducin activation by rhodopsin without a covalent bond to the 11-cis-retinal chromophore. Science 251,558-560. Zhukovsky, E. A., Robinson, P. R., & Oprian, D. D. (1992). Changing the location of the Schiffbase counterion in rhodopsin. Biochemistry 31, 10400-10405. Zhukovsky, E. A., & Oprian, D. D. (1989). Effect of carboxylic acid side chains on the absorption maximum of visual pigments. Science 246, 928-930. Zozulya, S., & Stryer, L. (1992). Calcium-myristoyl protein switch. Proc. Natl. Acad. Sci. USA 89, 11569-11573. Zuckerman, R., & Cheasty, J. E. (1986). A 48 kDa protein arrests cGMP phosphodiesterase activation in retinal rod disk membranes. FEBS Lett. 207, 35--41.

198

KLAUS PETER HOFMANN and MARTIN HECK

Zuckerman, R., & Cheasty, J.E. (1988). Sites ofarrestin action during the quench phenomenon in retinal rods. FEBS Lett. 238, 379-384. Zvyaga, T.A., Min, K.C., Beck, M., & Sakmar, T.P. (1993). Movement of the retinylidene Schiffbase counterion in rhodopsin by one helix turn reverses the pH dependence of the metarhodopsin I to metarhodopsin 11transition. J. Biol. Chem. 268, 4661-4667.

MICROBIAL SENSORY RHODOPSINS

John L. Spudich and David N. Zacks

I. II.

III. IV.

V.

Scope of the Review .............................. Sensory Rhodopsins o f Archaea . . . . . . . . . . . . . . . . . . . . . . . . . A. Sensory Rhodopsin I (SR-I) . . . . . . . . . . . . . . . . . . . . . . . . B. Sensory Rhodopsin II (SR-II) . . . . . . . . . . . . . . . . . . . . . . . C. Pharaonis SR-II . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D. Sensory Rhodopsins of New Halobacterial Isolates . . . . . . . . . . . . Evolutionary and Mechanistic Relatedness of Microbial Rhodopsins and Visual Pigments . . . . . . . . . . . . . . . . . . . . . . . . Sensory Rhodopsins o f Unicellular Eukaryotes . . . . . . . . . . . . . . . . . A. C h l a m y d o m o n a s reinhardtii . . . . . . . . . . . . . . . . . . . . . . . . B. E u g l e n a gracilis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Spermatozopsis similis . . . . . . . . . . . . . . . . . . . . . . . . . . . Eubacterial Rhodopsin-Like Photoactivity: The Case of Photoactive Yellow Pigment . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Note Added in Proof . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Biomembranes Volume 2A, pages 199-226 Copyright © 1996 by JAI Press Inc. All rights of reproduction in any form reserved. ISBN: 1-55938-659-2

199

200 201 201 207 208 208 208 209 209 218 219 219 220 220 220 226

200

JOHN L. SPUDICH and DAVID N. ZACKS

!. SCOPE OF THE REVIEW After the elucidation of the retinal-based chromophore of bovine rod rhodopsin by George Wald in 1958, similar retinylidene proteins have been demonstrated to be ubiquitous in visual systems throughout the animal kingdom (Applebury and Hargrave, 1986). The defining characteristics of these photoreceptor proteins (or "rhodopsins," using this term in a general sense) are: the presence of 7-transmembrane helices forming an intemal pocket for the chromophore retinal (vitamin A aldehyde); the covalent attachment of retinal via a protonated Schiffbase linkage to a lysine residue in the center of the seventh (i.e. carboxyl-terminal) helix; and the interaction of the photoactivated receptor with a transducer protein (G-proteins in bovine rod outer segments) to relay the signal of retinal photoisomerization to signal transduction components. Retinal was found to be used as a chromophore in the microbial world over 20 years ago when the photoactive retinylidene pigment bacteriorhodopsin (BR) was discovered in the archaeon Halobacterium salinarium (previously known as H. halobium) (Oesterhelt and Stoeckenius, 1971). Bacteriorhodopsin and the later found halorhodopsin (HR) (Schobert and Lanyi, 1982) function as light-driven ion pumps for energy production (see Chapter 1, this volume) rather than as sensory receptors requiring interaction with transduction machinery. The first of the microbial sensory rhodopsins (now called sensory rhodopsin I (SR-I)) was detected in H. salinarium, where it functions as a phototaxis receptor controlling the cells' motility behavior (Bogomolni and Spudich, 1982). A second sensory rhodopsin (SR-II or phoborhodopsin) mediates phototaxis 1 in a different spectral range in the same organism (Takahashi et al., 1985b). Sensory rhodopsinlike pigments have also been detected in related halobacterial isolates (Otomo et al., 1992; Soppa et al., 1993), and in another archaeon, Natronobacterium pharaonis (Scharf et al., 1992b). Evidence for a retinylidene phototaxis receptor in the unicellular eukaryotic alga Chlamydomonas reinhardtii was provided in 1984 (Foster et al., 1984) extending the family of retinylidene photoreceptors for the first time to eukaryotic microbes. Suggestive evidence for retinal-containing pigments in several other unicellular eukaryotes has been gathered over the past few years. Several recent reviews have been published on the halobacterial (archaeal) receptors (Spudich and Bogomolni, 1988; Oesterhelt and Marwan, 1990; Bogomolni and Spudich, 1991). Several reviews on aspects of Chlamydomonas are available (Nultsch and H/ider, 1988; Hegemann and Harz, 1993; Witman, 1993). In this review we will begin by summarizing recent results on SR-I, the best characterized of these receptors, and its transducer protein, followed by a discussion of other archaeal receptors. We will then review Chlamydomonas and other unicellular eukaryotes, where the molecular information is much more limited, but the physiological information is rich and complex.

Microbial Sensory Rhodopsins

201

Ii. SENSORY RHODOPSINS OF ARCHAEA A. Sensory Rhodopsin I (SR-I) Physiological Function in Color-discrimination SR-I is unusual among known photosensory receptors in that it exists in two spectrally distinct forms, each of which can be photoactivated to generate sensory signals (Spudich and Bogomolni, 1984). The thermally stable species SR587 (~max= 587 nm) (Figure 1) is an attractant receptor; its photoexcitation suppresses directional reorientation (swimming reversals), thereby favoring cell migration into higher-intensity regions of orange-red light. This spectral region is beneficial to the cells since they contain 2 ion-pumping rhodopsins BR and HR which absorb maximally at 568 nm and 578 nm, respectively. Photoexcitation of SR587 produces a red-shifted species, 8610 which thermally converts into a blue-shifted species, $560,which then converts to a relatively stable form $373 (~max = 373 nm) (Figure 1). $373is the longest-living intermediate species in the SR-I photocycle, and it exists in significant concentrations in the photo-

SR 587

8610

80ms

90tJs

b

$510

t

270ps

1

$560

$373 Figure 1. Photochemical reaction cycle of SR-I. Wavy arrows indicate light reactions, other arrows indicate thermal reactions for which half-lives are shown, and subscripts indicate absorption maxima of photointermediates. Photoexcitation of SR587generates an attractant signal, while photoexcitation of S373generates a repellent signal to the flagellar motor. From Spudich and Bogomolni (1984) and Bogomolni and Spudich (1987).

202

JOHN L. SPUDICH and DAVID N. ZACKS

stationary state established by continuous 587 nm light (Figure 1). $373 is photochemically reactive and functions as a repellent receptor; its photoexcitation induces swimming reversals and migration away from regions of high UV-blue light intensity. Such higher energy light is avoided presumably because of its potential damage to protein and DNA. The dual receptor function of SR-I based on its photochromic reactions provides a mechanism for color sensing by the swimming cell (Spudich and Bogomolni, 1984). The color-discrimination mechanism of SR-I has been extensively confirmed by mutant analysis (Sundberg et al., 1985) and correlations of swimming response behavior with SR587and S373photoexcitation (Spudich and Bogomolni, 1984; Takahashi et al., 1985a; Wolff et al., 1986; Marwan and Oesterhelt, 1990; Yan and Spudich, 1991). Recently two laboratories independently confirmed the dual function of SR-I by targeted deletion of the SR-I apoprotein gene and reintroduction on a plasmid which resulted in loss and restoration, respectively, of the color-sensitive responses (Krebs et al.,1993; Ferrando-May et al., 1993a). The SR-i Transducer Phototaxis mutant analysis identified a methylated membrane protein with an apparent molecular mass of 97 kDa, expression of which tightly correlated with the SR-I protein of 25 kDa (Spudich et al., 1988). The methylation linkage (reversible carboxylmethyl esterification (Spudich et al., 1988; Alam et al., 1989)) of the 97-kDa M r protein is of the same type as that found in the chemotaxis signal transducers of eubacteria (e.g., Escherichia coli), chemoreceptive proteins that transmit signals from the membrane to a cytoplasmic sensory pathway (reviewed by Stock et al., 1990; Bourret et al., 1991). SR-I attractant and repellent signals modulate methyl group turnover in vivo, as do chemoattractant and repellent eubacterial chemotaxis transducers (Alam et al., 1989; Spudich et al., 1989). Based on these findings the methyl-accepting protein was postulated to be the postreceptor transducer for SR-I signals (Spudich et al., 1989). Analysis of taxis mutants and revertants (Sundberg et al., 1985, 1990; Spudich et al., 1988, 1989) and antigenic crossreactivity of the 97-kDa M r protein with eubacterial transducers (Alam and Hazelbauer, 1991) further strengthened the analogy to the chemotaxis system. Partial sequence of the proposed transducer protein was obtained, allowing the identification and cloning of its gene (Yao and Spudich, 1992). The predicted amino acid sequence yields a size of 57 kDa for the methylated protein indicating an aberrant electrophoretic migration on SDS/polyacrylamide gels, as occurs with other acidic halophilic proteins. Analysis of the encoded primary structure (discussed below) and results from expression studies confirmed the role of this protein as a transducer of SR-I signals and hence it was named HtrI (halobacterial transducer for sensory rhodopsin/).

Microbial SensoryRhodopsins

203

Organization of the Receptor and Transducer Genes The genes encoding the apoprotein ("opsin") of SR-I and HtrI, sopI (Blanck et al., 1989) and htrI (Yao and Spudich, 1992), respectively, are located in tandem in an operon-like arrangement on the H. salinarium chromosome. The sopI gene initiator codon overlaps the termination codon of htrI and the htrI-sopI pair is preceded by a putative promoter region (Yao and Spudich, 1992) and the pair is expressed as a single transcriptional unit (Ferrando-May et al., 1993b). When the two proteins are expressed in H. salinarium from a plasmid carrying the putative promoter region and the gene pair they restore phototaxis in a mutant containing a deletion in the htrI-sopI region (Yao and Spudich, 1992; Ferrando-May et al., 1993b). Evidence from H. salinarium phototaxis and chemotaxis mutant studies (Sundberg et al., 1985; 1990; Spudich et al., 1988; 1989) and the close analogies with other bacterial taxis transduction systems strongly suggest that HtrI interacts with cytoplasmic signal transduction components that are shared with the transducers of SR-II and chemotaxis receptors in the cell (Spudich et al., 1989). Hence SR-I and HtrI appear to be the only components specific to SR-I signals. Integration of stimuli through SR-I, SR-II and chemotaxis receptors (Spudich and Stoeckenius, 1979) therefore is expected to occur in post-HtrI steps in the pathway.

Structure of SR-i and Its Transducer 2-D folding models of SR-I (Blanck et al., 1988) and HtrI (Yao and Spudich, 1992) have been developed from hydropathy analysis of their gene-predicted amino acid sequences. An approximate tertiary structure of SR-I has been derived from comparison of its sequence with that of BR for which near-atomic resolution structural information has been obtained from electron diffraction (Henderson et al., 1990). Several lines of evidence support the use of the BR structure as a first approximation to that of SR-I, especially in the retinal-binding cavity. Hydropathy analysis predicts a 7-transmembrane helical structure for SR-I, and although SR-I exhibits only 26% amino acid sequence identity with that of BR, it is >80% identical in the putative retinal-binding pocket (Blanck et al., 1989). Activation of SR-I, like that of BR, requires all-trans/13-cis-retinal isomerization (Yan et al., 1990). The retinal-binding pockets of BR and SR-I are closely similar in their electrostatic and hydrophobic interactions with the chromophore, as shown by the similar spectroscopic shifts of the BR and SR-I analog pigments generated with synthetic retinal analogs (Yan et al., 1991a). In the SR-I model, the protein folds into 7-transmembrane helical segments which form an internal cavity in which all-trans retinal is attached via a protonated Schiffbase linkage (Fodor et al., 1989) to the e-amino group of Lys205 (Figure 2). In native membranes BR (Muccio and Cassim, 1979) and HR (Hasselbacher et al., 1988) exhibit a biphasic ellipticity in their main circular dichroism (CD) bands, which has been interpreted to derive from their trimeric organization in the

204

JOHN L. SPUDICH and DAVID N. ZACKS

A

I I I

B

ty,

s,, ! IG~. Arg =11 il

I I teu

teu IAla II

t,p.

I vo, "

P~! I Ala

Gly I

I vail

VoI 4s

AI

GI

A

teu1 ~

Gly

I

. tea lyr

<~'~'+1,,

.. IIAla" lhr I Iv,"

Ala

Asp

Met I

I

le

Arg Arg

r Val ~=

Set

tie

lie •

I Tyr Ala

1

Val

"

Is', Asp

lie Gly

Met

G

Pro Asp

Val

Pro Glu

I Li

IGIy I ~I

VoI

I

";;e Gin

GIAS ny

/hr Asn VoI lie Val

G!y tys VoI AIo l, teu /al t ,~a J Thr lhr Gly Asp

AlaThr V~ lu Mel Val

Phe

ly, I

teu Ala

Tyr

~, i I L/

Asn tl"

i Gly

teu

T_vr v°' Ty'II

I!

'le His l~ro

I I

A laL.~.J L,

[eu

I i teun~~ LPreo~

GIr I I

Pro Pro Ala Ser Pro AlaAI a Glu Glu Asp Thr Ser Gin Ala His

Leu

I Leo Girl Set ,3.

vo..,vo,,,ol,,. <JI

Pro t

!1 t,u I < Ilvo, + II I I

F

Ivpi~"~"~ t

Ala A/ao,, ,,° Ser

Gly

,2

Val Ala

Ira,

I Pr9"~/

Val I I

/k,"

tr,

I I

t,u"

I

I vo,"° I!

Gly

r,o

!1

E

Ar~

Se,

At,

D

Val Val

Asp Gly Set teu I His P~o

Leu I I

!I i

C

~. Lei~]~VolPh Val

teu~

P:o c,,

Tyr

Vat

Ala leu

AIo Gly

lhr

1~7

c,j,

Ala

AIo

lie Gly Glu Ala

lhr

Figure 2. Two-dimensional folding model of SR-I. The large boxes represent the seven a-helical transmembrane segments (A-G). The retinal-binding lysine 205 (square) and other residues suggested to be in the retinal binding pocket (diamonds) are based on homology with BR. Redrawn from Henderson et al. (1990); sequence from Blanck et al. (1989).

membrane. SR-I lacks this feature in its CD spectrum (Hasselbacher et al., 1988) indicating that it differs in its organization in the membrane from BR and HR. The deduced protein sequence of HtrI contains 536 amino acid residues and predicts two transmembrane helices near the N terminal that would anchor the protein to the membrane. Beyond this hydrophobic region of 46 residues, the remainder of the protein is hydrophilic. The C-terminal 270 residues contain a region homologous to the signaling domains of eubacterial chemotaxis transducers (e.g., Eseheriehia call Tsr protein), flanked by two regions homologous to the methylation domains of this transducer family. The protein differs from E. call Tsr in that it does not have an extramembranous-receptor binding domain, and that it has a more extended cytoplasmic region. Chromophore cross-linking measurements suggested physical proximity of SR-I and HtrI (Spudich et al., 1988). Compelling evidence for SR-I~trI interaction has been obtained in recent expression studies which demonstrate HtrI alters SR-I photochemical reactions in native membranes and the accessibility of the retinylidene Schiff base to the extramem-

205

Microbial Sensory Rhodopsins

branous environment (Spudich and Spudich, 1993; Olson and Spudich, 1993). Our current view of the structure of the SR-I/HtrI complex is depicted in Figure 3. It is not known whether the transmembrane or cytoplasmic regions of HtrI or both contain the interaction sites with the receptor protein. Recently we have expressed a form of HtrI deleted of residues 248-500, which contain the signaling and methylation domains (Yao et al., 1994). This shortened transducer does not signal to the flagellar motor, as expected, but it does modulate the photoreactions of the receptor as described for full length transducer (Spudich and Spudich, 1993). Existence of the signaling/methylation motif in HtrI extends the family of eubacterial signal transducers to the archaea, indicating an early origin for this type of signal transducer. In eubacterial chemotaxis, the signaling domain controls a two-component regulatory system that is a member of a widespread family of cytoplasmic phosphotransfer systems. Such a cytoplasmic transduction system has

-

Htr I

MEMBRANE

7

CYTOPLASM

J

METHYLATION L~

C

REGIONS

SIGNALING DOMAIN

Figure 3. Models for SR-I and Htrl disposition in the membrane. Interaction of SR-I and Htrl may occur within the membrane and/or between their extramembranous domains. Reproduced with permission from Spudich (1993).

206

JOHN L. SPUDICH and DAVID N. ZACKS

not been demonstrated in archaea, but the existence of a similar signaling domain on HtrI strongly suggests that this system occurs. In this regard, phenotypes of phototaxis and chemotaxis mutants ofH. salinarium exhibit similarities to mutants in E. coli chemotaxis (Sundberg et al., 1985, 1990; Spudich et a1.,.1988, 1989). Transducer-control of Proton Transfers in SR-I SR-I is structurally similar to the electrogenic proton pump BR, and like BR, SR-I appears to deprotonate and later reprotonate the Schiff base nitrogen during its photocycle. Nevertheless, early studies showed its photoreactions do not translocate protons nor result in membrane hyperpolarization (Spudich and Spudich, 1982; Bogomolni and Spudich, 1982; Ehrlich et al., 1985). These studies used functionally active SR-I, that is, SR-I complexed with its transducer HtrI. The recent availability of transducer-free SR-I and SR-I containing mutations in ionizable residues near the protonated Schiff base has led to a series of studies which reveal a fundamental similarity between SR-I and BR despite their different functions (Spudich and Spudich, 1993; Olson and Spudich, 1993; Rath et al., 1994; Bogomolni et al., 1994). The picture emerging from UV-Vis kinetic absorption spectroscopy, FTIR, and pH and membrane potential probes is as follows: Photoreactions of transducer-free SR-I do vectorially translocate protons across the membrane in the same direction as BR. In SR-I this pumping is suppressed by interaction with transducer which diverts the proton movements into an electroneutral path. A key step in this diversion is strongly suggested by the data: transducer interaction appears to raise the pK a of the aspartyl residue in SR-I (D76) which corresponds to the primary proton-accepting residue in the BR pump (D85). The main observations supporting this interpretation are: 1. Removal of transducer alters the photochemical reactions of the receptor so that the thermal decay of $373, which requires reprotonation of the retinal attachment site in the photoactive center of the protein, becomes highly pH-sensitive (reprotonation proceeds more rapidly in proportion to the proton concentration; Spudich and Spudich, 1993). This observation was the first to indicate a proton path from the medium to the SR-I photoactive center. 2. In transducer-flee SR-I, formation and decay of $373 are accompanied by release and uptake, respectively, of protons in the medium (Olson and Spudich, 1993). Transducer binding blocks this proton exchange (Olson et al., 1992; Olson and Spudich, 1993). 3. The proton release and uptake is vectorial (from the cytoplasmic to the extracellular side of the membrane) if a residue of pK a = 7.2 is deprotonated; transducer interaction raises the pK a of this residue more than 1 pH unit (Bogomolni et al., 1994). 4. Fourier transform infrared spectroscopy shows D76, the aspartyl residue in SR-I in a corresponding position as the aspartyl proton-acceptor near the BR chromophore critical for BR proton ejection, is neutral in transducer complexed

Microbial SensoryRhodopsins

207

SR-I (Rash et al., 1994). The properties of D76N indicate D76 is the residue with pK a = 7.2 in (3) (Bogomolni et al., 1994). 5. The latent proton-translocation capacity of SR-I may reflect the evolution of the SR-I sensory signaling mechanism from the proton pumping mechanism of BR. An attractive hypothesis is that the electroneutral proton path in transducer-complexed SR-I is important to activation of the transducer.

B. Sensory Rhodopsin Ii (SR-II) A second sensory rhodopsin (SR-II or phoborhodopsin) in H. salinarium was detected by flash photolysis and behavioral analysis (Takahashi et al., 1985b). SR-I, BR, and HR are not produced during exponential phase aerobic growth. When oxygen and respiratory substances are ample, the cells avoid sunlight by synthesizing SR-II, a repellent receptor, as their only rhodopsin (Tomioka et al., 1986b). SR-II absorbs maximally at 487 nm, the energy peak of the solar spectrum at the earth's surface (Takahashi et al., 1990). Hence, its wavelength sensitivity has evolved strategically to be maximally sensitive for seeking the dark. When the cells become anaerobic, cellular content of SR-II is reduced and biosynthesis of BR and HR is induced, enabling light to be used as an energy source. Production of SR-I is also induced enabling the cells to migrate into illuminated regions where the ion pumps will be maximally activated. The absorption spectrum of the dark-adapted state of SR-II, SR-II487, exhibits vibrational fine structure, an unusual feature for retinylidene proteins (Takahashi et al., 1990; Imamoto et al., 1991). Retinal/apoprotein interactions responsible for this feature (primarily due to retinal ring/chain coplanarity) and other absorption properties of SR-II487 have been analyzed with retinal analogs (Takahashi et al., 1990). Retinal isomerization in SR-II has been characterized by chromophore extractions and analog studies and the receptor appears to undergo all-trans to 13 cis photoisomerization like the other archaeal rhodopsins (Yan et al., 1990; Scharf et al., 1992a). Photoreaction cycle imermediates of SR-II have been trapped at cryogenic temperatures and a sequence of thermal states obtained by stepwise warming of the sample. The technique was used to identify early intermediates in the SR-II photocycle (Schichida et al., 1988; Imamoto et al., 1991). At room temperature later intermediates have been resolved (Tomioka et al., 1986a; Wolffet al., 1986, Spudich et al., 1986; Scherrer et al., 1987). Like that of BR and SR-I, the SR-II photocycle at room temperature exhibits intermediates with K-, L- and M-like absorption, but no L-like intermediate is evident by low temperature spectrophotometry. An O-like intermediate is also observed as in BR. Kinetic analysis of flash photolysis transients by Scharf et al. (1992b) reveals complexity in the early parts of the SR-II photocycle which leads to a model with a branch and back reactions. An important property of the SR-II photocycle is that, unlike the ion pumps BR and HR, it contains intermediates with long (102103 msec) lifetimes. The long-lived interme-

208

JOHN L. SPUDICH and DAVID N. ZACKS

diates of SR-II, like that of SR-I, appear to be the signaling states (i.e., signal-generating conformations) of the receptor (Yan et al., 199 lb). The SR-II protein has nearly the same molecular mass as SR-I, electrophoretically migrating slightly faster (23,000 Mr) than SR-I (25,000 Mr) in highly resolving gels in a radiolabeling study (Spudich et al., 1986). A similar M r has been observed in independent radiolabeling experiments (Scherrer et al., 1987; Scharf et al., 1992a). Photoactivated SR-II modulates the methylation/demethylation system in wildtype and in a mutant which lacks HtrI, suggesting a similar methyl-accepting protein ("HtrlI") exists for SR-II signal transduction (Spudich et al., 1989). Further evidence for such a protein is that an SR-II-deficient mutant lacks a protein present in its SR-II ÷ parent shown to cross-react with a broad specificity anti-methyl-accepting-transducer antibody (Alam and Hazelbauer, 1991).

C. Pharaonis SR-II An SR-II-like pigment (pSR-II) was identified by flash photolysis of membranes from the haloalkalophile Natronobacterium pharaonis (Bivin and Stoeckenius, 1986). Although N. pharaonis and H. salinarium are only distantly related, biochemical and photochemical properties of pSR-II and SR-II are very similar, including the presence of fine structure in the pSR-II absorption spectrum (Scharf et al., 1992b; Hirayama et al., 1992). Repellent phototaxis responses measured in iV. pharaonis exhibit a wavelength-dependence consistent with pSR-II function as the receptor (Scharf et al., 1992b). Chromophore extraction (Imamoto et al., 1992) shows that like SR-I! (Takahashi et al., 1990; Scharf et al., 1992a) pSR-II contains an all-trans retinal configuration in the dark and photoisomerizes to 13-cis as do also SR-I, BR, and HR.

D. Sensory Rhodopsins of New Halobacterial Isolates Otomo et al. (1992) examined halobacterial strains isolated from crude solar salts from Mexico and Australia. SR-I and SR-II-like photocycling activities were detected by flash photolysis. Three isolates exhibited flash-induced absorbance changes characteristic of the SR587---~$373transition and very slow $373 decay (> 15-fold retarded compared to H. salinariurn SR-I). Also a gene 71% identical to sopI was recently identified in a Red Sea isolate Halobacterium sp. strain SGI (Soppa et al., 1993). Studies of the growing family of sensory rhodopsin variants can be expected to provide information regarding their evolution and structure/function relationships.

I!!. EVOLUTIONARY AND MECHANISTIC RELATEDNESS OF MICROBIAL RHODOPSINS AND VISUAL PIGMENTS The evolutionary relationships among archaeal rhodopsins, microbial eukaryotic rhodopsins, and higher animal visual pigments may be difficult to determine given

Microbial SensoryRhodopsins

209

the extreme phylogenetic differences. Sequence comparisons between the archaeal family and visual pigments (and related 7-transmembrane receptors) reveal no obvious primary structure homology (Henderson and Schertler, 1990; Soppa et al., 1993), however the secondary and tertiary structural motifs are quite similar and may be indicative of a progenitor protein. Several protein structure/function relationships in the two best characterized sensory retinylidene proteins, SR-I and bovine rhodopsin, are remarkably similar (compare for example Ganter et al., 1989 and Yan et al., 199 l a; Hofmann, 1995 and Spudich and Spudich, 1993). If the mechanistic similarities derive from convergent evolution they evidently reflect fundamental constraints on the design of retinal-containing sensory receptors. This is especially significant when the nonsensory 2 retinylidene proteins BR and HR differ in a particular property from SR-I and rhodopsin. In all of these proteins retinal binds in a protonated Schiffbase linkage to a lysine residue placed in the middle of the C-terminal (7th) transmembrane helix. Although the photoisomerization occurs across different double bonds in SR-I and rhodopsin (13-trans to 13-cis and 11-cis to 11-trans, respectively), retinal/apoprotein interactions in the sensory proteins are similar in that: (1) unlike in BR thermal isomerization of retinal is restricted, that is, only one configuration of the isomerizing double bond is tolerated in the pocket in the nonphotoactivated protein in the dark (13-trans in SR-I and 11-cis in rhodopsin); and (2) unlike in BR, formation of functional photoproducts requires steric interactions between retinal polyene chain methyl groups and protein residues; specifically apoprotein interaction with a methyl group near the photo-isomerizing double bond (the 13-methyl in SR-I and the 9-methyl in rhodopsin); and (3) formation of the signaling conformations of SR-I and rhodopsin ($373 and Metarhodopsin-II38 o, respectively) each require Schiffbase deprotonation, and, moreover both couple with membrane anchored transducers (HtrI and transducin) which influence the Schiffbase protontransfer reactions. Further understanding of similarities and differences between the SR-I/HtrI and rhodopsin/transducin signal relay mechanisms may well reveal fundamental principles in the design of retinylidene receptors and more generally, of 7-transmembrane receptors.

IV. SENSORY RHODOPSINS OF UNICELLULAR EUKARYOTES A. Chlamydomonasreinhardtii Introduction C. reinhardtii is a motile, unicellular alga that exhibits two behavioral responses to light which allow the cells to access photosynthetically optimal environments. One, phototaxis, is the orientation of the cells' swimming direction along the axis of a light beam. 1 The second light induced behavioral change is called the photo-

210

JOHN L. SPUDICH and DAVID N. ZACKS

phobic response (a.k.a. "stop" or "shock" response) and consists of a brief cessation in forward swimming on temporal changes in light intensity. Early photobiologists and microscopists qualitatively described these phenomena (Jennings, 1915; Buder, 1919), but did not define the receptors controlling these behaviors, and the field of Chlamydomonas phototaxis progressed slowly for almost half a century. Then, in the late 1960s interest renewed, with laboratories investigating the physiology of these photoresponses, primarily the calcium dependence of photoinduced behavioral changes. This literature reviewed by Feinleib (1980) and Nultsch (1983). The nature of the receptor for phototaxis was finally defined in 1984 by Foster and coworkers who provided evidence for the existence of a Chlamydomonas rhodopsin. This set off another wave of research, this time concentrating on the physical characteristics of the photoreceptor itself. Currently, we are at the threshold of understanding the receptor and transduction events involved in Chlamydomonas reinhardtii responses to light. Harz et al. (1992) and Witman (1993) review the phototransduction cascade. In this section we will review the physical properties of the photoreceptors controlling these two behavioral responses, the behavioral relationship between the two responses, and clarify several points which have caused a certain amount of confusion in the field.

Receptor Identification The first clues imo the nature of the photoreceptor came in 1971 when Nultsch and coworkers measured a phototaxis action spectrum from fluence response curves (responses to different intensities, or doses, of light) obtained at different wavelengths. Movements of cell populations were monitored in a "phototaxigraph" which records the cells' absorption of non-actinic, infrared light in two parts of a cuvette. Directional actinic light causes the cells to accumulate towards either end of the cuvette. This asymmetric cell distribution between the different parts of the cuvette results in a differential absorbance of the monitoring beams that is proportional to the cells' phototaxis response. The action spectrum thus obtained showed a major peak at 503 nm with a secondary peak at 443 nm. The shape of this spectrum was consistent with the presence ofa retinylidene photoreceptor, but did not exclude the possibility of a ravin and/or pterin based pigment. Foster and Smyth (1980), in their review of algal ultrastructure as it relates to phototaxis, suggested a new approach to measuring action spectra based on the threshold for receptor activation, rather than the more traditional finite-response or criterion-response action spectra used by Nultsch's group. The latter type of action spectra is based on the light intensity, at various wavelengths, required to elicit a criterion response. The threshold action spectrum, however, extrapolates the minimum light intensity at each wavelength required to elicit a just discemable response. The advantage of this technique is that the threshold intensity is a more accurate measure of the photoreceptor absorption, eliminating complications presented by the absorbance of screening pigments present in the cell, particularly the eyespot. They recalculated the action spectrum for phototaxis by extrapolating the fluence

Microbial Sensory Rhodopsins

211

response curves of Nultsch et al. (1971) to zero light intensity, and by doing so eliminated the minor peak at 443 nm observed in the finite-criterion action spectrum. This new threshold action spectrum fit well the absorption spectrum of bovine rhodopsin, suggesting a retinal-based photoreceptor for phototaxis. A test of the hypothesis that Chlamydomonas contains a rhodopsin came from the same group of Foster and coworkers (1984) using the general pigment-deficient mutant strain FN68 that lacks both the phototaxis and photophobic responses. The cells' ability to undergo phototaxis was shown to be restored upon the addition of retinal to the cell suspension, and the phototaxis threshold action spectrum maxima shifted depending on the isomer or analog used to reconstitute the response. These two facts: that the cells' maximum wavelength sensitivity depends on the analog used, and that the maxima are red shifted compared to pure retinal's absorption maximum, are consistent with the characteristics of all other known rhodopsins (reviewed by Balogh-Nair and Nakanishi, 1990), strongly indicating that the retinal was reconstituting a photoreceptor, and not effecting the response by some secondary mechanism. Further evidence supporting the existence of a rhodopsin in Chlamydomonas came from the hydroxylamine bleaching experiments of Hegemann et al. (1988). Hydroxylamine is known to remove the retinal moiety from archaeal and eukaryotic rhodopsins. Hegemann and coworkers showed that incubation of either wild-type cells or retinal-reconstituted FN68 cells with hydroxylamine followed by exposure to light resulted in a decrease ofphotosensitivity. This reduction in sensitivity could be reversed by washing out the hydroxylamine, and in the case of the FN68 cells, also adding back retinal, suggesting that the hydroxylamine only affected the interaction of the chromophore with the opsin. These findings are consistent with hydroxylamine bleaching experiments done in other rhodopsins (Oesterhelt et al., 1974). The experiments of Foster and coworkers described earlier used a petri-dish population migration assay to measure and quantify the phototaxis response. In this assay, actinic light is delivered from the side onto a uniform distribution of cells spread over a petri dish. If the cells accumulate towards either end of the dish, then they produce a clear zone. The degree of cleating on the side facing the light minus the degree of clearing on the side opposite the light is taken as an index of phototaxis. This index taken over time gives the phototactic rate of the population. This assay provides an easy method for determining general phototaxis capabilities of cell populations, but cannot provide quantitative measurements of cell motility or individual cell responses. These limitations must be considered when evaluating the conclusions of these papers regarding the chromophore properties discussed in the next section.

ChromophoreProperties In Foster and coworkers' experiments (1984), restoration of the phototaxis response sensitivity with the different natural isomers of retinal: 9-cis, 11-cis, and

212

JOHN L. SPUDICH and DAVID N. ZACKS

all-trans was maximal with 11-cis retinal. The suggestion was made that the native chromophore for the phototaxis photoreceptor in Chlamydomonas is 11-cis retinal, similar to the chromophore for visual rhodopsin (Wald, 1968) but unlike the archaeal rhodopsins which have all-trans retinal as their chromophore (Ottolenghi and Sheves, 1989). In 1987, Foster and coworkers reported evidence from reconstitution studies with analogs that the retinal chromophore is in a planar conformation across the C6-C7 single bond which links the retinal's polyene chain to its ionone ring. This is the conformation of the archaeal rhodopsins' chromophores (Bogomolni and Spudich, 1991), but not of the visual rhodopsins' chromophores (Birge, 1990). The primary transduction event in known retinylidene proteins is the isomerization of the chromophore across one of the double bonds in the polyene chain. In visual rhodopsins 11-cis retinal isomerizes to all-trans retinal (Birge, 1990) and in the archaeal rhodopsins all-trans retinal isomerizes to 13-cis retinal (Bogomolni and Spudich, 1991). Foster et al. (1988b, 1989) incorporated over 20 analogs of retinal and reported that they all restored phototaxis in Chlamydomonas reinhardtii, as measured with the petri-dish assay. They concluded that no retinal isomerization was necessary since analogs which were prevented from isomerizing across any or all of the chromophore's double bonds restored phototaxis. Even hexanal--which does not have any carbon-carbon double bonds----restored light sensitivity. The authors reported shifted action spectra peaks, indicating that the analogs were binding to the opsin apoprotein and forming a functional photoreceptor. These findings were extended in later work from the same group (Nakanishi et al., 1988), where they put forth the hypothesis that charge redistribution and not isomerization was the mechanism of light energy transduction in the phototaxis photoreceptor of C. reinhardtii (this hypothesis is discussed in greater detail in Foster et al., 1991). The hypothesis and the evidence supporting it are in direct contrast to the dogma in the field of visual science that chromophore isomerization is necessary for photosensitivity, and indicated, therefore, a novel mechanism of photosensory transduction. However, there are several arguments based on further information and more highly resolving phototaxis assays which suggest reevaluation of this conclusion. The mutant strain FN68 exhibits a property which suggests an alternative explanation for some of the results. Light exposure alone, without the addition of exogenous retinal, restores the phototaxis response (Foster et al., 1988a). The mechanism by which this effect occurs is unknown, but it is not inhibited by the protein synthesis inhibitors cycloheximide and/or chloramphenicol (Foster et al., 1988a). Foster's group concluded that light itself induces the synthesis of retinal, and that a retinylidene protein is the photoreceptor responsible for the auto-induction of photosensitivity since the action spectrum of the light induced photosensitivity shifted to the same extent as that of phototaxis when a retinal analog was added. The petri-dish phototaxis assay used in all these experiments exposes the cells to 10 minutes of light, an amount of time sufficient to restore considerable phototaxis sensitivity in the cells (Beckmann and Hegemann, 1991). Thus, the possibility

Microbial Sensory Rhodopsins

213

arises that the non-isomerizing analogs found to restore phototaxis actually serve as the chromophore for photoinduction of phototaxis, and the observed shifts in phototaxis action spectra maxima actually correspond to the photoinduction receptor. An alternative mechanism of energy transduction that does not involve chromophore isomerization may occur with the photoinduction phenomenon (an interesting possibility in and of itself) but isomerization may still be required for photobehavior. In the mutant strain FN68 both the photoinduction of photosensitivity and photobehavior are restored by retinal addition, making independent analysis of each photoresponse more difficult. Recently, several groups have reexamined the chromophore properties of the photoreceptors using assays with higher resolution than that provided by the petri-dish test. One technique employed is a computerized videomicrographic motion analysis system for monitoring photophobic responses in Chlamydomonas. In this assay individual cells are monitored by infrared, dark field microscopy and their swimming behavior is digitally recorded and analyzed as a function of time by computer. This type of assay was first developed for monitoring light induced reversals in the archaeon H. salinarium (Takahashi and Kobatake, 1982; Sundberg et al., 1986), and applied to the description of Chlamydomonas photophobic responses by Hegemann and Bruck (1989). The primary advantage this type of assay has over a population assay is that the behavioral responses of individual cells to light stimuli are measured. Using a computerized system allows for the analysis of motility and swimming parameters of large numbers of cells, thus providing more accurate population behavioral data as a function of individual cell motion. Lawson et al. (1991) used computerized cell tracking and motion analysis to study the photophobic responses in a second pigment-deficient mutant CC-2359. This mutant was shown to be insensitive to photophobic stimuli over a wide range of light intensities, and did not show regeneration of sensitivity upon light exposure. The photophobic stimuli used were 5 to 20 msec pulses of 500 nm light, as opposed to 10 min exposure periods used in the petri-dish assay, greatly reducing possible light effects on the retinal isomers and analogs themselves. The group tested the ability of various retinal isomers and analogs to reconstitute the photophobic response. Their experiments showed that all-trans retinal reconstituted the photophobic response most efficiently, followed, in decreasing order of efficacy, by the 13-cis, 11-cis and 9-cis isomers. This indicated that the polyene chain configuration of the natural chromophore for the photophobic receptor is all-trans, in contrast to the conclusions of Foster et al. (1984) for phototaxis. An analog with the C6-C7 bond "locked" in a trans configuration was able to reconstitute photosensitivity, whereas an analog with the same bond "locked" in a cis configuration was not. These data strongly suggested that the polyene chain is planar and in a trans conformation with respect to the ionone ring. As mentioned before, a central dogma of visual science is that signal transduction by rhodopsins is initiated by the retinal chromophore's isomerization on light stimulation. Lawson et al. (1991) tested the ability of 13-trans locked and 13-cis

214

JOHN L. SPUDICH and DAVID N. ZACKS

locked retinal analogs to reconstitute photophobic responses in CC-2359. Neither of these analogs were able to restore sensitivity to photophobic stimuli, and both these analogs competitively inhibited all-trans retinal from reconstituting the response. Inhibition occurred regardless of whether the locked retinal analog had already been added to the cells and their response reconstituted, or whether the analogs were preincubated with the cells before retinal addition. These experiments solidified the idea that the retinal was binding to a receptor protein, in the same manner that a ligand binds to receptor, and suggested that the retinal/apoprotein interaction (presumably including a Schiff base linkage) is more dynamic than previously thought for retinylidene proteins. Dissociation and inhibition constants were determined for the all-trans retinal, 13-trans locked retinal, and the 13-cis locked retinal of 2.5 x 10-11 M, 5.2 x 10-:° M, and 5.4 x 10-9 M, respectively. The authors state, however., that these numbers represent only apparent in vivo dissociation and inhibition constants, and that these values may well be scaled by some factor, but that the relative order of affinities should not be affected. The two major conclusions from this workman all-trans chromophore, and 13-14 double bond photoisomerization--distinctly differ from the conclusions of Foster and coworkers, as described above, and suggested that the chromophore of the photophobic response receptor is more similar to the chromophores of the archaeal rhodopsins than to the chromophores of other known eukaryotic rhodopsins. These two points have been independently confirmed by other groups. Hegemann et al. (1991) reconstituted the same strain CC-2359 with several different retinal isomers and analogs, and monitored phototaxis in a taxigraph (similar in principle to Nultsch's apparatus described above) and in a light scattering apparatus. The light scattering assay is an indirect method for examining whole cell population responses to light stimuli, which correlates well with the photophobic and phototaxis responses, depending on the stimulus intensity (Uhl and Hegemann, 1990). Their results also show greatest reconstitution efficiency with the all-trans isomer of retinal, followed by the 13-cis, 11-cis, 9-cis and 7-cis isomers. Analogs lacking the ionone ring, but with at least 3 conjugated double bonds in the polyene chain also restored photosensitivity, but at much lower efficiencies, suggesting a "minimal" polyene chain chromophore structure capable of reconstituting phototaxis. Analog molecules prevented from isomerizing around any of the polyene chain double bonds, by the introduction of a phenyl ring into the polyene system, do not reconstitute responses, providing further support for the role ofisomerization in phototransduction. Hegemann et al. (1991) also investigated the role of each of the methyl groups at the C9 and C 13 positions using analogs lacking them. Chromophores lacking the C9 methyl group were capable of reconstituting photosensitivity, but not chromophores lacking the C 13 methyl group. Bovine rhodopsin requires the C9 methyl group for activation (Ganter et al., 1989), whereas SR-I requires the C 13 methyl group (Yan et al., 1991a). Hence, the C 13 methyl requirement further indicates the chromophore environment of the Chlamydomonas photoreceptors more closely resembles the archaeal chromophore environment than that of visual pigments.

Microbial Sensory Rhodopsins

215

Takahashi et al. ( 1991) developed a motion analysis system for the simultaneous study of photophobic and phototaxis responses. Their reconstitution experiments also showed that both responses were more efficiently restored by all-trans retinal than by 11-cis retinal, in strain CC-2359, and that analogs prevented from isomerizing around the C 13-C 14 double bond were ineffective in reconstituting photosensitivity. Further work by this group (Takahashi et al., 1992) used analogs known to slow the photocycle of the archaeal SR-I and SR-II, and found that they caused an inversion of the sign of phototaxis in strain FN68. That is to say, FN68 cells, which are normally only negatively phototactic when reconstituted with all-trans retinal, became only positively phototactic when these analogs were incorporated into the photoreceptor. This is the first evidence, albeit indirect, that there is a photocycle in the photoreceptor, and that the period of this cycle may be working in conjunction with the cell's rotation and eyespot shading to control the direction of phototaxis. If one of these factors is shifted out of phase, then the sign of the response changes accordingly. Chromophore extraction experiments (Derguini et al., 1991; Beckmann and Hegemann, 1991) have shown the presence of all-trans retinal but not of 11-cis retinal. These studies together provide very convincing evidence that the photophobic and phototaxis receptors contain chromophores that are in an all-trans configuration, 6-s-trans conformation, and that isomerization across the C13-C14 double bond is required for behavioral responses to occur. The differences between these conclusions and those made by Foster's group have been usually explained away as due to strain differences, different types of assays employed, and different responses examined. To help resolve these discrepancies direct comparison of the chromophore properties of the photophobic and phototaxis receptors in strain FN68 was undertaken (Zacks et al., 1993). Analogs and isomers of retinal were compared in their ability to reconstitute both photoresponses in the same population of cells. Behavior was analyzed with a computerized videomicroscopic motion analysis system similar to the one used by Lawson et al. (1991) for photophobic response analysis and modified to also allow a direct examination of large number of individual cells' phototaxis responses. Both photoresponses are most effectively reconstituted by all-trans retinal, followed by the other isomers in the same order as described in Lawson et al. ( 1991) and Hegemann et al. (1991). Neither response in this strain is reconstituted by the analogs prevented from C13-C14 cis/trans isomerization, consistent with the results in strain CC-2359. A difference between the chromophore properties of the two responses appears, however, in the inhibition of all-trans retinal reconstitution by the nonisomerizable 13 trans-locked analog. The photophobic response reconstitution is inhibited, as is the case in strain CC-2359 (Lawson et al., 1991). The phototaxis response however, does not decrease even at very high concentrations of this analog. A second difference found between the two responses is in their desensitization to repetitive light stimuli. Cells become desensitized to photophobic stimuli after repetitive flashes of light (Hegemann and Bruck, 1989) or after several seconds exposure to continuous light (Zacks et al., 1993), whereas the phototaxis response remains fully sensitive.

216

JOHN L. SPUDICH and DAVID N. ZACKS

Relationship Between the Photophobic and Phototaxis Responses High speed cinematographic analysis of Chlamydomonas swimming behavior shows that as the cells swim they rotate around their longitudinal axis forming a helical path (Kamiya and Witman, 1984; Ruffer and Nultsch, 1985). The cell's eyespot region, which is thought to contain the photoreceptors for phototaxis (Melkonian and Robenek, 1980; Foster and Smyth, 1980), normally faces the outside of the helix (Kamiya and Witman, 1984; Ruffer and Nultsch, 1985). Ruffer and Nultsch (1991) concluded that the helical swimming pattern is important for phototaxis orientation because this type of movement generates periodic illumination and shading of the photoreceptor region by the eyespot. Their conclusion was strongly supported by their demonstration that increases and decreases in light intensity cause asymmetric changes in the two flagella's beat pattems which allow the cell to orient relative to a light beam. It is not known whether phototaxis and photophobic responses are mediated by the same photoreceptor. One possibility is that each response has its own receptor and transduction mechanism. The function of the photophobic desensitization in this case would be simply to prevent photophobic reactions from interfering with phototaxis. A second possibility, however, is that the two light induced behaviors are controlled by the same photoreceptor. This would be possible if the flagellar beat changes which cause photophobic orientation are the residual of the photophobic responses after desensitization (i.e., "miniphotophobic" reactions which cause brief reorienting motions rather than full stops). Photophobic desensitization, in this case, would need to cause a precise grading of the photoreceptor output to yield at different light intensities the asymmetric flagellar beat patterns observed by Ruffer and Nultsch (1991) instead of full-fledged stop responses. Supporting the notion of a graded output, a continuum in the response to photophobic stimuli was observed by Lawson et al. ( 1991) who reported that low intensity flashes cause cell tuming, whereas higher intensity flashes generate stops. Zacks and Spudich (1994) report three specific experimental observations which support the graded output model, which are: (1) The desensitization of the photophobic response obeys Weber's Law (i.e., at each light intensity (I) in the cells' dynamic range responses to photophobic stimuli are desensitized to a point where the relative contrast (AI/I) generated by the eyespot yields a constam photophobic signal regardless of the signal expected from the absolute contrast (AI) at the eyespot; (2) Cells exhibiting faster rates of desensitization of photophobic responses have a faster onset of phototaxis; and (3) Cells orienting to a light beam exhibit a loss of orientation when the light is suddenly shifted to a lower intensity, due to the more highly desensitized state of the cells relative to the lower intensity, and the consequent deviation of the relative contrast from its constant value. These data are consistent with there being only one receptor for both photobehaviors and with the photophobic response providing the elemental course corrections causing phototaxis orientation.

Microbial SensoryRhodopsins

217

Pigment Localization Ultrastructural examination of the Chlamydomonas cells has suggested an intimate relationship between the eyespot structure and the photoreceptor. The eyespot has been prolSosed by Foster and Smyth (1980) to function as a quarter-wave plate interference reflector. In their model, the eyespot serves to reflect and concentrate light onto a plane, in which the receptor is presumably located, parallel to the eyespot surface. Recently, Kreimer and Melkonian (1990) have tested the reflection hypothesis in different algal species using laser confocal scanning microscopy, and confirm increased reflection intensity caused by the eyespot. This group has extended their study (Kreimer et al., 1992) to the eyespot reflectance in wild-type and eyespot deficient mutants of C. reinhardtii, again confirming the eyespot's role as a reflector. Electron microscopy of the FN68 and CC-2359 mutants reveals an eyespot lacking pigment granules in the former and no detectable eyespot structure in the latter (M. Lawson and E Satir, personal communication). Since both strains exhibit photophobic and phototaxis responses, a fully developed eyespot structure is not essential for either behavior, although it probably increases photodetection efficiency and contrast. In C. reinhardtii the outer chloroplast membrane and the cell membrane are positioned in very close proximity to the eyespot and overlay the structure in parallel planes. According to the theory of Foster and Smyth (1980) the eyespot increases the light intensity at these membranes. Melkonian and Robenek (1980) observed structural specializations in these membranes, and suggested that they may represent the photoreceptors and/or transduction apparatus involved in phototaxis. In vivo microspectrophotometric measurements of this region in wild-type cells show an absorbance peak at 500 nm with a bandwidth that fits standard absorption spectra of rhodopsins (Crescitelli et al., 1992).

Pigment Enrichment In vitro studies on Chlamydomonas photoreceptors have proven difficult due to the lack of convenient enrichment procedures. To date, only two reports on pigment enrichment have been published. The first report, at a meeting poster session (Starace and Foster, 1989), describes reconstitution of light bleachable eyespot absorbance at 498 nm after addition of 11-cis or 9-cis retinal but not after addition ofall-trans retinal. Beckmann and Hegemann (1991) report.in detail the enrichment of membrane fractions from wild-type cells that show absorbance at 495 nm, with a shoulder at 465 nm. This absorbance is bleached by light, and restored by addition of all-trans retinal. The shoulder may derive from fine structure of the pigment spectrum, as is the case of the archaeal SR-II (Takahashi et al., 1990), or may be due to the presence of multiple pigments or pigment species. Radioactive retinal labeling of this fraction showed only one labeled band on an SDS-PAGE gel, with an apparent molecular mass of 32,000 kDa.

218

JOHN L. SPUDICH and DAVID N. ZACKS

A genetic approach toward rhodopsin identification has been reported by Martin et al. (1986), where bovine opsin and Drosophila opsin cDNA probes were used to label C. reinhardtii genomic DNA. Several fragments of DNA were labelled, but rhodopsin-encoding genes in these fragments have not been reported.

B. Euglena gracilis Euglena gracilis is another motile, unicellular green alga that exhibits phototaxis and photophobic responses. The swimming motion of this organism is different than that of Chlamydomonas, because of a different flagellar architecture, and this results in a different manifestation of the light induced responses. Temporal changes in light intensity result in a type of tumbling behavior that decays much more slowly than the stop response of Chlamydomonas. One consequence of this response is that when cells move across a light dark boundary a ttmable is induced, causing them to accumulate in the area. The receptors responsible for these photoresponses have been the focus of several studies, and much of the evidence points to a flavin and/or pterin chromophore (Doughty and Diehn, 1980; Lenci et al., 1983; Ghetti et al., 1985, Schmidt et al., 1990). Recently, however, there have been some indications that a rhodopsin may be present in Euglena, and may serve as the photoreceptor for at least one of these photoresponses. Evidence for a retinylidene photoreceptor first appeared in 1989 when Gualtieri and coworkers reported a technique for isolating the paraflagellar swelling of Euglena gracilis. The absorption spectrum of these swellings taken with a microspectrophotometer showed a maximum at 500 nm with a bandwidth of approximately 80 nm. The absorbance is distinct from the absorbance of the eyespots and of the chloroplasts, and the authors suggested that this absorbance could derive from a rhodopsin molecule. Further work from this group (Barsanti et al., 1992) showed the elimination of paraflagellar swellings when the cells were grown up in medium containing nicotine. Nicotine is known to inhibit the biosynthetic pathway of carotenoids. Nicotine-treated Euglena cells did not photoaccumulate in an illuminated region of a petri dish, suggesting that perhaps the nicotine inhibited the production of a carotenoid chromophore for the photoreceptor. Carotenoids have been known to exist in Euglena for many years (Krinsky and Goldsmith, 1960). Batra and Tollin (1964) developed a method for enriching eyespot fractions from Euglena, and extracted the carotenoids lutein, cryptoxanthin, and 13-carotene. Suspensions of these eyespot fractions showed absorption maxima at 490,462, and 436 nm, which corresponded to the action spectrum profile for phototaxis. The suggestion was made that perhaps the eyespots contain the photoreceptor, but no further evidence was available to support this conjecture. In 1992, Gualtieri and coworkers extracted all-trans retinal from Euglena cells, and suggested that this chromophore derives from a rhodopsin-like photoreceptor. Unfortunately, retinal-deficient mutants of Euglena are not available, precluding the type of retinal reconstitution experiments that were done with Chlamydomonas.

Microbgal Sensory Rhodopsins

219

C. Spermatozopsis similis A third green alga, Spermatozops& simil&, has been studied spectroscopically, and evidence was found for a retinylidene pigment. Kreimer et al. (1991a,b) reported a technique for the enrichment ofeyespots and associated membranes. The absorption spectrum of this enriched sample is complex, due to the variety of carotenoids present in the eyespot, but a rapidly photobleachable peak at approximately 550 nm is resolved by difference absorption spectroscopy. Absorbance difference spectra between bleached versus unbleached samples show a shape characteristic of rhodopsins, with a peak at approximately 460 nm. Reconstitution of a bleached sample with 9-c& retinal yielded a species with a maximal absorbance at 540 nm. Both bleaching and reconstitution spectra taken over time form single isosbestic points at 400 and 480 nm, respectively. Isolated, non-bleached eyespot fractions contain extractable 11-c& and all-trans retinal, suggesting that retinal is the chromophore for this photoreceptor. These data strongly suggest the presence of a retinylidene protein, but further work needs to be done to correlate these findings to behavioral responses in the organism.

V. EUBACTERIAL RHODOPSIN-LIKE PHOTOACTIVITY: THE CASE OF PHOTOACTIVE YELLOW PIGMENT So far we have seen that retinylidene proteins exist in the archaea and in eukaryotes, both multi- and unicellular, but what about eubacteria? Retinylidene 7-transmembrane receptors have not been demonstrated in eubacteria, but a pigment implicated in phototaxis reception by the halophilic, phototropic eubacterium Ectothiorhodospira halophila exhibits an SR-I-like photoreaction. A search for soluble cytochromes and ferrodoxins in E. halophila uncovered the presence of large quantities ofa photoactive yellow pigment (PYP) (Meyer, 1985). This pigment is a protein of 15.3 kDa, with a maximal absorbance at 446 nm and an extinction coefficient of 48,000 M-lcm-1. PYP exhibits photocycling, with kinetics similar to sensory rhodopsin I ofH. salinarium (Meyer et al., 1987, 1989, 1991). The function of this pigment in photosensory reception has not yet been proven, but recent work on the motility behavior of this organism shows a repellent response wavelength-dependence with a peak matching the absorbance maximum of PYP (Sprenger et al., 1993). McRee and coworkers (1986, 1989) reported the crystallization and X-ray diffraction of PYP. The structure is totally different from that of rhodopsins (Henderson et al., 1990), and is characterized by a "13-clam" structure. This structure consists of two layers ofl3-pleated sheets, forming a pocket into which the chromophore inserts, as opposed to the 7 c~-helical structure of rhodopsins, but similar to the structure of known soluble carotenoid and lipid binding proteins. The exact chromophore species was not determined, but is believed to be similar to all-trans retinal (McRee et al., 1989), but probably not identical (Hoff, Van Beeumer, and Hellingwerf, unpublished results, as cited in Sprenger et al., 1993). Although PYP differs greatly in structure from rhodopsins,

220

JOHN L. SPUDICH and DAVID N. ZACKS

the structural basis o f its SR-I-like p h o t o c y l e m a y p r o v i d e insight into the p h o t o t r a n s f o r m a t i o n s in the r h o d o p s i n family.

ACKNOWLEDGMENTS The authors were supported by National Institutes of Health grant GM27750 (JLS), National Science Foundation grant MCB-9219854 (JLS), and NIH training grant T32 BM07288 from the National Institute of General Medical Sciences (DNZ).

NOTES 1. In keeping with common usage the term"phototaxis" is used in two different senses in this review. For discussion of archaeal behavior the term is used in its general sense as any motility behavior in response to light. For discussion of algae where a finer distinction is important, we use the more restrictive definition ofphototaxis as orientation of swimming direction with respect to a beam of light. 2. Photoexcitation of BR has been shown to cause an attractant response.in energy-depleted cells (Bibikov et al., 1991, 1993: Yan et al., 1992). This appears to be a secondary consequence of membrane electrical potential changes due to proton pumping rather than to primary sensory signaling, such as that mediated by sensory rhodopsins.

REFERENCES Alam, M., & Hazelbauer, G. L. (1991). Structural features of methyl-accepting taxis proteins conserved between archaebacteria and eubacteria revealed by antigenic cross-reaction. J. Bact. 173, 58375842. Alam, M., Lebert, M., Oesterhelt, D., & Hazelbauer, G. L. (1989). Methyl-accepting taxis proteins in Halobacterium halobium. EMBO J. 8, 631-639. Applebury, M. L., & Hargrave, P. A. (1986). Molecular biology of the visual pigments. Vision Res. 26, 1881-1895. Balogh-Nair, V., & Nakanishi, K. (1990). Visual pigment and bacteriorhodopsin analogs. In Chemistry and Biology of Synthetic Retinoids. Eds: Dawson, M., & Okamura, W., pp. 147-176. CRC Press, Boca Raton, Florida. Barsanti, L., Passarelli, V., Lenzi, P., & Gualtieri, P. (1992). Elimination of photoreceptor (paraflagellar swelling) and photoreception in Euglena gracilis by means of the carotenoid biosynthesis inhibitor nicotine. J. Photochem. Photobiol. 13, 135-144. Batra, P., & Tollin, G. (1964). Phototaxis in Euglena. I. Isolation of the eyespot granules and identification of the eyespot pigments. Biochim. Biophys. Acta. 79, 371-378. Beckmann, M., & Hegemann, P. (1991). In vitro identification of rhodopsin in the green alga Chlamydomonas. Biochemistry 30, 3692-3697. Bibikov, S. I., Gfishanin, R. N., Kaulen, A. D., Marwan, W., Oesterhelt, D., & Skulachez, V. P. (1993). Bacteriorhodopsin is involved in halobacterial photoreception. Proc. Natl. Acad. Sci. USA 90, 9446-9450. Bibikov, S. I., Grishanin, R. N., Marwan, W., Oesterhelt, D., & Skulachev, V. P. (1991). The proton pump bacteriorhodopsin is a photoreceptor for signal transduction in Halobacterium halobium. FEBS Lett. 295, 223-226. Birge, R. (1990). Nature of the primary photochemical events in rhodopsin and bacteriorhodopsin. Biochim. Biophys. Acta 1016, 293-327. Bivin, D. B., & Stoeckenius, W. (1986). Photoactive retinal pigments in haloalkaliphilic bacteria. J. Gen. Micrbiol. 132, 2167-2177.

Microbial SensoryRhodopsins

221

Blanck, A., Oesterhelt, D., Ferrando, E., Schegk, E. S., & Lottspeich, E (1989). Primary structure of sensory rhodopsin I, a prokaryotic photoreceptor. EMBO J. 8, 3963-3971. Bogomolni, R. A., & Spudich, J. L. (1982). Identification of a third rhodopsin-like pigment in phototactic Halobacterium halobium. Proc. Natl. Acad. Sci. USA 79, 6250-6254. Bogomolni, R. A., & Spudich, J. L. (1991). Archaebacterial rhodopsins: Sensory and energy transducing membrane proteins. In Sensory Receptors and Signal Transduction. Modem Cell Biology. Vol. 10. Ed. J. L. Spudich and B. Satir. pp. 227-249. Bogomolni, R. A., Stoeckenius, W., Szundi, I., Perozo, E., Olson, K. D., & Spudich, J. L. (1994). Removal of transducer HtrI allows electrogenic proton translocation by sensory rhodopsin I. Proc. Natl. Acad. Sci. USA 91, 10188-10192. Bourret, R. B., Borkovich, K. A., & Simon, M. I. (1991). Signal transduction pathways involving protein phosphorylation in prokaryotes. Annu. Rev. Biochem. 60, 401-441. Buder, J. (1919). Zur kennmis der phototaktischen Richtungsbewegungen. Jahrb. Wiss. Bot. 58, 102-220. Crescitelli, F., James, T. W., Erikson, J. M., Loew, E. R., & McFarland, W.N. (1992). The eyespot of Chlamydomonas reinhardtii: A comparative microspectrophotometric study. Vision Res. 32(9), 1593-1600. Derguini, F., Mazer, P., & Nakanishi, K. (1991). All-trans retinal is the chromophore bound to the rhodopsin of the alga Chlamydomonas reinhardtii. Photochem. Photobiol. 54(6), 1017-1021. Doughty, M., & Diehn, B. (1980). Flavins as photoreceptor pigments for behavioral responses in motile microorganisms, especially in flagellated alga Euglena sp. Struct. Bond. 41, 45-70. Ehrlich, B. E., Schen, C. R., & Spudich, J. L. (1984). Bacterial rhodopsins monitored with fluorescent dyes in vesicles and in vivo. J. Memb. Biol. 82, 89-94. Feinleib, M. E. (1980). Photomotile responses in flagellates. In Phoreception and Sensory Transduction in Aneural Organisms. Ed. F. Lenci & G. Colombetti, pp. 45-68. Plenum Press. New York. Ferrando-May, E., Brustmann, B., & Oesterhelt, D. (1993a). A C-terminal truncation results in high-level expression of the functional photoreceptor sensory rhodopsin I in the archaeon Halobacterium salinarium. Mol. Microbiol. 9, 943-953. Ferrando-May, E., Krah, M., Marwan, W., & Oesterhelt, D. (1993b). The methyl-accepting transducer protein HtrI is functionally associated with the photoreceptor sensory rhodopsin I in the archaeon Halobacterium salinarium. The EMBO Journal 12, 2999-3005. Fodor, S. E A., Gebhard, R., Lugtenburg, J., Bogomolni, R. A., & Mathies, R. A. (1989). Structure of the retinal chromophore in sensory rhodopsin I from resonance Raman spectroscopy. J. Biol. Chem. 264, 18280-18283. Foster, K., & Smyth, R. (1980). Light antennas in phototactic algae. Microbiol. Rev. 44, 572-630. Foster, K., Saranak, J., Patel, N., Zarrilli, G., Okabe, M., Kline, T., & Nakanishi, K. (1984). A rhodopsin is the functional photoreceptor for phototaxis in the unicellular eukaryote Chlamydomonas. Nature. 311,756-759. Foster, K., Saranak, J., van der Steen, R., & Lugtenburg, J. (1987). Retinal in Chlamydomonas rhodopsin is in a planar 6-s-trans conformation as shown by in vivo incorporation of 6-7 locked retinal analogs. Invest. Optham. Vis. Sci. 28(supplement), $253. Foster, K., Saranak, J., & Zarrilli, G. (1988a). Autoregulation ofrhodopsin synthesis in Chlamydomonas reinhardtii. Proc. Natl. Acad. Sci. USA 85, 6379-6383. Foster, K., Saranak, J., Derguini, E, Rao, J., Zarrilli, G., Okabe, M., Fang, J., Shimizu, N., & Nakanishi, K. (1988b). Rhodopsin activation: A novel view suggested by in vivo Chlamydomonas experiments. J. Amer. Chem. Soc. 110, 6588. Foster, K., Saranak, J., Derguini, F., Zarrilli, G., Johnson, R., Okabe, M., & Nakanishi, K. (1989). Activation of Chlamydomonas rhodopsin in vivo does not require isomerization of retinal. Biochemistry 28, 819-824. Foster, K., Saranak, J., & Dowben, P. (1991). Spectral sensitivity, structure and activation of eukaryotic rhodopsins: activation spectroscopy of rhodopsin analogs in Chlamydomonas. J. Photochem. Photobiol. 8, 385-408.

222

JOHN L. SPUDICH and DAVID N. ZACKS

Ganter, U. N., Schmid, E. D., Perez-Sala, D., Rando, R. R., & Seibert, E (1989). Removal ofthe 9-methyl group of retinal inhibits signal transduction in the visual process. A Fourier transform infrared and biochemical investigation. Biochemistry 28, 5954-5962. Ghetti, F., Colombetti, G., Lenci, E, Campani, E., Polacco, E., & Quaglia, M. (1985). Fluorescence of Euglena gracilis photoreceptor pigment: An in vivo microspectrophotometric study. Photochem. Photobiol. 42(1), 29-33. Gualtieri, P., Barsanti, L., & Passarelli, V. (1989). Absorption spectrum of a single isolated paraflagellar swelling of Euglena gracilis. Biochim. Biophys. Acta. 993,293-296. Gualtieri, P., Pelosi, P., Passarelli, V., & Barsanti, L. (1992). Identification of a rhodopsin photoreceptor in Euglena gracilis. Biochim. Biophys Acta. 1117(1), 55-59. Harz, H., Nonnengasser, C., & Hegemann, P. (1992). The photoreceptor current of the green alga Chlamydomonas. Phil. Trans. Royal Soc.: Biol. Sci.(B). 338(1238), 39-52. Hasselbacher, C. A., Spudich, J. L., & Dewey, T. G. (1988). Circular dichroism of halorhodopsin: Comparison with bacteriorhodopsin and sensory rhodopsin I. Biochemistry 27, 2540-2546. Hegemann, E, & Bruck, B. (1989). Light-induced response in Chlamydomonas reinhardtii: Occurrence and adaptation phenomena. Cell Mot. Cytoskel. 14, 501-515. Hegemann, P., Gartner, W., & Uhl, R. (1991). All-trans retinal constitutes the functional chromophore in Chlamydomonas rhodopsin. Biophys. J. 60, 1477-1489. Hegemann, P., & Harz, H. (1993). Photoreception in Chlamydomonas. In Signal Transduction, Prokaryotic and Simple Eukaryotic Systems. J. Kurjan & Talor, B. L., Eds. Academic Press: New York, 1993, pp. 279-307. Hegemann, E, Hegemann, U., & Foster, K. (1988). Reversible bleaching of Chlamydomonas reinhardtii rhodopsin in vivo. Photochem. Photobiol. 48, 123-128. Henderson, R., Baldwin, J. M., Ceska, T. A., Zemlin, E, Beckmann, E., & Downing, K. H. (1990). Model for the structure ofbacteriorhodopsin based on high-resolution electron cryo-microscopy. J. Mol. Biol. 213,899-929. Henderson, R., & Schertler, G. (1990). The structure ofbacteriorhodopsin and its relevance for the visual opsins and other seven-helix G-protein coupled receptors. Philos. Trans. Roy. Soc. London 326, 379-389. Hirayamal J., Imamoto, Y., Shichida, Y., Kamo, N., Tomioka, H., & Yoshizawa, T. (1992). Photocycle of phoborhodopsin from haloalkaliphilic bacterium (Natronobacterium pharaonis) studied by low-temperature spectrophotometry. Biochemistry 31, 2093-2098. Hofmann, K. P., & Heck, M. (1996)-Chapter 3. Imamoto, Y., Shichida, Y., Hirayama, J., Tomioka, H., Kamo, N., & Yoshizawa, T. (1992). Chromophore configuration ofpharaonis phoborhodopsin and its isomerization on photon absorption. Biochemistry 31, 2523-2528. Imamoto, Y., Schichida, Y., Yoshizawa, T., Tomioka, H., Takahashi, T., Fujikawa, K., Kamo, N., & Kobatake, Y. (1991). Photoreaction cycle ofphoborhodopsin studied by low-temperature spectrophotometry. Biochemistry. 30, 7416-7424. Jennings, H. S. (1915). Behavior of the Lower Organisms. Columbia Univ. Press. New York. Kamiya, R., & Witman, G. B. (1984). Submicromolar levels of calcium control the balance of beating between the two flagella in demembranated models of Chlamydomonas. J. Cell Biol. 98, 97-107. Krebs, M. P., Spudich, E. N., Khorana, H. G., & Spudich, J. L. (1993). Synthesis ofa gene for sensory rhodopsin I and its functional expression in Halobacterium halobium. Proc. Natl. Acad. Sci. USA 90, 3486-3490. Kreimer, G., & Melkonian, M. (1990). Reflection confocal laser scanning microscopy of eyespots in flagellated green algae. Eur. J. Cell Biol. 53, 101-111. Kreimer, G., Brohsonn, U., & Melkonian, M. (1991a). Isolation and partial characterization of the photoreceptive organelle for phototaxis of a flagellate green alga. Eur. J. Cell Biol. 55, 318-327. Kreimer, G., Mamer, F., Brohsonn, U., & Melkonian, M. (1991b). Identification of 11-cis and all-trans retinal in the photoreceptive organelle of a flagellate green alga. FEBS. Lett. 293(1, 2), 49-52.

Microbial SensoryRhodopsins

223

Kreimer, G., Overlander, C., Sineshchekov, O. A., Stolzis, H., Nultsch, W., & Melkonian, M. (1992). Functional analysis of the eyespot in Chlamydomonas reinhardtii mutant ey627, mt. Planta 188, 513--521. Krinsky, N., & Goldsmith, T. (1960). The carotenoids of the flagellated alga, Euglena gracilis. Arch. Biochem. Biophys. 91, 271-279. Lawson, M., Zacks, D. N., Derguini, F., Nakanishi, K., & Spudich, J. L. (1991). Retinal analog restoration ofphotophobic responses in a blind Chlamydomonas reinhardtii mutant: Evidence for an archaebacterial-like chromophore in a eukaryotic rhodopsin. Biophys. J. 601,490-1498. Lenci, F., Colombetti, G., & H/ider, D. (1983). Role of flavin quenchers and inhibitors in the sensory transduction of the negative phototaxis in the flagellate, Euglena gracilis. Curr. Microbiol. 9, 285-290. Martin, R. L., Wood, C., Baehr, W., & Applebury, M. L. (1986). Visual pigment homologies revealed by DNA hybridization. Science 232, 1266-1269. Marwan, W., & Oesterhelt, D. (1990). Quantitation of photochromism of sensory rhodopsin-I by computerized tracking of Halobacterium halobium cells. J. Mol. Biol. 215, 277-285. Marwan, W., & Oesterhelt, D. (1987). Signal formation in the halobacterial photophobic response mediated by a fourth retinal protein (P480). J. Mol. Biol. 195, 333--342. Melkonian, M., & Robenek, M. (1980). Eyespot membranes of Chlamydomonas reinhardtii: A freezefracture study. J. Ultrastruct. Res. 72, 90-102. Meyer, T. E. (1985). Isolation and characterization of soluble cytochromes, ferredoxins and other chromophoric proteins from the halophilic phototrophic bacterium Ectothiorhodospira halophila. Biochim. Biophys. Acta 806, 175-183. Meyer, T. E., Tollin, G., Hazzard, J. H., & Cusanovich, M. A. (1989). Photoactive yellow protein from the purple phototrophic bacterium, Ectothiorhodospira halophila. Quantum yield ofphotobleaching and effects of temperature, alcohols, glycerol, and sucrose on kinetics of photobleaching and recovery. Biophys. J. 56, 559-564. Meyer, T. E., Yakali, E., Cusanovich, M. A., & Tollin, G. (1987). Properties of a water-soluble, yellow protein isolated from a halophilic phototrophic bacterium that has photochemical activity analogous to sensory rhodopsin. Biochemistry 26, 418--423. Meyer, T. E., Tollin, G., Causgrove, T. P., Cheng, P., & Blankenship, R. E. (1991). Picosecond decay kinetics and quantum yield of fluorescence of the photoactive yellow protein from the halophilic purple phototrophic bacterium, Ectothiorhodospira halophila. Biophys. J. 59, 988-991. McRee, D. E., Meyers, T. E., Cusanovich, M. A., Parge, H. E., & Getzoff, E. D. (1986). Crystallographic characterization ofa photoactive yellow protein with photochemistry similar to sensory rhodopsin. J. Biol. Chem. 261(29), 13850-13851. McRee, D. E., Tainer, J. A., Meyer, T. E., van Beeumen, J., Cusanovich, M. A., & Getzoff, E. D. (1989). Crystallographic structure of a photoreceptor protein at 2.4 A resolution. Proc. Natl. Acad. Sci. USA 86, 6533-6537. Muccio, D., & Cassim, J. Y. (1979). Interpretation of the absorption and circular dichroic spectra of oriented purple membrane. Biophys. J. 26, 427-440. Nakanishi, K., Derguini, F., Rao, J., Zarrilli, G., Okabe, M., Lien, T., Johnson, R., Foster, K., & Saranak, J. (1989). Theory of rhodopsin activation: Probable charge redistribution of excited state chromophore. Pure and Appl. Chem. 61, 361-364. Nultsch, W. (1983). The photocontrol of movement of Chlamydomonas. In The Biology of Photoreception. Ed. D. Cosens and D. Vince-Prue. Soc. Exp. Biol. Symp. XXXVI. pp. 521-539. Nultsch, W., & H/ider, D.-P. (1988). Photomovement in motile microorganisms-II. Photochem. Photobiol. 47, 837-869. Nultsch, W., Throm, G., & Rimscha, I. (1971). Phototaktische Untersuchungen an Chlamydomonas reinhardtii dangeard in homokontinuierlicher Kultur. Arch. Mikrobiol. 80, 351-369. Oesterhelt, D., & Marwan, W. (1990). Signal transduction in Halobacterium halobium. Symp. Soc. Gen. Microbiol. 46, 21 9-239.

224

JOHN L. SPUDICH and DAVID N. ZACKS

Oesterhelt, D., Schuhmann, L., & Gruber, H. (1974). Light-dependent reaction of bacteriorhodopsin with hydroxylamine in cell suspensions of Halobacterium halobium: Demonstration of an apo-membrane. FEBS. Lett. 44(3), 257-261. Oesterhelt D., & Stoeckenius, W. (1971). Rhodopsin-like protein from the purple membrane of Halobacteium halobium. Nature[New Biol.] 233, 149-152. Olson, K. D., Deval, P., & Spudich, J. L. (1992). Absorption and photochemistry of sensory rhodopsi.n-I: pH effects. Photochem. Photobiol. 56, 1181-1187. Olson, K. D., & Spudich, J. L. (1993). Removal of the transducer protein from sensory rhodopsin I exposes sites of proton release and uptake during the receptor photocycle. Biophysical. J. 65, 2578-2586. Otomo, J., Tomioka, H., & Sasabe, H. (1992). Bacterial rhodopsins of newly isolated halobacteria. J. Gen. Microbiol. 138, 1027-1037. Ottolenghi, M., & Sheves, M. (1989). Synthetic retinals as probes for the binding site and photoreactions in rhodopsins. J. Membrane Biol. 112, 193-212. Rath, P., Olson, K. D., Spudich, J. L., & Rothschild, K. J. (1994). The Schiff base counterion in bacteriorhodopsin is protonated in sensory rhodopsin I: Spectroscopic and functional characterization of the mutated proteins D76N and D76A. Biochemistry 33, 5600-5606. Riaffer, U., & Nultsch, W. (1985). High-speed cinematographic analysis of the movement of Chlamydomonas. Cell. Motil. 5, 251-263. Riffler, U., & Nultsch, W. (1991). Flagellar photoresponses of Chlamydomonas cells held on micropipettes II: change in flagellar beat pattern. Cell Motil. Cytoskeleton 18, 269-278. Scharf, B.. Hess, B., & Engelhard, M. (1992a). Chromophore of sensory rhodopsin II from Halobacterium halobium. Biochemistry 31, 12486-12492. Scharf, B., Pevec, B., Hess, B., & Engelhard, M. (1992b). Biochemical and photochemical properties of the photophobic receptors from Halobacterium halobium and Natronobacterium pharaonis. Eur. J. Biochem. 206, 359-366. Scherrer, P., McGinnis, K., & Bogomolni, R. A. (1987). Biochemical and spectroscopic characterization of the blue-green photoreceptor in Halobacterium halobium. Proc. Natl. Acad. Sci. USA 84, 402-406. Schmidt, W., Galland, P., Senger, H., & Furuya, M. (1990). Microspectro-photometry of Euglena gracilis. Planta 182, 375-381. Schobert, B., & Lanyi, J. K. (1982). Halorhodopsin is a light driven chloride pump. J. Biol. Chem. 257, 10306-10313. Soppa, J., Duschl, J., & Oesterhelt, D. (1993). Bacterioopsin, haloopsin, sensory opsin I of the halobacterial isolate Halobacterium sp. strain SG 1: three new members of a growing family. J. Bact. 175, 2720-2726. Sprenger, W. W., Hoff, W. D., Armitage, J. P., & Hellingwerf, K. J. (1993). The eubacteriumEctothiorhodospira halophila is negatively phototactic, with a wavelength dependence that fits the absorption spectrum of the photoactive yellow protein. J. Bact. 175, 3096-3104. Spudich, E. N., Hasselbacher, C. A., & Spudich, J. L. (1988). A methyl-accepting protein associated with bacterial sensory rhodopsin I. J. Bact. 170, 4280-4285. Spudich, E. N., Takahashi, T., & Spudich, J. L. (1989). Sensory rhodopsins I and II modulate a methylation/demethylation system in Halobacterium halobium phototaxis. Proc. Natl. Acad. Sci. USA 20, 7746--7750. Spudich, E. N., & Spudich, J. L. (1982). Control of transmembrane ion fluxes to select halorhodopsindeficient and other energy transduction mutants of Halobacterium halobium. Proc. Natl. Acad. Sci. USA 79, 4308-4312. Spudich, E. N., & Spudich, J. L. (1993). The photochemical reactions of sensory rhodopsin I are altered by its transducer. J. Biol. Chem. 268, 16095-16097. Spudich, J. L. (1993). Color sensing in the Archaea: a eukaryotic-like receptor coupled to a prokaryotic transducer. J. Bact. 175, 7755-7761.

Microbial Sensory Rhodopsins

225

Spudich, J. L., & Bogomolni, R. A. (1984). The mechanism of colour discrimination by a bacterial sensory rhodopsin. Nature 312, 509--513. Spudich, J. L., & Bogomolni, R. A. (1988). Sensory rhodopsins of halobacteria. Annu. Rev. Biophys. Chem. 17, 193--215. Spudich, J. L., & Stoeckenius, W. (1979). Photosensory and chemosensory behavior of Halobacterium halobium. J. Photobiochem. Photobiophys. 1, 43-53. Spudich, J. L., McCain, D. A., Nakanishi, K., Okabe, M., Shimizu, N., Rodman, H., Honig, B., & Bogomolni, R. A. (1986). Chromophore/protein interaction in bacterial sensory rhodopsin and bacteriorhodopsin. Biophys. J. 49, 479-483. Starace, D., & Foster. K. (1989). Absorption spectroscopy of the rhodopsin of Chlamydomonas reinhardtii. Biophys. J. 55, 379a. Stock, J. B., Stock, A. M., & Mottonen, J. M. (1990). Signal transduction in bacteria. Nature 344, 395-400. Sundberg, S. A., Alam, M., Lebert, M., Spudich, J. L., 0esterhelt, D., & Hazelbauer, G. L. (1990). Characterization of mutants of Halobacterium halobium defective in taxis. J. Bacteriology 172, 2328-2335. Sundberg, S., Alam, M., & Spudich, J. L. (1986). Excitation signal processing times in Halobacterium halobilum phototaxis. Biophys. J. 50, 895-900. Sundberg, S.A., Bogomolni, R. A., & Spudich, J. L. (1985). Selection and properties of phototaxis-deficient mutants of Halobacterium halobium. J. Bact. 164, 282-287. Takahashi, T., & Kobatake, Y. (1982). Computer-linked automated method for measurement of the reversal frequency in phototaxis of Halobacterium halobium. Cell Struct. Funct. 7, 183-192. Takahashi, T., Mochizuki, Y., Kamo, N., & Kobatake, Y. (1985a). Evidence that the long-lifetime photointermediate of S-Rhodopsin is a receptor for negative phototaxis in Halobacterium halobium. Biochem. Biophys. Res. Commun. 127, 99-105. Takahashi, T., Tomioka H., Kamo, N., & Kobatake, Y. (1985b). A photosystem other than PS370 also mediates the negative phototaxis of Halobacterium halobium. FEMS Microbiol. Lett. 28, 161164. Takahashi, T., Yan, B., Mazur, P., Derguini, F., Nakanishi, K., & Spudich, J.L. (1990). Color regulation in the archaebacterial phototaxis receptor phoborhodopsin (sensory rhodopsin II). Biochemistry 29, 8467--8474. Takahashi, T., Yoshihara, K., Watanabe, M., Kubota, M., Johnson, R., Derguini, F., & Nakanishi, K. (1991). Photoisomerization of retinal at 13-ene is important for phototaxis of Chlamydomonas reinhardtii: Simultaneous measurements of phototactic and photophobic responses. Biochem. Biophys. Biol. Res. Comm. 178(3), 1273-1279. Takahashi, T., Kubota, M., Watanabe, M., Yoshihara, K., Derguini, F., & Nakanishi, K. (1992). Diversion of the sign ofphototaxis in a Chlamydomonas reinhardtii mutant incorporated with retinal and its analogs. FEBS Lett. 314(3), 275-279. Tomioka, H., Takahashi, T., Kamo, N., & Kobatake, Y. (1986a). Flash spectrophotometric identification of a fourth rhodopsin-like pigment in Halobacterium halobium. Biochem. Biophys. Res. Commun. 139, 389-395. Tomioka H., Takahashi, T., Kamo, N., & Kobatake, Y. (1986b). Action spectrum of the photoattractant response of Halobacterium halobium in early logarithmic growth phase and the role of sensory rhodopsin. Biochim. Biophys. Acta. 884, 578-584. Uhl, R., & Hegemann, P. (1990). Probing visual transduction in a plant cell. Optical recordings of rhodopsin-induced structural changes from Chlamydomonas reinhardtii. Biophys. J. 58, 1295-1302. Wald, G. (1968). The molecular basis of visual excitation. Nature 219, 800--807. Witman, G. B. (1993). Chlamydomonas phototaxis. Trends Cell Biol. 3,403--408. Wolff, E. K., Bogomolni, R. A., Scherrer, P., Hess, B., & Stoeckenius, W. (1986). Color discrimination in halobacteria: Spectroscopic characterization of a second sensory receptor coveting the bluegreen region of the spectrum. Proc. Natl. Acad. Sci. USA 83, 7272-7276.

226

JOHN L. SPUDICH and DAVID N. ZACKS

Yan, B., Cline, S. W., Doolittle, W. E, & Spudich, J. L. (1992). Transformation ofa BOP-HOP-SOP-I -SOP-IF Halobacterium halobium mutant to BOP+: Effectsofbacteriorhodopsin photoactivation on cellular proton fluxes and swimmingbehavior. Photochem. Photobiol. 56, 553-561. Yan, B., Nakanishi, K., & Spudich, J. L. (199l a). Mechanism of activation of sensory rhodopsin-I: Evidence for a steric trigger. Proc. Natl. Acad. Sci. USA 88, 9412-9416. Yan, B., Takahashi, T., & Spudich, J. L. (1991b). Identification of signaling states of a sensory receptor by modulation of lifetimes of stimulus-induced conformations: The case of sensory rhodopsin II. Biochemistry 30, 10686-10692. Yan, B., & Spudich, J. L. (1991). Evidence the repellent receptor form of sensory rhodopsin I is an attractant signaling state. Photochem. Photobiol. 54, 1023-1026. Yan, B., Takahashi,T., Johnson, R., Derguini, E, Nakanishi, K., & Spudich, J. L. (1990). All-trans/13-cis isomerization of retinal is required for phototaxis signaling by sensory rhodopsins in ttalobacterium halobium. Biophys. J. 57, 807-814. Yao, V. J., Spudich, E. N., & Spudich, J. L. (1994). Identification of distinct domains for signaling and receptor interaction of the sensory rhodopsin I transducer, Htrl. J. Bacteriology 176, 6931-6935. Yao, V. J., & Spudich, J. L. (1992). Primary structure of an archaebacterial transducer, a methyl-accepting protein associated with sensory rhodopsin I. Proc. Natl. Acad. Sci. USA 89, 11915-11919. Zacks, D. N., Derguini, E, Nakanishi, K., & Spudich, J. L. (1993). Comparativestudy ofphototactic and photophobic receptor chromophore properties in Chlamydomonas reinhardtii. Biophys. J. 65, 508-518. Zacks, D. N., & Spudich, J. L. (1994). Gain setting in Chlamydomonas reinhardtii: Mechanism of phototaxis and the role of the photophobic response. Cell Motil. Cytoskeleton 29, 225-230.

NOTE ADDED IN PROOF Several advances have been made since the preparation of this review. These include the cloning and sequencing of SR-II and HtrII-encoding genes from Natronobacterium pharaonis and Haloarcula vallismortis [Seidel et al. (1995) Proc. Natl. Acad. Sci. USA 92, 3036-3040] and of a histidine kinase gene (cheA) required for taxis by Halobacterium salinarium [Rudolph and Oesterhelt (1995) EMBO J. 14, 667-673]. An SR-I residue (aspartyl-201) critical for attractant signaling has been identified [Olson et al. (1995) Proc. Natl. Acad. Sci. USA 92, 3185-3189] and further progress has been reported on the control of proton transfers in SR-I by Htri [Spudich (1994) Cell 79, 747-750, and references therein]. An in vitro approach has been established for study of SR-I/HtrI interaction based on rapid high-yield purification and liposome reconstitution of polyhistidinetagged SR-I [Krebs et al. (1995) Protein Expression and Purification 6, 780-788]. Also the chromophore of PYP has been demonstrated to be a thio ester linked p-coumaric acid [Hoff et al. (1994) Biochemistry 33, 13959-13962; Baca et al. (1994) Biochemistry 33, 14369-14377] and the PYP crystal structure has been redetermined and shows the protein has an tx/13 fold, resembling eukaryotic proteins involved in signal transduction [Borgstahl et al. (1995) Biochemistry 34, 6278-6287]. For discussion of these and other recent advances, the reader is referred to Spudich, J.L., Zacks, D.N., and Bogomolni, R.A. (1996) Microbial Sensory Rhodopsins: Photochemistry and Function, Israel Journal of Chemistry, in press.

ALPHA-ADREN ERG IC RECEPTORS

David B. Bylund

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Classification of Alpha Adrenergic Receptors . . . . . . . . . . . . . . . . . Alpha- 1 Adrenergic Receptors . . . . . . . . . . . . . . . . . . . . . . . . . A. Subtypes of Alpha-1 Adrenergic Receptors . . . . . . . . . . . . . . . . B. Alpha-1 Adrenergic Receptor Functions . . . . . . . . . . . . . . . . . . IV. Alpha-2 Adrenergic Receptors . . . . . . . . . . . . . . . . . . . . . . . . . A. Subtypes of Alpha-2 Adrenergic Receptors . . . . . . . . . . . . . . . . B. Peripheral Alpha-2 Adrenergic Receptor Functions . . . . . . . . . . . . C. Central Alpha-2 Adrenergic Receptor Functions . . . . . . . . . . . . . . V. Correlation of Receptor Binding and Function . . . . . . . . . . . . . . . . . VI. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I.

II. III.

!.

227 228 229 229 230 231 231 234 236 237 238 238

INTRODUCTION

Many tissues and organ systems have marked responses to norepinephrine released from sympathetic nerves and to epinephrine, a circulating hormone. The effects of these catecholamines are mediated through adrenergic receptors. Historically adrenergic receptors have been divided into two major types, alpha and beta (Ahlquist,

Biomembranes Volume 2A, pages 227-244 Copyright © 1996 by JAI Press Inc. All rights of reproduction in any form reserved. ISBN: 1-55938-659-2

227

228

DAVID B. BYLUND

1948). Subsequently, both the beta and alpha types were further subdivided into beta-l, beta-2, alpha-1 and alpha-2 subtypes (for a more complete historical perspective, see Bylund, 1988b). More recently, it has become clear that a more useful classification scheme is based on three major types---alpha-1, alpha-2 and beta--each of which is further divided imo three or more subtypes (Bylund, 1988a). The rationale for this current classification scheme of three major adrenergic receptor types is based on three lines of evidence. First, the difference in affinity of drugs that show selectivity is on the order of 1,000- to 10,000-fold among the three major receptor types (alpha-1, alpha-2, and beta), whereas the difference in affinities among subtypes (e.g., alpha-lA, alpha-lB, and alpha-lC) is generally 10- to 100-fold. Second, the family of guanine nucleotide binding proteins (G-proteins) which couples the receptor to the second messenger systems is different for each of the major receptor types: Gq couples to alpha-l; Gi to alpha-2; and G s to beta. Finally, the predicted amino acid sequence of the adrenergic receptors indicate that structurally alpha-1 and alpha-2 receptors are no more closely related to each other than they are to the beta receptor. The alpha-1 and beta adrenergic receptors both have short third cytoplasmic loops and long cytoplasmic C-terminals, whereas the alpha-2 adrenergic receptors (along with the muscarinic cholinergic receptors) have long third cytoplasmic loops and short cytoplasmic C-terminals. Thus, from a structural point of view the alpha- 1 receptor may be more closely related to the beta receptor than it is to the alpha-2 receptor (Bylund, 1992). In any case, in spite of their names, the alpha- 1 and alpha-2 are not more closely related to each other than either is to the beta receptor. Therefore, alpha-l, alpha-2 and beta should be considered as three separate, major types of adrenergic receptors. It is the purpose of this paper to review selected aspects of the alpha- 1 and alpha-2 adrenergic receptors. For these two major adrenergic receptor types there are several reviews and books available to the interested reader including those on alpha adrenergic receptors (Harrison et al., 1991; Lomasney et al., 1991a; Ruffolo et al., 1995; Hieble et al., 1995; Ruffolo et al., 1991; Bylund, 1992; Summers and McMartin, 1993; Bylund et al., 1994), on alpha-1 adrenergic receptors (Ruffolo, 1987; Minneman, 1988; Bylund et al., 1995) and on alpha-2 adrenergic receptors (Limbird, 1988; Byltmd, 1988a; Maze and Tranquilli, 1991).

II. CLASSIFICATION OF ALPHA ADRENERGIC RECEPTORS The idea that there may be more than one type of alpha adrenergic receptor dates from the mid 1960s and was based on the finding that the pharmacological characteristics, as defined by both agonists and antagonists, of alpha adrenergic receptors in the vas deferens and rabbit intestine were different (Rossum, 1965). However, nearly 10 years went by before the terms alpha-1 and alpha-2 were used to differentiate among various alpha adrenergic receptors (Delbarre and Schmitt, 1973; Langer, 1974). For a while it was thought that alpha receptors could be differentiated on the basis of their localization, that is whether they were pre- or

Alpha Adrenergic Receptors

229

postsynaptic. However, a major conceptual advance was made in 1977 by Pettinger who realized that the anatomical localization of a receptor was not necessarily a useful criteria for classifying receptors and suggested instead that there were at least two pharmacologically distinct types of alpha adrenergic receptors (Berthelson and Pettinger, 1977). This pharmacological definition of the alpha-1 and alpha-2 adrenergic receptors has proven to be very useful and has been supported by both second messenger and molecular cloning data.

III. ALPHA-1 ADRENERGIC RECEPTORS A. Subtypesof Alpha-1 Adrenergic Receptors Building on an initial report that both phentolamine and WB 4101 inhibited [3H]prazosin binding in rat frontal cortex in a manner consistent with more than a single type of receptor (Battaglia et al., 1983), Morrow and Creese (1986) formally proposed a definition of alpha- 1A and alpha- 1B receptor binding sites. The essential features of this classification scheme were that [3H]prazosin had similar affinities for the two receptor subtypes whereas the antagonists WB 4101 and phentolamine were approximately 40- to 20-fold more potent at the alpha-lA site as opposed to the alpha- 1B site (Morrow and Creese, 1986). At about the same time Johnson and Minneman (1987) observed that the alkylating agent chlorethylclonidine potentially and selectively inactivated only half of the [t25I]HEAT binding to alpha-1 adrenergic receptors in rat cerebral cortex, whereas other alkylating agems inactivated all of the binding sites. Additional evidence indicated that the sites inactivated by chlorethylclonidine represented the alpha-1B adrenergic receptor subtype (Han et al., 1987), the same subtype having low affinity for WB 4101 (Minneman et al., 1988; Wilson and Minneman, 1989). More recemly it has been shown that 5-methylurapidil and (+)-nigoldapine differentiate two alpha-1 adrenergic receptor subtypes that correspond to the alpha-lA and alpha-1B subtypes (Hanft and Gross, 1989; Gross et al., 1989). To date only these two alpha-1 adrenergic receptor subtypes have been clearly identified in tissues by pharmacological means. Three alpha-1 adrenergic receptor subtypes have been identified by molecular cloning. Since there is not yet good agreement between the pharmacologically defined subtypes, and the molecularly defined subtypes, the pharmacologically defined subtypes will be designated by uppercase letters, whereas molecularly defined subtypes will have lowercase letters. The alpha- 1b adrenergic receptor from the DDT cell (hamster smooth muscle) was cloned first (Cotecchia et al., 1988), followed by a novel alpha- 1 adrenergic receptor from a bovine brain cDNA library which has been named the alpha- 1c subtype (Schwinn et al., 1990). Recent evidence suggests that this clone corresponds to the pharmacological alpha- 1A subtype (Ford et al., 1994). More recently an alpha- 1 adrenergic receptor subtype has been cloned from rat cerebral cortex cDNA libraries (Lomasney et al., 1991b; Perez et al., 1991). Although this was initially said to represent the alpha-la adrenergic receptor

230

DAVID B. BYLUND

subtype, a more careful evaluation of its pharmacological characteristics has revealed that it is not the pharmacologically defined alpha-1A adrenergic receptor subtype and has thus been called the alpha-1 d. Accordingly, it appears that there are three alpha-1 receptors: alpha-lA, -1B, and -1D, but not an alpha-lC (Ford et al., 1994; Hieble et al., 1995). In addition, a prazosin insensitive alpha-1 subtype may also exist (Muramatsu et al., 1995). Binding studies have also indicated the existence of additional subtypes. For example, the rat parotid gland acinar cell appears to have two alpha-1 receptors, one of which is an atypical alpha-1 subtype that differs from subtypes previously characterized in the rat (Porter et al., 1992). In addition to an alpha-lA site, which has high affinity for the subtype-selective antagonists 5-methylurapidil and WB 4101, radioligand binding studies revealed an additional site which has low affinities for these antagonists. This site, however, is not CEC sensitive, and thus can not be classified as either an alpha-lB nor alpha-lc subtype. The species orthologues of these three receptor subtypes have been cloned from" the human: the alpha-ld (Bruno et al., 1991); the alpha-lb (Ramarao et al., 1992); and the alpha-1 c (Hirasawa et al., 1993). In contradistinction to the other alpha adrenergic receptors, and to most other G-protein coupled receptors, both the human and rat alpha-lb receptors contain a single large intron of at least 16 kilobases (Ramarao et al., 1992; Gao and Kunos, 1993). The organization of the genes for the alpha- 1b receptor in the two species is well conserved in that the intron interrupts the coding sequence at the same place in the putative sixth transmembrane domain, both genes display certain features of housekeeping genes, and both genes are present as a single copy in the genome.

B. Alpha-1 Adrenergic Receptor Functions A major function of the sympathetic nervous system is to regulate blood pressure under a variety of conditions, and the alpha-1 adrenergic receptors are responsible for mediating much of this regulation. In nearly all vascular beds, postsynaptic alpha-1 adrenergic receptors mediate vasoconstriction. These alpha-1 adrenergic receptors are located postjunctionally and respond primarily to neuronally released norepinephrine. In most, but not all cases, the response to alpha-1 adrenergic receptor stimulation is more prominent than the response to activation of postsynaptic alpha-2 adrenergic receptors. Several exceptions to this general rule include the canine saphenous vein (Fowler et al., 1984; Alabaster et al., 1985; Ruffolo and Zeid, 1985), and perhaps the pulmonary vein (Shebuski et al., 1986). In peripheral tissues other than the vasculature, the role of alpha-1 adrenergic receptors is less well understood. In the heart, for example, the predominant postjunctional adrenergic receptor is the beta adrenergic receptor which mediates both the inotropic and the chronotropic responses. Alpha-1 adrenergic receptors exist in cardiac tissue but their physiological role is unknown at present (Terzic et al., 1993). It is possible that myocardial alpha-1 adrenergic receptor may serve as

Alpha Adrenergic Receptors

231

a reserve mechanism to maintain myocardial responsiveness to catecholamine under conditions where beta adrenergic receptor function is compromised (Ask et al., 1987). Myocardial alpha- 1 adrenergic receptors may also play an important role in the induction of ventricular arrhythmias (Benfey, 1982), particularly during ischemia or reprefusion (Culling et al., 1987). The kidney receives dense noradrenergic innervation which extends to all portions ofthe nephron to the vasculature. An increase in sympathetic activity produces an alpha-1 adrenergic receptor-mediated antinatriuretic response in dogs (Osborn et al., 1983) and rabbits (Hesse and Johns, 1984). Results from experiments with reverse transcription combined with polymerase chain reaction, as well as radioligand binding data, indicate that the alpha-1A, alpha-lB and alpha-1 d subtypes are differentially expressed in rat nephron segments (Feng et al., 1993). In the central nervous system alpha-1 adrenergic receptors appear to be widely distributed in the brain and spinal cord. However, precise roles for alpha-1 adrenergic receptors have not yet been defined. The major electrophysiological response to activation of alpha-1 adrenergic receptors is a decrease in potassium conductants which produces a slowly developing depolarization (McCormick and Prince, 1988). Activation of cortical alpha-1 adrenergic receptors produces an excitatory effect on spontaneous neuronal activity (Szabadi and Bradshaw, 1987).

IV. ALPHA-2 ADRENERGIC RECEPTORS A. Subtypesof Alpha-2 Adrenergic Receptors The potential for alpha-2 adrenergic receptor subtypes has been evident for many years. However, a consistent and useful definition of alpha-2 adrenergic receptor subtypes has only more recently been developed based on pharmacological characteristics as defined mainly by radioligand binding studies. Several alpha-2 adrenergic antagonists including prazosin, oxymetazoline, and ARC-239 were found to have significantly different affinities for inhibiting [3H]yohimbine binding in tissues such as neonatal rat lung and rat kidney (subsequently defined as alpha-2B) as compared to the human platelet (classified as alpha-2A). Cell lines in continuous culture were then identified that had pharmacological characteristics of either the alpha-2A (HT29 cells) or alpha-2B (NG 108-15 cells) receptors (Bylund et al., 1988). Additional studies using the attenuation of the inhibition of cyclic AMP production as a functional assay confirmed the radioligand binding studies (Bylund and Ray-Prenger, 1989). The characteristics of the alpha-2A subtype include a high affinity for rauwolscine and oxymetazoline and a low affinity for prazosin and ARC-239. The pharmacological characteristics for the alpha-2B receptor include a low affinity for oxymetazoline and a relatively high affinity for prazosin and spiroxatrine. The high affinity for prazosin was interpreted originally by some investigators to indicate that the alpha-2B receptor might be an alpha-1

232

DAVID B. BYLUND

adrenergic receptor. However, the affinity of prazosin at the alpha-1 adrenergic receptor is approximately 50- to 100-fold higher than it is at the alpha-2B receptor. Subsequently a third pharmacological subtype, the alpha-2C, was identified by radioligand binding studies in an opossum kidney (OK) cell line (Blaxall et al., 1991; Murphy and Bylund, 1988). The alpha-2C subtype is characterized b y a 5-fold higher affinity for rauwolscine as compared to the alpha-2A and alpha-2B subtypes and a relatively high affinity for prazosin and ARC-239. It is best differentiated from the alpha-2B receptor by a 15-fold higher affinity for BAM 1303 and a 10-fold higher affinity for WB 4101. The Y79 human retinoblastoma cell line also appears to express the alpha-2C adrenergic receptor (Gleason and Hieble, 1992). Some workers unfortunately have used the high affinity for prazosin as the sole or major criterion for identifying the alpha-2B adrenergic receptor subtype. This has led to some considerable confusion in the literature in that in some articles the alpha-2C receptor is identified as the alpha-2B subtype. This confusion illustrates the principle that it is inappropriate to classify receptors based on one or even two drugs. The alpha-2D adrenergic receptor was originally identified in the rat "submaxillary" gland on the basis of the unique pharmacological profile (Michel et al., 1989), although it was not given a name at that time. The alpha-2 adrenergic receptor in the bovine pineal gland was also shown to have a pharmacological profile that was significantly different from that for the alpha-2A, alpha-2B, and alpha-2C subtypes, but was similar to that in the rat "submaxillary" gland. This receptor in the bovine pineal was identified as the alpha-2D subtype (Simonneaux et al., 1991). It has a low affinity for rauwolscine, 15- to 20-fold lower than the alpha-2A and alpha-2B and 50-fold lower than the alpha-2C, as well as a modestly lower affinity for SKF104078. In addition to rauwolscine, BAM 1303 and the mianserin isomers clearly differentiate it from the alpha-2A and alpha-2D receptor subtypes. The previous report that adult rat submandibular glands do not contain a detectable level of alpha-2 adrenergic receptor binding (Bylund and Martinez, 1980) is at odds with the report of alpha-2 receptors in the "submaxillary" gland. Michel et al. (1989) report that they obtained their "submaxillary" gland from Pelfreeze. On inquiry we discovered that Pelfreeze does not separate the sublingual from the submandibular gland. Using an antagonist radioligand we were unable to detect any alpha-2 adrenergic receptor binding in the rat submandibular gland obtained from Pelfreeze after removal of the sublingual gland (unpublished data). By contrast, we found a high level of alpha-2 adrenergic receptor binding in the sublingual gland using [3H]rauwolscine as the radioligand (Bmax = 380 fmol/mg/ protein; K D= 15 nM). Further investigations of the pharmacological characteristics of the alpha-2 adrenergic receptor in the sublingual gland were consistent with the conclusion that the rat sublingual gland receptor can be classified as an alpha-2D adrenergic receptor. Thus, the results obtained in the study of Michel et al. (1989) in the "submaxillary" gland were due to the alpha-2 adrenergic receptors in the sublingual gland and not in the submandibular gland.

Alpha Adrenergic Receptors

233

The cloning, sequencing, and expression of the gene coding for the human platelet alpha-2 adrenergic receptor was first reported in 1987 (Kobilka et al., 1987). The gene coding for the human platelet alpha-2 adrenergic receptor was found to be localized to chromosome 10 and thus the cloned receptor was named alpha-2C 10. The pharmacological characterization of this expressed receptor was consistent with the pharmacology of the alpha-2 receptor found in the human platelet. Because the platelet is one of the prototypic tissues for the alpha-2A receptor, the alpha-2-C10 clearly corresponds to the alpha-2A subtype. An alpha-2 receptor having a high sequence similarity to the alpha-2-C 10 receptor has been isolated from the pig (Guyer et al., 1990). This clone appears to express pharmacological characteristics of an alpha-2A receptor, although only a limited pharmacological profile was reported. Molecular cloning has also identified receptors with high sequence similarity to the human alpha-2-C 10 receptor in the rat (Chalberg et al., 1990; Lanier et al., 1991) and mouse (Link et al., 1992). The pharmacological characteristics of the rat and mouse clones are significantly different from the human and pig clones but are similar to those in the rat salivary gland and the bovine pineal. Thus they have been identified as containing the alpha-2D rather than the alpha-2A subtype. However, the rodent clones are apparently species orthologues of the human and pig clones. The human alpha-2-C 10 and the porcine orthologues have cysteine at position 201 whereas the rat, mouse and bovine orthologues have serine at position 201. Link and colleagues (1992) showed that a serine to cysteine mutation at position 201 may be responsible for the low affinity binding of the alpha-2 adrenergic receptor antagonist yohimbine, one of the characteristics of alpha-2D pharmacology. However, the affinity of other antagonists including rauwolscine, WB 4101 and SKF104078 is not altered by the mutation at residue 201 (Bylund, unpublished data). The chicken pineal alpha-2 adrenergic receptor has alpha-2A subtype pharmacology (Bylund et al., 1988) however it has a serine at position 201 (Blaxall et al., 1993). This suggests that other amino acids may be responsible for the differences between alpha-2A and alpha-2D pharmacology. A decision as to whether or not the alpha-2D should be identified as a subtype separate from the alpha-2A remains controversial. Whereas it is clearly pharmacologically different, it is also genetically very similar. Furthermore, immunoprecipitation data indicates a high degree of similarity between the human alpha-2-C 10 adrenergic receptor and the alpha-2 adrenergic receptors expressed in rat submaxillary gland and bovine pineal (Kurose et al., 1993). Referring to the rodent and bovine clones as "alpha-2A/D" would emphasize both the differences in the pharmacological characteristics as well as the similarities in the genetic sequence. The alpha-2B adrenergic receptor has been cloned from the human (Weinshank et al., 1990; Lomasney et al., 1990), the rat (Zeng et al., 1990) and mouse (Chruscinski et al., 1992). The human clone has been designated as alpha-2-C2, reflecting its localization to chromosome 2, and the rat clone has been designated

234

DAVID B. BYLUND

as RNG alpha-2, reflecting the lack of N-linked glycosylation in the amino terminal portion of the receptor. The alpha-2C adrenergic receptor has been cloned from the human (Regan et al., 1988), the rat (Lanier et al., 1991; Voigt et al., 1991; Flordellis et al., 1991), the mouse (Link et al., 1992) and the opossum (Blaxall et al., 1994). Unfortunately, in one report the clone was misidentified as an alpha-2B adrenergic receptor (Flordellis et al., 1991). The opossum OK cell alpha-2C adrenergic receptor shows only a 64% identity to the human alpha-2-C4, whereas the human and rat orthologues have a 90% identity. Nevertheless, the pharmacological characteristics of the human and opossum receptors are indistinguishable. This may be due to the fact that within the transmembrane regions, the opossum receptor has an 89% identity to human alpha-2-C4, and if conservative substitutions are included, the overall similarity becomes 99%. The amino terminus and third cytoplasmic loop are the least-conserved regions. The lower sequence identity of the OK cell alpha-2C adrenergic receptor to the human, rat, and mouse orthologues may be a reflection of the greater evolutionary separation between metatherian (marsupial) and eutherian (placental) mammals. The opossum parathyroid hormone receptor (Jiappner et al., 1991) has 78% identity to the rat parathyroid receptor (Abou-Samra et al., 1992) and the opossum serotonin 5-HT1B receptor has 82% identity to the human receptor (Cerutis et al., 1994). Many workers in the field doubt that additional alpha-2 adrenergic receptor subtypes will be found beyond the three genetic (or four pharmacological) subtypes that have been identified to date. However, there is some evidence suggesting that additional subtypes and/or heterogeneity do exist (Hieble et al., 1991; Uhlen and Wikberg, 1991; Uhlen et al., 1992). The molecular basis for this additional apparent heterogeneity is not yet known. Several of the widely used alpha-2 adrenergic drugs including clonidine and idazoxan also bind with high affinities to nonadrenergic sites which have been designated as imidazoline binding sites (Ernsberger et al., 1987). These binding sites are not part of the alpha-2 adrenergic receptor family because of their very low aff'mity for the catecholamines, epinephrine and norepinephrine. The 11 imidazoline sites have high affinity for clonidine, moxonidine and oxymetazoline whereas the 12 imidazoline sites have high affinity for idazoxan, cirazoline and guanabenz (Lanier et al., 1993; MacKinnon et al., 1993; Piletz and Sletten, 1993; Molderings et al., 1993). The I2 sites appear to be related to MAO (Raddatz et al., 1995).

B. Peripheral Alpha-2 Adrenergic Receptor Functions Essentially all sympathetically innervated tissues possess presynaptic alpha-2 adrenergic receptors which mediate the inhibition of norepinephrine release. In these tissues alpha-2 adrenergic receptor antagonists may produce effects due to the enhancement of neurotransmitter release. In addition, many tissues also have postsynaptic alpha-2 receptors.

Alpha Adrenergic Receptors

235

The vasoconstriction caused by neuronally released norepinephrine as well as circulating epinephrine has been classically ascribed to mediation through the alpha-1 adrenergic receptor. More recently, however, it has become clear that in many vascular beds postsynaptic alpha-2 adrenergic receptors also may play an important role in mediating vasoconstriction. Indeed, in some vessels such as the dog saphenous vein, the alpha-2 adrenergic-mediated vasoconstriction may be more important than the alpha-1 adrenergic receptor-mediated vasoconstriction (Fowler et al., 1984; Alabaster et al., 1985; Ruffolo and Zeid, 1985). As a general rule it appears that alpha-1 receptors are located postjunctionally whereas the postsynaptic alpha-2 adrenergic receptors are localized extrajunctionally. Postsynaptic vascular alpha-2 adrenergic receptors are thought to reside extrajunctionally because they are probably not activated by norepinephrine released from sympathetic nerve, but can be stimulated by administered epinephrine or norepinephrine (Langer and Shepperson, 1982; Wilffert et al., 1982). The physiological role of extrajunctional alpha-2 adrenergic receptors is not fully understood. It has been suggested, however, that the importance of the alpha-2 receptor may lie in responses to stress and/or in certain pathological states such as hypertension and possibly congestive heart failure (Bolli et al., 1984; Jie et al., 1986; Brodde and Michel, 1992). Further support for the concept that alpha-2 adrenergic receptors may be involved in certain disease state comes from studies in the coronary arterial circulation (Heusch and Deussen, 1983; Seitelberger et al., 1988), the pulmonary arterial circulation (Hyman and Kadowitz, 1986; Shebuski et al., 1986). An exception to the general rule that alpha-1 adrenergic receptors are more prominent than alpha-2 adrenergic receptors is the cutaneous circulation, where it appears that alpha-2 adrenergic receptors play a predominant role in thermal regulation (Flavahan et al., 1987a). In several tissues the alpha-2 adrenergic receptors appear to be localizedjunctionally rather than extrajunctionally. These include the hepatic portal system (Segstro and Greenway, 1986) and the saphenous vein (Flavahan et al., 1987b). There is a relatively high density of alpha-2 adrenergic receptors in the kidney (twice as high as alpha-1 adrenergic receptors in the rat) (Sanchez and Pettinger, 1981), but the physiological function of these alpha-2 adrenergic receptors in renal function is not yet clear. In the gastrointestinal tract, alpha-2 adrenergic receptors mediate several responses at different levels, such as regulation of gastric and intestinal mobility and secretions, some of which may be centrally mediated (Nagata and Osumi, 1993). Stimulation of alpha-2 adrenergic receptors promotes net sodium and chloride absorption (Field and McColl, 1973). This results in a net water absorption and thus clonidine produces an inhibition of watery diarrhea (McArthur et al., 1982). In the pancreas alpha-2 adrenergic receptor antagonists increase glucose-stimulated insulin release (Nakaki et al., 1980). Thus it appears that insulin secretion is under tonic inhibition of pancreatic alpha-2 adrenergic receptors, but the precise role of these receptors is unclear. Both circulating epinephrine released from the

236

DAVID B. BYLUND

adrenal medulla and sympathetic release of norepinephrine may contribute to the alpha-2-mediated tonic suppression of insulin release. Lower urinary tract function is under control of the autonomic nervous system and both alpha- 1 and alpha-2 adrenergic receptors appear to mediate the contractile response. Although functional studies originally suggested that only alpha-1 adrenergic receptor mediate the contractile response (Tsujimoto et al., 1986; Kunisawa et al., 1985), more recent results suggest a predominance of alpha-2 adrenergic receptor in the female (Latifpour et al., 1990; Yoshida et al., 1991). Functional studies suggest the presence of alpha-2 adrenergic receptors on vascular endothelium which cause a relaxation mediated by an endothelium derived relaxing factor, nitric oxide (Cocks and Angus, 1983; Angus et al., 1986; Bockman et al., 1993; Richard et al., 1990). The recent report that norepinephrine-induced release of nitric oxide is enhanced in mineraiocorticoid hypertension (Bockman et al., 1992), suggests that endothelium alpha-2 adrenergic receptors may play an enhanced role in regulating vascular tone under pathological conditions. In general, however, these studies have not been particularly productive. Alpha-2 adrenergic receptor-mediated lowering of intraocular pressure in animals and humans was demonstrated as early as 1966 (Makabe, 1966; Hasslinger, 1969). This appears to be the result of suppressing aqueous flow (Lee et al., 1984; Gharagozloo et al., 1988; Toris et al., 1995), although whether the receptors responsible for this action are presynaptic or postsynaptic is not yet entirely clear. The ability of epinephrine to induce aggregation of human platelets and to potentiate aggregation induced by other agents such as thrombin and ADP is well known. This function is mediated by alpha-2 adrenergic receptors. In general, however, the plasma concentrations of epinephrine are much lower than those required to induce platelet aggregation in vitro. It is possible that the physiological control of platelet aggregation is mediated by the actions of multiple agems including epinephrine, each of which is present at levels below that which would individually cause platelet aggregation. Radioligand binding assays of human platelet alpha-2 adrenergic receptors has been used in numerous clinical studies in an attempt to understand the role of alpha-2 adrenergic receptors in various pathological and physiological states.

C. Central Alpha-2 Adrenergic Receptor Functions Alpha-2 adrenergic receptors are widely distributed in the central nervous system and appear to be directly involved in several central functions. Stimulation of central alpha-2 adrenergic receptors in the ventrolateral medulla induces a reduction in sympathetic outflow to the periphery and a resulting reduction in arterial blood pressure. The hypotensive action of alpha-2 adrenergic receptor agonists appears to be due to stimulation of postsynaptic alpha-2 receptors rather than presynaptic receptors. The imidazoline receptors also may participate in the mediation of the antihypertensive effect of many alpha-2 adrenergic receptor drugs.

Alpha Adrenergic Receptors

237

For nearly two decades alpha-2 adrenergic receptor agonists have been widely used by veterinarians to achieve dose-dependent sedation, analgesia and muscle relaxation in a variety of domesticated and wild animal species. More recently this is being extended to human medicine (Hayashi and Maze, 1993). The site and mechanism of action for these effects remain controversial. Alpha-2 adrenergic receptors also seem to play a role in the regulation of affect or mood, and some alpha-2 adrenergic antagonists appear to be effective antidepressants. Alpha-2 adrenergic agonists also appear to improve memory under certain conditions. Opiate and alpha-2 adrenergic receptor agonists act through independent receptors within the locus coeruleus but produce similar depressant effects on net cell activity (Aghajanian, 1978). Since opiates and alpha-2 adrenergic receptor agonists do not exhibit cross tolerance, alpha-2 adrenergic agonists have become a mainstay in the management of acute detoxification from opiate addiction (Gold et al., 1978).

V. CORRELATION OF RECEPTOR BINDING A N D FUNCTION One of the more difficult issues facing investigators who study receptors is the correlation of receptor binding with receptor function. Of particular difficulty is establishing that the receptor identified by radioligand binding studies is the same receptor identified in functional studies. Frequently investigators feel it is sufficient to show that the pharmacological characteristics of the receptors are the same. However, studies with the submandibular gland indicated that this approach is not sufficient. We have shown that the alpha-2 adrenergic receptor labeled in radioligand binding studies by [3H]clonidine is localized postsynaptically (Bylund and Martinez, 1981). This conclusion is based on the following evidence. (1) Partial destruction of the presynaptic nerve terminal with 6-hydroxydopamine did not decrease the density of alpha-2 adrenergic receptors following subsequent reserpine administration; (2) duct ligation which results in atrophy of the gland markedly decrease the density of receptors following subsequent reserpine administration; (3) surgical denervation resulted in an appearance of high density of alpha-2 adrenergic receptors; and (4) the changes in alpha-2 adrenergic receptor binding paralleled the changes in the postsynaptic beta adrenergic receptor (Bylund and Martinez, 1981). By contrast functional studies such as those involving the release of [3H]norepinephrine, indicate that the alpha-2 adrenergic receptors which modulate inhibition of [3H]norepinephrine release are localized presynaptically (Filinger et al., 1978). Thus, the alpha-2 receptors which are identified functionally and appear to be localized presynaptically, may not be the same alpha-2 adrenergic receptors identified in this tissue by radioligand binding studies and which are postsynaptically localized. This is in spite of data indicating that the pharmacological characteristics of the submandibular gland presynaptic receptor as identified by functional studies

238

DAVID B. BYLUND

appears to be of the alpha-2D subtype (Smith et al., 1992) the same subtype identified in binding studies (Michel et al., 1989; Bylund, 1992). These results demonstrate that considerable care is needed in attempting to correlate receptor binding and function. In fact, no study has yet clearly identified presynaptic alpha-2 adrenergic receptors based on radioligand binding studies. This may be due to a much lower density of the presynaptic receptor which is beyond the detection limit of current assay systems.

Vi.

CONCLUSION

Adrenergic receptors are classified into three main types, the alpha-1, alpha-2, and beta adrenergic receptor types. Each of these types is further divided into three or more subtypes. For the alpha-2 adrenergic receptor subtypes there is good agreement between the pharmacologically defined subtypes and the subtypes identified by molecular cloning. By contrast, it is not yet clear how the various cloned alpha- 1 adrenergic receptors fit into the existing pharmacological classification scheme. Alpha-1 and alpha-2 adrenergic receptors are known to mediate a multitude of peripheral and central functions, although the particular subtype (i.e., alpha-lA, -1B and alpha-2A, -2B or -2C) involved is generally not known. A determination of this subtype specificity for each of these functions is an important challenge for the rest of this decade.

REFERENCES Abou-Samra, A.-B., JOppner, H., Force, T., Freeman, M. W., Kong, X.-F., Schipani, E.,. Urena, E, Richards, J., Bonventre, J. V., Ports, J. T., Jr., Kronenberg, H. M., & Segre, G. V. (1992). Expression cloning of a commonreceptor for a parathyroidhormone and parathyroidhormone-relatedpeptide from rat osteoblast-likecells: A singlereceptor stimulatesintracellularaccumulationof both cAMP and inositol trisphosphates and increases intracellular free calcium. Proc. Natl. Acad. Sci. USA 89, 2732-2736. Aghajanian, G. K. (1978). Tolerance of locus coeruleus neurons to morphine and suppression of withdrawal response by clonidine. Nature 276, 186-188. Ahlquist, R. P. (1948). A study of adrenotropic receptors. Am. J. Physiol. 153,586-600. Alabaster, V. A., Keir, R. F., & Peters, C. J. (1985). Comparisonofactivity ofalpha-adrenoceptoragonists and antagonists in dog and rabbit saphenous vein. Naunyn-Schmiedeberg'sArch. Pharmacol. 330, 33--36. Angus, J. A., Cocks, T. M., & Satoh, K. (1986). The et-adrenoceptors on endothelial cells. Fed. Proc. 45, 2355-2359. Ask, J. A., Stene-Larsen, G., Helle, K. B., & Besch, F. (1987). Functional ot-adrenoceptors in human atrial preparations in the presence of 13-receptorblockade. Acta Physiol. Scand. 131,439--445. Battaglia, G., Shannon, M., Borgundvaag, B., & Titeler, M. (1983). Properties of 3H-prazosin labeled oq-adrenergic receptors in rat brain and porcine neurointermediate lobe tissue. J. Neurochem. 41, 538-542. Berthelson, S., & Pettinger, W. A. (1977). A functionalbasis for classificationofct-adrenergic receptors. Life Sci. 21,595--606. Benfey, B. G. (1982). Function of myocardial ot-adrenoceptors. Life Sci. 31,101-112.

Alpha Adrenergic Receptors

239

Blaxall, H. S., Cerutis, D. R., Hass, N. A., & Bylund, D. B. (1994). Clonirig and expression of the alpha-2C adrenergic receptor from the OK cell line. Mol. Pharmacol. Blaxall, H. S., Heck, D. A., & Bylund, D. B. (1993). Molecular determinants of the alpha-2D adrenergic receptor subtype. Life Sci. 53, 255-259. Blaxall, H. S., Murphy, T. J., Baker, J. C., Ray, C., & Bylund, D. B. (1991). Characterization of the alpha-2C adrenergic receptor subtype in the opossum kidney and in the OK cell line. J. Pharmacol. Exp. Ther. 259, 323-329. Bockman, C. S., Jeffries, W. B., & Abel, P. W. (1993). Binding and functional characterization of alpha-2 adrenergic receptor subtypes on pig vascular endothelium. J. Pharmacol. Exp. Ther. 267, 11261133. Bockman, C. S., Jeffries, W. B., Pettinger, W. A., & Abel, P. W. (1992). Enhanced release ofendothelium-derived relaxing factor in mineralocorticoid hypertension. Hypertension 20, 304-313. Bolli, P., Eme, P., Block, L. H., Ji, B. H., Kiowski, W., & Buhler, F.R. (1984). Adrenaline induces vasoconstriction through post-junctional u2-adrenoceptors and this response is enhanced in patients with essential hypertension. J. Hypertension 2, 115-118. Brodde, O.-E., & Michel, M. C. (1992). Adrenergic receptors and their signal transduction mechanisms in hypertension. J. Hypertension 10, S 133-S145. Bruno, J. F., Whittaker, J., Song, J., & Berelowitz, M. (1991). Molecular cloning and sequencing of a cDNA encoding a human alA adrenergic receptor. Biochem. Biophys. Res. Comm. 179, 14851490. Bylund, D. B. (1985). Heterogeneity of alpha-2 adrenergic receptors. Pharmacol. Biochem. Behav. 22, 835--843. Bylund, D. B. (1988a). Subtypes of a2-adrenoceptors: pharmacological and molecular biological evidence converge. TIPS 9, 356-361. Bylund, D. B. (1988b). Alpha-2 adrenergic receptors: A historical perspective. In: The Alpha-2 Adrenergic Receptors (L. Limbird, ed.), pp. 1-13 Humana Press, Clifton, N.J. Bylund, D. B. (1992). Subtypes of eq-and Gt2-adrenergic receptors. FASEB J. 6, 832--839. Bylund, D. B., Eikenberg, D. C., Hieble, J. P., Langer, S. Z., Lefkowitz, R. J., Minneman, K. P., Molinoff, P. B., Ruffolo, R. R., Jr., & Trendelenburg, A. U. (1994). IV. International Union of Pharmacology Nomenclature of Adrenoceptors. Pharmacol. Rev. 46, 121-136. Bylund, D. B., & Martinez, J. R. (1980). a2-Adrenergic receptors appear in rat salivary glands after reserpine treatment. Nature 285, 229-230. Bylund, D. B., & Martinez, J. R. (1981). Postsynaptic localization of u2-adrenergic receptors in rat submandibular gland. J. Neurosci. 1, 1003-1007. Bylund, D. B., & Ray-Prenger, C. (1989). Alpha-2A and alpha-2B adrenergic receptor subtypes: attenuation of cyclic AMP production in cell lines containing only one receptor subtype. J. Pharmacol. Exp. Ther. 251,640-644. Bylund, D. B., Ray-Prenger, C., & Murphy, T. J. (1988). Alpha-2A and alpha-2B adrenergic receptor subtypes: antagonist binding in tissues and cell lines containing only one subtype. J. Pharmacol. Exp. Ther. 245, 600-607. Bylund, D. B., Regan, J. W., Faber, J. E., Hieble, J. P., Triggle, C. R., & Ruffolo, R. R., Jr. (1995). Vascular a-adrenoceptors: From the gene to the human. Can. J. Physiol. Pharmacol. 73,533-543. Cerutis, D. R., Hass, N. A., Iversen, L. J., & Bylund, D. B. (1994). The cloning and expression of an OK cell cDNA encoding a 5-HT m receptor. Mol. Pharmacol., in press. Chalberg, S. C., Duda, T., Rhine, J. A., & Sharma, R. K. (1990). Molecular cloning, sequencing and expression of an ot2-adrenergic receptor complementary DNA from rat brain. Mol. Cell. Biochem. 97, 161-172. Chruscinski, A. J., Link, R. E., Daunt, D. A., Barsh, G. S., & Kobilka, B. K. (1992). Cloning and expression of the mouse homolog of the human et2-C2 adrenergie receptor. Biochem. Biophys. Res. Comm. 186, 1280-1287.

240

DAVID 13. BYLUND

Cocks, T. M., & Angus, J. A. (1983). Endothelium-dependent relaxation of coronary arteries by noradrenaline and serotonin. Nature 305, 627-630. Cotecchia, S., Schwinn, D. A., Randall, R. R., Lefkowitz, R. J., Caron, M. G., & Kobilka, B. K. (1988). Molecular cloning and expression of the cDNA for the hamster Otl-adrenergic receptor. Proc. Natl. Acad. Sci. USA 85, 7159-7163. Culling, W., Penny, W. J., Cunliffe, G., Flores, N. A., & Sheridan, D. J. (1987). Arrhythmogenic and electrophysiological effects of alpha adrenoceptor stimulation during myocardial ischaemia and reperfusion. J. Mol. Cell. Cardiol. 19, 251-258. Delbarre, B., & Schmitt, H. (1973). A further attempt to characterize sedative receptors activated by clonidine in chickens and mice. Eur. J. Pharmacol. 22, 355-359. Emsberger, P., Meeley, M. P., Mann, J. J., & Reis, D. J. (1987). Clonidine binds to imidazole binding sites as well as alpha2-adrenoceptors in the ventrolateral medullar. Eur. J. Pharmacol. 134, 1-13. Feng, E, Abel, E W., Scofield, M., Liu, F., Wolff, D. W., & Jeffries, W. B. (1993). Heterogeneous expression of ctl-adrenoceptor subtypes among rat nephron segments. Mol. Pharmacol. 44, 926-933. Field, M., & McColl, I. (1973). Ion transport in rabbit ileal mucosa. III. Effects of catecholamines. Am. J. Physiol. 225, 852-857. Filinger, E. J., Langer, S. Z., Perec, C. J., & Stefano, F. J. E. (1978). Evidence for the presynaptic location of the a-adrenoceptors which regulate noradrenaline release in the rat submaxillary gland. Naunyn Schmiedeberg's Arch. Pharmacol. 304, 21-26. Flavahan, N. A., Cooke, J. P., Shepherd, J. T., & Vanhoutte, E M. (1987a). Human postjunctional alpha- 1 and alpha-2 adrenoceptors: differential distribution in arteries of the limbs. J. Pharmacol. Exp. Ther. 241, 361-365. Flavahan, N. A., Miller, V. M., Aarhus, L. L., & Vanhoutte, P. M. (1987b). Denervation augments alpha-2 but not alpha-1 adrenergic responses in canine saphenous veins. J. Pharmacol. Exp. Ther. 240, 589-593. Flordellis, C. S., Handy, D. E., Bresnahan, M. R., Zannis, V. I., & Gavras, H. (1991). Cloning and expression of a rat brain CtEB-adrenergicreceptor. Proc. Natl. Acad. Sci. USA 88, 1019-1023. Ford, A. P. D. W., Williams, T. J., Blue, D. R., & Clarke, D. E. (1994). txl-Adrenoceptor classification: sharpening Occam's razor. Trends Pharmacol. Sci. 15, 167-170. Fowler, P. J., Grous, M., Price. W., & Matthews, W. D. (1984). Pharmacological differentiation of postsynaptic alpha adrenoceptors in the dog saphenous vein. J. Pharmacol. Exp. Ther. 229, 712-718. Gao, B., & Kunos, G. (1993). Isolation and characterization of the gene encoding the rat ala-adrenergic receptor. Gene 131,243-247. Gharagozloo, N. Z., Relf, S. J., & Brubaker, R. F. (1988). Aqueous flow is reduced by the alpha-adrenergic agonist, apraclonidine hydrochloride (ALO 2145). Ophthalmology 95, 1217-1220. Gleason, M. M., & Hieble, J. P. (1992). The ct2-adrenoceptors of the human retinoblastoma cell line (Y79) may represent an additional example of the a2C-adrenoceptor. Br. J. Pharmacol. 107, 222-225. Gold, M. S., Redmond, D. E., Jr., & Kleber, H. D. (1978). Clonidine blocks acute opiate withdrawal symptoms. Lancet 2, 599-602. Gross, G., Hanft, G., & Mehdom, H. M. (1989). Demonstration of tXlA- and otlB-adrenoceptor binding sites in human brain tissue. Eur. J. Pharmacol. 169, 325-328. Guyer, C. A., Horstman, D. A., Wilson, A. L., Clark, J. D., Cragoe, E. J., Jr., & Limbird. L. E. (1990). Cloning, sequencing, and expression of the gene encoding the porcine (x2-adrenergic receptor. Allosteric modulation by Na ÷, H÷, and amiloride analogs. J. Biol. Chem. 265, 17307-17317. Han, C., Abel, E W., & Minneman, K. E (1987). Heterogeneity of ctl-adrenergic receptors revealed by chloroethylclonidine. Mol. Pharmacol. 32, 505-510. Hanft, G., & Gross, G. (1989). Subclassification of al-adrenoceptor recognition sites by urapidil derivatives and other selective antagonists. Br. J. Pharmacol. 97, 691-700.

Alpha Adrenergic Receptors

241

Harrison, J. K., Pearson, W. R., & Lynch, K. R. (1991). Molecular characterization of otl- and ct2-adrenoceptors. TIPS 12, 62--67. Hasslinger, C. (1969). Catapres (2-[2,6-dichlorophenylamino]-2-imidazoline hydrochloride}---a new drug lowering intraocular pressure. Klin. Mbl. Augenheilk. 154, 95-105. Hayashi, Y., & Maze, M. (1993). Alpha2 adrenoceptor agonists and anaesthesia. Brit. J. Anaes. 71, 108-118. Hesse, I. F. A., & Johns, E. J. (1984). The subtype of cz-adrenoceptor involved in the neural control of renal tubular sodium reabsorption in the rabbit. J. Physiol. 352, 527-538. Heusch, G., & Deussen, A. (1983). The effects of cardiac sympathetic nerve stimulation on perfusion of stenotic coronary arteries in the dog. Circ. Res. 53, 8-15. Hieble, J. P., Bondinell, W. E., & Ruffolo, R. R., Jr. (1995). or- and 13-Adrenoceptors: From the gene to the clinic. 1. Molecular biology and adrenoceptor subclassification. J. Med. Chem. 38, 3415-3444. Hieble, J. P., Bylund, D. B., Clarke, D. E., Eikenburg, D. C., Langer, S. Z., Lefkowitz, R. J., Minneman, K. P., & Ruffolo, R. R., Jr. (1995). International Union of Pharmacology. X. Recommendation for nomenclature of czl-adrenoceptors: Consensus update. Pharmacol. Rev. 47, 267-270. Hieble, J. P., Sulpizio, A. C., Edwards, R., Chapman, H., Young, P., Roberts, S. P., Blackburn, T. P., Wood, M. D., Shah, D. H., Demarinis, R. M., & Ruffolo, R. R., Jr. (1991). Additional evidence for functional subclassification of alpha-2 adrenoceptors based on a new selective antagonist, SK&F 104856. J. Pharmacol. Exp. Ther. 259, 643--652. Hirasawa, A., Horie, K., Tanaka, T., Takagaki, K., Mural, M., Yano, J., & Tsujimoto, G. (1993). Cloning, functional expression and tissue distribution of human cDNA for the atlc-adrenergic receptor. Biochem. Biophys. Res. Comm. 195, 902-909. Hyman, A. L., & Kadowitz, P. J. (1986). Enhancement of~t- and 13-adrenoceptorresponses by elevations in vascular tone in pulmonary circulation. Am. J. Physiol. 250, H 1109-H 1116. Jie, K., van Brummelen, P., Vermey, P., Timmermans, P. B. M. W. M., & van Zwieten, P. A. (1986). Alpha 1- and alpha2-adrenoceptor mediated vasoconstriction in the forearm of normotensive and hypertensive subjects. J. Cardiovasc. Pharmacol. 8, 190-196. Johnson, R. D.. & Minneman, K. P. (1987). Differemiation of ¢xl-adrenergic receptors linked to phosphatidylinositol turnover and cyclic AMP accumulation in rat brain. Mol. Pharmacol. 31, 239-246. JOppner, H., Abou-Samra, A.-B., Freeman, M., Kong, X. F., Schipani, E., Richards, J., Kolakowski, L. F., Jr., Hock, J., Ports, J. R., Jr., Kronenberg, H. M., & Segre, G. V. (1991). A G protein-linked receptor for parathyroid hormone and parathyroid hormone-related peptide. Science 254, 10241026. Kobilka, B. K., Matsui, H., Kobilka, T. S., Yang-Feng, T. L., Francke, U., Caron, M. G., Lefkowitz, R. J., & Regan, J. W. (1987). Cloning, sequencing, and expression of the gene coding for the human platelet ot2-adrenergic receptor. Science 238, 650-656. Kunisawa, Y., Kawabe, K., Niijima, T., Honda, K., & Takeraka, T. (1985). A pharmacological study of c~-adrenergic subtypes in smooth muscle of human urinary bladder base and prostatic urethra. J. Urol. 134, 396-398. Kurose, H., Arriza, J. L., & Lefkowitz, R. J. (1993). Characterization of et2-adrenergic receptor subtype-specific antibodies. Mol. Pharmacol. 43,444-450. Langer, S. Z. (1974). Presynaptic regulation of catecholamine release. Biochem. Pharmacol. 23, 1793-1800. Langer, S. Z., & Shepperson, N. B. (1982). Postjunctional cq- and c~2-adrenoceptors: preferemial innervation of otl-adrenoceptors and the role of neuronal uptake. J. Cardiovasc. Pharmacol. 4, $8-S13. Lanier, S. M., Downing, S., Duzie, E., &Homcy, C. J. (1991). Isolation of rat genomic clones encoding subtypes of the ot2-adrenergic receptor. Identification of a unique receptor subtype. J. Biol. Chem. 266, 10470-10478.

242

DAVID B. BYLUND

Lanier, S. M., Ivkovic, B., Singh, I., Neumeyer, J. L., & Bakthavachalam, V. (1993). Visualization of multiple imidazoline/guanidinium-receptive sites. J. Biol. Chem. 16047-16051. Latifpour, J., Kondo, S., O'Hollaren, B., Morita, T., & Weiss, R. M. (1990). Autonomic receptors in urinary tract: Sex and age differences. J. Pharmacol. Exp. Ther. 253, 661--667. Lee, D. A., Topper, J. E., & Brubaker, R. E (1984). Effect ofclonidine on aqueous humor flow in normal human eyes. Exp. Eye Res. 38, 239-246. Limbird, L. E. (ed.) (1988). The Alpha-2 Adrenergic Receptors, Humana Press, Clifton. N.J. Link, R., Daunt, D., Barsh, G., Chruscinski, A., & Kobilka, B. (1992). Cloning of two mouse genes encoding a2-adrenergic receptor subtypes and identification of a single amino acid in the mouse a2-C 10 homolog responsible for an interspecies variation in antagonist binding. Mol. Pharmacol. 42, 16-27. Lomasney, J. W., Cotecchia, S., Lefkowitz, R. J., & Caron, M. G. (1991a). Molecular biology of a-adrenergic receptors: implications for receptor classification and for structure-function relationships. Biochim. Biophys. Acta 1095, 127-139. Lomasney, J. W., Cotecchia, S., Lorenz, W., Leung, W.-Y., Schwinn, D. A., Yang-Feng, T. L., Brownstein, M., Lefkowitz, R. J., & Caron, M. G. (1991b). Molecular cloning and expression of the cDNA for the alA-adrenergic receptor. The gene for which is located on human chromosome 5. J. Biol. Chem. 266, 6365-6369. Lomasney, J. W., Lorenz, W., Allen, L. F., King, K., Regan, J. W., Yang-Feng, T. L., Caron, M. G., & Lefkowitz, R. J. (1990). Expansion of the t~z-adrenergic receptor family: Cloning and characterization of a human ct2-adrenergic receptor subtype, the gene for which is located on chromosome 2. Proc. Natl. Acad. Sci. USA 87, 5094-5098. MacKinnon, A. C., Stewart, M., Olverman, H. J., Spedding, M., & Brown, C. M. (1993). [3H]p-Aminoclonidine and [3H]idazoxan label different populations of imidazoline sites on rat kidney. Eur. J. Pharmacol. 232, 79-87. Makabe, R. (1966). Ophthalmological studies with dichlorophenyamine-Imidazolines. Dtsch. Med. Wochenschr. 91, 1686-1696. Maze~ M., & Tranquilli, W. (1991). Alpha-2 adrenoceptor agonists: defining the role in clinical anesthesia. Anesthesiology 74, 581-605. McArthur, K. E., Anderson, D. S., Durbin, T. E., Orloff, M. J., & Dharmsathaphom, K. (1982). Clonidine and lidamidine to inhibit watery diarrhea in a patient with lung cancer. Ann. Intem. Med. 96, 323--325. McCormick, D. A., & Prince, D. A. (1988). Noradrenergic modulation of firing pattern in guinea pig and cat thalamic neurons, in vivo. J. Neurophysiol. 59, 978-996. McGrath, J. C., Brown, C. M., & Wilson, V. G. (1989). Alpha-adrenoceptors: A critical review. Med. Res. Rev. 9, 407-533. Michel, A. D., Loury, D. N., & Whiting, R. L. (1989). Differences between the ct2-adrenoceptor in rat submaxillary gland and the aZA- and a2B-adrenoeeptor subtypes. Br. J. Pharmacol. 98, 890-897. Minneman, K. P. (1988). ot1-Adrenergic receptor subtypes, inositol phosphates, and sources of cell Ca 2+. Pharmacol. Rev. 40, 87-119. Minneman, K. E, Han, C., & Abel, P. W. (1988). Comparison of cq-adrenergic receptor subtypes distinguished by chloroethylclonidine and WB 4101. Mol. Pharmaeol. 33, 509-514. Molderings, G. J., Moura, D., Fink, K., Brnisch, H., & Grthert, M. (1993). Binding of [3H]elonidine to Ii-imidazoline sites in bovine adrenal medullary membranes. Naunyn-Schmiedeberg's Arch. Pharmacol. 348, 70-76. Morrow, A. L., & Creese, I. (1986). Characterization of cq-adrenergic receptor subtypes in rat brain: a reevaluation of [3H]WB 4101 and [3H]prazosin binding. Mol. Pharmacol. 29, 321-330. Muramatsu, I., Oshita, M., Ohmura, T., Kigoshi, S., Akino, H., Gobara, M., & Okada, K. (1994). Pharmacological characterization of ctl-adrenoceptor subtypes in the human prostate: Functional and binding studies. Br. J. Urol. 74, 572-578.

Alpha Adrenergic Receptors

243

Murphy, T. J., & Bylund, D. B. (1988). Characterization of alpha-2 adrenergic receptors in the OK cell, an opossum kidney cell line. J. Pharmacol. Exp. Ther. 244, 571-578. Nagata, M., & Yoshitsugu, O. (1993). Central ct2-adrenoceptor-mediated inhibition of gastric motility in rats. Japan. J. Pharmacol. 62, 329-330. Nakaki, T., Nakadate, T., & Kato, R. (1980). ct-Adrenoceptors modulating insulin release from isolated pancreatic islet cells. Naunyn-Schmiedeberg's Arch. Pharmacol. 313, 151-153. Osborn, J. L., Holdaas, H., Thames, M. D., & Di Bona, G. F. (1983). Renal adrenoceptor mediation of antinatriuretic and renin secretion responses to low frequency renal nerve stimulation in the dog. Circ. Res. 53,298-305. Perez, D. M., Piascik, M. T., & Graham, R. M. (1991). Solution-phase library screening for the identification of rare clones: isolation of an OqD-adrenergic receptor cDNA. Mol. Pharmacol. 40, 876-883. Piletz, J. E., & Sletten, K. (1993). Nonadrenergic imidazoline binding sites on human platelets. J. Pharmacol. Exp. Ther. 267, 1493-1502. Porter, J. E., Dowd, F. J., & Abel, P. W. (1992). Atypical alpha-1 adrenergic receptors in the rat parotid gland acinar cell. J. Pharmacol. Exp. Ther. 263, 1062-1067. Raddatz, R., Parini, A., & Lanier, S. M. (1995). Imidazoline/guanidinium binding domains on monoamine oxidases---Relationship to subtypes of imidazoline-binding proteins and tissuespecific interaction of imidazoline ligands with monoamine oxidase B. J. Biol. Chem. 270, 27961-27968. Ramarao, C. S., Denker, J. M., Perez, D. M., Gaivin, R. J., Riek, R. P., & Graham, R. M. (1992). Genomic organization and expression of the human OtlB-adrenergic receptor. J. Biol. Chem. 267, 2193621945. Regan, J. W., Kobilka, T. S., Yang-Feng, T. L., Caron, M. G., Lefkowitz, R. J., & Kobilka, B. K. (1988). Cloning and expression of a human kidney cDNA for an ot2-adrenergic receptor subtype. Proc. Natl. Acad. Sci. USA 85, 6301-6305. Richard, V., Tanner, F. C., Tschudi, M., & Luscher, T. F. (1990). Different activation of L-arginine pathway by bradykinin, serotonin, and clonidine in coronary arteries. Am. J. Physiol. 259, H1433--H1439. Rossum, J. M. (1965). Different types of sympathomimetic a-receptors. J. Pharm. Pharmacol. 17, 202-216. Ruffolo, R., Jr. (ed.) (1987). The Alpha-1 Adrenergic Receptors, Humana Press, Clifton, N.J. Ruffolo, R., Jr. (ed.)(1991). ct-Adrenoceptors: molecular biology, biochemistry and pharmacology, S. Karger AG. Switzerland. Ruffolo, R. R., Bondinell, W., & Hieble, J. P. (1995). ct- and 13-adrenoceptors: From the gene to the clinic. 2. Structure-activity relationships and therapeutic applications. J. Med. Chem. 38, 36813716. Ruffolo, R. R., Jr., & Zeid, R. L. (1985). Relationship between ct-adrenoceptor occupancy and response for the ctl-adrenoceptor agonist, cirazoline, and the ot2-adrenoceptor agonist, B-HT 933, in canine saphenous vein. J. Pharmacol. Exp. Ther. 235, 636--643. Ruffolo, R. R., Jr., Nichols, A. J., Stadel, J. M., & Hieble, J. P. (1991). Structure and function of et-adrenoceptors. Pharmacol. Rev. 43, 475--505. Sanchez, A., & Pettinger, W. A. (1981). Dietary sodium regulation of blood pressure and renal ot1- and et2-receptors in WKY and SHR. Life Sci. 29, 2795-2802. Schwinn, D. A., & Lomasney, J. W. (1992). Pharmacologic characterization of cloned otl-adrenoceptor subtypes: selective antagonists suggest the existence of a fourth subtype. Eur. J. Pharmacol. 227, 433-436. Schwinn, D. A., Lomasney, J. W., Lorenz, W., Szklut, P. J., Fremeau, R. T., Jr., Yang-Feng, T. L., Caron, M. G., Lefkowitz, R. J., & Cotecchia, S. (1990). Molecular cloning and expression of the cDNA for a novel eq-adrenergic receptor subtype. J. Biol. Chem. 265, 8183-8189.

244

DAVID B. BYLUND

Segstro, R., & Greenway, C. V. (1986). Alpha receptor subtype mediating sympathetic mobilization of blood from the hepatic venous system in anesthetized cats. J. Pharmacol. Exp. Ther. 236, 224-229. Seitelberger, R., Guth, B. D., Heusch, G., Lee, J. D., Katayama, K., & Ross, J., Jr. (1988). Intracoronary a2-adrenergic receptor blockade attenuates ischemia in conscious dogs during exercise. Circ. Res. 62, 436-442. Shebuski, R. J., Fujita, T., & Ruffolo, R. R., Jr. (1986). Evaluation of alpha-1 and alpha-2 adrenoceptor-mediated vasoconstriction in the in situ, autoperfused, pulmonary circulation of the anesthetized dog. J. Pharmacol. Exp. Ther. 238, 217-223. Simonneaux, V., Ebadi, M., & Bylund, D. B. (1991). Identification and characterization of alpha-2D adrenergic receptors in bovine pineal gland. Mol. Pharmacol. 40, 235-241. Smith K., Connaughton, S., & Docherty, J. R. (1992). Investigations ofprejunctional a2-adrenoceptors in rat atrium, vas deferens and submandibular gland. Eur. J. Pharmacol. 211, 251-256. Summers, R. J., & McMartin, L. R. (1993). Adrenoceptors and their second messenger systems. J. Neurochem. 60, 11-21. Szabadi, E., & Bradshaw, C. M. (1987). Alpha- 1 adrenergic receptors in the central nervous system: In The Alpha- 1 Adrenergic Receptor (Ruffolo, R. R., ed.), pp. 405-433, Humana Press, Clifton, NJ. Terzic, A., Pucrat, M., Vassort, G., & Vogel, S. M. (1993). Cardiac al-adrenoceptors: An overview. Pharmacol. Rev. 45, 147-175. Timmermans, P. B. M. W. M., & van Zwieten, P. A. (1982). a2-Adrenoceptors: classification, localization, mechanisms, and targets for drugs. J. Med. Chem. 25, 1390-1401. Toffs, C. B., Tafoya, M. E., Camras, C. B., & Yablonski, M. E. (1995). Effects of apraclonidine on aqueous humor dynamics in human eyes. Ophthalmology 102, 456-461. Tsujimoto, G., Timmins, P. V., & Hoffman, B. B. (1986). ct-Adrenergic receptors in the rabbit bladder base smooth muscle: al-adrenergic receptor mediate contractile responses. J. Pharmacol. Exp. Ther. 236, 384-389. Uhlen, S., & Wikberg, J. E. S. (1991). Delineation of three pharmacological subtypes of a2-adrenoceptors in the kidney. Br. J. Pharmacol. 104, 657-664. Uhlen, S., Xia, Y., Chhajlani, V., Felder, C. C., & Wikberg, J. E. S. (1992). [3H]-MK-912 binding delineates two a2-adrenoceptor subtypes in rat CNS one of which is identical with the cloned pA2d a2-adrenoceptor. Br. J. Pharmacol. 106, 986-995. Voigt, M. M., McCune, S. K., Kanterman, R. Y., & Felder, C. C. (1991). The rat ct2-C4 adrenergic receptor gene encodes a novel pharmacological subtype. FEBS Lett. 278, 45-50. Weinshank, R. L., Zgombick, J. M., Macchi, M., Adham, N., Lichtblau, H., Branchek, T. A., & Hartig, P. R. (1990). Cloning, expression, and pharmacological characterization of a human a2B-adrenergic receptor. Mol. Pharmacol. 38, 681-688. Wilffert, B., Timmermans, P. B. M. W. M., & van Zwieten, P. A. (1982). Extrasynaptic location of alpha-2 and noninnervated beta-2 adrenoceptors in the vascular system of the pithed normotensive rat. J. Pharmacol. Exp. Ther. 221,762-768. Wilson, K. M., & Minneman, K. P. (1989). Regional variations in t~l-adrenergic receptor subtypes in rat brain. J. Neurochem. 43, 1782-1786. Yoshida, M., Latifpour, J., Nishimoto, T., & Weiss, R. M. (1991). Pharmacological characterization of alpha adrenergic receptors in the young and old female rabbit urethra. J. Pharmacol. Exp. Ther. 257, 1100-1108. Zeng, D., Harrison, J. K., D' Angelo, D. D., Barber, C. M., Tucker, A. L., Lu, Z., & Lynch, K. R. (1990). Molecular characterization of a rat Ct2B-adrenergic receptor. Proc. Natl. Acad. Sci. USA 87, 3102-3106.

Biomembranes

A Multi-Volume Treatise Edited by A.G. Lee, Department of Biochemistry, University of Southampton "Progress in understanding the nature of the biological membranes has been very rapid over a broad front, but still pockets of ignorance remain. Application of the techniques of molecular biology has provided the sequences of a very large number of membrane proteins, and led to the discovery of superfamilies of membrane proteins of related structure. In turn, the identification of these supeffamilies has led to new ways of thinking about membrane processes. Many of these processes can now be discussed in molecular terms, and unexpected relationships between apparently unrelated phenomena are bringing a new unity to the study of biological membranes. The quantity of information available about membrane proteins is now too large for any one person to be familiar with anything but a small part of the primary literature. A series of volumes concentrating on molecular aspects of biological membranes therefore seems timely. The hope is that, when complete, these volumes will provide a convenient introduction to the study of a wide range of membrane functions." n From the Preface

Volume 1, General Principles 1995, 279 pp. ISBN 1-55938-658-8

$97.50

CONTENTS: Preface, A.G. Lee. The Functional Role of Lipids in Biological Membranes, David B. Fenske, Myrna A. Monck, Michael J. Hope, and lan R. Cullis. Principles of Membrane Protein Structure, M.S.P. Sansom and lan D. Kerr. Fatty Acidand Isoprenoid-Linked Membrane Proteins, Marco Parenti and Anthony I. Magee. The Biosynthesis of Membrane Proteins, David Stephens, Sunita Kulkami, and Brian Austen. Specificity of Lipid-Protein Interactions, Derek Marsh. Effects of Lipid-Protein Interactions on Membrane Functions, A.G. Lee. General Principles of Membrane Transport, lan C. West. Membrane Signaling Systems, C.U.M. Smith.

JAI PRESS INC.

55 Old Post Road No. 2 - P.O. Box 1678 Greenwich, Connecticut 06836-1678 Tel: (203) 661- 7602 Fax: (203) 661-0792

.1 A_

P R E S S

J A l P R E S S

Advances in Developmental Biochemistry Edited by Paul Wassarman, Department of Cell and Developmental Biology, Roche Institute of Molecular Biology, Volume 1, 1992, 256 pp. ISBN 1-55938-347-X

$97.50

CONTENTS: Introduction. Organelle Assembly and Function

in the Amphibian Germinal Vesicle, Joseph G. Gall. DNA Replication and the Role of Transcriptional Elements During Animal Development, Melvin L. DePamphilis. Transcriptional Regulation During Early Drosophila Development, K. Prakash, Joanne Topoi. C.R. Dearolf, and Carl S. Parker. Translational Regulation of Maternal Messenger RNA, L. Dennis Smith. Gut Esterase Expression in the Nematode Caenorhabditis elegans, James D. McGhee. Transcriptional Regulation of Crystallin Genes: Cis Elements, Trans-factors and Signal Transduction Systems in the Lens, Joram Piatigorksy and Peggy S. Zelenka. Subject Index. Volume 2, 1993, 237 pp. ISBN 1-55938-609-6

$97.50

CONTENTS: Preface, Paul M. Wassarman. Drosophila Ho-

meobox Genes, William McGinnis. Structural and Functional Aspects of Mammalian HOX Genes, Denis Duboule. Developmental Control Genes in Myogenesis of Vertebrates, Hans Henning-Amold. Mammalian Fertilization: Sperm Receptor Genes and Glycoproteins, Paul M. Wassarman.The Fertilization Calcium Signal and How It Is Triggered, Michael Whitaker. Subject Index. Volume 3, 1994, 188 pp. ISBN 1-55938-853-6

$97.50

CONTENTS: Preface, Paul M. Wassarman. Expression and

Function of Protein Kinases During Mammalian Gametogenesis, Deborah L. Chapman and Debra J. Wolgemuth. Regulation of the Dopa Decarboxylase Gene During Drosophila Development, Martha J. Lundell and Jay Hirsh. Transcription Factors in Mammalian Development: Murine Homeobox Genes, S. Steven Potter. Expression and Function of C-Mos in Mammalian Germ Cells, Geoffrey M. Cooper. Regulation of Pigmentation During Mammalian Development, Friedrich Beermann, Ruth Ganss, and Gunther Schutz. Index.

Advances in Developmental Biology Edited by Paul Wassarman, Department of Cell and Developmental Biology, Roche Institute of Molecular Biology Volume 1, 1992, 192 pp. ISBN 1-55938-348-8

$97.50

CONTENTS: Introduction. Y Chromosome Function in Mammalian Development, Paul S. Burgoyne. A Super Family of Putative Developmental Signalling Molecules Related to the Proto-Oncogene Wnt-1/int-1, Andrew P. McMahon and Kenneth R. Howard. Pattern Formation in Caenorhabditis Elegans, Min Han and Paul W. Stemberg. Gap Junctional Communication During Mouse Development, Norton B. Gilula, Miyuki NishL and Nalin Kumar. Lens Differentiation and Oncogenesis in Transgenic Mice, Heiner WestphaL Subject Index.

Volume 2, 1993, 195 pp. ISBN 1-55938-582-0

$97.50

CONTENTS: Preface, Paul M. Wassarman. The Sry Gene and Sex Determination in Mammals, Blanche Capel and Robin LovelI-Badge. Molecular and Genetic Studies of Human X Chromosome Inactivation, Carolyn J. Brown and Huntington F. Willard. Genomic Imprinting and Regulation of Mammalian Development, Colin L. Stewart. Cell Interactions in Neural Crest Cell Migration, Marianne Bronner-Fraser. Enzymes and Morphogenesis: Alkaline Phosphatase and Control of Cell Migration, Saul L. Zackson. Subject Index.

Volume 3, 1994, 194 pp. ISBN 1-55938-865-X

$97.50

CONTENTS: Preface, Paul M. Wassarman. Mechanisms of Neurogenesis in Drosophila melanogaster, Jose A. CamposOrtega. The Role of Growth Factors in Mammalian Pregastrulation Development, Daniel A. Rappolee and Zena Werb. Retinoid Signaling in Mouse Embryos, Elwood Linney and Anthony-Samuel LaMantia. RNA Localization During Oogenesis in Drosophila, Elizabeth R. Gavis and Ruth Lehmann. Actin as a Tissue-Specific Marker in Studies of Ascidian Development and Evolution, William R. Jeffery. Index.

.1 A [ P R

S S

J A l

Advances In Cell and Molecular Biology of Membranes and Organelles (Previosly published as Advances In Cell and Molecular Biology of Membranes)

Edited by Alan M. Tartakoff, Institute of Pathology, Case Western Reserve University Volume 1, Endosomes and Lysosomes: A Dynamic Relationship 1993, 434 pp. ISBN 1-55938-362-3

$97.50

Edited by Brian Storrie, Department of Biochemistry and Nutrition, Virginia Polytechnic Institute and State University and Robert F. Murphy, Department of Biological Sciences, Camegie Mellon University

P R E S S

CONTENTS: Preface, Brian Storrie and Robert F. Murphy. Models of Endosome and Lysosome Traffic, Robert F. Murphy. Endocytic Receptors, Michael G. Roth. Functions of the Mannose 6-Phosphate Receptors, Bemard Hoflack and Peter LobeL Chemistry of Lysosomal Cysteine Proteinases, Robert W. Mason and Donna Wilcon. Mechanism and Regulation of Autophagic Degradation of Cellular Proteins, William A. Dunn, Jr.. Cell-Free Systems for Endocytosis, William A. Braell. Genetic Analysis of Membrane Traffic in Mammalian Cells, Penelope A. Colbaugh and Rockford K. Draper. Plasma Membrane Lipid Transport in Cultured Cells: Studies Using Lipid Analogs and Model Systems, Michael KovaL Endosomes, Lysosomes, and Trans-Golgi-Related Systems in Conventional Neurons and the Grof Retina: Shards and Suppositions, Eric Holtzman, Eliene Augenbraun, Robert St. Jules, and Maria Santa-Hernandez. The Role of Endocytosis in Epidermal Growth Factor Signaling, Bryan K. McCune, William R. Huckle, and H. Shelton Earp. Membrane Traffic Through the Late Stages of the Yeast Secretory, Eric A. Whitters, Henry B. Skinner, and Vytas A. Bankaitis. Regulation of Lysosomal Trafficking and Function During Growth and Development of Dictyostelium discoideum, James A. Cardelli. Towards an Understanding of the Inheritance of Mammalian Lysosomes and Yeast Vacuoles, Brian Storrie. Volume 2, Membrane Transport in Protozoa 1993, 483 pp. 2 Part Set Set ISBN 1-55938-628-2 Edited by Helmut Plattner, Fakult~t f(~rBiologie, Universit&t Konstanz

$195.00

PART A - CONTENTS: Preface. Involvement of the TransGolgi Network, Coated Vesicles, Vesicle Fusion and Secretory Product Condensation in the Biogenesis of Pseudomicrothorax Trichocysts, Robert K. Peck, Barbara Swiderski and Anne-Marie TourmeL Early Steps of the Secretory Pathway in Paramecium: Ultra-structural, Immunocyto-Chemical and Genetic Analysis of Trichocyst Biogenesis, Nicole Garreau de Loubresse. Calcium and Trichocyst Exocytosis in Paramecium: Genetic and Physiological Studies, Jean Cohen and Daniel Kerboeuf. Exocytotic Events During Cell Invasion by Apicomplexa, Jean Francois Dubremeta and Roff Entzeroth. Pathways of Lysosomal Enzyme Secretion in Tetrahymena, Amo Tiedtke, Thomas Kiy, Christian VosskEihlerand Leif Rasmussen. Synchronization of Different Steps of the Secretory Cycle in Paramecium Tetraurelia: Trichocyst Exocytosis, Exocytosis-Coupled Endocytosis and Intracellular Transport, Helrout Plattner, Gerd Knoll, and Regina Pape. The Ciliary Membrane and its Engagement In Conjugation, Jason Wolfe. Ciliary and Plasma Membrane Proteins in Paramecium: Description, Localization and Intracellular Transit, Yvonne Capdeville, Ren6e Charret, Claude Anthony, Julienne Delorme, Pierre Nahon and Andre. Adoutte. PART B - - CONTENTS: Endocytosis and Intracellular Trans-

port of Variant Surface Glycoproteins in Trypanosomes, Michael Duszenko and Andreas Seyfang. A Comparative Survey on Phagosome Formation in Protozoa, Klaus Hausmann and Renate Radek. Endosomal Membrance Traffic of Ciliates, Richard D. Allen and Agnes K. Fok. Membrane Flow in the Digestive Cycle in Paramecium, Agnes K. Fok and Richard D. Allen. Signal Coupling During Endocytosis in Amoeba proteus, Robert D. Prusch. Membrane Recycling and Turnover in Large, Free-Living Amoebae, Kwang W. Jeon. Food Uptake and Digestion in Amoebae, Wilhelm Stockem and Melpo Christofidou-Solomidou. The Lysosomal System in Malaria Parasites, Christian Slomianny. Membrane and Microtubule Dynamics in Heliozoa, Toshinobu Suzaki and Yoshinobu Shigenaka. The Host-Symbiont-lnterface in Ciliate-Algae Associations: Inhibition of Membrane Fusion, Wemer Reisser. Lipid Composition of Membranes Involved in Membrane Traffic in Tetrahymena, Shigenobu Umeki and Yoshinori Nozawa. Volume 3 Signal Transduction Through Growth Factor Receptors

1994, 223 pp. ISBN 1-55938-344-5

$97.50

Edited by: Yasuo Kitagawa, BioSciences Center, Laboratory of Organogenesis, Nagoya University and Ryuzo Sasaki, Faculty of Agriculture, Department of Food Science and Technology, Kyoto University

.1 A l P E S S

J A l

CONTENTS: The Hepatocyte Growth Factor/c-MET Signaling Pathway, D. P. Bottario, A. M.-L. Chan, J. S. Rubin, E. Gak, E. Fortney, J. , Schind/er, M. Chedid, and S., A. Aaronson. Insulin Receptor, Y. Ebina, H. Hayashi, F. Kanai, S. Kamohara, and Y. Nishioka. Interleukin-3 Receptor: Structure and Signal Transduction, T. Kitamura, and A. Miyajima. Interleukin-5 Receptor, K. Takatsu. Interleukin-6 Receptor and Signal Transduction, T. Matsuda, T. Nakajima, T. Kaisho, K. Nakajima, and 7. Hirano. Receptor for Granulocyte ColonyStimulating Factor, S. Nagata, & R. Fukunaga. Receptor for Granulocyte/Macrophage Colony-stimulating Factor, K. Kurata, T. Yokota, A. Miyajima, & K. Arai. Perspectives On The Structure And Mechanisms of Signal Transduction by The Erythropoietin Receptor , S. S. Jones. Interleukin-1 Signal Transduction, J. E. Sims, T. A. Bird, J. G. Giri, and K. S. Dower. Volume 4, Protein Export and Membrane Biogenesis 1995, 276 pp. $97.50 ISBN 1-55938-924-9

P R E

S S

Edited by Ross E. Dalbey, Department of Chemistry, The Ohio State University CONTENTS: Introduction to the Series, Alan M. Tartakoff. Preface, Ross E. Dalbey. Membrane Protein Assembly, Paul Whitley and Gunnar yon Heijne. Membrane Insertion of Small Proteins: Evolutionary and Functional Aspects, Dorothee Kiefer and Andreas Kuhn. Protein Translocation Genetics, Koreaki Ito. Biochemical Analyses of Components Comprising the Protein Translocation Machinery of Escherichia coil, Shin-ichi Matsuyama and Shoji Mizushima. Pigment Protein Complex Assembly in Rhodobacter sphaeroides and Rhodobacter capsulatus, Amy R. Vargas and Samuel Kaplan. Identification and Reconstitution of Anion Exchange Mechanisms in Bacteria, Atul Varadhachary and Peter C. Maloney. Helix Packing in the C-Terminal Half of Lactose Permease, H. Ronald Kaback, Kirsten Jung, Heinrich Jung, Jianhua Wu, Gilbert C. Prive, and Kevin Zen. Export and Assembly of Outer Membrane Proteins in E. coil, Jan Tommassen and Hans de Cock. StructureFunction Relationships in the Membrane Channel Porin, Georg E. Schulz. Role of Phospholipids in Escherichia coil Cell Function, William Dowhan. Mechanism of Transmembrahe Signaling in Osmoregulation, Alfaan A. Rampersaud. Index.

JAI PRESS INC.

55 Old Post Road No. 2 - P.O. Box 1678 Greenwich, Connecticut 06836-1678 Tel: (203) 661- 7602 Fax: (203) 661-0792