EDITORIAL Cause and effect Identifying the causative agent of an infectious disease is a notoriously tricky, but vitally important, business. When faced with new potential pathogens, the principles fostered by Koch’s postulates are as relevant today as they were over a century ago.
…Koch’s postulates, and the scientific rigour they foster, still provide an invaluable service in efforts to prove that a microorganism is the cause of a disease.
All microbiologists interested in infectious diseases are aware of the various statistics used to emphasize the threat microbial infection poses to human health, statistics that are regurgitated daily for conference audiences, funding bodies and journalists. One particularly interesting statistic, however, states that new infectious diseases are estimated to be emerging at the rate of one per year. In light of the potential ramifications of this statistic for global public health, a relevant question that requires consideration is: what constitutes a new infectious disease? In the early days of microbiology, the attention of scientists and clinicians was primarily focused on diseases that presented with ‘textbook’ clinical symptoms and signs, that is, easily recognizable and distinguishable conditions. Many of the infectious diseases that preoccupy today’s practitioners do not possess such obvious hallmark symptoms or signs. Patients can remain asymptomatic for decades, but these individuals act as reservoirs for some of our most deadly infections. Modern clinical advances, including transplantationrelated immunosuppression, have blurred the distinction between commensals and pathogens, and this fuzziness prevents a consensus being reached about what actually constitutes a pathogen1. These issues aside, more than 120 years after they were first proposed, Koch’s postulates still remain the gold standard for any investigation that sets out to prove the aetiology of an infectious disease. Although advances in microbiology since the nineteenth century have led to modern reintepretations2,3, Koch’s postulates, and the scientific rigour they foster, still provide an invaluable service in efforts to prove that a microorganism is the cause of a disease. Take a recent case in point. Patients with chronic granulomatous disease (CGD) are subject to recurrent infections with a range of bacterial and fungal pathogens. In many cases, however, patients present with clinical symptoms that are indicative of an infection for which no obvious pathogen can be identified. In these cases, is disease caused by a known microorganism that eludes detection, or is a novel pathogen to blame? A recent study published in PLoS Pathogens4 describes the story of one CGD patient suffering from recurring lymphadenitis of unknown aetiology. The paper reports on the isolation and characterization of a new bacterium,
414 | JUNE 2006 | VOLUME 4
Granulobacter bethesdensis, from the lymph nodes of this patient. More importantly, by applying Koch’s postulates with commendable rigour, the authors were able to convincingly establish a causal link between this novel organism and the infection in the patient — the first reported case of an invasive human disease caused by the Acetobacteraceae. Another recent example provides more food for thought. Infection of the root canals of human teeth (commonly polymicrobial in nature), can lead to inflammation and the destruction of the periradicular tissues. A paper recently published in the Journal of Clinical Microbiology reports, for the first time, the detection and identification of an archaeal phylotype in a proportion of infected root canals5. This is an intriguing finding because although Archaea are recognized as a component of the human microbiota, it is generally assumed that this domain does not cause human disease. In this example, however, does an association of Archaea with an infected site prove that these organisms are actively contributing to disease? Are any of Koch’s postulates fulfilled? By demonstrating an association between Archaea and infection, these findings certainly support a hypothesis that members of this domain can be human pathogens. An association, however, represents only one step on the journey to proof, and more research will be required before causality can be conclusively established in this case. In the fight against newly emerging infectious diseases, the accurate identification of the offending aetiological agent will be essential if preventative and therapeutic measures are to be implemented effectively. It will serve the microbiological community well if Koch’s postulates remain the gold standard to aim for when determining microbial cause and effect. 1. Falkow, S. What is a pathogen? ASM News 63, 359–365 (1997). 2. Falkow, S. Molecular Koch’s postulates applied to bacterial pathogenicty — a personal recollection 15 years later. Nature Rev. Microbiol. 2, 67–72 (2004). 3. Fredericks, D. N. & Relman, D. A. Sequence-based identification of microbial pathogens: a reconsideration of Koch’s postulates. Clin. Microbiol. Rev. 9, 18–33 (1996). 4. Greenberg, D. E. et al. A novel bacterium associated with lymphadenitis in a patient with chronic granulomatous disease. PLoS Pathogens 2, e28 (2006). 5. Vianna, M. E. et al. Identification and quantification of archaea involved in primary endodontic infections. J. Clin. Microbiol. 44, 1274–1282 (2006).
www.nature.com/reviews/micro © 2006 Nature Publishing Group
RESEARCH
HIGHLIGHTS I N N AT E I M M U N I T Y
Finding flagellin Bacterial flagellin is a known ligand for the Toll-like receptor TLR5. However, several recent papers have now revealed that in addition to the TLR5 pathway, which responds to extracellular flagellin, host macrophages can respond to cytosolic flagellin through members of the NOD-like receptor (NLR) family. The recognition of pathogenassociated molecular patterns by host TLRs is a key component of innate immunity and much has been learned about TLRs and their signalling pathways over the past decade.
More recently, attention has turned to the role of non-TLRs in innate immunity, including the cytoplasmic NLR family. Details of the NLR signalling pathways are beginning to emerge and NLRs are known to be involved in secretion of the proinflammatory cytokine interleukin1β (IL-1β) by macrophages. IL-1β is produced initially as a zymogen that is activated for secretion by caspase 1. In Salmonella typhimurium infection, the NLR protein Ipaf was known to be involved in caspase 1 activation and IL-1β secretion but, until now, the S. typhimurium ligand for Ipaf was unknown. Two independent groups led by Gabriel Núñez and Alan Aderem investigated the nature of the innate immune response to S. typhimurium infection. Both groups confirmed that Ipaf was required for IL-1β production and caspase 1 activation in macrophages. Additionally, they both found that S. typhimurium mutants that either lack or have mutated flagella did not stimulate caspase 1 activation or IL-1β secretion, suggesting that flagellin is the S. typhimurium ligand for Ipaf. As flagellin is also a known ligand for TLR5, the involvement of TLRs was examined — both groups found that S. typhimurium could stimulate caspase 1 activation and IL-1β secretion in TLR5-deficient macrophages, and in wild-type macrophages, and in addition Franchi et al. found normal levels of caspase 1 activation
and IL-1β secretion in tolerant macrophages that are refractory to TLR stimulation. Taken together, these results suggest that macrophages sense flagellin through a TLR5independent pathway that relies on the cytoplasmic sensor Ipaf. Further confirmation that Ipaf senses flagellin in the cytosol independently of TLR5 comes from the fact that both groups also demonstrated that purified flagellin delivered to the cytosol triggered caspase 1 activation in wild-type but not Ipaf-deficient macrophages. The mechanism by which flagellin accesses the cytosol during infection remains to be completely elucidated, however genetic evidence presented by Miao et al. suggests that it is transferred directly into the host cell cytoplasm by the virulence-associated type III secretion system. These results are echoed by results published recently in two independent papers, one in PLoS Pathogens and one in Journal of Experimental Medicine, which indicate that an NLR is also involved in cytosolic sensing of Legionella pneumophila flagellin through a TLR5-independent, caspase-1-dependent pathway. Sheilagh Molloy ORIGINAL RESEARCH PAPERS Franchi, L. et al. Cytosolic flagellin requires Ipaf for activation of caspase-1 and interleukin 1β in Salmonellainfected macrophages. Nature Immunol. 30 April 2006 (doi:10.1038/ni1346) | Miao, E. A. et al. Cytoplasmic flagellin activates caspase-1 and secretion of interleukin 1β via Ipaf. Nature Immunol. 30 April 2006 (doi:10.1038/ni1344) | Ren, T. et al. Flagellin-deficient Legionella mutants evade caspase 1 and Naip5-mediated macrophage immunity. PLoS Pathogens 3 e18 (2006) | Molofsky, A. B. et al. Cytosolic recognition of flagellin by mouse macrophages restricts Legionella pneumophila infection. J. Exp. Med. 203, 1093–1104 (2006)
RESEARCH HIGHLIGHTS ADVISORS ADRIANO AGUZZI University Hospital of Zürich, Zürich, Switzerland NORMA ANDREWS Yale University School of Medicine, New Haven, CT, USA ARTURO CASADEVALL The Albert Einstein College of Medicine, Bronx, NY, USA
RITA COLWELL University of Maryland Biotechnology Institute, Baltimore, MD, USA STANLEY FALKOW Stanford University School of Medicine, Stanford, CA, USA TIMOTHY FOSTER Trinity College, Dublin, Ireland
NATURE REVIEWS | MICROBIOLOGY
KEITH GULL University of Oxford, Oxford, UK NEIL GOW University of Aberdeen, Aberdeen, UK HANS-DIETER KLENK Philipps University, Marburg, Germany
BERNARD MOSS NIAID, National Institutes of Health, Bethesda, MD, USA JOHN REX AstraZeneca, Cheshire, UK DAVID ROOS University of Pennsylvania, Philadelphia, PA, USA
PHILIPPE SANSONETTI Institut Pasteur, Paris, France CHIHIRO SASAKAWA University of Tokyo, Tokyo, Japan ROBIN WEISS University College London, London, UK
VOLUME 4 | JUNE 2006 | 415
RESEARCH HIGHLIGHTS A DA P T I V E I M M U N I T Y
Private repertoire blues
…it is the unique make-up … of the cross-reactive memory CD8+ T cells … that determines the response to heterologous infection…
The diversity of the mammalian immune system is one of its most powerful attributes. Activation of cell-mediated immunity by foreign antigens induces a diverse range of T cells, each clone bearing a T-cell receptor (TCR) with unique specificity. But sometimes, for reasons that have so far been unclear, the T-cell response to viral infection is restricted, consisting of only a limited range of T-cell clones. New work by Liisa Selin, Raymond Welsh, Markus Cornberg and colleagues reveals that the unique composition of an individual’s memory T-cell pool may lie at the heart of this dysfunctional immune response. The authors suspected that activation of cross-reactive memory T cells might be implicated in generating a restricted T-cell response and they used a mouse model of virus
infection to investigate this further. They first showed that epitopes from lymphocytic choriomeningitis virus (LCMV NP205–212) and Pichinde virus (PV NP205–212) are crossreactive: NP205-specific CD8+ T cells that are activated by primary infection with LCMV or PV produced IFN-γ in response to the heterologous NP205 peptide. Sequence analysis of the Vβ portion of the TCR (a region of the TCR that varies among T cells) showed that NP205-specific CD8+ T cells activated during acute LCMV or PV infection have a broad and diverse TCR repertoire. TCRs from NP205-specific CD8+ T cells obtained from LCMV-infected mice revealed hundreds of different TCR Vβ sequences or clonotypes. Once acute infection with LCMV or PV resolves, a pool of memory CD8+
B AC T E R I A L P H Y S I O LO GY
Precious metal Riboswitches are conformational switches in complex folded RNA domains that are induced by binding to specific small molecules, and which lead to a switch in gene-regulatory function. In a recent issue of Cell, Eduardo Groisman and his team at the Howard Hughes Medical Institute in St Louis, USA, report on the first
example of a riboswitch that senses, and responds to, a metal ion. Magnesium (Mg2+) ions are abundant in biological systems and are required for a wide variety of cellular processes and structures. Cells maintain cytoplasmic Mg2+ levels within a narrow range, but little is known about the mechanisms through which
…the first example of a riboswitch that senses, and responds to, a metal ion.
416 | JUNE 2006 | VOLUME 4
T cells is established, including the cross-reactive NP205-specific CD8+ T cells. The authors went on to show that these cross-reactive T cells can have a profound influence on the memory T-cell repertoire. In PV-immune mice that are infected with LCMV, proliferation of cross-reactive PV NP205-specific memory CD8+ T cells results in a dominant NP205 response. But in contrast to the diverse CD8+ T-cell population induced by NP205 in acute infection, the TCR repertoire of NP205-specific CD8+ T cells after heterologous LCMV challenge is limited and dominated by specific Vβ clonotypes. These results indicate that only a small subset of the cross-reactive NP205-specific CD8+ T cells proliferate after heterologous infection. This oligoclonal TCR repertoire can favour the generation of viral escape mutants as shown by the isolation of an LCMV NP205-variant virus from PV-immune mice 8 months after LCMV challenge. Using adoptive transfer experiments, Cornberg et al. revealed
cells sense and maintain these Mg2+ levels. Salmonella enterica serovar Typhimurium (S. typhimurium) contains the Mg2+-responding PhoP–PhoQ twocomponent regulatory system and several Mg2+ transporters. One of these transporters is MgtA, which mediates Mg2+ influx. The mgtA gene is regulated by the transcriptional activator PhoP in response to extracytoplasmic Mg2+ sensed by the sensor kinase PhoQ. In addition to the PhoP–PhoQ system, it was suspected that mgtA is regulated by another, Mg2+-sensing, system, as mgtA transcription still responded to Mg2+ in an S. typhimurium strain lacking PhoQ and harbouring a Mg2+-independent PhoP mutant. This Mg2+-sensing mechanism is what Cromie et al. set out to investigate. First, they used RT-PCR to examine the levels of mgtA mRNA in wild-type S. typhimurium strains that had been grown in the presence of different Mg2+ concentrations. They found that the mgtA transcript reaches the coding region only when S. typhimurium experiences very low Mg2+ concentrations, despite mgtA transcription being initiated at Mg2+
www.nature.com/reviews/micro
RESEARCH HIGHLIGHTS that it is the unique make-up (or the private specificity) of the crossreactive memory CD8+ T cells of individual mice that determines the response to heterologous infection, although so far the mechanisms that favour the growth of some of these cells over others remain unknown. The cross-reactive profiles of private T-cell repertoires might explain why some patients mount oligoclonal T-cell responses to HIV or hepatitis C virus infection and seem to be prone to the development of viral escape mutants. Finally, these results have important implications for vaccine design, as, in certain individuals, immunizing peptides could generate a pool of cross-reactive memory CD8+ T cells that respond to subsequent heterologous virus challenge with a restricted T-cell repertoire.
IN BRIEF B AC T E R I A L P H Y S I O LO GY
Definition of the bacterial N-glycosylation site consensus sequence Kowarik, M. et al. EMBO J. 13 April 2006 (doi:10.1038/ sj.emboj.7601087)
In eukaryotes, the consensus sequence for N-linked glycosylation is N–X–S/T (where X represents any amino acid except proline). In this work, Kowarik et al. investigated the substrate requirements for bacterial N-glycosylation in vivo and found the Campylobacter jejuni protein glycosylation machinery recognizes the consensus sequence D/E–Y–N–X–S/T (where X and Y can be any amino acid except proline). In addition to having an extended consensus sequence, Kowarik et al. demonstrated that the bacterial glycosylation system also shows increased stringency and specificity compared with eukaryotes. These results have implications for the biotechnological production of glycoproteins in bacteria. VIROLOGY
Role of the α/β interferon response in the acquisition of susceptibility to poliovirus by kidney cells in culture
Shannon Amoils
Yoshikawa, T. et al. J. Virol. 80, 4313–4325 (2006)
ORIGINAL RESEARCH PAPER Cornberg, M. et al . Narrowed TCR repertoire and viral escape as a consequence of heterologous immunity. J. Clin. Invest. 13 April 2006 (doi:10.1172/JCI27804)
concentrations 100-fold higher. Further experiments showed that the 5′-UTR region of mgtA is necessary, and sufficient, to respond to Mg2+. In a next step, Groisman and colleagues used the Mfold program to predict the secondary structure that might be adopted by the mgtA 5′-UTR. They identified two potential mutually exclusive structures: one including two stem-loop structures, A plus B, and an alternative stem-loop structure, C. Using S. typhimurium mutant strains where mgtA transcription initiated at different positions, it was found that stem loop B might be the structure formed when Mg2+ levels are high, and might be responsible for transcription stopping before the mgtA coding region. On the other hand, formation of stem loop C, the alternative to stem loops A plus B, might be favoured when Mg2+ levels are low. Synthesizing the full-length mgtA 5′UTR, and treating the RNA with RNases and chemicals to probe its structure after incubation in the presence of different Mg2+ concentrations, Cromie et al. showed that Mg2+ can modify the structure of the mgtA 5′-UTR.
NATURE REVIEWS | MICROBIOLOGY
Stem loop A was found to be crucial for Mg2+ sensing, and for the Mg2+promoted changes taking place in stem loops B and C. Finally, using an in vitro transcription system, they demonstrated that the mgtA 5′-UTR responds to Mg2+ by affecting the ability of RNA polymerase to stop transcription. So, the expression of MgtA is controlled by Mg2+ in two steps — the initiation of mgtA transcription by PhoP, which is controlled by PhoQ in response to extracytoplasmic Mg2+, and the early stopping of mgtA transcription by its 5′-UTR riboswitch, which responds to cytoplasmic Mg2+ — in contrast to most genes that are controlled by riboswitches, which are typically only regulated by their respective riboswitches. Although this is the first evidence of a cation-responsive riboswitch, it is likely that additional ion-regulated riboswitches exist. Annie Tremp ORIGINAL RESEARCH PAPER Cromie, M. J., Shi, Y., Latifi, T. & Groisman, E. A. An RNA sensor for intracellular Mg2+. Cell 125, 71–84 (2006)
In vivo, poliovirus, the causative agent of poliomyelitis, has a strict neurotropism, replicating at only a few sites including the brain and spinal cord. Fortunately for researchers working on poliovirus, the pioneers of early laboratory techniques for poliovirus cultivation, including John Enders and Renato Dulbecco, discovered that this neurotropism was not strictly observed in vitro and that cell lines of monolayer cultures derived from almost any primate tissue are susceptible to poliovirus infection. However, the molecular basis for the acquisition of poliovirus susceptibility by cultured cells has always been unknown. Now, reporting in the Journal of Virology, Yoshikawa et al. have shown that changes in the antiviral activity of the interferon response are likely to be the most important factors determining the acquisition of poliovirus susceptibility. ANTI-INFECTIVES
Co-expression of virulence and fosfomycin susceptibility in Listeria: molecular basis of an antimicrobial in vitro–in vivo paradox Scortti, M. et al. Nature Med. 23 April 2006 (doi:10.1038/nm1396)
The so-called ‘in vitro–in vivo’ paradox for antibiotics refers to discrepancies between the resistance to a particular compound that is observed using in vitro susceptibility tests and the treatment outcomes that are observed for this compound in the clinic; that is, antibiotics that have poor efficacy in vitro can be successfully used to treat infections in vivo. In this Brief Communication in Nature Medicine, the authors present a molecular mechanism that could explain the paradox in fosfomycin treatment of Listeria monocytogenes infection. In in vitro susceptibility tests L. monocytogenes is resistant to fosfomycin, yet Scortti et al. demonstrate that in mice fosfomycin does have anti-listerial activity. They found that L. monocytogenes uptake of fosfomycin is dependent on the Hpt transporter and its regulator PrfA. PrfA is only weakly expressed under in vitro conditions but is highly induced in vivo, therefore explaining why L. monocytogenes is resistant to fosfomycin in vitro, but not in vivo.
VOLUME 4 | JUNE 2006 | 417
RESEARCH HIGHLIGHTS PA R A S I T O L O G Y
E N V I R O N M E N TA L M I C R O B I O L O G Y
Who’s HOSTing who? How the obligate intracellular pathogen Toxoplasma gondii acquires the nutrients that are necessary for its survival and growth has long been an open question. Reporting in Cell, Isabelle Coppens and colleagues now describe a novel mechanism by which T. gondii gains access to key molecules: sequestration of organelles of the host-cell endocytic pathway. Infection of host cells with T. gondii involves internalization of the parasite into a membrane-bound compartment known as the parasitophorous vacuole, which — in contrast to most other pathogen-containing compartments — does not fuse with host-cell endocytic organelles. T. gondii therefore avoids the degradation that would result from sequential fusion with compartments of the endocytic pathway; however, in this way, it is not exposed to the proteins and lipids, such as cholesterol (for which T. gondii is auxotrophic), that enter the host cell through this pathway. Coppens and colleagues, however, had previously observed that cholesterol is somehow transferred from host endocytic organelles to the parasitophorous vacuole, so they set out to examine the localization of host endocytic organelles in T. gondiiinfected cells, using fluorescence microscopy and transmission electron microscopy. After infection with T. gondii, host endolysosomes were found to accumulate around the parasitophorous vacuole, which was also shown to be surrounded by networks of host microtubules. This cytoskeletal reorganization was found to result in the formation of invaginations — referred to as host organelle-sequestering tubulo-structures (HOST) — of the membrane of the parasitophorous vacuole. The authors propose that the tubular structure of HOST allows them to function as conduits for the delivery of endolysosomes to the parasitophorous vacuole. Furthermore, HOST were found to be coated with parasite proteins, including the granule protein GRA7, which can constrict the HOST conduits, thereby preventing exit of the host organelles and sequestering them within the parasitophorous vacuole. The authors postulate that the contents of these organelles would then undergo hydrolysis, and cholesterol and other important lowmolecular-weight molecules would be released across the membrane of the intact organelles and absorbed by the parasite. Not only is this an unexpected finding of a novel process of nutrient acquisition by a parasite, but it is also a unique mechanism for the unidirectional transport and sequestration of host-cell organelles. Davina Dadley-Moore ORIGINAL RESEARCH PAPER Coppens, I. et al. Toxoplasma gondii sequesters lysosomes from mammalian hosts in the vacuolar space. Cell 125, 261–274 (2006) FURTHER READING Gruenberg, J. & and Gisou van der Goot, F. Toxoplasma: Guess who’s coming to dinner. Cell 125, 226–228 (2006) WEB SITE Isabelle Coppens’s homepage: http://faculty.jhsph.edu/?F=Isabelle&L=Coppens
418 | JUNE 2006 | VOLUME 4
A global unculture… Two reports recently published in Nature shed new light on two processes only recently identified as being major contributors to global carbon and nitrogen cycling. Both processes — the anaerobic oxidation of ammonium and methane — were thought to be non-existent in nature but are, in fact, catalysed by unrelated microorganisms that have yet to be grown in pure culture. The obligate anaerobic ammonium oxidation (anammox) reaction, which uses nitrite as the primary electron acceptor, is
accountable for up to 50% of oceanic nitrogen loss. Although identified in 1999, the planctomycete-like bacteria that catalyse this process have proven difficult to study owing to their extremely slow growth. Now, Strous et al. have applied the advantages of environmental genomics to gain insight into the unusual biology of these important microorganisms. Anammox bacteria have a unique prokaryotic organelle (the ‘anammoxosome’) surrounded by ladderane lipids that contain hydrazine oxidoreductase,
The photograph shows Marc Strous extracting a mud sample from the Twentekanaal, a canal in the Netherlands. These anoxic sediments were used as the inoculum for the enrichment of an anaerobic methane-oxidizing microbial consortium. Photograph courtesy of Radboud University Nijmegen.
F U N G A L PAT H O G E N E S I S
Global control Researchers have identified the regulator of the phase transition from the non-pathogenic hyphal form to the pathogenic yeast form in dimorphic fungal pathogens, according to a paper published recently in Science. Six dimorphic ascomycetes including Blastomyces dermatitidis and Histoplasma capsulatum are typically found in the environment but can cause disease in humans if spores are inhaled. Temperature is the key environmental cue for fungal dimorphic switching, with the transition occurring with a shift from 25°C to 37°C, but the key outstanding question in the field has been what regulates this phase transition. In this work, Julie Nemecek, Marcel Wüthrich
and Bruce Klein set out to identify the master regulator that controls the switch from the hyphal to the yeast form. Previous work in the Klein laboratory had shown that Agrobacterium tumefaciens DNA could be transformed randomly into the B. dermatitidis genome and so Nemecek et al. devised an insertional mutagenesis assay in which A. tumefaciens DNA was transformed into a B. dermatitidis reporter strain. Blue/white selection identified seven transformants and the transformant chosen for further analysis had an 86% reduction in transcription of the yeast-phasespecific reporter gene (BAD1) compared with the parental strain. Further analysis of the mutant
www.nature.com/reviews/micro
RESEARCH HIGHLIGHTS an enzyme that combines nitrite and ammonia in a one-to-one mechanism. Using an anoxic laboratory bioreactor that contained a complex microbial community dominated by the anammox bacterium Kuenenia stuttgartiensis, the authors sequenced over 1 gigabase of extracted DNA. These sequences data allowed the researchers to assemble the genome of K. stuttgartiensis, one of the first complete genome sequences of an organism not available in pure culture. Analysis of the genome revealed the presence of over 200 genes directly involved in anammox catabolism and respiration, as well as candidate genes responsible for ladderane biosynthesis and biological hydrazine metabolism. This large number of genes is also indicative of an unexpected degree of metabolic versatility in the microorganism. Environmental sequence analysis was also used to analyse the genomes of a microbial consortium that coupled the direct oxidation of methane to denitrification of nitrate in the absence of oxygen. In this separate study, Marc Strous and colleagues enriched for microbial life derived from the anoxic sediments of a freshwater canal over a
16-month period. Analysis of the culture revealed the presence of two different types of microorganism, a bacterium belonging to a novel phylum without any documented cultured species, and an archaeon distantly related to marine methanotrophic archaea. Interestingly, similar sequences were also identified in freshwater ecosystems from different global locations, strongly suggesting that the process of microbial anaerobic methane oxidation coupled to denitrification has worldwide ecological significance. The contribution of microorganisms to global biogeochemical cycles is well established; however, important aspects pertaining to their biology have remained ill-defined or overlooked. These two studies represent significant advances in addressing these gaps and clearly show the power of genomics in elucidating the biology of environmentally important, yet ‘unculturable’, microorganisms.
phenotype revealed a pleiotropic set of defects, including the inability to undergo the switch from the hyphal to the yeast form. The authors went on to map the site of insertion in the B. dermatitidis genome and identified the disrupted gene as an open-reading frame (ORF) encoding a 1274-residue protein. Sequence analysis revealed that this ORF encodes a hybrid histidine kinase with homology to the Saccharomyces cerevisiae histidine kinase SLNI. Knocking out the ORF by allelic replacement induced the same pleiomorphic defects as observed in the original transformant and so the authors named the gene DRK1, for dimorphism-regulating histidine kinase. In addition to having a homologue in the S. cerevisiae genome, DRK1 is also conserved in the other dimorphic fungal pathogens for which extensive sequence information is available, including H. capsulatum and Coccidioides posadasii. RNA interference was used to silence DRK1
expression in B. dermatitidis and H. capsulatum and it was found that in H. capsulatum, as in B. dermatitidis, DRK1 is involved in dimorphic switching and virulence-gene expression. Finally, in vivo analysis showed that the pathogenicity of DRK1-silenced strains of both B. dermatitidis and H. capsulatum was reduced in murine models of infection. This study represents a milestone in research for those interested in dimorphic fungal pathogens as it not only identifies DRK1 as a key global regulator that controls the hyphalto-yeast transition, virulence-gene expression and pathogenicity but also finally confirms that the transition to the yeast form is a requirement for pathogenicity. Sheilagh Molloy
NATURE REVIEWS | MICROBIOLOGY
David O’Connell ORIGINAL RESEARCH PAPERS Strous, M. et al. Deciphering the evolution and metabolism of an anammox bacteria from a community genome. Nature 440, 790–794 (2006) | Raghoebarsing, A. A. et al. A microbial consortium couples anaerobic methane oxidation to denitrification. Nature 440, 918–921 (2006)
ORIGINAL RESEARCH PAPER Nemecek, J. C., Wüthrich, M. & Klein, B. S. Global control of dimorphism and virulence in fungi. Science 312, 583–588 (2006)
Immuno-electron microscopy showing non-fusogenic disruption of the EEV outer membrane, which exposes the IMV particle to the cell surface. Image reproduced with permission from Law et al. © (2006) National Academy of Sciences, USA.
VIROLOGY
A novel state of undress Enveloped viruses penetrate target cells by fusing their single lipid membrane with the target cell membrane and releasing the infective naked viral core. But some virus forms such as the extracellular enveloped virus (EEV) of Vaccinia virus have two lipid membranes, and therefore cell entry requires the removal of both lipid barriers. New work from Geoffrey Smith’s laboratory at Imperial College London has uncovered a unique way in which EEV sheds its outer lipid membrane, providing novel insights into virus entry and also a new target for antiviral therapy. The authors studied the binding and entry of EEV to cells by immuno-electron microscopy and observed that the outer EEV membrane ruptures at the site of cell contact. This disruption, however, takes place in the absence of membrane fusion and the shed EEV outer membrane remains draped over the underlying single-enveloped virion, which is called an intracellular mature virus (IMV). The single membrane of IMV subsequently fuses with the target cell membrane and the virus core enters the cell. Because disruption of the EEV outer membrane takes place only on contact with a cell, and not a synthetic substrate, the authors proposed that outer membrane dissolution required the interaction between ligands on the virus and the target cell surface. A series of experiments identified the virus glycoproteins and, and cellular surface polyanions, known as glycosaminoglycans, as the required ligands. This is the first example of the removal of a virus membrane without fusion, which Smith and colleagues term ligand-dependent nonfusogenic viral membrane dissolution. The authors went on to show that a soluble polyanion such as heparin can be used to rupture the EEV outer envelope in vitro, exposing the IMV and allowing neutralization by antiIMV monoclonal antibodies. The combined administration of polyanions and anti-IMV antibodies protected mice against disease with Vaccinia virus, demonstrating the therapeutic potential of this strategy. The EEV virion is the form of Vaccinia virus that spreads within the host and therefore an increased understanding of its entry mechanism and the identification of a strategy to target EEV entry are important advances in this field. Shannon Amoils ORIGINAL RESEARCH PAPER Law, M. et al. Ligand-induced and nonfusogenic dissolution of a viral membrane. Proc. Natl Acad. Sci. USA 11, 5989–5994 (2006)
VOLUME 4 | JUNE 2006 | 419
N E W S & A N A LY S I S DISEASE WATCH | IN THE NEWS Gut feeling Signalling through Toll-like receptors (TLRs) is a key component of the innate immune response, but the precise role of TLR-based innate recognition at mucosal surfaces such as the gastrointestinal tract is difficult to assess. In a recent paper in Journal of Experimental Medicine, Mathias Hornef and colleagues have begun to address this by investigating the role of lipopolysaccharide (LPS) signalling through the TLR4–MD2 complex, using an assay involving primary intestinal epithelial cells (IECs) from fetal, newborn and adult mice. They found that although all cells examined expressed the TLR4 complex, the ability to respond to agonists varied with the developmental stage. Only fetal IECs were able to activate the signalling pathways downstream of TLR4 when challenged with LPS, interleukin-1β or tumour necrosis factor; newborn IECs spontaneously released inflammatory cytokines, but this was not enhanced by the presence of an agonist, and adult IECs were non-responsive. Further investigations comparing vaginally born mice with mice born by caesarean section revealed that the spontaneous activation of newborn IECs and concomitant acquisition of LPS resistance requires vaginal delivery and oral exposure to LPS. The authors conclude that: “This adaptive processing might be crucial to facilitate postnatal microbial colonization and subsequent development of the astonishingly stable, lifelong symbiosis.” JEM
World health report card As part of the 2006 World Health Day, the WHO have published their annual World Health Report. The theme of this year’s report was ‘working together for health’ and it highlighted the fact that although the global population is growing, the number of healthcare workers is not increasing sufficiently to allow national healthcare systems to either provide essential medical interventions such as immunization and access to treatment for HIV/AIDS, malaria and tuberculosis, or to respond effectively to a healthcare crisis (such as SARS or avian influenza). In fact, there is a chronic shortage of healthcare workers — including doctors, nurses and midwives — in 57 countries worldwide, including 36 countries in sub-Saharan Africa. More than 4 million additional workers are needed to plug the gap, according to the report, which includes
a 10-year plan to address the crisis. WHO Assistant Director-General Timothy Evans also commented specifically on the impact of the ‘brain drain’ of qualified healthcare professionals from developing countries, warning that industrialized nations “are likely to attract even more foreign staff because of their aging populations, who will need more long-term care”. The report recommends that 50% of all new donor funds for health should be spent on strengthening healthcare systems. WHO
untreated, can lead to irreversible blindness. The WHO recommends that trachomaaffected countries implement a national strategy based on the SAFE principles to eliminate the disease; SAFE involves eyelid surgery; antibiotics; facial cleanliness and environmental changes. Last month’s In the News reported on the successful results of a clinical trial on the use of a single oral dose of azithromycin as a treatment for the disease, and the free availability of azithromycin forms a core part of the SAFE strategy. WHO
Contact lens warning Post exposure Marburg vaccine?
The US FDA and CDC have issued an alert to healthcare professionals and their patients about the risk of corneal fungal infections in wearers of soft contact lenses. As of the beginning of May, almost 200 cases of suspected Fusarium corneal infection were being investigated. US company Bausch & Lomb voluntarily suspended sales of its lens cleaner ReNu With MoistureLoc in April after it was linked with the outbreak, but no direct link has yet been proven. However, news that a Fusarium outbreak in Hong Kong at the end of 2005 had also been linked with the Bausch & Lomb solution prompted some criticism that the company had not acted quickly enough. As we went to press, news had begun to emerge that Fusarium infections had also been reported in Europe. CDC
More trachoma progress At their 10th meeting held recently in Geneva, The Alliance for the Global Elimination of Blinding Trachoma by the Year 2020 (GET2020) reviewed the encouraging progress that has been made towards their goal of eliminating trachoma by the year 2020. Trachoma is associated with Chlamydia trachomatis infection of the eye and, if left
420 | JUNE 2006 | VOLUME 4
The results of further work by Heinz Feldmann and colleagues on their live attenuated vaccine against the filovirus Marburg virus were reported recently in the Lancet. The work suggests that the vaccine, which is based on recombinant vesicular stomatitis virus (rVSV), could be used prophylactically after exposure to the virus, in addition to providing protection before exposure. In an efficacy assessment study carried out in a rhesus-macaque model of Marburg haemorrhagic fever, eight animals received an intramuscular dose of 1,000 plaque-forming units (pfu) of Marburg virus. Twenty to thirty minutes later, five animals received 1 x 107 pfu of the rVSVbased vaccine, given as four intramuscular injections in different anatomical sites, and the three control animals were treated with the vector alone. All of the five animals that received the vaccine survived for at least 80 days; by contrast, the three animals in the control group all died by day 12. The authors suggest that the 20–30-minute time interval between exposure to the virus and administration of the vaccine would be sufficient to respond to accidental needlestick exposure in a laboratory or healthcare situation. Lancet
Two new kids on the block Chronic granulomatous disease (CGD) is an inherited primary immune deficiency that is caused by a defect in the NADPH-oxidase complex in phagocytic cells, and CGD patients are therefore more susceptible to infection with various catalase-producing microorganisms, including Staphylococcus aureus and Aspergillus spp. In addition, patients often present with symptoms, such as inflammation of the lymph nodes, for which the causative infectious
www.nature.com/reviews/micro © 2006 Nature Publishing Group
organism cannot be identified; in such cases, broad-spectrum antimicrobials are usually given to cover all options. Now, Greenberg et al. report in PLoS Pathogens on the identification of a new Gram-negative bacterium associated with lymph-node inflammation in a CGD patient. The organism — named Granulobacter bethesdensis — is a new genus and species within the Acetobacteraceae, and is the first member of this family to be associated with invasive disease. Since the paper was accepted for publication, the researchers have isolated G. bethesdensis from two additional CGD patients. In a separate publication in a recent issue of Journal of Clinical Microbiology, Archaea have been linked to infectious diseases for the first time, with the identification of a Methanobrevibacter oralis-like archaeal phylotype in human endodontic infection. Eurekalert
Genetic test for HCV prognosis Up to 20% of individuals infected with the hepatitis C virus (HCV) go on to develop cirrhosis, a chronic condition that can result in liver failure or hepatocellular carcinoma. It has long been known that the risk of developing cirrhosis is associated with clinical risk factors, such as being male, being over 40 years of age at the time of infection, and alcohol abuse, but genetic risk factors have been difficult to find. Now, Celera have presented data which indicate that a genetic test to predict which individuals infected with HCV will go on to develop cirrhosis could be available in the not-toodistant future. Speaking at an international conference for liver researchers in April, the company announced the results of a study to identify genetic predictors that involved ~1,500 HCV-infected individuals and has identified a total of seven single-nucleotide
polymorphisms that can predict the risk of developing cirrhosis more accurately than the existing clinical risk factors. Celera press release
Leishmaniasis vaccine update There are four main forms of leishmaniasis, ranging from the relatively mild cutaneous leishmaniasis to visceral leishmaniasis, which can be fatal if left untreated and which affects an estimated 500,000 individuals annually, with 60,000 deaths. All leishmaniases are caused by parasitic protozoa belonging to the genus Leishmania, which are transmitted by phlebotomine sandflies and, so far, there is no effective vaccine for any of the leishmaniases. Progress does, however, seem to have been made recently, according to a paper in ACS Chemical Biology that featured heavily on the newswires. Carbohydrate epitopes are becoming increasingly attractive for vaccine formulations, but require a suitable carrier and adjuvant. Immunostimulating reconstituted influenza virosomes (IRIVs) combine both these properties and have already been used successfully in vaccine construction. Now, IRIV technology has been used to create a leishmaniasis vaccine formulation with the cap tetrasaccharide from the Leishmania donovani lipophosphoglycan, and preliminary results in mice show that a strong protective response was elicited. Eurekalert
Microbicides in the news Microbicides moved higher up the news agenda towards the end of April, as reports came in from Microbicides 2006 in Cape Town — the first time that this biennial conference has been held in Africa. Stories highlighted included the results of Phase I and Phase II clinical trials of new microbicide products, and there was also a focus on the need for research into the acceptability of microbicide use.
Avian influenza Human cases of infection with the H5N1 virus continued to be reported throughout April, in countries including Egypt, China and Indonesia, and the total number of confirmed cases in 2006 by 5 May was 62, with 38 deaths. In terms of the general media however, these cases received
NATURE REVIEWS | MICROBIOLOGY
comparatively little attention and instead news items focused mainly on efforts to develop a vaccine to be used in the event of a pandemic, with the US Federal Government announcing it was awarding $1 billion in contracts to five pharmaceutical companies for vaccine development. Speaking at an avian influenza conference influenza expert Robert Webster commented that the H5N1 virus is the worst flu virus he’s seen. In the United Kingdom, there was initial alarm when avian influenza was detected on a poultry farm in Norfolk. The 35,000 chickens on the farm were slaughtered as a precaution, but tests have confirmed that the virus was a lowpathogenicity H7N3 strain. Promed Mail.
Outbreak news Mumps. A mumps outbreak in the Midwest of America has been hitting the headlines recently as it is the worst outbreak of mumps in the country for 20 years. The state of Iowa has been badly hit, with the number of recorded cases reaching more than 1,500. Mass immunization clinics for 18–22-year olds have been organized. Cholera. A cholera outbreak in Angola that began in February has been exacerbated by recent heavy rains, and more than 1,000 deaths had been reported by the beginning of May, with more than 20,000 infections, according to Médicins san Frontières (MSF). The head of the MSF mission in Angola commented that the outbreak had yet to reach its peak and that the number of deaths could double. In the News was compiled with the assistance of David Ojcius, University of California, Merced, USA. David’s links to infectious disease news stories can be accessed on Connotea (http://www. connotea.org), under the username ojcius.
VOLUME 4 | JUNE 2006 | 421 © 2006 Nature Publishing Group
REVIEWS Listeria monocytogenes: a multifaceted model Mélanie Hamon*‡§, Hélène Bierne*‡§ and Pascale Cossart*‡§
Abstract | The opportunistic intracellular pathogen Listeria monocytogenes has become a paradigm for the study of host–pathogen interactions and bacterial adaptation to mammalian hosts. Analysis of L. monocytogenes infection has provided considerable insight into how bacteria invade cells, move intracellularly, and disseminate in tissues, as well as tools to address fundamental processes in cell biology. Moreover, the vast amount of knowledge that has been gathered through in-depth comparative genomic analyses and in vivo studies makes L. monocytogenes one of the most well-studied bacterial pathogens.
*Institut Pasteur, Unité des Interactions Bactéries Cellules, Paris 75015, France. ‡ Institut National de la Santé et de la Recherche Médicale, U604, Paris 75015, France. § Institut National de la Recherche Agronomique, USC2020, Paris 75015, France. Correspondence to P.C. e-mail:
[email protected] doi:10.1038/nrmicro1413
The Gram-positive bacterium Listeria monocytogenes is a ubiquitous pathogen that thrives in diverse environments such as soil, water, various food products, humans and animals. The disease caused by this bacterium, listeriosis, is acquired by ingesting contaminated food products and mainly affects immunocompromised individuals, pregnant women and newborns. Listeriosis manifests as gastroenteritis, meningitis, encephalitis, mother-to-fetus infections and septicaemia, resulting in death in 25–30% of cases. The diverse clinical manifestations of infection with L. monocytogenes reflect its ability to cross three tight barriers in the human host. Following ingestion, L. monocytogenes crosses the intestinal barrier by invading the intestinal epithelium, thereby gaining access to internal organs. During severe infections, crossing the blood–brain barrier results in infection of the meninges and the brain, and in pregnant women, crossing the fetoplacental barrier leads to infection of the fetus1. L. monocytogenes infection has been a useful model for evaluation of the cellular interactions that are crucial for the initiation of the host T-cell response. However, this aspect of listerial biology is well documented and has recently been reviewed elsewhere2. This Review highlights the many ways in which L. monocytogenes manipulates its mammalian host, and it focuses on the lessons this bacterium has taught us in cell biology, bacterial pathophysiology, virulence-factor regulation, and bacterial adaptation to the host cytosol. Indeed, the ability of L. monocytogenes to invade and replicate in different cell types has been extensively studied and has revealed the sophisticated relationship between the bacterium and its host. The breadth of information gathered on the elaborate mimicries that are used by L. monocytogenes to subvert host processes
NATURE REVIEWS | MICROBIOLOGY
has made it an exceptional tool for the study of cellular processes such as actin-based motility, growth-factormediated signalling, endocytosis and cellular adhesion. Furthermore, available animal models, genetic tools and genomics have facilitated the compilation of information on different aspects of L. monocytogenes biology and have made this bacterium one of the most useful model organisms for the study of bacterial pathogenesis and pathophysiology.
An intracellular bacterium and a cell biologist L. monocytogenes is a facultative intracellular bacterium. Its life cycle reflects its remarkable adaptation to intracellular survival and multiplication in macrophages and other cell types1,3,4 (FIG. 1). Similar to the situation for most pathogens, the invasion of macrophages by L. monocytogenes is a passive process, but entry into non-professional phagocytes is induced by binding of the bacterial surface proteins internalin A (InlA) and InlB to receptors on the host cell. Both of these invasins are necessary and sufficient for bacterial entry into cell types such as enterocytes, hepatocytes, fibroblasts, epithelial cells and endothelial cells, but InlAmediated entry is restricted to the smaller number of cell types that express its receptor. Entry of L. monocytogenes into mammalian cells is a dynamic process that requires actin polymerization and membrane remodelling, and is an excellent example of how a bacterium can manipulate host-cell signalling and endocytic pathways to its advantage. L. monocytogenes can also harness the actinpolymerization machinery in the cytoplasm to facilitate intracellular and intercellular movement. New mechanisms by which L. monocytogenes manipulates the host cell are emerging through the use of microarray analyses that aim to determine the genes that are specifically activated by bacterial entry into the host cell5,6. VOLUME 4 | JUNE 2006 | 423
© 2006 Nature Publishing Group
REVIEWS Listeria monocytogenes
a
a
b b
c 0.5 µm
c
Phagosome Lysis of phagosome and replication in cytosol
d d
0.5 µm
F-actin
0.5 µm
e e
Double-membraned vacuole
f f
0.5 µm
Lysis of vacuole 0.5 µm
Figure 1 | Schematic representation and electron micrographs of the Listeria monocytogenes life cycle. a | L. monocytogenes induces its entry into a non-professional phagocyte. b | Bacteria are internalized in a vacuole (also known as a phagosome). c,d | The membrane of the vacuole is disrupted by the secretion of two phospholipases, PlcA and PlcB, and the pore-forming toxin listeriolysin O. Bacteria are released into the cytoplasm, where they multiply and start to polymerize actin, as observed by the presence of the characteristic actin tails (see Supplementary information S3 (figure)). e | Actin polymerization allows bacteria to pass into a neighbouring cell by forming protrusions in the plasma membrane. f | On entry into the neighbouring cell, bacteria are present in a double-membraned vacuole, from which they can escape to perpetuate the cycle. F-actin, filamentous actin. Electron micrographs a–c,e–f are reproduced with permission from REF. 113 © (1998) European Molecular Biology Organization, and d is reproduced with permission from REF. 30 © (1992) Elsevier.
InlB: subverting cellular-signalling and endocytic pathways. Binding of InlB to its cellular receptor Met results in the entry of L. monocytogenes into different cell types. Met is a protein tyrosine kinase, and the endogenous ligand of this receptor is hepatocyte growth factor (HGF) 7 (FIG. 2) . In vivo, Met is expressed mainly by cells of epithelial origin, whereas HGF is produced mainly by fibroblasts and stromal cells. The binding of HGF to Met activates cellular survival and proliferation signals, and it induces cytoskeletal rearrangements that function in cellular motility and differentiation. Binding of InlB activates
424 | JUNE 2006 | VOLUME 4
the protein-tyrosine-kinase activity of Met, as well as the phosphatidylinositol 3-kinase (PI3K) and the Ras–mitogen-activated protein kinase (MAPK) pathways, all of which are required for the uptake process7–9. Interestingly, although InlB and HGF both bind and activate Met, InlB does not strictly mimic HGF. Indeed, the kinetics of InlB-induced signalling are different from those of HGF-induced signalling7, and at an equal concentration, InlB seems to induce a more potent activation of the Ras–MAPK pathway than does HGF10. Differences in signalling might be explained by the finding that InlB also binds gC1qR
www.nature.com/reviews/micro © 2006 Nature Publishing Group
REVIEWS a HGF-mediated signalling
b InlB-mediated signalling Listeria monocytogenes InlB
HGF
GAGs Met
Cbl Rac1
GRB2
Survival
CDC42
Dynamin
Akt
CD2AP CIN85 Cbl EPS15 Ub
Ras
SOS Cytoskeletal rearrangements Degradation
Clathrin
PI3K Cbl GAB1 SHC
ABI
PLCγ
Rac1 CDC42
SP
GAB1 SHC
In
PtdIns(3,4,5)P3
N-W A
Ub
PLCγ
PI3K
HR S A3
Clathrinmediated endocytosis
Akt
Cbl
PtdIns(4,5)P2
GRB2
GG
Recycling
PtdIns(3,4,5)P3
Cortactin
In
Met
gC1qR
Out PtdIns(4,5)P2
WAVE
Out
F-actin
Ras Arp2/3
Nucleus
Proliferation
Figure 2 | Met signalling induced by hepatocyte growth factor (HGF) and internalin B (InlB). a | Phosphorylation of Met leads to the recruitment and activation of many transducers, which in turn recruit cytosolic signalling proteins. Signalling mediated by HGF activates survival and proliferation signals, and it induces cytoskeletal rearrangements that are important for cellular motility and differentiation. On stimulation with HGF, the endocytosis of Met, similar to most signalling receptors, is an important regulatory mechanism that downregulates the cell-surface expression of the activated receptor. b | Met signalling mediated by the Listeria monocytogenes protein InlB induces cytoskeletal rearrangements that are important for bacterial entry into non-phagocytic cells. Clathrin components of the endocytic machinery are also recruited to the site of entry. The link between the cytoskeletal machinery (shown on the right) and the endocytic machinery (shown on the left) is still unclear. InlB, through the GW repeats at its C terminus, also binds gC1qR (the receptor for the globular part of complement component C1q) and glucosaminoglycans (GAGs), which are negatively charged polysaccharides that are present at cell surfaces. Both components might contribute to entry of L. monocytogenes by modulating the interaction of InlB with Met. ABI, Abl interactor 1; Arp, actin-related protein; CD2AP, CD2-associated protein; CIN85, Cbl-interacting protein of 85 kDa; EPS15, epidermal-growth-factor-receptor substrate 15; F-actin, filamentous actin; GAB1, GRB2-associated binding protein 1; GGA3, Golgi-localized, γ-ear-containing, ADP-ribosylation-factor-binding protein 3; GRB2, growth-factor-receptor-bound protein 2; HRS, HGF-regulated tyrosine-kinase substrate; N-WASP, neural Wiskott–Aldrich syndrome protein; PI3K, phosphatidylinositol 3-kinase; PLCγ, phospholipase C-γ; PtdIns(3,4,5)P3, phosphatidylinositol-3,4,5-trisphosphate; PtdIns(4,5)P2, phosphatidylinositol-4,5-bisphosphate; SHC, SRC-homology-2domain-containing transforming protein C; SOS, son of sevenless; Ub, ubiquitin; WAVE, Wiskott–Aldrich syndrome protein (WASP)-family verprolin homologous protein.
(the receptor for the globular part of complement component C1q), which might therefore function as a co-receptor for InlB11. Furthermore, InlB and HGF do not compete for binding to Met7, indicating that they might bind distinct sites on Met. These results are consistent with the fact that InlB and HGF have no sequence homology and are structurally unrelated. Crystal structures of InlB bound to Met could provide valuable information about the molecular mechanism that underlies the activation of Met by InlB. The signalling pathways that are activated by InlB ultimately lead to cytoskeletal rearrangements and entry of L. monocytogenes. How activation of the PI3K and the Ras–MAPK pathways leads to cytoskeletal rearrangements has been extensively studied, and many of the proteins that are crucial for invagination and internalization have been identified. Local actin remodelling at the site of InlB attachment is mediated by the recruitment and activation of the actin-nucleation complex, Arp2/3, which promotes actin nucleation and polymerization (discussed later). The mechanism of
NATURE REVIEWS | MICROBIOLOGY
Arp2/3 activation seems to be cell-type dependent, but it involves a combination of the small GTPases Rac and CDC42 and proteins of the Wiskott–Aldrich syndrome protein (WASP) family, which includes neural WASP (N-WASP) and WAVE 12. Proteins of the Ena/VASP family (enabled homologue/vasodilator-stimulatedphosphoprotein family), which promote actin-filament elongation, are also central to the process. In addition, cofilin, which is essential for depolymerization of actin filaments, functions successively as a stimulator and a downregulator of actin rearrangements that occur during the internalization process 12,13. All of these components that are recruited by the binding of InlB to Met also have a role in growth-factor-receptor activation. Therefore, although there are differences between the kinetics of signalling mediated by InlB and HGF, the machinery that is recruited to the site of activation seems to be the same for both molecules, showing the utility of L. monocytogenes as a tool to study cellular signalling by Met or other growth-factor receptors.
VOLUME 4 | JUNE 2006 | 425 © 2006 Nature Publishing Group
REVIEWS Recently, the study of InlB-induced internalization revealed an unexpected mechanism used by L. monocytogenes during host-cell entry. L. monocytogenes invades epithelial cells by subverting clathrin-mediated endocytosis14 (FIG. 2b; see Supplementary information S1 (figure)), a process that is used by mammalian cells to take up nutrients and to recycle membrane proteins. Owing to their size, which is usually 1–3 µm, bacteria were thought to enter cells through a mechanism related to phagocytosis (that is, an actin-dependent mechanism), which differs from ‘endocytosis’, a process that is thought to internalize particles no larger than 120 nm and that was considered to be actin independent until recently. Therefore, the finding that L. monocytogenes can induce its internalization by using clathrin-dependent machinery was surprising and indicated that clathrin-coated structures can engulf much larger particles than previously thought. Cell-surface expression of Met is downregulated by HGF, and similarly, soluble InlB induces the degradation of Met, through monoubiquitylation and clathrin-mediated endo cytosis 14. Furthermore, as shown using RNA-interferencemediated knockdown, important components of the endocytic machinery are required for internalization of L. monocytogenes. Although the underlying molecular mechanisms have not been defined, other invasive bacteria had previously been reported to use clathrinmediated entry, implying that this mechanism is not restricted to Listeria species and could be a commonly used mechanism for bacterial entry 15–17. Although the endocytic machinery is important for entry of L. monocytogenes, this bacterium might exploit other mechanisms, because inhibitors of endocytosis reduce bacterial entry but do not completely abolish it14. Plasma-membrane microdomains known as lipid rafts have also been shown to be important for the entry of L. monocytogenes18. Because lipid rafts are usually associated with clathrin-independent endocytic pathways, this indicates either that L. monocytogenes exploits more than one endocytic mechanism or that the separation among the classes of endocytosis is not as well demarcated as conventionally thought. Further work is necessary to decipher how lipid rafts, clathrin-mediated endocytosis and actin-mediated phagocytosis combine to enable listerial entry and cellular invasion.
Adherens junctions Together with tight junctions and desmosomes, these are specialized structures that allow epithelial cells to adhere to each other, and they have epithelial cadherin (E-cadherin) as a major component.
InlA: exploiting intercellular junctions. Similar to InlB, InlA induces local cytoskeletal rearrangements in the host cell to stimulate uptake of L. monocytogenes by epithelial cells. InlA is a covalently linked bacterial cell-wall protein that binds the host epithelial-cell protein E-cadherin19 (FIG. 3). E-cadherin is a transmembrane protein that belongs to a large family of cell–cell adhesion molecules that are required for the correct formation of adherens junctions between epithelial cells. E-cadherin is localized at these cellular junctions, where its intracellular domain forms a complex with the cytoskeleton through the catenins (which are cadherin-binding proteins), and its extracellular domain is in contact with E-cadherin molecules on neighbouring cells 20. InlA binds the extracellular
426 | JUNE 2006 | VOLUME 4
domain of E-cadherin, but it is the intracellular domain of E-cadherin that is essential for the cytoskeletal rearrangements that are required for bacterial entry21. Because all of the components of the endogenous machinery of cell–cell junctions seem to be recruited on binding of InlA, L. monocytogenes is a good system for the study of cellular adhesion and identification of components that are involved in this process. Although it is known that assembly and attachment of the E-cadherin, α-catenin and β-catenin complex to the cytoskeleton is central to both intercellular adhesion and L. monocytogenes entry21, neither the mechanisms that hold cells together nor the molecular mechanisms that are required for InlA-dependent entry are as yet fully understood. Until recently, the accepted model of intercellular adhesion proposed that α-catenin anchors the E-cadherin–β-catenin complex to the actin cytoskeleton, providing a stable structure that maintains tissue integrity. However, it has become apparent that the dynamics of this process are more complex. As was recently shown, α-catenin cannot simultaneously interact with actin filaments and the E-cadherin–β-catenin complex, indicating that α-catenin is not an actin anchor but is, instead, an actin-filament organizer22,23. Further investigation is required to understand this mechanism fully; however, the dynamic interactions between the cadherin–catenin complex and the underlying actin cytoskeleton are consistent with the findings that L. monocytogenes can regulate and rearrange actin structures at intercellular junctions through adhesion to E-cadherin, and further emphasize the validity of the listerial model for analysing events at intercellular junctions. The validity of using L. monocytogenes as a tool to study cell–cell junction formation was shown by a recent study of InlA-dependent entry that identified the protein ARHGAP10 (Rho GTPase-activating protein 10) as a novel cellular component that is involved in the recruitment of α-catenin to cell–cell junctions. This study also showed that ARHGAP10 was essential for listerial entry (see Supplementary information S2 (figure))24. Overexpression of ARHGAP10 disrupted the cytoskeleton and increased the local concentration of α-catenin, indicating that ARHGAP10 has a direct role in regulation of the dynamics of cell–cell junction formation. Furthermore, ARHGAP10 was shown to control the activity of RhoA and CDC42, two proteins that regulate cell–cell junction formation. Components that generate the tension that is required to hold neighbouring cells together, that is, myosin VIIA and its ligand vezatin25, have been found to be important for L. monocytogenes entry, indicating that these components might generate the force that is necessary for engulfment of the bacterium through phagocytosis26. These findings also indicate that the tension that holds two cells together could be similar to the tension that is exerted during phagocytosis, as if each cell is attempting to engulf its neighbour. An analogous process known as ‘frustrated phagocytosis’ occurs when macrophages adhere to immune-complex-coated surfaces27.
www.nature.com/reviews/micro © 2006 Nature Publishing Group
REVIEWS a Intercellular junctions
b InlA-induced entry F-actin
Arp2/3
α
G-actin α
β
α
β
Formin
β
Listeria monocytogenes
In InlA Out
E-cadherin E-cadherin Out
In Vezatin
Out
Catenins β
β
In
Myosin VII
β α
β α
β α
α ARHG AP10 ARF6
β α
α
Catenins Formin
G-actin
Arp2/3 F-actin
Figure 3 | Adherens junction and internalin A (InlA)-induced bacterial entry. a | Adherens junctions hold adjacent cells together through the transmembrane protein epithelial cadherin (E-cadherin). The intracellular domain of E-cadherin recruits α-catenin and β-catenin, and α-catenin bridges the actin cytoskeleton and E-cadherin. Formins, which interact directly with α-catenin, are also essential for forming actin cables at cell–cell junctions, although the mechanism by which they achieve this is not understood. b | The receptor for the Listeria monocytogenes protein InlA is E-cadherin. Many components that are important for adherens junctions are recruited to the site of bacterial entry, where the cytoskeletal rearrangements that are required for invasion occur. ARF6, ADP-ribosylation factor 6; ARHGAP10, Rho GTPase-activating protein 10; Arp, actin-related protein; F-actin, filamentous actin; G-actin, globular actin.
Harnessing the actin-polymerization machinery. Following internalization into a host-cell vacuole, L. monocytogenes lyses the membrane-bound phagosome (discussed later) and escapes into the cytoplasm, where it can polymerize host actin and propel itself through the cell and into neighbouring cells28. The ability to spread from cell to cell without coming in contact with the extracellular milieu allows the bacterium to propagate through tissues and avoid contact with circulating antibodies or other extracellular bactericidal compounds. At the leading edge of moving cells, the control of actin polymerization is a complex mechanism triggered by intricate signalling pathways that take place at the plasma membrane. L. monocytogenes bypasses these signalling pathways and directly nucleates actin constitutively, making it a simple and effective model for studying the dynamics of actin polymerization. L. monocytogenes polymerizes actin asymmetrically along its surface, producing an actin tail that propels the bacterium through the cytoplasm28,29 (see Supplementary
NATURE REVIEWS | MICROBIOLOGY
information S3 (figure)). Polymerization of host actin is mediated by the bacterial surface protein ActA, the first protein that was identified to have functions that promote actin nucleation30–32. It was later found that ActA mimics its eukaryotic counterparts, proteins of the WASP family (which includes N-WASP and WAVE)33. The N-terminal region of ActA, which is essential for actin-based motility 34, has homology to the C-terminal region of WASP, which binds the actin-nucleation complex Arp2/3 (REF. 35). Accordingly, both ActA and WASP-family proteins function as nucleation-promoting factors (NPFs) for the Arp2/3 complex and are involved in forming a ternary complex that is composed of Arp2/3, an NPF and actin. The Arp2/3 complex consists of seven proteins, including two proteins that are related to actin, Arp2 and Arp3, which (as a result of their structural similarity to actin) are thought to function as a template for polymerization. Discovery of the Arp2/3 complex ensued from studies of ActA partners, thus the role of Arp2/3 as a key actin-nucleating complex was consequently revealed.
VOLUME 4 | JUNE 2006 | 427 © 2006 Nature Publishing Group
REVIEWS a
b
N
Listeria monocytogenes InlB
InlA
Pro16
C
Out In
Met E-cadherin Guinea pig and rabbit
Met E-cadherin Mouse
Met E-cadherin Human
Figure 4 | Host specificity of Listeria monocytogenes proteins internalin A (InlA) and InlB. a | InlA and InlB can bind and induce entry of L. monocytogenes into human cells that express the respective cell-surface receptors, epithelial cadherin (E-cadherin) or Met. However, a single amino-acid change in E-cadherin (at position 16; see b) prevents InlA from binding mouse E-cadherin, and for unknown reasons, InlB cannot recognize or activate guinea pig or rabbit Met. b | A diagrammatic representation of the crystal structure of the leucine-rich-repeat region of InlA (purple) bound to E-cadherin (green) is shown. The position of the crucial proline residue at position 16 of E-cadherin is indicated. It is this residue that determines intermolecular recognition and therefore host specificity. The crystal-structure representation is reproduced with permission from REF. 114 © (2002) Elsevier.
Filopodia Rod-like cell-surface projections that are composed of actin filaments. They are found on various cell types and have sensory or exploratory functions.
Lamellipodia Thin actin-rich structures that form protrusions at the edge of the cell and are essential for cellular motility.
The L. monocytogenes model has also been useful for understanding the importance and the function of VASP, the other main ligand of ActA. VASP binds the central proline-rich domain of ActA and promotes efficient actin-based motility34,36–38, highlighting the importance of this protein in actin polymerization. Only recently, however, have the complex molecular mechanisms of VASP activity begun to emerge. VASP seems to promote listerial motility by recruiting the actin-binding protein profilin, which promotes polymerization at actin-filament barbed ends39. VASP also seems to induce faster growth of the actin network at the bacterial surface by causing the release of Arp2/3 from ActA40. In addition, VASP seems to decrease Y-branch formation, thereby increasing parallel alignment of actin filaments41. Although their mechanistic role is not fully elucidated, proteins of the Ena/VASP family are important in the formation of actin fibres, filopodial tips and the lamellipodial leading edge42. Intracellular and intercellular movement using actin polymerization is not restricted to L. monocytogenes. A growing number of intracellular pathogens, including Rickettsia species, Shigella species, mycobacteria, Burkholderia pseudomallei and vaccinia virus, show this feature during infection (see Supplementary information S3 (figure), and for recent reviews, see REFS 43,44).
A paradigm in pathophysiology As a pathogen that displays such interesting features as strong T-cell activation, a sophisticated relationship with its host and crossing of protective barriers, L. monocytogenes has more recently also emerged as a model to study the pathophysiology of a complex bacterial infection. Findings from in vitro work have been
428 | JUNE 2006 | VOLUME 4
used to bypass stringent host species specificity and generate relevant model systems to study infection in vivo. In this respect, L. monocytogenes is a good example of how in vitro studies can help to generate animal models that more closely reflect human infection. For many years, the animal model used to study L. monocytogenes infection was intravenous infection of mice. This model provides a dose-dependent infection with dissemination of the bacteria into organs and was crucial in the discovery of cell-mediated immunity 3. However, recent molecular evidence showed that the mouse model is inadequate to study the crossing of barriers that is characteristic of listeriosis. It had long been known that oral inoculation of mice (rather than intravenous infection), which most closely reflects the human mode of infection, is not efficient because only small numbers of L. monocytogenes cross the mouse intestinal barrier. The reason for this was uncovered by molecular in vitro studies that showed that a single amino-acid difference in the mouse cellular receptor for InlA, E-cadherin, prevented it from binding InlA, thereby showing the inadequacy of the mouse model for study of the invasive role of InlA45 (FIG. 4). Consequently, a novel animal system was developed to study the crossing of the intestinal barrier: a transgenic mouse that expresses human E-cadherin on the surface of enterocytes46. This model showed that InlA has a key role in disease, because it is essential for crossing of the intestinal barrier. In this model system, InlA could interact with E-cadherin, and a wild-type strain of L. monocytogenes was able to cause disease through oral inoculation. So far, E-cadherin has been found only at cell–cell junctions and on the basolateral face of epithelial cells, so the mechanism by which L. monocytogenes gains access to E-cadherin was enigmatic. Two hypotheses were put forward to explain how InlA could target E-cadherin 46. The first hypothesis proposed that L. monocytogenes could gain access to E-cadherin at the tips of intestinal microvilli, where apoptotic epithelial cells slough off. The second hypothesis proposed a synergy between InlA- and InlB-dependent internalization, because activation of Met by HGF has been shown to stimulate the disassembly of junctions between epithelial cells46. Recent results have shown that L. monocytogenes invades the intestinal epithelium at sites of cell extrusion at the tips of villi47 and that the contribution of InlB to crossing of the intestinal barrier is insignificant in vivo48. Whether the mechanism of intestinal invasion is used to cross other barriers is unknown. Synergy between InlA and InlB could still be important for crossing of the placental barrier49. Because the transgenic mice described here express human E-cadherin only on enterocytes, it is not possible to study the role of InlA in deeper tissues. The generation of transgenic mice that express E-cadherin on all cells is in progress. At present, the role of InlA can also be studied in guinea pigs or rabbits, because E-cadherin is recognized by InlA in these animals. However, recent studies show a species specificity for InlB: InlB does not recognize or activate guinea-pig
www.nature.com/reviews/micro © 2006 Nature Publishing Group
REVIEWS or rabbit Met. Therefore, the role of InlA and InlB in infection cannot be studied using these animal models48. A human model remains the best model system for studying listerial infection, and human explants have been successfully used to determine the mechanism by which L. monocytogenes crosses the maternofetal barrier49. Human placental villus explants, together with primary or immortalized trophoblastic cells, were used to show that InlA is a key bacterial protein that is required for crossing of the human maternofetal barrier, although InlA is not essential for this role in the pregnant guinea-pig model49,50. Crossing of the blood– brain barrier is still poorly understood. However, the optimization of animal models should help to decipher this crucial step. Other model systems are being developed to identify the host factors that are required for the intracellular survival of L. monocytogenes and possibly of other intracellular pathogens. Drosophila melanogaster has attracted attention as a model because of the many genetic and immunological studies that have been carried out using this organism, and it has been successfully used to test L. monocytogenes virulence51. Moreover, genome-wide RNA-interference screens in D. melanogaster S2 cells (which are macrophage-like cells) have revealed many new host factors that are important for entry into the host cell, escape from the vacuole and intracellular growth of L. monocytogenes52,53. Another noteworthy organism that has been shown to support a listerial infection is Caenorhabditis elegans54. a P
plcA
P1 and P2
prfA
b Ribosome subunits
SD prfA mRNA
ATG prfA
High temperature Low temperature
No PrfA
PrfA Virulence-gene expression
Figure 5 | The PrfA regulator. a | Schematic representation of the prfA region. During exponential phase, prfA is mostly transcribed as a bicistronic product from the promoter (P) upstream of plcA. By contrast, during stationary phase, prfA is mostly transcribed as a monocistronic product from P1 and/or P2. b | Mechanism that controls the thermoregulated expression of PrfA in the promoter upstream of prfA. At low temperatures (≤30°C), a secondary structure forms in the untranslated region of prfA, and this prevents ribosome binding and therefore expression of PrfA. At high temperatures (≥37°C), melting of the prfA untranslated region allows ribosomes to bind and PrfA expression to occur. SD, Shine–Dalgarno sequence. This figure is modified with permission from REF. 58 © (2002) Elsevier.
NATURE REVIEWS | MICROBIOLOGY
Novel regulatory mechanisms L. monocytogenes is a facultative intracellular pathogen that can live both inside and outside its host. This bacterium has therefore evolved sophisticated regulatory mechanisms to ensure that virulence factors are optimally expressed when they are required. These regulatory mechanisms might prove to be general mechanisms used by other bacteria. PrfA: a tightly regulated protein. Most of the virulence proteins that have been identified in L. monocytogenes are under the control of one transcriptional regulator, PrfA, which itself is tightly regulated by environmental conditions. During exponential growth, prfA is mainly transcribed as a bicistronic mRNA from the plcA promoter. By contrast, during stationary phase, a monocistronic mRNA is preferentially transcribed from a promoter upstream of prfA55–57 (FIG. 5a). Similar to many other pathogens, L. monocytogenes can sense conditions in the mammalian host and respond by expressing virulence genes. A novel regulation of PrfA by temperature, owing to the structure of the upstream untranslated region of the prfA mRNA, was recently discovered58 (FIG. 5b). At low temperature (30°C), the prfA leader transcript controls translation of the downstream mRNA by forming a secondary structure that masks the ribosome-binding site. At mammalian host temperature (37°C), this structure partially melts to expose the ribosome-binding site, thereby allowing translation to occur. Fusion of the prfA leader transcript to the gene that encodes green fluorescent protein (GFP) also resulted in thermoregulation of GFP. This mechanism might be used by other bacteria, as has previously been suggested for the Yersinia pestis activator protein LcrF59. Such post-transcriptional regulation of prfA allows rapid expression of the encoded transcription factor and therefore efficient transcription of virulence factors as soon as the bacterium enters the host. Other environmental conditions — such as osmolarity, iron concentrations, pH, the presence of fermentable sugars, stress (through σB), and conditions in the host-cell intracellular compartment — have been shown to regulate prfA and PrfA-controlled genes through mechanisms that are not completely understood60,61. Post-translational regulation of PrfA by a putative cofactor is suggested by its structure, which resembles that of the cyclic-AMP receptor62,63. The number of mechanisms that regulate PrfA is probably indicative of the importance of this crucial virulence factor during infection. Listeriolysin O: a pH-sensing protein. L. monocytogenes thrives in the cytoplasm of numerous cell types. Following internalization, bacteria escape from membrane-bound phagosomes by secreting two phospholipases, PlcA and PlcB, and the pore-forming toxin listeriolysin O (LLO), thereby gaining access to the cytoplasm (FIG. 1). Although LLO is a member of a large family of cholesterol-dependent cytolysins that are secreted by numerous Gram-positive bacteria, L. monocytogenes is the only pathogen that secretes this type of toxin inside the host cell. Therefore, secretion of LLO must
VOLUME 4 | JUNE 2006 | 429 © 2006 Nature Publishing Group
REVIEWS Premature unfolding
Neutral pH
Soluble monomer D1 Membrane-bound monomers D2 D3
D4
Pre-pore oligomer
Formation of the pre-insertion β-sheet Pore oligomer
Low pH
Out
In
Figure 6 | Listeriolysin O (LLO) pore-forming mechanism. At low (acidic) pH, the soluble LLO monomer interacts with the host-cell plasma membrane, presumably by binding cholesterol. On contact with the membrane, structural rearrangements in one monomer expose residues that can form hydrogen bonds with other monomers, thereby allowing oligomerization into a pre-pore complex. Following oligomerization, two α-helical bundles from each monomer extend to form transmembrane β-hairpins (red) that punch through the membrane. Pores formed by LLO and other cholesterol-binding proteins can be 250–300 Å in diameter. At neutral pH, domain 3 (D3) of the monomer prematurely unfolds, rendering the protein unable to form pores. This figure is modified with permission from REF. 115 © (2005) American Society for Microbiology.
Two-component regulatory system A two-protein signaltransduction system that is important for the bacterial response to environmental changes. It consists of a membrane-bound sensor protein kinase and a transcriptional-response regulator.
Elongation factor A protein that allows tRNAs to bind the ribosome and is essential for elongation of the polypeptide chain.
be tightly regulated, because the bacterium needs to balance efficient escape from the vacuole against prevention of host-cell damage to allow its intracellular survival. Unlike other pore-forming toxins, the activity of LLO is optimal at acidic pH (<6). So, LLO is fully active in the acidic environment of the phagosome, but it is less active at the neutral pH of the host-cell cytoplasm. An L. monocytogenes strain that was constructed to express the pore-forming toxin perfringolysin O, which is active at neutral pH, can also escape the vacuole but is toxic to the host cell and is avirulent in mice, highlighting the importance of pH regulation of LLO activity 64. The molecular mechanism that underlies the optimum pH for LLO activity was only recently identified: LLO was found to be stable at acidic pH, and to unfold and aggregate at neutral pH65 (FIG. 6). At acidic pH, pore formation occurs when LLO monomers oligomerize into large complexes that penetrate the membrane by extending transmembrane β-hairpins66. At neutral pH, acidic amino-acid residues in the transmembrane domain initiate irreversible denaturation of the β-hairpins of LLO, leading to inactivation of the pore-forming function of LLO65. The elaborate structure of LLO therefore allows this protein to sense and respond to its environment.
430 | JUNE 2006 | VOLUME 4
In addition to its pore-forming ability, LLO has the surprising ability to induce potent signalling in the host cell. It has been shown to activate NF-κB67, MAPK68, phosphatidylinositol 69,70, calcium 71–73 and proteinkinase-C 74 signalling pathways. As a pore-forming toxin with cholesterol as the only identified ligand, this signalling function is puzzling. LLO could have a signalling function at neutral pH, as has been suggested for another member of the cholesterol-dependent-cytolysin family, Streptococcus intermedius intermedilysin, which binds the host cellular receptor CD59 (REF. 75). However, it has been difficult to separate the signalling function of LLO from its pore-forming ability. For example, MAPK activation occurs on treatment of host cells with LLO, but it also occurs after exposure to a mild detergent, indicating that signalling could be the result of membrane damage rather than a specific activity of LLO68. Clearly, the signalling activity of LLO requires further investigation.
A model of bacterial adaptation One particularly interesting feature of L. monocytogenes is its capacity to thrive both inside and outside the host. This ability to colonize a broad range of ecosystems is reflected in the genome of the organism, which contains an unusually large number of regulatory and transport proteins (11.6% of all predicted genes)76. With one-quarter of its transport proteins dedicated to carbohydrate transport, L. monocytogenes has the largest number of phosphotransferase systems that have been described for bacteria so far. Furthermore, the genome contains 16 putative two-component regulatory systems, more than are found in other pathogenic bacteria and comparable to the number found in ubiquitous bacteria such as Bacillus subtilis. Inactivation of several of these two-component regulatory systems has revealed their importance for the resistance of L. monocytogenes to various stresses and for survival in the host77–82. In addition to these conventional bacterial two-component regulatory systems, which phosphorylate histidine and aspartic-acid residues, L. monocytogenes also uses a eukaryotic-like serine/threonine protein phosphatase system, known as Stp, to regulate the translation elongation factor EF-Tu and, consequently, virulence83. Outside the host, L. monocytogenes is particularly well adapted to grow at low temperatures, making it difficult for the food industry to rely on refrigeration to control listerial contamination84–87. More remarkable is the ability of L. monocytogenes to survive in the host. The resistance of L. monocytogenes to acidic conditions88,89 and to bile salts90–92 makes this pathogen particularly adept at infecting the gastrointestinal tract. Consistent with the many bile-resistance genes in the L. monocytogenes genome, a recent study using in vivo bioluminescence imaging (a non-invasive procedure that allows bioluminescent L. monocytogenes to be traced through the body of a mouse) has shown an intriguing new reservoir for L. monocytogenes, the gall bladder, where bile is stored and concentrated93. Whether this organ is indeed colonized during human listeriosis is unknown.
www.nature.com/reviews/micro © 2006 Nature Publishing Group
REVIEWS InlG
490
InlE
499
Lmo1136
539
InlH
548
LPXTG
Lmo0610
589
Lmo1289
593
Lmo1290
598
Lmo0514
605
Lmo2026
626
Lmo0331
633
Lmo0732
638
Lmo0801
646
InlA
800
InIF
821
Lmo0171
832
InlJ
851
Lmo2396
940
Lmo0327
1349
InlI
Secreted
GW
1778 InlB
630
InlC
296
Lmo2445
300
Lmo2027 Lmo2470
367 388
Lmo0549 Signal peptide
673 LPXTG motif
Conserved inter-repeat region
LRR region
Sorting signal
GW module
Figure 7 | The internalin family of proteins. Homologous regions of internalin-family members include an N-terminal signal-peptide sequence, followed by several leucinerich tandem repeats (LRRs). Several internalins contain an inter-repeat region, which is structurally related to an immunoglobulin-like domain. The internalin family can be divided into three subfamilies on the basis of their association with bacteria. The first subfamily includes internalins that contain an LPXTG amino-acid motif (where X denotes any amino acid), and these are covalently anchored to the cell wall. The second subfamily includes internalin B (InlB), which is loosely associated with the bacterial surface through its GW modules (which contain the amino-acid motif GW). The third subfamily contains five internalins that are predicted to be secreted, because they do not have any surfaceanchoring domains. It should be noted that the number of LRRs present in the proteins Lmo2445 and Lmo2470 is unclear, because these proteins contain degenerate LRRs. The number of amino acids in each protein is indicated. This figure is modified with permission from REF. 101 © (2002) Elsevier.
Survival and replication in the cytosol of many types of host cells, including macrophages and epithelial cells, is one of the distinguishing features of infection with L. monocytogenes3,4. Other extracellular bacteria and intracellular pathogens that normally reside inside a vacuole cannot replicate in the cytosol of the host cell94, highlighting that L. monocytogenes has evolved specific mechanisms to grow in the host-cell cytosol. Although intracytosolic survival remains poorly understood, two key proteins for intracellular growth have been
NATURE REVIEWS | MICROBIOLOGY
identified: the sugar-uptake system (Hpt) and lipoate protein ligase A1 (LplA1). Hpt (which is homologous to the mammalian translocase of glucose-6-phosphate) is required for optimal intracellular replication 95. Interestingly, transcription of hpt is regulated by PrfA and is therefore activated upon entry into the cytosol of the host cell. Expression of Hpt allows the bacterium to use glucose-1-phosphate, which is an available carbon source in the host-cell cytoplasm. Other bacterial pathogens such as Escherichia coli, Salmonella enterica, Shigella flexneri and Chlamydia trachomatis have a homologue of hpt, indicating that this transporter system could have a general role in intracellular survival. The other factor that has been identified to be important for intracellular growth, LplA1, catalyses the formation of a covalent link between lipoic acid and specific protein targets96. A mutant that lacks LplA1 cannot replicate in the cytoplasm of macrophages and has a 300-fold decrease in virulence96. LplA1 seems to be important for lipoylation and full activity of the pyruvate-dehydrogenase enzyme complex in the host-cell cytosol, where lipoic acid is mainly absent. LplA1 has therefore been proposed to be important for the scavenging of lipoyl groups from host molecules. Interestingly, several reports indicate that the expression of many L. monocytogenes genes is upregulated in the cytosol. A library of Tn917-lacZ L. monocytogenes mutants was screened for higher lacZ expression when present in the cytosol of macrophages than when grown in rich broth. This approach allowed the identification of several L. monocytogenes genes, including plcA and prfA, that are activated after infection of host cells. Recently, two transcriptomic studies have advanced the understanding of adaptation to the host-cell cytosol by showing that L. monocytogenes turns on ~500 genes for survival in this cellular compartment98,99. Further analysis of the genes that were identified could elucidate the metabolism of this bacterium in the host-cell cytosol.
Genomics reveals new virulence factors Major listerial virulence factors have been identified by classical genetics; that is, the search for a mutant with a particular phenotype, followed by gene identification and characterization 1. The genetic basis of L. monocytogenes virulence can now be deciphered by exploitation of the genome sequences of L. monocytogenes and the closely related nonpathogenic species Listeria innocua76, using three main strategies: first, the analysis of genes that are present in L. monocytogenes and absent from L. innocua; second, the study of genes that encode potentially interesting proteins; and third, the use of wholegenome DNA arrays to screen a large population of strains from different Listeria species and to identify genes that are present in, and specific to, all virulent L. monocytogenes strains100. A first glance at the global genome structure indicates that, in contrast to many other bacterial genomes, the L. monocytogenes genome (strain EGD-e) contains only three insertion sequences and five bacteriophages, VOLUME 4 | JUNE 2006 | 431
© 2006 Nature Publishing Group
REVIEWS Chaperone A protein that assists other proteins to fold correctly.
1.
2.
3.
4.
5.
none of which seems to have a role in acquisition of virulence genes. Moreover, except for genes of the major virulence locus, virulence genes seem to be dispersed on the chromosome and are not concentrated in pathogenicity islands. One of the most striking features of the L. monocytogenes genome with respect to virulence is the exceptionally large number of genes that encode surface proteins (4.7% of all predicted genes101). These proteins are among the most likely candidates to interact with the host and therefore to be virulence factors. These proteins can be divided into three families. The largest family is composed of 68 lipoproteins, and the second largest family contains 41 LPXTG proteins (including InlA), which are anchored to the cell wall by a sortase (an enzyme that is involved in the covalent linkage of Grampositive bacterial proteins to the bacterial surface)102–104. Inactivation of sortase A (SrtA) or Lsp, a signal peptidase that is involved in maturation of lipoproteins, attenuates virulence, thereby pinpointing the importance of these two classes of surface protein in L. monocytogenes infection103,105,106. The third family of surface proteins includes proteins that are non-covalently attached to the bacterial surface by their C-terminal domains. This family includes GW proteins (such as InlB and Ami), which contain modules of 80 amino acids that contain the dipeptide Gly–Trp (also known as GW modules107) in their C-terminal region. An important outcome from genome-sequence analysis was the discovery of a SecA2dependent pathway, which allows proteins that lack a typical signal-peptide sequence to be targeted to the bacterial surface or to be secreted108. Comparison of the L. monocytogenes and L. innocua genomes has proved an efficient approach to identify new virulence factors, including bile-salt hydrolase (Bsh)90, two LPXTG proteins (Vip and InlJ) and a GW protein (Auto)105,109,110. The GW protein Auto is a novel surfaceassociated cell-wall hydrolase that is required for entry into eukaryotic cells. Its contribution to listerial infection is still unclear but might also be linked to the release of immunologically active cell-wall components that could interact with components of the innate immune system109. The LPXTG protein Vip, which is positively regulated by the transcriptional activator PrfA, is required for bacterial entry into some eukaryotic cells in vitro and
Khelef, N. et al. The Prokaryotes: An Evolving Electronic Resource for the Microbiological Community [online] (eds Dworkin, M., Falkow, S., Rosenberg, E., Schleifer, K.-H. & Stackebrandt, E.) (Springer, New York, 2005)
. Pamer, E. G. Immune responses to Listeria monocytogenes. Nature Rev. Immunol. 4, 812–823 (2004). Mackaness, G. B. Cellular resistance to infection. J. Exp. Med. 116, 381–406 (1962). First in-depth analysis of the interaction between L. monocytogenes and its host, both in mice and in cultured mouse macrophages. Racz, P., Tenner, K. & Mero, E. Experimental Listeria enteritis. I. An electron microscopic study of the epithelial phase in experimental Listeria infection. Lab. Invest. 26, 694–700 (1972). McCaffrey, R. L. et al. A specific gene expression program triggered by Gram-positive bacteria in the cytosol. Proc. Natl Acad. Sci. USA 101, 11386–11391 (2004).
6.
for the infectious process in vivo110. Vip interacts with the host-cell endoplasmic-reticulum chaperone gp96, a protein that is involved in Toll-like-receptor signalling111. The LPXTG protein InlJ, the function of which remains to be elucidated, is a leucine-rich-repeat-containing protein that is structurally related to InlA and InlB105. These three proteins are members of the internalin family, which is a large family of proteins that contain leucine-rich repeats (FIG. 7). The gene that encodes InlJ, similar to the genes that encode five other proteins of the internalin family (InlA, InlB, InlE, InlH and InlI), is conserved in the genomes of pathogenic serovars of L. monocytogenes (that is, 1/2a, 1/2b, 1/2c and 4b) and absent from all other Listeria species100,105, and this is consistent with a role for InlJ in virulence.
Conclusions and future perspectives The study of L. monocytogenes infection highlights the sophisticated relationship between this bacterium and its host, and reveals their long co-evolution and reciprocal adaptation. For decades, L. monocytogenes has been a tool for immunologists. It has now also become a paradigm in bacterial pathogenesis and cellular microbiology. The availability of the genome sequence of five L. monocytogenes serovars76,100,112 and L. innocua76, and soon of Listeria ivanovii and Listeria welshimeri, will provide further insight into the molecular basis of the pathogenesis determinants of Listeria species. Comparative genomics could also reveal genetic loci that confer specific pathogenic traits to epidemic strains: for example, loci that facilitate adaptation to different environments and that influence the onset of infection. It should be noted that it is often difficult to unravel the function of virulence factors that are identified by post-genomic methods and reverse genetics. To this end, more sophisticated in vivo studies — such as measurement of cytokine production, non-invasive in vivo imaging techniques or infection of ex vivo tissue explants — need to be used to understand the role of these bacterial factors in the interplay between the bacterium and the host. Applications of these novel techniques to the study of Listeria species will probably continue to show that this bacterium is an invaluable model in the fields of cellular microbiology, bacterial pathogenesis and cell biology.
Cohen, P. et al. Monitoring cellular responses to Listeria monocytogenes with oligonucleotide arrays. J. Biol. Chem. 275, 11181–11190 (2000). 7. Shen, Y., Naujokas, M., Park, M. & Ireton, K. InIBdependent internalization of Listeria is mediated by the Met receptor tyrosine kinase. Cell 103, 501–510 (2000). Identifies the cellular receptor for InlB and illustrates the exploitation of a receptor-tyrosinekinase pathway for bacterial entry. 8. Tang, P., Sutherland, C. L., Gold, M. R. & Finlay, B. B. Listeria monocytogenes invasion of epithelial cells requires the MEK-1/ERK-2 mitogen-activated protein kinase pathway. Infect. Immun. 66, 1106–1112 (1998). 9. Ireton, K. et al. A role for phosphoinositide 3-kinase in bacterial invasion. Science 274, 780–782 (1996). 10. Copp, J., Marino, M., Banerjee, M., Ghosh, P. & van der Geer, P. Multiple regions of internalin B contribute to its ability to turn on the Ras–mitogenactivated protein kinase pathway. J. Biol. Chem. 278, 7783–7789 (2003).
432 | JUNE 2006 | VOLUME 4
11.
12.
13.
14.
15.
Braun, L., Ghebrehiwet, B. & Cossart, P. gC1q-R/p32, a C1q-binding protein, is a receptor for the InlB invasion protein of Listeria monocytogenes. EMBO J. 19, 1458–1466 (2000). Bierne, H. et al. WASP-related proteins, Abi1 and Ena/ VASP are required for Listeria invasion induced by the Met receptor. J. Cell Sci. 118, 1537–1547 (2005). Bierne, H. et al. A role for cofilin and LIM kinase in Listeria-induced phagocytosis. J. Cell Biol. 155, 101–112 (2001). Veiga, E. & Cossart, P. Listeria hijacks the clathrindependent endocytic machinery to invade mammalian cells. Nature Cell Biol. 7, 894–900 (2005). Shows that the endocytic machinery is subverted for L. monocytogenes entry and therefore proposes that clathrin can help engulf large particles such as bacteria. Harvey, H. A., Jennings, M. P., Campbell, C. A., Williams, R. & Apicella, M. A. Receptor-mediated endocytosis of Neisseria gonorrhoeae into primary human urethral epithelial cells: the role of the
www.nature.com/reviews/micro © 2006 Nature Publishing Group
REVIEWS
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
asialoglycoprotein receptor. Mol. Microbiol. 42, 659–672 (2001). Van Nhieu, G. T., Krukonis, E. S., Reszka, A. A., Horwitz, A. F. & Isberg, R. R. Mutations in the cytoplasmic domain of the integrin β1 chain indicate a role for endocytosis factors in bacterial internalization. J. Biol. Chem. 271, 7665–7672 (1996). Wyrick, P. B. et al. Entry of genital Chlamydia trachomatis into polarized human epithelial cells. Infect. Immun. 57, 2378–2389 (1989). Seveau, S., Bierne, H., Giroux, S., Prevost, M. C. & Cossart, P. Role of lipid rafts in E-cadherin- and HGF-R/Met-mediated entry of Listeria monocytogenes into host cells. J. Cell Biol. 166, 743–753 (2004). Mengaud, J., Ohayon, H., Gounon, P., Mege, R. M. & Cossart, P. E-cadherin is the receptor for internalin, a surface protein required for entry of L. monocytogenes into epithelial cells. Cell 84, 923–932 (1996). Identifies E-cadherin as the first cellular receptor for L. monocytogenes and reveals a novel heterophilic interaction with E-cadherin. Perez-Moreno, M., Jamora, C. & Fuchs, E. Sticky business: orchestrating cellular signals at adherens junctions. Cell 112, 535–548 (2003). Lecuit, M. et al. A role for α- and β-catenins in bacterial uptake. Proc. Natl Acad. Sci. USA 97, 10008–10013 (2000). Drees, F., Pokutta, S., Yamada, S., Nelson, W. J. & Weis, W. I. α-Catenin is a molecular switch that binds E-cadherin–β-catenin and regulates actin-filament assembly. Cell 123, 903–915 (2005). Yamada, S., Pokutta, S., Drees, F., Weis, W. I. & Nelson, W. J. Deconstructing the cadherin–catenin– actin complex. Cell 123, 889–901 (2005). Sousa, S. et al. ARHGAP10 is necessary for α-catenin recruitment at adherens junctions and for Listeria invasion. Nature Cell Biol. 7, 954–960 (2005). Kussel-Andermann, P. et al. Vezatin, a novel transmembrane protein, bridges myosin VIIA to the cadherin–catenins complex. EMBO J. 19, 6020–6029 (2000). Sousa, S. et al. Unconventional myosin VIIa and vezatin, two proteins crucial for Listeria entry into epithelial cells. J. Cell Sci. 117, 2121–2130 (2004). Takemura, R., Stenberg, P. E., Bainton, D. F. & Werb, Z. Rapid redistribution of clathrin onto macrophage plasma membranes in response to Fc receptor–ligand interaction during frustrated phagocytosis. J. Cell Biol. 102, 55–69 (1986). Tilney, L. G. & Portnoy, D. A. Actin filaments and the growth, movement, and spread of the intracellular bacterial parasite, Listeria monocytogenes. J. Cell Biol. 109, 1597–1608 (1989). First characterization of actin nucleation by L. monocytogenes and its importance for cell–cell spread during infection. Dabiri, G. A., Sanger, J. M., Portnoy, D. A. & Southwick, F. S. Listeria monocytogenes moves rapidly through the host-cell cytoplasm by inducing directional actin assembly. Proc. Natl Acad. Sci. USA 87, 6068–6072 (1990). Kocks, C. et al. L. monocytogenes-induced actin assembly requires the actA gene product, a surface protein. Cell 68, 521–531 (1992). Domann, E. et al. A novel bacterial virulence gene in Listeria monocytogenes required for host cell microfilament interaction with homology to the proline-rich region of vinculin. EMBO J. 11, 1981–1990 (1992). References 30 and 31 identify ActA as the bacterial protein that is required for actin polymerization by L. monocytogenes. Welch, M. D., Iwamatsu, A. & Mitchison, T. J. Actin polymerization is induced by Arp2/3 protein complex at the surface of Listeria monocytogenes. Nature 385, 265–269 (1997). Shows that the Arp2/3 actin-nucleation complex initiates ActA-dependent polymerization. Machesky, L. M. et al. Scar, a WASp-related protein, activates nucleation of actin filaments by the Arp2/3 complex. Proc. Natl Acad. Sci. USA 96, 3739–3744 (1999). Lasa, I., David, V., Gouin, E., Marchand, J. B. & Cossart, P. The amino-terminal part of ActA is critical for the actin-based motility of Listeria monocytogenes; the central proline-rich region acts as a stimulator. Mol. Microbiol. 18, 425–436 (1995). May, R. C. et al. The Arp2/3 complex is essential for the actin-based motility of Listeria monocytogenes. Curr. Biol. 9, 759–762 (1999).
36. Smith, G. A., Theriot, J. A. & Portnoy, D. A. The tandem repeat domain in the Listeria monocytogenes ActA protein controls the rate of actin-based motility, the percentage of moving bacteria, and the localization of vasodilator-stimulated phosphoprotein and profilin. J. Cell Biol. 135, 647–660 (1996). 37. Niebuhr, K. et al. A novel proline-rich motif present in ActA of Listeria monocytogenes and cytoskeletal proteins is the ligand for the EVH1 domain, a protein module present in the Ena/VASP family. EMBO J. 16, 5433–5444 (1997). 38. Loisel, T. P., Boujemaa, R., Pantaloni, D. & Carlier, M. F. Reconstitution of actin-based motility of Listeria and Shigella using pure proteins. Nature 401, 613–616 (1999). 39. Geese, M. et al. Contribution of Ena/VASP proteins to intracellular motility of Listeria requires phosphorylation and proline-rich core but not F-actin binding or multimerization. Mol. Biol. Cell 13, 2383–2396 (2002). 40. Samarin, S. et al. How VASP enhances actin-based motility. J. Cell Biol. 163, 131–142 (2003). 41. Plastino, J., Olivier, S. & Sykes, C. Actin filaments align into hollow comets for rapid VASP-mediated propulsion. Curr. Biol. 14, 1766–1771 (2004). 42. Krause, M., Bear, J. E., Loureiro, J. J. & Gertler, F. B. The Ena/VASP enigma. J. Cell Sci. 115, 4721–4726 (2002). 43. Gouin, E., Welch, M. D. & Cossart, P. Actin-based motility of intracellular pathogens. Curr. Opin. Microbiol. 8, 35–45 (2005). 44. Stevens, J. M., Galyov, E. E. & Stevens, M. P. Actindependent movement of bacterial pathogens. Nature Rev. Microbiol. 4, 91–101 (2006). 45. Lecuit, M. et al. A single amino acid in E-cadherin responsible for host specificity towards the human pathogen Listeria monocytogenes. EMBO J. 18, 3956–3963 (1999). 46. Lecuit, M. et al. A transgenic model for listeriosis: role of internalin in crossing the intestinal barrier. Science 292, 1722–1725 (2001). 47. Pentecost, M., Otto, G., Theriot, J. A. & Amieva, M. R. Listeria monocytogenes invades the epithelial junctions at sites of cell extrusion. PLoS Pathog. 2, e3 (2006). 48. Khelef, N., Lecuit, M., Bierne, H. & Cossart, P. Species specificity of the Listeria monocytogenes InlB protein. Cell. Microbiol. 8, 457–470 (2006). Shows host specificity for InlB, thereby challenging the validity of some of the animal models that are used to study listerial infection and emphasizing the need to develop new models. 49. Lecuit, M. et al. Targeting and crossing of the human maternofetal barrier by Listeria monocytogenes: role of internalin interaction with trophoblast E-cadherin. Proc. Natl Acad. Sci. USA 101, 6152–6157 (2004). 50. Bakardjiev, A. I., Stacy, B. A., Fisher, S. J. & Portnoy, D. A. Listeriosis in the pregnant guinea pig: a model of vertical transmission. Infect. Immun. 72, 489–497 (2004). 51. Mansfield, B. E., Dionne, M. S., Schneider, D. S. & Freitag, N. E. Exploration of host–pathogen interactions using Listeria monocytogenes and Drosophila melanogaster. Cell. Microbiol. 5, 901–911 (2003). 52. Cheng, L. W. et al. Use of RNA interference in Drosophila S2 cells to identify host pathways controlling compartmentalization of an intracellular pathogen. Proc. Natl Acad. Sci. USA 102, 13646–13651 (2005). 53. Agaisse, H. et al. Genome-wide RNAi screen for host factors required for intracellular bacterial infection. Science 309, 1248–1251 (2005). References 52 and 53 use a novel system to identify, on a large scale, host factors that are important for supporting a listerial infection. 54. Thomsen, L. E., Slutz, S. S., Tan, M. W. & Ingmer, H. Caenorhabditis elegans is a model host for Listeria monocytogenes. Appl. Environ. Microbiol. 72, 1700–1701 (2006). 55. Mengaud, J. et al. Pleiotropic control of Listeria monocytogenes virulence factors by a gene that is autoregulated. Mol. Microbiol. 5, 2273–2283 (1991). 56. Freitag, N. E., Rong, L. & Portnoy, D. A. Regulation of the prfA transcriptional activator of Listeria monocytogenes: multiple promoter elements contribute to intracellular growth and cell-to-cell spread. Infect. Immun. 61, 2537–2544 (1993). 57. Leimeister-Wachter, M., Haffner, C., Domann, E., Goebel, W. & Chakraborty, T. Identification of a gene
NATURE REVIEWS | MICROBIOLOGY
58.
59.
60.
61.
62.
63.
64.
65.
66.
67.
68.
69.
70.
71.
72.
73.
74.
75.
that positively regulates expression of listeriolysin, the major virulence factor of Listeria monocytogenes. Proc. Natl Acad. Sci. USA 87, 8336–8340 (1990). Johansson, J. et al. An RNA thermosensor controls expression of virulence genes in Listeria monocytogenes. Cell 110, 551–561 (2002). Characterizes a novel regulatory mechanism that is encoded by the upstream untranslated region of prfA mRNA and that controls the expression of this important virulence transcription factor according to the temperature. Hoe, N. P. & Goguen, J. D. Temperature sensing in Yersinia pestis: translation of the LcrF activator protein is thermally regulated. J. Bacteriol. 175, 7901–7909 (1993). Schwab, U., Bowen, B., Nadon, C., Wiedmann, M. & Boor, K. J. The Listeria monocytogenes prfAP2 promoter is regulated by σB in a growth phase dependent manner. FEMS Microbiol. Lett. 245, 329–336 (2005). Kreft, J. & Vazquez-Boland, J. A. Regulation of virulence genes in Listeria. Int. J. Med. Microbiol. 291, 145–157 (2001). Ripio, M. T., Dominguez-Bernal, G., Lara, M., Suarez, M. & Vazquez-Boland, J. A. A Gly145Ser substitution in the transcriptional activator PrfA causes constitutive overexpression of virulence factors in Listeria monocytogenes. J. Bacteriol. 179, 1533–1540 (1997). Eiting, M., Hageluken, G., Schubert, W. D. & Heinz, D. W. The mutation G145S in PrfA, a key virulence regulator of Listeria monocytogenes, increases DNA-binding affinity by stabilizing the HTH motif. Mol. Microbiol. 56, 433–446 (2005). Jones, S. & Portnoy, D. A. Characterization of Listeria monocytogenes pathogenesis in a strain expressing perfringolysin O in place of listeriolysin O. Infect. Immun. 62, 5608–5613 (1994). Schuerch, D. W., Wilson-Kubalek, E. M. & Tweten, R. K. Molecular basis of listeriolysin O pH dependence. Proc. Natl Acad. Sci. USA 102, 12537–12542 (2005). Structural studies of the L. monocytogenes toxin LLO reveal the molecular mechanism that controls the optimal pH for its activity. Shatursky, O. et al. The mechanism of membrane insertion for a cholesterol-dependent cytolysin: a novel paradigm for pore-forming toxins. Cell 99, 293–299 (1999). Kayal, S. et al. Listeriolysin O-dependent activation of endothelial cells during infection with Listeria monocytogenes: activation of NF-κB and upregulation of adhesion molecules and chemokines. Mol. Microbiol. 31, 1709–1722 (1999). Tang, P., Rosenshine, I., Cossart, P. & Finlay, B. B. Listeriolysin O activates mitogen-activated protein kinase in eucaryotic cells. Infect. Immun. 64, 2359–2361 (1996). Sibelius, U. et al. Listeriolysin is a potent inducer of the phosphatidylinositol response and lipid mediator generation in human endothelial cells. Infect. Immun. 64, 674–676 (1996). Sibelius, U. et al. The listerial exotoxins listeriolysin and phosphatidylinositol-specific phospholipase C synergize to elicit endothelial cell phosphoinositide metabolism. J. Immunol. 157, 4055–4060 (1996). Dramsi, S. & Cossart, P. Listeriolysin O-mediated calcium influx potentiates entry of Listeria monocytogenes into the human Hep-2 epithelial cell line. Infect. Immun. 71, 3614–3618 (2003). Tsuchiya, K. et al. Listeriolysin O-induced membrane permeation mediates persistent interleukin-6 production in Caco-2 cells during Listeria monocytogenes infection in vitro. Infect. Immun. 73, 3869–3877 (2005). Wadsworth, S. J. & Goldfine, H. Listeria monocytogenes phospholipase C-dependent calcium signaling modulates bacterial entry into J774 macrophage-like cells. Infect. Immun. 67, 1770–1778 (1999). Wadsworth, S. J. & Goldfine, H. Mobilization of protein kinase C in macrophages induced by Listeria monocytogenes affects its internalization and escape from the phagosome. Infect. Immun. 70, 4650–4660 (2002). Giddings, K. S., Zhao, J., Sims, P. J. & Tweten, R. K. Human CD59 is a receptor for the cholesteroldependent cytolysin intermedilysin. Nature Struct. Mol. Biol. 11, 1173–1178 (2004).
VOLUME 4 | JUNE 2006 | 433 © 2006 Nature Publishing Group
REVIEWS 76. Glaser, P. et al. Comparative genomics of Listeria species. Science 294, 849–852 (2001). Comparative analysis of the genome sequences of L. monocytogenes (which is pathogenic) and L. innocua (which is non-pathogenic), greatly advancing the understanding of Listeria species pathogenicity. 77. Williams, T., Bauer, S., Beier, D. & Kuhn, M. Construction and characterization of Listeria monocytogenes mutants with in-frame deletions in the response regulator genes identified in the genome sequence. Infect. Immun. 73, 3152–3159 (2005). 78. Autret, N., Raynaud, C., Dubail, I., Berche, P. & Charbit, A. Identification of the agr locus of Listeria monocytogenes: role in bacterial virulence. Infect. Immun. 71, 4463–4471 (2003). 79. Mandin, P. et al. VirR, a response regulator critical for Listeria monocytogenes virulence. Mol. Microbiol. 57, 1367–1380 (2005). 80. Cotter, P. D., Emerson, N., Gahan, C. G. & Hill, C. Identification and disruption of lisRK, a genetic locus encoding a two-component signal transduction system involved in stress tolerance and virulence in Listeria monocytogenes. J. Bacteriol. 181, 6840–6843 (1999). 81. Dons, L. et al. Role of flagellin and the two-component CheA/CheY system of Listeria monocytogenes in host cell invasion and virulence. Infect. Immun. 72, 3237–3244 (2004). 82. Christiansen, J. K., Larsen, M. H., Ingmer, H., Sogaard-Andersen, L. & Kallipolitis, B. H. The RNAbinding protein Hfq of Listeria monocytogenes: role in stress tolerance and virulence. J. Bacteriol. 186, 3355–3362 (2004). 83. Archambaud, C., Gouin, E., Pizarro-Cerda, J., Cossart, P. & Dussurget, O. Translation elongation factor EF-Tu is a target for Stp, a serine–threonine phosphatase involved in virulence of Listeria monocytogenes. Mol. Microbiol. 56, 383–396 (2005). 84. Annous, B. A., Becker, L. A., Bayles, D. O., Labeda, D. P. & Wilkinson, B. J. Critical role of anteiso-C15:0 fatty acid in the growth of Listeria monocytogenes at low temperatures. Appl. Environ. Microbiol. 63, 3887–3894 (1997). 85. Angelidis, A. S. & Smith, G. M. Role of the glycine betaine and carnitine transporters in adaptation of Listeria monocytogenes to chill stress in defined medium. Appl. Environ. Microbiol. 69, 7492–7498 (2003). 86. Ko, R., Smith, L. T. & Smith, G. M. Glycine betaine confers enhanced osmotolerance and cryotolerance on Listeria monocytogenes. J. Bacteriol. 176, 426–431 (1994). 87. Junttila, J. R., Niemela, S. I. & Hirn, J. Minimum growth temperatures of Listeria monocytogenes and non-haemolytic Listeria. J. Appl. Bacteriol. 65, 321–327 (1988). 88. Cotter, P. D., Gahan, C. G. & Hill, C. Analysis of the role of the Listeria monocytogenes F0F1-ATPase operon in the acid tolerance response. Int. J. Food Microbiol. 60, 137–146 (2000). 89. Cotter, P. D., Gahan, C. G. & Hill, C. A glutamate decarboxylase system protects Listeria monocytogenes in gastric fluid. Mol. Microbiol. 40, 465–475 (2001). 90. Dussurget, O. et al. Listeria monocytogenes bile salt hydrolase is a PrfA-regulated virulence factor involved in the intestinal and hepatic phases of listeriosis. Mol. Microbiol. 45, 1095–1106 (2002). 91. Begley, M., Sleator, R. D., Gahan, C. G. & Hill, C. Contribution of three bile-associated loci, bsh, pva, and btlB, to gastrointestinal persistence and bile
92.
93.
94.
95.
96.
97.
98.
99.
100.
101.
102.
103.
104.
105.
106.
107.
108.
109.
tolerance of Listeria monocytogenes. Infect. Immun. 73, 894–904 (2005). Sleator, R. D., Wemekamp-Kamphuis, H. H., Gahan, C. G., Abee, T. & Hill, C. A PrfA-regulated bile exclusion system (BilE) is a novel virulence factor in Listeria monocytogenes. Mol. Microbiol. 55, 1183–1195 (2005). Hardy, J. et al. Extracellular replication of Listeria monocytogenes in the murine gall bladder. Science 303, 851–853 (2004). Goetz, M. et al. Microinjection and growth of bacteria in the cytosol of mammalian host cells. Proc. Natl Acad. Sci. USA 98, 12221–12226 (2001). Chico-Calero, I. et al. Hpt, a bacterial homolog of the microsomal glucose-6-phosphate translocase, mediates rapid intracellular proliferation in Listeria. Proc. Natl Acad. Sci. USA 99, 431–436 (2002). O’Riordan, M., Moors, M. A. & Portnoy, D. A. Listeria intracellular growth and virulence require host-derived lipoic acid. Science 302, 462–464 (2003). Klarsfeld, A. D., Goossens, P. L. & Cossart, P. Five Listeria monocytogenes genes preferentially expressed in infected mammalian cells: plcA, purH, purD, pyrE and an arginine ABC transporter gene, arpJ. Mol. Microbiol. 13, 585–597 (1994). Chatterjee, S. S. et al. Intracellular gene expression profile of Listeria monocytogenes. Infect. Immun. 74, 1323–1338 (2006). Joseph, B. et al. Identification of Listeria monocytogenes genes contributing to intracellular replication by expression profiling and mutant screening. J. Bacteriol. 188, 556–568 (2006). An important study that identifies the listerial genes that are required to sustain growth in the cytoplasm of the host cell. Doumith, M. et al. New aspects regarding evolution and virulence of Listeria monocytogenes revealed by comparative genomics and DNA arrays. Infect. Immun. 72, 1072–1083 (2004). Cabanes, D., Dehoux, P., Dussurget, O., Frangeul, L. & Cossart, P. Surface proteins and the pathogenic potential of Listeria monocytogenes. Trends Microbiol. 10, 238–245 (2002). Dhar, G., Faull, K. F. & Schneewind, O. Anchor structure of cell wall surface proteins in Listeria monocytogenes. Biochemistry 39, 3725–3733 (2000). Bierne, H. et al. Inactivation of the srtA gene in Listeria monocytogenes inhibits anchoring of surface proteins and affects virulence. Mol. Microbiol. 43, 869–881 (2002). Pucciarelli, M. G. et al. Identification of substrates of the Listeria monocytogenes sortases A and B by a nongel proteomic analysis. Proteomics 5, 4808–4817 (2005). Sabet, C., Lecuit, M., Cabanes, D., Cossart, P. & Bierne, H. LPXTG protein InlJ, a newly identified internalin involved in Listeria monocytogenes virulence. Infect. Immun. 73, 6912–6922 (2005). Reglier-Poupet, H. et al. Maturation of lipoproteins by type II signal peptidase is required for phagosomal escape of Listeria monocytogenes. J. Biol. Chem. 278, 49469–49477 (2003). Braun, L. et al. InlB: an invasion protein of Listeria monocytogenes with a novel type of surface association. Mol. Microbiol. 25, 285–294 (1997). Lenz, L. L., Mohammadi, S., Geissler, A. & Portnoy, D. A. SecA2-dependent secretion of autolytic enzymes promotes Listeria monocytogenes pathogenesis. Proc. Natl Acad. Sci. USA 100, 12432–12437 (2003). Cabanes, D., Dussurget, O., Dehoux, P. & Cossart, P. Auto, a surface associated autolysin of Listeria
434 | JUNE 2006 | VOLUME 4
110.
111.
112.
113.
114.
115.
monocytogenes required for entry into eukaryotic cells and virulence. Mol. Microbiol. 51, 1601–1614 (2004). Cabanes, D. et al. Gp96 is a receptor for a novel Listeria monocytogenes virulence factor, Vip, a surface protein. EMBO J. 24, 2827–2838 (2005). Li, Z., Dai, J., Zheng, H., Liu, B. & Caudill, M. An integrated view of the roles and mechanisms of heat shock protein gp96–peptide complex in eliciting immune response. Front. Biosci. 7, d731–d751 (2002). Nelson, K. E. et al. Whole genome comparisons of serotype 4b and 1/2a strains of the food-borne pathogen Listeria monocytogenes reveal new insights into the core genome components of this species. Nucleic Acids Res. 32, 2386–2395 (2004). Cossart, P. & Lecuit, M. Interactions of Listeria monocytogenes with mammalian cells during entry and actin-based movement: bacterial factors, cellular ligands and signaling. EMBO J. 17, 3797–3806 (1998). Schubert, W. D. et al. Structure of internalin, a major invasion protein of Listeria monocytogenes, in complex with its human receptor E-cadherin. Cell 111, 825–836 (2002). Tweten, R. K. Cholesterol-dependent cytolysins, a family of versatile pore-forming toxins. Infect. Immun. 73, 6199–6209 (2005).
Acknowledgements Our sincere apologies to colleagues whose work has not been cited in this manuscript because of space constraints and because of the focus of this Review, which does not claim to be comprehensive. We thank E. Gouin for critical assessment of the manuscript and E. Veiga for help with FIG. 2. Work in the laboratory of P.C. is supported by the Institut National de la Recherche Agronomique, the Institut Pasteur, the Institut National de la Santé et de la Recherche Médicale, the Ministère de l’Education Nationale et de la Recherche Scientifique et Technique, and the Association par la Recherche sur le Cancer. M.H. is supported by a Pasteur Foundation fellowship. P.C. is an international research scholar of the Howard Hughes Medical Institute.
Competing interests statement The authors declare no competing financial interests.
DATABASES The following terms in this article are linked online to: Entrez Genome Project: http://www.ncbi.nlm.nih.gov/ entrez/query.fcgi?db=genomeprj Bacillus subtilis | Burkholderia pseudomallei | Chlamydia trachomatis | Escherichia coli | Listeria innocua | Listeria ivanovii | Listeria monocytogenes | Listeria welshimeri | Shigella flexneri | Yersinia pestis UniProtKB: http://ca.expasy.org/sprot InlA | InlB | LLO | PrfA
FURTHER INFORMATION Bacteria–Cells Interactions Unit, Institut Pasteur: http://www.pasteur.fr/recherche/unites/uibc/welcome.html ListiList: http://genolist.pasteur.fr/ListiList/ Microbiology in Motion: http://www.nature.com/nrmicro/animation/index.html The Prokaryotes: An Evolving Electronic Resource for the Microbiological Community: http://link.springer-ny.com/link/service/books/10125/
SUPPLEMENTARY INFORMATION See online article: S1 (figure) | S2 (figure) | S3 (figure) Access to this links box is available online.
www.nature.com/reviews/micro © 2006 Nature Publishing Group
REVIEWS
Signalling and oxidant adaptation in Candida albicans and Aspergillus fumigatus Neeraj Chauhan*, Jean-Paul Latge‡ and Richard Calderone*
Abstract | Candida species and Aspergillus fumigatus were once thought to be relatively benign organisms. However, it is now known that this is not the case — Candida species rank among the top four causes of nosocomial infectious diseases in humans and A. fumigatus is the most deadly mould, often having a 90% mortality rate in immunocompromised transplant recipients. Adaptation to stress, including oxidative stress, is a necessary requisite for survival of these organisms during infection. Here, we describe the latest information on the signalling pathways and target proteins that contribute to oxidant adaptation in C. albicans and A. fumigatus, which has been obtained primarily through the analysis of mutants or inference from genome annotation. This review focuses on three questions concerning signalling and oxidant adaptation in Candida albicans and Aspergillus fumigatus. First, can correlates be made between the role and regulation of antioxidant proteins — which have largely been defined by in vitro studies — and the survival of these organisms in phagocytic cells? Second, the paradigm for innate immuneprotection against invasive candidiasis, which was derived from both basic science and clinical observations, is that neutrophils and not monocytes are more important in killing C. albicans. Are there new data to support this conclusion? And third, does probing the antioxidant mechanisms of these fungi provide hints for improving patient care? Before answering these questions, we first briefly describe the magnitude of the healthcare problems caused by these two fungal pathogens, and their pathogenesis.
*Georgetown University Medical Center, Department of Microbiology & Immunology, Washington DC 20057, USA. ‡ Aspergillus Unit, Institut Pasteur, 28, rue du Dr Roux, 75015 Paris, France. Correspondence to R.C. e-mail:
[email protected] doi:10.1038/nrmicro1426
Invasive candidiasis and aspergillosis Candidiasis is the fourth most common cause of nosocomial infectious disease (NID) in the US, and published data indicate a similar pattern worldwide1–6. Invasive aspergillosis (IA) is a less common cause of NID, but mortality has been reported to be as high as 90% in some studies2,4,7–10. Over the past decade, IA, which is most often caused by A. fumigatus, has emerged as the most serious life-threatening infectious complication of intensive remission-induction chemotherapy and allogeneic bone marrow transplants, and in patients with various haematological malignancies8,9. Patients with solid
NATURE REVIEWS | MICROBIOLOGY
organ transplants, AIDS, chronic pulmonary diseases including cystic fibrosis, and the rare inherited chronic granulomatous disease are also at risk. The magnitude of NID caused by these fungi is reflected in patient costs4. During 1998, the cost of treatment of invasive fungal infections, estimated to be approximately 62,000 cases per year in the US, was $2.6 billion, and the average perpatient attributable cost was $31,200. Given the clinical importance of candidiasis and IA, efforts continue to identify new drug targets and therefore new anti-fungal drugs, to improve diagnostic assays and to provide alternative therapeutic options such as the use of passive immune anti-Candida antibodies or vaccines for the treatment of high-risk patients11,12.
C. albicans and A. fumigatus: pathogenesis Our understanding of the pathogenesis of these organisms has been partly accelerated by the availability of molecular tools and the application of bioinformatics. For example, the genomes of both C. albicans and A. fumigatus have been sequenced, and reverse-genetic approaches are available to assess gene function13–16. However, classical genetic analysis of each organism is not generally performed, either because mating has not been identified (A. fumigatus) or because mating is prevented by a diploid genome in which most strains are self-sterile (C. albicans)17. Instead, for both organisms, non-sexual mechanisms might be crucial to achieve genetic variation — a subject better understood in C. albicans than in A. fumigatus18. VOLUME 4 | JUNE 2006 | 435
© 2006 Nature Publishing Group
REVIEWS A a c f
b
b
d
e B
Ciliated epithelium
T cell
Cytokines PMN
AM
Alveolus
Figure 1 | Candidiasis and invasive aspergillosis. A | Disseminated candidiasis can originate at a gastrointestinal site. Candida albicans enters epithelial microvilli through persorption of yeast cells or by germination (a,c). In both cases, organisms enter the vasculature (b,d) for dissemination into tissues such as the kidney (e). Typically, it localizes in the cortex (f) where it grows as hyphae/pseudohyphae. A vigorous host response occurs at this site consisting of both mononuclear and polymorphonuclear leukocytes. Virulence factors (adhesins, morphogenesis, switch phenotypes, antioxidant proteins and invasive enzymes) promote the invasion of the organism. B | Aerosols of Aspergillus fumigatus conidia are inhaled and travel to the alveoli. In the healthy host, alveolar macrophages (AM) phagocytose and kill the organism after swelling of the conidium, an essential pre-germination stage. The production of reactive oxygen intermediates by AM is required to eliminate the organism, but polymorphonuclear cells (PMNs) also contribute. In the immunosuppressed patient, reduced numbers of PMNs and inefficient AM allow growth of the fungus. Consequently, the conidia germinate and escape from the AM. Direct invasion of the ciliated epithelium has also been reported (upper panel).
436 | JUNE 2006 | VOLUME 4
The complex nature of the host–pathogen interactions involved in candidiasis and IA continues to be assessed. In candidiasis, immunity seems to be sitespecific; innate immunity plays a much greater part in resistance to invasive disease whereas T-cell immunity protects against oral mucosal candidiasis19. Likewise, innate immunity, including phagocytic cell functions, is crucial for protection against IA, but so far we do not have a complete picture of immunity to these diseases. The pathogenesis of the invasive diseases caused by both opportunistic fungi is illustrated in FIG. 1. C. albicans is a commensal of mucosal surfaces, and invasive disease can be initiated from the gut mucosa when predisposing conditions are present. Invasion begins either through persorption of the organism by epithelial cells or by adherent yeast cells germinating through the intestinal microvilli20. Once within the epithelia, the organism crosses the endothelial layer and enters the bloodstream by undefined mechanisms that might be growth-form specific (FIG. 1). All growth forms (yeast and hyphae/ pseudohyphae) are required for disease. Intuitively, yeast forms might be thought most suitable for dissemination of the organism to tissues and, because of their larger size and manner of growth, filamentous forms thought more suitable for invasion, but this hypothesis is probably too simplistic an explanation. Once in the bloodstream, the fungus disseminates to tissues such as the kidney, one of the key target organs of Candida species. In the kidney, growth of the organism is usually, but not always, limited to the cortex following infiltration of host cells (FIG. 1). There, phagocytic cells have an important role in limiting or allowing disease to occur. Despite the need for an immunocompromising condition to predispose patients to invasive disease, virulence factors, such as adherence proteins, the yeast– hyphal phase transition and invasive enzymes (secreted aspartyl proteases and lipases) seem to be required for fully virulent strains21. By contrast, A. fumigatus is an environmental pathogen that we constantly inhale. Following inhalation by a healthy host, A. fumigatus conidia are phagocytosed and killed by alveolar macrophages that produce powerful reactive oxygen intermediates (ROIs)22 (FIG. 1). If conidial germination occurs before or after phagocytosis, neutrophils migrate to the site of germination. Upon contact, the neutrophils degranulate, release higher amounts of ROIs than are found in alveolar macrophages and kill the mycelium. Oxidative mechanisms such as the myeloperoxidase (MPO) system are either inefficient or are activated later in infection. Such findings were initially made in mice and have been confirmed in humans with congenital diseases. For example, transgenic mice with X-linked chronic granulomatous disease (X-CGD) — an inherited immunodeficiency caused by a defect in the NADPH oxidase complex in phagocytic cells that results in defective production of the superoxide anion and its metabolites — results in a dramatic increase in mortality due to IA23. By contrast, mice with an MPO deficiency, resulting in a lack of oxidants from hydrogen peroxide (H2O2), including hypochlorous acid (HOCl), are not susceptible to IA, suggesting that the superoxide
www.nature.com/reviews/micro © 2006 Nature Publishing Group
REVIEWS Table 1 | From spot-plate assays to phagocytes — proteins that protect against oxidants Protein
Survival in peroxide?
Survival in PMN?
Survival in MNC?
Transcribed in PMN?
Virulence?
Refs
Ssk1p
Yes
Yes
ND
ND
Yes
41,55,130
Cap1p
Yes
Yes
ND
Yes
Yes
43,90
Sod5p
Yes
Yes
No
Yes
Yes
43,47
Cta1p (Cct1p, Cat1p)
Yes
Yes
ND
ND
Yes
42,46
Sod1p
Yes
ND
Yes
ND
Yes
44
Chk1p
Yes
Yes
ND
ND
Yes
48,68,69
ND
Yes
ND
Yes
107–109
Candida albicans
Aspergillus fumigatus Alb1
Yes
Cat1, Cat2, CatA
Yes
No
No
ND
No
103
Gliz, Glip
Yes
Yes
ND
ND
ND
110,111
MNC, mononuclear cells; ND, not determined; PMN, polymorphonuclear cells.
produced by NADPH oxidase is more important in protection than the HOCl produced by MPO24. Also, patients with MPO deficiency are not known to be predisposed to IA, whereas CGD patients are at high risk of developing various infections, including IA. However, neutrophils from both CGD and MPO-deficiency patients are defective in damaging A. fumigatus hyphae, implying that other factors might explain the lack of IA in patients with an MPO deficiency25,26. These data suggest that, at least in patients, the efficient oxidative pathway is MPO independent, a conclusion that is supported by the observation that alveolar macrophages that lack MPO are still efficient in killing conidia22. Studies have found the cidal mechanisms of ROIs to be growth-form specific, in that oxidative mechanisms kill conidia, whereas oxidative-burst and non-oxidative mechanisms kill hyphae27. In contrast to A. fumigatus, for C. albicans, MPO has an essential role in protection (especially at high inoculum levels), early host defence and macrophage activation27–30. However, other studies describe a requirement for non-oxidative mechanisms including proteases or polycations to kill Candida31–35.
Oxidant adaptation in phagocytic cells Interactions with phagocytes occupy a large portion of the disease cycle of each pathogen, whether in the blood, tissues, alveoli or at mucosal surfaces. So, it follows that to survive in this environment requires, among other things, induction of the signalling pathways that sense oxidants (input) and the production of downstream antioxidants (output) to detoxify these oxidants. But is there proof of this? Do in vitro studies of antioxidants in these fungi have any relevance to what occurs in phagocytes in vivo, given the paradigm that neutrophils are more protective than mononuclear phagocytes36,37? In the following sections, we focus on oxidative adaptation and do not attempt to describe other events that occur during the interactions of these pathogens with phagocytes. For instance, Toll-like receptors have been proposed to regulate the balance between polymorphonuclear cell NATURE REVIEWS | MICROBIOLOGY
(PMN)-dependent protection and immunopathology in both organisms, and readers are directed to other sources for this information19,38–40. In regard to how relevant in vitro studies are, there are many genes for which the role in oxidant adaptation in vitro (usually determined by spot-plate assays) is known but whether they have a role in survival in phagocytes has only been investigated for a few candidates. Included in this category are the C. albicans genes that encode the response regulator Ssk1p, the histidine kinase Chk1p, three antioxidant proteins (Cat1p, Sod2p and Sod5p) and the transcription factor Cap1p41–48 (TABLE 1). The correlations between in vitro and ex vivo (human pooled phagocytes) functions were made by constructing mutants and then testing each for in vitro oxidant susceptibility and ex vivo survival in phagocytes. Importantly, each mutant was avirulent in a murine model of invasive candidiasis (TABLE 1). We interpret the mutant data to mean that there is a correlation, albeit with a limited number of genes, between oxidant resistance in vitro and survival in neutrophils and, importantly, with virulence. Below, we describe each of these gene products, and those of A. fumigatus, and their functions in greater detail. Comparisons are also drawn to the model yeast Saccharomyces cerevisiae when possible.
Ssk1p and the HOG MAPK pathway One of the major signal transduction pathways in C. albicans, which senses oxidative stress and initiates the transcriptional regulation that allows cells to adapt, is the high osmolarity glycerol (HOG) mitogen-activated protein kinase (MAPK) pathway, which is also involved in sensing osmotic stress49. The HOG MAPK protein was initially described in S. cerevisiae and assigned a function in osmoadaptation50 (FIG. 2). Upstream of the HOG MAPK pathway are multi-step phosphorelay proteins that determine whether this pathway is activated. In unstressed cells, a membrane-bound sensor protein, Sln1p, autophosphorylates on a histidine residue within VOLUME 4 | JUNE 2006 | 437
© 2006 Nature Publishing Group
REVIEWS a No stress
Ssk2p
ATP
P D P
H
D
P
Ssk1p
H
Sln1p
Pbs2p
Ypd1p Hog1p No response
b Oxidative/osmotic stress No phosphotransfer D
H Sln1p
P
D
H Ypd1p
Ssk2p
Ssk1p P Hog1p P Adaptive response
Pbs2p
P Hog1p
c Oxidative/osmotic stress P Sln1p
P
P
Ypd1p
Ypd1p P Skn7p
Ste2p
P Ste2p
Sho1p
Skn7p-dependent genes P Hog1p
Pbs2p
P
P Hog1p
Adaptive response
Figure 2 | The high osmolarity glycerol (HOG) mitogen-activated protein kinase (MAPK) pathway in fungi. a | In unstressed conditions, Sln1p is autophosphorylated on histidine (H) and aspartate (D) residues. The protein then phosphorylates Ypd1p, a phosphohistidine intermediate protein, on a histidine residue; Ssk1p acts as a kinase to phosphorylate its aspartate residue (D). Downstream phosphotransfer to the HOG MAPK pathway is inhibited due to the inability of phosphorylated Ssk1p to activate the Ssk2p MAPK kinase kinase (MAPKKK). All phosphotransfers occur within the H box domain (Sln1p and Ypd1p) or receiver domain (Sln1p and Ssk1p). b | In cells subject to osmotic or oxidant stress, upstream phosphorylation does not occur such that Ssk1p activates the Ssk2p MAPKKK. Subsequent phosphotransfer to Hog1p through Pbs2p results in its translocation to the nucleus, allowing the transcriptional changes required for cells to adapt to stress. c | Skn7p also provides oxidant adaptation functions. The Sho1p branch pathway of Saccharomyces cerevisiae also adapts cells to osmotic stress through Pbs2p and Hog1p, bypassing the need for the upstream proteins mentioned in a and b. The current model for S. cerevisiae is that Skn7p is always in the nucleus. In hypoosmotic stress, Ypd1p (either phosphorylated via phosphotransfer from Sln1p or dephosphorylated) is translocated to the nucleus. Ypd1p then phosphorylates Skn7p, which leads to activation of Skn7-dependent gene expression132. Phosphorylated Ypd1p also phosphorylates Ssk1p to negatively regulate the HOG MAP kinase cascade as described.
438 | JUNE 2006 | VOLUME 4
a domain referred to as an H-box, and the phosphate is transferred to an aspartate residue within its receiver domain. Subsequently, the phosphate is transferred to a histidine residue in Ypd1p, a histidine-containing phosphotransfer (HPt) protein, and then to Ssk1p on an aspartate residue. Phosphorylated Ssk1p is unable to induce activation of Ssk2p, the MAPK kinase kinase of the HOG MAPK cascade. However, under osmotic (S. cerevisiae) or oxidative/osmotic (C. albicans) stress, phosphotransfer through Sln1p-Ypd1p-Ssk1p does not occur, and unphosphorylated Ssk1p activates Ssk2p, which in turn activates Pbs2p (a MAPK kinase) and finally Hog1p (a MAPK). Phosphorylated Hog1p then translocates to the nucleus and initiates the transcription of genes associated with adaptation (FIG. 2). The negative regulation by upstream phosphorylation prevents cells from transcribing genes in the absence of stress, not a trivial saving in cellular energy considering that the expression of 10% of S. cerevisiae genes is altered when cells encounter high osmotic conditions49. Another mechanism for osmoadaptation in S. cerevisae and oxidant adaptation in C. albicans involves the Sho1p sensor, which interacts with Pbs2p to activate Hog1p, but activation of this pathway in S. cerevisiae requires greater osmolarity50,51 (FIG. 2). There are functional and regulatory differences between the C. albicans HOG MAPK pathway and the S. cerevisiae pathway. For example, deletion of SLN1 or YPD1 in S. cerevisiae is lethal — there is no off switch for Hog1p phosphorylation as Ssk1p is not phosphorylated and Hog1p is always activated; lethality is due to an overproduction of glycerol, an osmolyte produced in response to osmotic stress to maintain cell turgor50. In C. albicans, deletion of sln1 results in a mild-to-moderate sensitivity to osmotic or peroxide shock, but the mutation is not lethal (N.C. and R.C., unpublished observations and REF. 52). At present, the basis for this difference in regulation is not understood, but could be explained by the presence of a robust Sho1p-mediated or other adaptation pathway that bypasses Ssk1p and Sln1p in C. albicans (FIG. 2). No functional inferences can be made for C. albicans YPD1, as a deletion mutant has not been reported53. Other differences focus on C. albicans Ssk1p, which contributes to oxidant adaptation and is also directly or indirectly part of the transcriptional regulation of cellwall components41,54. For example, the ssk1Δ mutant is sensitive to several oxidants, and the in vitro sensitivity correlates with its reduced survival in human PMNs and avirulence41,55 (TABLE 1). Transcriptional analysis of an ssk1 mutant has shown that approximately 13% of the genes, transcription of which is altered in the mutant, encode stress-adaptation proteins 41. The number of transcriptional changes is remarkable given the fact that these data are obtained from cells grown in unstressed media, indicating that the mutant probably has difficulties adapting to its own oxidant wastes. Also, approximately 12% of changes occur in genes associated with cell-wall functions, of which one of the major hostrecognition proteins, Als1p, is downregulated, which correlates with the reduced adherence of the mutant to human oesophageal tissue41,56–58.
www.nature.com/reviews/micro © 2006 Nature Publishing Group
REVIEWS Histidine kinases Chk1p N
C
Nik1p N
Sln1p
C
N
C
Phospho-histidine intermediate Ypd1p N
Response regulators Ssk1p N
C
C
Skn7p N
C Coiled coil
Histidine phosphotransfer
Response regulator
Ser/Thr kinase
Transmembrane domain
His kinase A (phosphoacceptor)
Heat shock factor-DNA binding
Histidine kinase, DNA gyrase B-, phytochrome-like ATPases
Histidine kinase, adenylyl cyclase, methyl-binding protein, phosphatase
GAF signalling
Figure 3 | The two-component signal proteins of Candida albicans and their domains. The proteins are grouped as histidine kinases, a phospho-histidine intermediate and response regulators. Both Chk1p and Ssk1p are needed for peroxide resistance, survival in polymorphonuclear cells and virulence.
The phosphohistidine intermediate protein Ypd1p of S. cerevisiae also interacts with a second response regulator, Skn7p, which in S. cerevisiae also provides antioxidant function and cell-wall biosynthesis regulation50. As with Ssk1p, Skn7p is a response-regulator protein and, at least in S. cerevisiae, Skn7p acts as a transcription factor independent of the HOG pathway. As such, Skn7p is functionally more similar to the response regulators of bacteria, which do not have MAPK pathways59. As with S. cerevisiae Ssk1p, C. albicans Skn7p is required for adaptation to oxidative stress in vitro60–62. However, the C. albicans ssk1 mutant is sensitive to a wide range of oxidants, including menadione, whereas the skn7 mutant is not, and the ssk1 mutant is avirulent whereas the skn7 mutant retains virulence61. Of the C. albicans HOG-MAPK-pathway proteins, functions have been determined for Pbs2p and Hog1p but not for Ssk2p63–67. A pbs2 mutant of C. albicans was sensitive to both osmotic and oxidative stress, as was a hog1 mutant, but interestingly, the pbs2 mutant lost viability more quickly than the hog1 mutant under oxidative stress66. Both Hog1p and Pbs2p also have a role in cell-wall biogenesis, as determined by the susceptibility of hog1 and pbs2 mutants to cell-wall inhibitors such as Congo red and calcofluor white. The association between Hog1p and cell-wall biosynthesis has not been established for S. cerevisiae Hog1p. As is true for the ssk1 and hog1 mutants, co-regulation of oxidant adaptation and cell-wall biosynthesis is observed either as an increase in sensitivity to cell-wall inhibitors or through changes in the transcription of genes associated with each mutant41,65. In addition to
NATURE REVIEWS | MICROBIOLOGY
Sln1p, Ssk1p, Ypd1p and Skn7p, C. albicans has two other histidine kinases (HKs), Chk1p and Nik1p (FIG. 3), both of which are required for virulence, morphogenesis and cell-wall biosynthesis48,68–73. Furthermore, the CHK1 mutant is oxidant-sensitive, killed more readily by human neutrophils than wild-type cells and is avirulent68–70 (TABLE 1) .However, of the upstream phosphorelay and HOG-MAPK-pathway proteins, Ssk1p, Pbs2p and Hog1p are required for virulence, but virulence has only been linked to survival in neutrophils for Ssk1p. Three homologues of the HOG-MAPK-pathway proteins have been studied in A. fumigatus. Similar to C. albicans, the Sln1p homologue TscB is not involved in reactive-oxidant sensing as a tscB mutant is as sensitive to reactive oxidants as the parental strain74. This observation can be explained by the functional redundancy in A. fumigatus HKs. A survey of the A. fumigatus genome has shown the presence of 13 HKs, grouped in three subfamilies — a finding that might be associated with the need to resist multiple external environmental insults16. The high number of HK genes in A. fumigatus is typical of filamentous ascomycetes72. Another HK which has been studied, fos1, the Nik1p homologue of Neurospora crassa, is required for cell-wall assembly and virulence75,76, but its role in oxidant adaptation has not been investigated. Also, similar to C. albicans, recent studies have shown that the Skn7p homologue in A. fumigatus has a role in protection against oxidative stress (R.C. & J.-P.L., unpublished observations). Hog1p (SakA) mutants of A. fumigatus are also oxidant sensitive74,77. The Hog1p homologue in A. fumigatus is also associated with morphogenesis as it has a role in nutrient sensing and conidial germination77. Therefore, it is apparent that the signal transduction functions of the HOG pathway are similar in these two human opportunistic pathogens despite their very distant taxonomy. S. cerevisiae is often presented as a model organism, but this does not seem to be the case for oxidant-signalling pathways.
Cap1p and oxidative-stress adaptation S. cerevisiae Yap1p is a basic leucine zipper (bZIP) transcription factor belonging to the AP-1 family that is required for oxidative-stress tolerance and mediates pleiotropic drug and metal resistance49,78–81. Yap1p is localized to the nucleus in response to the presence of oxidants82,83. Among the genes that are transcriptionally regulated by Yap1p are GPX2 (glutathione peroxidase), TRX2 (thioredoxin) and GSH1 (glutathione biosynthesis)84–87. The Yap1p homologue in C. albicans is Cap1p. CAP1 partially complements the S. cerevisiae yap1 mutant, and C. albicans cells lacking CAP1 are hypersensitive to oxidative stress induced by diamide or H2O2 (REFS 88, 89). Cap1p also has a role in multi-drug resistance, and its transcription is induced in the presence of human neutrophils43,88. Similar to Yap1p, Cap1p localizes to the nucleus in response to oxidative stress89. CAP1 transcription increases when phagocytosed by human PMNs, and deletion of this gene decreases survival in PMNs and virulence43,90 (TABLE 1). The relationship between Cap1p, Hog1p and the global transcription response VOLUME 4 | JUNE 2006 | 439
© 2006 Nature Publishing Group
REVIEWS that follows the introduction of oxidants and other stress conditions has recently been reported for C. albicans67. The regulation of oxidant adaptation varied according to the concentration of peroxide to which cells were exposed. Inactivation of HOG1 had multiple effects on the global transcription response during peroxide stress, but only about 15% of the genes for which transcription was altered were dependent on Hog1p. In light of this observation, it has been suggested that Cap1p is another major player in adaptation of C. albicans to oxidant stress. The current belief is that Hog1p acts alone, or in parallel with, other signalling pathways. For example, deletion of HOG1 has no effect on Cap1p translocation, and deletion of CAP1 does not inhibit Hog1p phosphorylation. With regard to oxidant adaptation, Cap1p regulates adaptation to both low and high peroxide concentrations, and Hog1p regulates adaptation to high peroxide concentrations67. Other transcription factors that regulate oxidant adaptation have been described in S. cerevisiae, but their functions in C. albicans are either unknown (Sko1p), or they do not have a role in stress adaptation (Msn2p and Msn4p)91–94. The role of the yap1 and sko1 homologues of A. fumigatus is unknown.
Antioxidant enzymes As with most organisms, in addition to specific oxidative-response pathways, fungi also possess various antioxidant enzymes to counter the threat of oxidative damage, including superoxide dismutase (SOD), thioredoxin, glutathione reductase, catalase and glutathione peroxidase42,44–47,95–101. C. albicans has six SODs, the cytoplasmic Sod1p (CuZn SOD) and Sod3p (Mn-SOD), the mitochondrial Sod2p (Mn-SOD) and three other Cu-Zn SODs, Sod4p–6p (REFS 44,45,47,99,100). Of the SODs, the most studied with regard to their role in pathogenesis are Sod5p and Sod1p (TABLE 1). Upregulation of SOD5 was observed in the yeast-to-hyphal transition in the presence of nonfermentable carbon sources, and when cells were grown in oxidative and osmotic stress47. A SOD5 deletion mutant of C. albicans was sensitive to H2O2 when cells were grown in nutrient-limiting conditions and was also required for virulence in a mouse model of infection. However, the mutant could survive in macrophages to the same extent as wild-type cells47. Transcription profiling of C. albicans with PMNs from whole human blood indicated that SOD5 was upregulated following phagocytosis43. In this same study, survival of the sod5 mutant in PMN cells was significantly reduced compared with wild-type cells (TABLE 1). A sod1 mutant of C. albicans has also been examined and was shown to be sensitive to menadione but not H2O2, to be more sensitive to killing by cultured macrophages than wild-type cells, and to have reduced virulence in an invasive murine model44 (TABLE 1). A single catalase has been reported in C. albicans (designated CTA1, CCT1 or CAT1). Cta1p protects cells from peroxide stress and is required for virulence in a murine model of invasive candidiasis42,46 (TABLE 1). Furthermore, hyphae of a CTA1 mutant were damaged 440 | JUNE 2006 | VOLUME 4
by human neutrophils, indicating a protective effect for this gene against PMNs42. In A. fumigatus, one extracellular Cu-Zn SOD, two intracellular Mn-SODs and three catalases (one conidial and two mycelial) have been identified, as have glutathione transferases and peroxidases, and glutaredoxin and thioredoxin systems102–104. Single-gene (catA, conidia) and double-gene (cat1/2, mycelia) disruptions in catalases of A. fumigatus result in conidia and mycelia lacking catalase activity. These mutants are slightly more sensitive to H2O2 than wild-type cells. Infection with the cat1/2 double mutant results in a delay in infection and a reduction in the inflammatory reactions in an experimental model of IA, indicating that both mycelial catalases provide transient protection for the fungus103. The limited reduction of virulence of the A. fumigatus catalase mutants might be due to compensatory reactions by redundant genes of the glutathione pathway neutralizing exogenous H2O2. Mutants in the glutathione pathway or in SOD have not been constructed in this fungus, despite the exquisite sensitivity of A. fumigatus to the superoxide ion (10 µM for menadione). However, both Mn- and Cu-Zn SODs have been shown to be immunoreactive in humans105,106. As for other antioxidants, melanin pigmentation of conidia seems to protect the organism from being killed by oxidants in vitro, and from human monocytes by scavenging ROIs. A pigment-less mutant with a defect in alb1, encoding a polyketide synthase involved in melanin biosynthesis, was significantly less virulent in mice than wild type107–109. Avirulence was associated with greater binding of complement component C3 to the alb1 mutant compared with binding to wild-type conidia, which probably explains the increased phagocytosis by human neutrophils108. Recently, alb1 in A. fumigatus conidia was associated with inhibition of phagolysosome fusion in human monocyte-derived macrophages109. Gliotoxin (GT), a secondary metabolite that specifically inhibits NADPH oxidase, greatly impaired the killing activity of PMNs against A. fumigatus conidia, and the hyphae were not damaged110,111. The inhibition of O2– generation is dependent on the disulphide bridge of GT that interferes with oxidase activation but not catalysis of activated oxidase. GT mutants have been recently constructed through the disruption of the non-ribosomal peptide synthase GLIP or the transcription factor GLIZ, but their role in fungal virulence has not been clearly established.
The core stress response: signal integration In S. cerevisiae, there is a common environmental response (CER) pathway (also referred to as a general or core stress response) in which a common set of genes is induced by different types of stress, including heat shock, osmotic stress and oxidative stress112. This observation might indicate conservation of the signalling and/or transcriptional regulation that allows cells to adapt to different stresses. In S. cerevisiae, regulation of the core response to different stress conditions is linked to the general stressresponse transcription factors Msn2p and Msn4p, but there is no single regulatory pathway that adapts cells to
www.nature.com/reviews/micro © 2006 Nature Publishing Group
REVIEWS various types of stress112. Therefore, S. cerevisiae Hog1p seems to respond primarily to changes in osmolarity, whereas Hsf1p and Skn7p regulate temperature shock, and Skn7p and Yap1p regulate oxidative stress49,50. A CER pathway is also operative in C. albicans but only when cells are subjected to concentration-dependent conditions that activate the MAPK Hog1p. Therefore, for C. albicans the transcriptional changes observed at 0.4 mM H2O2, 0.3 M NaCl, or during heat shock (conditions which induce a CER in S. cerevisiae) show minimal overlap, indicating that a CER pathway does not occur113. By contrast, other data indicate that Hog1p does regulate a core stress response in C. albicans under more drastic stress conditions, such as 2–5 mM H2O2(REF 114). As
Neutrophil
Monocyte
+
+
Anti-oxidants ↑ Glyoxylate enzymes↑ Amino acid metabolism↑ Yeast cell phenotype
Anti-oxidants ↓ Glyoxylate enzymes↓ Amino acid metabolism↓ Hyphal phenotype
Phagocyte response Chemotaxis pseudopodia increase Membrane ruffling Toll receptor signalling Complement binding Inflammatory receptors/proteins Granule movement/activation Non-oxidative mechanisms, for example defensins
Figure 4 | The processes that precede and follow phagocytosis of Candida albicans yeast cells by neutrophils (left) and monocytes (right). Both types of phagocytes display chemotaxis and other changes associated with their function as responder cells (middle box). In the panels on the left and right, the transcriptional changes of C. albicans in either type of phagocyte are indicated. Neutrophils can inhibit the yeast-to-hyphae transition, whereas monocytes cannot. Fewer C. albicans genes encoding antioxidant functions are transcriptionally upregulated in monocytes (4 out of 18) than in neutrophils (16 out of 18). Likewise, a greater number of genes encoding amino acid biosynthesis functions and glyoxylate enzymes are upregulated in neutrophils than in monocytes. Interestingly, in neutrophils, downregulation of genes associated with protein biosynthesis (7 out of 7) was reported, whereas in monocytes no changes were observed. The data were extracted from REF 43.
NATURE REVIEWS | MICROBIOLOGY
stated above, S. cerevisiae Hog1p responds primarily to osmotic stress conditions, but activation of C. albicans Hog1p (measured by its phosphorylation and translocation to the nucleus) occurs under a range of conditions, including osmotic stress, menadione (which produces superoxide), H2O2, the heavy metals Cd2+ and As3+, caffeine, farnesol (which induces quorum sensing) and serum114. These functional differences intuitively relate to the needs of each organism: C. albicans must overcome the challenges of oxidative stress at mucosal surfaces or when engulfed by human phagocytes, whereas S. cerevisiae uses Hog1p to adapt to osmotic stress as it is confronted with high osmolarity in its environment. A feature of the CER pathway in S. cerevisiae is that prior exposure to a mild stress can protect the cells against a subsequent stronger challenge with the same or a different stress112. For C. albicans, such cross-protection is only partially observed. For example, a mild increase (two fold) in resistance to oxidative stress is induced by mild heat stress, but there is no improvement in survival if a mild oxidative stress is followed by a strong heat shock. However, if the primary stress condition activates Hog1p through phosphorylation, then subsequent exposure to high levels of a second, unrelated stress confers protection. So, Hog1p phosphorylation seems to regulate a cross-protective response114. In summary, there are several genes in these pathogens the contribution of which to oxidant adaptation in vitro can be correlated with their survival and virulence in phagocytes in vivo. Undoubtedly, there are other signal pathways and transcriptional regulators that are also involved. Part of the reason for the lack of data on other genes is that fresh neutrophils cannot be kept in culture, and harvesting these cells from volunteers yields insufficient numbers for multiple mutant analyses. However, the recent development of a neutrophil cell line (HL-60 myelomonocytic leukaemia cells), that when treated with dimethyl formamide gains anti-Candida activity, will be useful in future studies115.
New data on an old paradigm Next, we address the question of whether new data support the paradigm that neutrophils are more effective than monocytes and macrophages in their ability to kill C. albicans. To answer this question, we turn to a recent analysis that compared the gene-transcription profiles of C. albicans incubated with human blood cells 43. In this study, the influence of whole human blood on the transcription profile and growth of C. albicans was compared with the influence of enriched populations of erythrocytes, mononuclear cells (MNCs), PMNs, CD15+/— cells and plasma (FIG. 4). Several conclusions were drawn, including the observation that the PMN cell fraction closely mimicked the events that occurred in whole blood with respect to gene transcription and its effect on the growth of the organism. The authors therefore speculated that PMNs are the dominant cells responsible for the strong transcriptional response observed for C. albicans in whole blood. This was also reflected in varied growth profiles. For example, yeast forms predominated VOLUME 4 | JUNE 2006 | 441
© 2006 Nature Publishing Group
REVIEWS in whole blood whereas in erythrocytes, plasma or monocytes, hyphae were the predominant growth form; approximately 97% of C. albicans cells remained in the yeast form in PMNs (FIG. 4). With regard to the transcription profiling of antioxidants from infected PMNs, the expression of 16 out of 18 antioxidant genes was upregulated, whereas only 2 of the same 18 genes were upregulated in infected MNCs. Among the highest of the upregulated genes in PMN cell incubations were those involved in glutathione biosynthesis. The transcriptional changes in C. albicans in PMNs and whole blood were similar for amino-acid-encoding genes, the glyoxylate-cycle gene family, genes involved in protein synthesis and genes involved in glycolysis. In comparison, the MNC fractions more often resembled the erythrocyte and plasma fractions in both gene profiling and, as stated above, in C. albicans growth profiles43. In summary, gene profiling has established many events that occur following phagocytosis; these changes have been correlated microscopically with growth of the organism to show that, for example, C. albicans does not convert to hyphae in infected PMNs but does so in infected MNCs. Furthermore, C. albicans uses a more aggressive antioxidant response in PMNs versus MNCs43. Therefore, the transcriptional events that occur in human neutrophil and monocyte populations are substantially different, and these differences seem to correlate with a greater growth-inhibitory activity of PMNs compared with MNCs. So, the paradigm remains firmly entrenched, PMNs are the major contributors to innate immunity against C. albicans, supporting the well established observations on in vitro differences in the killing of C. albicans by neutrophils versus monocytes/macrophages, in addition to observations from patient data.
Practical applications for patients Can our knowledge of the oxidant-adaptation mechanisms in these organisms be applied to immune potentiation? In this regard, pretreatment of PMNs with interleukin-15 (IL-15) or tumour necrosis factor-α (TNF-α) led to an increase in the oxidative burst and hyphal damage to A. fumigatus116,117. Cytokines such as granulocyte colony-stimulating factor (G-CSF), granulocyte–macrophage colony-stimulating factor (GM-CSF), macrophage colony-stimulating factor (M-CSF), interferon-γ (IFN-γ) and TNF-α have also been shown to enhance the respiratory burst of phagocytes116–129. Glucocorticoids, which are associated with a predisposition to IA, have been shown to reduce ROI production in both alveolar macrophages and neutrophils22,125. As a result of cytokine treatment, the inhibitory effect of immunosuppressive therapies is attenuated and conidial and hyphal damage is increased. Such immunomodulatory therapy has been applied to CGD patients with some success126. The site of administration of the cytokine seems to be crucial, with the intranasal administration being most efficient for IFN-γ127,128. In candidiasis, pretreatment of C. albicans with a monoclonal antibody that targets the acid-labile 442 | JUNE 2006 | VOLUME 4
cell-wall oligomannan increased killing of the organism by neutrophils129. Therefore, passive transfer of antibodies might be protective in high-risk patients by enhancing killing of the organism by PMNs. The identification of new targets for drug discovery is also important, and one possibility is Ssk1p which, as we have described, has an important role in virulence and protection against ROIs in neutrophils130.
Conclusions Some of the signal-transduction pathways and the downstream antioxidants that they regulate following phagocytosis have been identified in C. albicans and A. fumigatus, but they probably account for only a fraction of the genetic machinery that contributes to the survival of these pathogens. Equally important, but not addressed in this review, is how these organisms adapt to the oxidants generated by their own cellular respiration or in a mixed microbiota environment. In the case of the human pathogenic fungi C. albicans and A. fumigatus, adaptation against stress such as ROIs enables invasive disease to occur if the immunity of the infected individual is compromised. This might indicate that antioxidant adaptation does not have a role in the survival of these organisms during encounters with host phagocytes. However, several observations, described above, argue that oxidant adaptation does have a role following phagocytosis. First, mutants of C. albicans deficient in antioxidant proteins (Sod5p, Cta1p and Sod2p (Alb1 in A. fumigatus)) or lacking signalling/transcriptional proteins (Ssk1p, Yap1p and Chk1p) are more susceptible to neutrophil killing or damage than the wild-type strains from which they were derived. Second, although neutrophils are quite candicidal, monocytes are not. The same picture has been seen for A. fumigatus, where hyphae are quickly killed by neutrophils (a few hours), whereas the killing of conidia by macrophages is much slower. So, survival in PMNs could depend more on antioxidant adaptation and, especially in neutropenic patients, monocytes would constitute the primary phagocytic cells involved in resistance. Third, the virulence of wild-type strains in animal models is well established, but virulence is abrogated in strains lacking antioxidant adaptive mechanisms (for example, SSK1)130. Additional studies on the contribution of antioxidant proteins to the survival of these fungi might need to focus on specific sites in the host. For example, unpublished observations indicate that the ssk1 mutant is less able to colonize the vaginal mucosa (R.C. & J-P.L., unpublished observations), but its survival at the gastrointestinal and oropharyngeal mucosa has not been reported. Are insights into patient management gained from these studies? Recently, infused neutrophils were shown to protect against murine neutropenic models of invasive candidiasis and IA131. Furthermore, the lack of some of the antioxidant regulatory proteins in the human host (for example, Ssk1p) suggests that new therapeutics that target these proteins could be a possibility. These examples serve to indicate the need for continued study of these important human–fungus interactions.
www.nature.com/reviews/micro © 2006 Nature Publishing Group
REVIEWS 1.
2.
3.
4.
5.
6.
7.
8.
9. 10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
Wisplinghoff, H. et al. Nosocomial bloodstream infections in US hospitals: analysis of 24,179 cases from a prospective nationwide surveillance study. Clin. Infect. Dis. 39, 309–317 (2004). Ascioglu, S. et al. Defining opportunistic fungal infections in immunocompromised patients with cancer and hematopoietic stem cell transplants: an international consensus. Clin. Infect. Dis. 34, 7–14 (2002). Stover, B. H. et al. Nosocomial infection rates in US children’s hospitals neonatal and pediatric intensive care units. Am. J. Infect. Control 29, 152–157 (2001). Wilson, L. S. et al. The direct cost and incidence of systemic fungal infections. Value Health 5, 26–34 (2002). Wisplinghoff, H. et al. Nosocomial blood stream infections in pediatric patients in US hospitals: epidemiology, clinical features, and susceptibilities. Pediatr. Infect. Dis. J. 22, 686–691 (2003). Wenzel, R. P. Nosocomial candidemia: risk factors and attributable mortality. Clin. Infect. Dis. 20, 153–154 (1995). One of the first studies on the estimation of mortality in the background of immunocompromised patients. Anderson, J. B. Antifungal-drug resistance: mechanisms and pathogen fitness. Nature Rev. Microbiol. 3, 547–556 (2005). Singh, N. & Paterson, D. L. Aspergillus infections in transplant recipients. Clin. Microbiol. Rev. 18, 44–69 (2005). Denning, D. W. Invasive aspergillosis. Clin. Infect. Dis. 26, 781–803 (1998). Yamazaki, T. et al. Epidemiology of visceral mycoses: analysis of data in the annual of the pathological autopsy cases in Japan. J. Clin. Microbiol. 37, 1732–1738 (1999). Casadevall, A., Dadachova, E. & Pirofski, L. Passive antibody therapy for infectious diseases. Nature Rev. Microbiol. 2, 695–703 (2004). Calderone, R. In Candida and Candidiasis Ch.1 (ed. Calderone, R.) 3–15 (ASM Press, Washington DC, 2002). Bruno, V. M. & Mitchell, A. Largescale gene function analysis in Candida albicans. Trends Microbiol. 12, 157–161 (2004). Denning, D. W., Anderson, M. J., Turner, G., Latge, J.-P. & Bennett, J. W. Sequencing the Aspergillus fumigatus genome. Lancet Infect. Dis. 2, 251–253 (2002). Jones, T. et al. The diploid genome sequence of Candida albicans. Proc. Natl Acad. Sci. USA 101, 7329–7334 (2004). Nierman, W. et al. Genomic sequence of the pathogenic and allergenic filamentous fungus Aspergillus fumigatus. Nature 438, 1151–1156 (2005). Magee, B. B. & Magee, P. T. Through a glass opaquely: the biological significance of mating in Candida albicans. Science 289, 310–313 (2004). Rustchenko, E. & Sherman, F. Genetic instability in Candida albicans. In Pathogenic Fungi in Humans and Animals (ed. Howard, D. H.) 723–776 (Marcel Dekker, New York, 2003). Romani, L. Innate immunity to fungi: the art of speed and specificity, In Pathogenic Fungi: Host Interactions and Emerging Strategies for Control (eds San Blas, G. & Calderone, R.) 167–214 (Calister Academic Press, Norfolk, 2004). Cole, G., Seshan, K. R., Lynn, K. T. & Franco, M. Gastrointestinal candidiasis: histopathology of Candida–host interactions in a murine model. Mycol. Res. 4, 385–408 (1993). Calderone, R. & Fonzi, W. Virulence attributes of Candida albicans. Trends Microbiol. 9, 327–335 (2001). Phillipe, B. et al. Killing of Aspergillus fumigatus is mediated by reactive oxygen intermediates. Infect. Immun. 71, 3034–3042 (2003). Dennis, C. G. et al. Effect of amphotericin B and micafungin combination on survival, histopathology and fungal burden in experimental aspergillosis in the p47phox–/– mouse model of chronic granulomatous disease. Antimicrob. Agents Chemother. 50, 422–427 (2006). Aratani, Y. et al. Relative contributions of myeloperoxidase and NADPH-oxidase to the early host defense against pulmonary infections with Candida albicans and Aspergillus fumigatus. Med. Mycol. 40, 557–563 (2002). Rex, J. H., Bennett, J. E., Gallin, J. I., Malech, H. L. & Melnick, D. A. Normal and deficient neutrophils can cooperate to damage Aspergillus fumigatus hyphae. J. Infect. Dis. 162, 523–528 (1990). Washburn, R., Gallin, J. & Bennett, J. Oxidative killing of Aspergillus fumigatus proceeds by parallel myeloperoxidase-dependent and -independent pathways. Infect. Immun. 55, 2088–2092 (1987).
27. Bellocchio, S. et al. TLRs govern neutrophil activity in aspergillosis. J. Immunol. 173, 7406–7415 (2004). 28. Aratani, Y. et al. Severe impairment in early host defense against Candida albicans in mice deficient in myeloperoxidase. Infect. Immun. 67, 1828–1836 (1999). 29. Aratani, Y. et al. Critical role of myeloperoxidase and nicotinamide adenine dinucleotide phosphate-oxidase in high-burden systemic infection of mice with Candida albicans. J. Infect. Dis. 185, 1833–1837 (2002). 30. Lefkowitz, S. S., Gelderman, M. P., Lefkowitz, D. L., Moguilevsky, N. & Bollen, A. Phagocytosis and intracellular killing of Candida albicans by macrophages exposed to myeloperoxidase. J. Infect. Dis. 173, 1202–1207 (1996). 31. Balish, E. et al. Susceptibility of germfree phagocyte oxidase- and nitric oxide synthase 2-deficient mice, defective in the production of reactive metabolites of both oxygen and nitrogen, to mucosal and systemic candidiasis of endogenous origin. Infect. Immun. 73, 1313–1320 (2005). 32. Diamond, R. D., Lyman, C. A. & Wysong, D. R. Disparate effects of interferon-γ and tumor necrosis factor-α on early neutrophil respiratory burst and fungicidal responses to Candida albicans hyphae in vitro. J. Clin. Invest. 87, 711–720 (1991). 33. Lehrer, R. I., Ganz, T., Szklarek, D. & Selsted, M. E. Modulation of the in vitro candidacidal activity of human neutrophil defensins by target cell metabolism and divalent cations. J. Clin. Invest. 81, 1829–1835 (1998). 34. Newman, S. L., Bhugra, B., Holly, A. & Morris, R. E. Enhanced killing of Candida albicans by human macrophages adherent to type 1 collagen matrices via induction of phagolysosomal fusion. Infect. Immun. 73, 770–777 (2005). 35. Reeves, E, P. et al. Killing activity of neutrophils is mediated through activation of proteases by K+ flux. Nature 416, 291–297 (2002). 36. Lehrer, R. I. & Cline, M. J. Interaction of Candida albicans with human leukocytes. J. Bacteriol. 98, 996–1004 (1969). 37. Cutler, J. E. & Poor, A. H. Effect of mouse phagocytes on Candida albicans in in vivo chambers. Infect. Immun. 31, 1110–1116 (1981). 38. Romani, L., Bistoni, F. & Puccetti, P. Fungi, dendritic cells, and receptors: a host perspective of fungal virulence. Trends Microbiol. 10, 508–514 (2002). 39. Balloy, V. et al. Involvement of Toll-like receptor 2 in experimental invasive pulmonary aspergillosis. Infect. Immun. 73, 5420–5425 (2005). 40. van Der Graaf, C. A., Netea, M. G., Verschueren, I., van der Meer, J. W. & Kullberg, B. J. Differential cytokine production and Toll-like receptor signaling pathways by Candida albicans blastoconidia and hyphae. Infect. Immun. 73, 7458–7464 (2005). 41. Chauhan, N. et al. The SSK1 of Candida albicans is associated with oxidative stress adaptation and cell wall biosynthesis. Euk. Cell 2, 1018–1024 (2003). 42. Wysong, D. R., Christin, L., Sugar, A. L., Robbins, P. W. & Diamond, R. D. Cloning and sequencing of a Candida albicans catalase gene and affects of disruption of the gene. Infect. Immun. 66, 1953–1961 (1998). 43. Fradin, C. et al. Granulocytes govern the transcriptional response, morphology and proliferation of Candida albicans in human blood. Mol. Microbiol. 56, 397–415 (2005). An exceptional presentation of data comparing the relative contribution of blood-cell populations in resistance to C. albicans. The paper clearly demonstrates that neutrophils are the primary mediators of innate immunity. 44. Hwang, C. S. et al. Copper- and zinc-containing superoxide dismutase (Cu/ZnSOD) is required for the protection of Candida albicans against oxidative stresses and the expression of its full virulence. Microbiology 148, 3705–3713 (2002). 45. Hwang, C. S., Baek, Y. U., Yim, H. S. & Kang, S. O. Protective roles of mitochondrial manganese-containing superoxide dismutase against various stresses in Candida albicans. Yeast 20, 929–941 (2003). 46. Nakagawa, Y., Kanbe, T. & Mizuguchi, I. Disruption of the human pathogenic yeast Candida albicans catalase gene decreases survival in mouse-model infection and elevates susceptibility to higher temperature and to detergents. Microbiol. Immunol. 47, 395–403 (2003). 47. Martchenko, M., Alarco, A. M., Harcus, D. & Whiteway, M. Superoxide dismutases in Candida albicans: transcriptional regulation and functional characterization of the hyphal-induced SOD5 gene. Mol. Biol. Cell 15, 456–467 (2004). 48. Kruppa, M. et al. The Chk1p of Candida albicans and its role in the regulation of cell wall synthesis. FEMS Yeast Res. 3, 289–299 (2003).
NATURE REVIEWS | MICROBIOLOGY
49. Moye-Rowley, S. Regulation of the transcriptional response to oxidative stress in fungi: similarities and differences. Euk. Cell 2, 381–389 (2003). 50. Hohmann, S. Osmotic stress signaling and osmoadaptation in yeasts. Microbiol. Mol. Biol. Rev. 66, 300–372 (2002). 51. Roman, E., Nombela, C. & Pla, J. The Sho1 adaptor protein links oxidative stress to morphogenesis and cell wall biosynthesis in the fungal pathogen Candida albicans. Mol. Cell. Biol. 25, 10611–10627 (2005). 52. Nagahashi, S., et al. Isolation of CaSLN1 and CaNIK1, the genes for osmosensing histidine kinase homologues, from the pathogenic fungus, Candida albicans. Microbiology 144, 425–432 (1998). Among the first papers that described the twocomponent proteins and their functions in human pathogenic fungi. 53. Calera, J. A., Herman, D & Calderone, R. Identification of Ypd1p, a gene of Candida albicans which encodes a two-component phosphohistidine intermediate protein. Yeast 16, 1053–1059 (2000). 54. Calera, J. A. & Calderone, R. Identification of a putative response regulator, two-component phosphorelay gene (CaSSK1) from Candida albicans. Yeast 15, 1243–1254 (1999). 55. Chen, D., Calderone, R., Richert, J. & Li. D. Deletion of the SSK1 response regulator gene in Candida albicans contributes to enhanced killing by human polymorphonuclear neutrophils. Infect. Immun. 73, 865–871 (2005). 56. Li, D., Bernhardt, J. & Calderone, R. Temporal expression of the Candida albicans genes CHK1 and CSSK1, adherence and morphogenesis in a model of reconstituted human esophageal epithelial candidiasis. Infect. Immun. 70, 1558–1565 (2002). 57. Bernhardt, J, Herman, D., Sheridan, M. & Calderone, R. Adherence and invasion studies of Candida albicans strains utilizing in vitro models of esophageal candidiasis. J. Infect. Dis. 184, 1170–1175 (2001). 58. Hoyer, L. The ALS gene family of Candida albicans. Trends Microbiol. 9, 176–180 (2001). 59. Stock, A. M. & West, A. H. Response regulators and their interactions with histidine protein kinases In Histidine Kinases in Signal Transduction 237–271 (Elsevier Science, USA, 2002). A concise and well explained review of the response-regulator proteins in bacteria, including the structural requirements for phosphotransfer. 60. Lee, J. et al. Yap1 and Skn7 control two specialized oxidative stress response regulons in yeast. J. Biol. Chem. 274, 16040–16046 (1999). 61. Singh, P. et al. SKN7 of Candida albicans: mutant construction and phenotype analysis. Infect. Immun. 72, 2390–2394 (2004). 62. Jamieson, D. J., Stephen, D. W. S. & Terriere, E. C. Analysis of the adaptive oxidative response of Candida albicans. FEMS Microbiol. Lett. 138, 83–88 (1996). 63. San Jose, C. et al. The mitogen-activated protein kinase homolog HOG1 gene controls glycerol accumulation in the pathogenic fungus Candida albicans. J. Bacteriol. 178, 5850–5852 (1996). 64. Alonso-Monge, R. et al. The Hog1 MAP kinase is essential in the oxidative stress response and chlamydospore formation in Candida albicans. Euk. Cell 2, 351–361 (2003). 65. Alonso-Monge, R., et al. Role of mitogen-activated protein kinase Hog1p in morphogenesis and virulence of Candida albicans. J. Bacteriol. 181, 3058–3068 (1999). 66. Arana, D. M., Nombela, C., Alonso-Monge, R. & Pla, J. The Pbs2 MAP kinase kinase is essential for the oxidative stress response in the fungal pathogen Candida albicans. Microbiology 151, 1033–1049 (2005). 67. Enjalbert, B. et al. Role of Hog1 stress-activated protein kinase in global transcriptional response to stress in the fungal pathogen Candida albicans. Mol. Biol. Cell 17, 1018–1032 (2006). A comprehensive, well written presentation of gene regulation during stress in C. albicans. 68. Yamada-Okabe, T. et al. Roles of three histidine kinase genes in hyphal development and virulence of the pathogenic fungus, Candida albicans. J. Bacteriol. 181, 7243–7247 (1999). 69. Torosantucci, A. et al. Deletion of the two-component histidine kinase gene (CHK1) of Candida albicans contributes to enhanced growth inhibition and killing by human neutrophils in vitro. Infect. Immun. 70, 985–987 (2002). 70. Calera, J. A. & Calderone, R. A. Histidine kinase, twocomponent signal transduction proteins of Candida albicans and the pathogenesis of candidosis. Mycoses 42 S49–S53 (1999).
VOLUME 4 | JUNE 2006 | 443 © 2006 Nature Publishing Group
REVIEWS 71. 72.
73. 74.
75.
76.
77.
78.
79. 80.
81. 82. 83. 84.
85.
86.
87.
88.
89.
90.
91.
92.
93.
94.
Santos, J. L. & Shiozaki, K. Fungal histidine kinases. Science STKE 98, 1–14 (2001). Catlett, N. L., Yoder, O. C. & Turgeon, B. G. Wholegenome analysis of two-component signal transduction genes in fungal pathogens. Euk. Cell 2, 1151–1161 (2003). A must-read presentation of two-component proteins in filamentous fungi. Li, D. et al. Transcription of CHK1 of Candida albicans is regulated by Ssk1p and Hog1p. Microbiology 150, 3305–3315 (2004). Chen, D., Sarfati, J., Latge, J-P. & Calderone, R. The SakA (Hog1) and TscB (Sln1) two-component proteins in oxidant adaptation in Aspergillus fumigatus. Med. Mycol. (In the press). Pott, G. B., Miller, T. K., Bartlett, J. A., Palas, J. S. & Selitrennikoff, C. P. The isolation of FOS-1, a gene encoding a putative two-component histidine kinase from Aspergillus fumigatus. Fungal Genet. Biol. 31, 55–67 (2000). Clemons, K. V., Miller, T. K., Selitrennikoff, C. P. & Stevens, D. A. fos-1, a putative histidine kinase as a virulence factor for systemic aspergillosis. Med. Mycol. 40, 259–262 (2002). Xue, T. et al. A mitogen-activated protein kinase that senses nitrogen regulates conidial germination and growth in Aspergillus fumigatus. Euk. Cell 3, 557–560 (2004). Wendler F. et al. Diazaborine resistance in the yeast Saccharomyces cerevisiae reveals a link between YAP1 and the pleiotropic drug resistance genes PDR1 and PDR3. J. Biol. Chem. 272, 27091–27098 (1997). Wu, A. et al. Yeast bZip proteins mediate pleiotropic drug and metal resistance. J. Biol. Chem. 268, 18850–18858 (1993). Wysocki, R. et al. Transcriptional activation of metalloid tolerance genes in Saccharomyces cerevisiae requires the AP-1-like proteins Yap1p and Yap8p. Mol. Biol. Cell 15, 2049–60 (2004). Moye-Rowley, W. S. Transcription factors regulating the response to oxidative stress in yeast. Antioxid. Redox Signal. 4, 123–140 (2002). Huh, W. K. et al. Global analysis of protein localization in budding yeast. Nature 425, 686–691 (2003). Kuge, S., Jones, N. & Nomoto, A. Regulation of yAP-1 nuclear localization in response to oxidative stress. EMBO J. 16, 1710–1720 (1997). Tanaka, T., Izawa, S. & Inoue, Y. GPX2, encoding a phospholipid hydroperoxide glutathione peroxidase homologue, codes for an atypical 2-Cys peroxiredoxin in Saccharomyces cerevisiae. J. Biol. Chem. 280, 42078–42087 (2005). Ross, S. J., Findlay, V. J., Malakasi, P. & Morgan, B. A. Thioredoxin peroxidase is required for the transcriptional response to oxidative stress in budding yeast. Mol. Biol. Cell. 11, 2631–2642 (2000). Dormer, U. H., Westwater, J., Stephen, D. W. & Jamieson, D. J. Oxidant regulation of the Saccharomyces cerevisiae GSH1 gene. Biochim. Biophys. Acta 1576, 23–29 (2002). Tsuzi D., Maeta K., Takatsume Y., Izawa S. & Inoue Y. Regulation of the yeast phospholipid hydroperoxide glutathione peroxidase GPX2 by oxidative stress is mediated by Yap1 and Skn7. FEBS Lett. 565, 148–154 (2004). Alarco, A. M. & Raymond, M. The bZip transcription factor Cap1p is involved in multidrug resistance and oxidative stress response in Candida albicans. J. Bacteriol. 181, 700–708 (1999). Zhang, X., De Micheli, M., Coleman, S. T., Sanglard, D. & Moye-Rowley, W. S. Analysis of the oxidative stress regulation of the Candida albicans transcription factor, Cap1p. Mol. Microbiol. 36, 618–629 (2000). Bahn, Y. S. & Sundstrom, P. CAP1, an adenylate cyclase-associated protein gene, regulates bud-hypha transition filamentous growth, and cAMP levels, is required for virulence of Candida albicans. J. Bacteriol 183, 3211–3223 (2001). Rep, M. et al. The Saccharomyces cerevisiae Sko1p transcription factor mediates HOG pathway-dependent osmotic regulation of a set of genes encoding enzymes implicated in protection from oxidative damage. Mol. Microbiol. 40, 1067–1083 (2001). Pascual-Ahuir, A., Posas, F., Serrano, R. & Proft, M. Multiple levels of control regulate the yeast cAMPresponse element-binding protein repressor Sko1p in response to stress. J. Biol. Chem. 276, 37373–37378 (2001). Martinez-Pastor, M. T. et al. The Saccharomyces cerevisiae zinc finger proteins Msn2p and Msn4p are required for transcriptional induction through the stress response element (STRE). EMBO J. 15, 2227–2235 (1996). Nicholls, S. et al. Msn2- and Msn4-like transcription factors play no obvious roles in the stress responses of the fungal pathogen Candida albicans. Euk. Cell 3, 1111–1123 (2004).
95. Rementeria, A. et al. Genes and molecules involved in Aspergillus fumigatus virulence. Rev. Iberoam. Micol. 22, 1–23 (2005). 96. Lushchak, V. I. & Gospodaryov, D. V. Catalases protect cellular proteins from oxidative modification in Saccharomyces cerevisiae. Cell. Biol. Int. 29, 187–192 (2005). 97. Avery, A. M. & Avery, S. V. Saccharomyces cerevisiae expresses three phospholipid hydroperoxide glutathione peroxidases. J. Biol. Chem. 276, 33730–33735 (2001). 98. Ikner, A. & Shiozaki, K. Yeast signaling pathways in the oxidative stress response. Mut. Res. 569, 13–27 (2005). 99. O’Brien, K. M., Dirmeier, R., Engle, M. & Poyton, R. O. Mitochondrial protein oxidation in yeast mutants lacking manganese-(MnSOD) or copper- and zinccontaining superoxide dismutase (CuZnSOD): evidence that MnSOD and CuZnSOD have both unique and overlapping functions in protecting mitochondrial proteins from oxidative damage. J. Biol. Chem. 279, 51817–51827 (2004). 100. Lamarre, C., LeMay, J. D., Deslauriers, N. & Bourbonnais, Y. Candida albicans expresses an unusual cytoplasmic manganese-containing superoxide dismutase (SOD3 gene product) upon the entry and during the stationary phase. J. Biol. Chem. 276, 43784–43791 (2001). 101. Urban, C. et al. The moonlighting protein Tsa1p is implicated in oxidative stress response and in cell wall biogenesis in Candida albicans. Mol. Microbiol. 57, 1318–1341 (2005). 102. Holdom, M. D., Lechenne, B., Hay, R. J., Hamilton, A. J. & Monod, M. Production and characterization of recombinant Aspergillus fumigatus Cu, Zn superoxide dismutase and its recognition by immune human sera. J. Clin. Microbiol. 38, 558–562 (2000). 103. Paris, S. et al. Catalases of Aspergillus fumigatus. Infect. Immun. 71, 3551–3562 (2003). 104. Tekaia, F. & Latge, J. P. Aspergillus fumigatus: saprophyte or pathogen. Curr. Opin. Microbiol. 8, 385–392 (2005). 105. Schmid-Grendelmeier, P. et al. IgE-mediated and T cellmediated autoimmunity against manganese superoxide dismutase in atopic dermatitis. J. Allergy Clin. Immunol. 115, 1068–1075 (2005). 106. Burns, C. et al. Identification, cloning, and functional expression of three glutathione transferase genes from Aspergillus fumigatus. Fungal Genet. Biol. 42, 319–327 (2005). 107. Jahn, B. et al. Characterization of a pigment less conidium mutant of Aspergillus fumigatus with altered conidial surface and reduced virulence. Infect. Immun. 65, 5110–5117 (1997). 108. Tsai, H. F., Chang, Y. C., Washburn, P. G., Wheeler, R. & Kwong-Chung, K. J. The developmental regulation of alb1 gene of Aspergillus fumigatus: its role in modulation of conidia morphogenesis and virulence. J. Bacteriol. 180, 3031–3038 (1998). 109. Brakhage, A. A. & Liebmann, B. Aspergillus fumigatus conidial pigment and cAMP signal transduction: significance for virulence. Med. Mycology 43,S75–S82 (2005). 110. Tsunawaki, S., Yoshida, L. S., Nishida, S., Kobayashi, T. & Shimoyama, T. Fungal metabolite gliotoxin inhibits assembly of the human respiratory burst NADPH oxidase. Infect Immun. 72, 3373–3382 (2004). 111. Sugui, J. et al. Disruption of the non-ribosomal peptide synthase gene, glip, in the Aspergillus fumigatus gliotoxin pathway. 2nd Symp. of Advances against Aspergillosis. Abstr. 18 (Athens, 2006). 112. Ruis, H. & Schuller, C. Stress signaling in yeast. Bioassays 17, 959–965 (1995). 113. Enjalbert, B., Nantel, A. & Whiteway, M. Stress-induced gene expression in Candida albicans: absence of a general stress response. Mol. Biol. Cell 14, 1460–1467 (2003). 114. Smith, D. A. et al. A conserved stress-activated protein kinase regulates a core stress response in the human pathogen Candida albicans. Mol. Biol. Cell 15, 4179–4190 (2004). 115. Mullick, A. et al. Gene expression in HL60 granulocytoids and human polymorphonuclear leukocytes exposed in Candida albicans. Infect. Immun. 72, 414–429 (2004). 116. Winn, R. M. et al. Effects of IL-15 on anti-fungal activity with IL-8 release by PMN in response to hyphae of Aspergillus fumigatus. J. Infect. Dis. 188, 585–590 (2003). 117. Rolides, E. et al. TNF-α enhances antifungal activity of PMN and monocytes against Aspergillus fumigatus. Infect. Immun. 66, 5999–6003 (1998).
444 | JUNE 2006 | VOLUME 4
118. Liles, W. C., Huang, J. E., van Burik, J. A. H., Bowden, R. A. & Dale, D. C. Granulocyte colony-stimulating factor administered in vivo augments neutrophilmediated activity against opportunistic fungal pathogens. J. Infect. Med. 175, 1012–1015 (1997). 119. Nemunaitis, J. Use of macrophage colony-stimulating factor in the treatment of fungal infections. Clin. Infect. Dis. 26, 1279–1281 (1998). 120. Roilides, E., Dignani, M. C., Anaissie, E. J. & Rex, J. H. The role of immunoreconstitution in the management of refractory opportunistic fungal infections. Med. Mycol. 36, 12–25 (1998). 121. Roilides, E., Dimitriadou-Georgiadou, A., Sein, T., Kadiltsoglou, I. & Walsh, T. J. Tumor necrosis factor α enhances antifungal activities of polymorphonuclear and mononuclear phagocytes against Aspergillus fumigatus. Infect. Immun. 66, 5999–6003 (1998). 122. Roilides, E. et al. Antifungal activity of elutriated human monocytes against Aspergillus fumigatus hyphae: enhancement by granulocyte–macrophages colonystimulating factor and interferon-γ. J. Infect. Dis. 170, 894–899 (1994). 123. Roilides, E., Uhlig, K., Venzon, D., Pizzo, P. A. & Walsh, T. J. Enhancement of oxidative response and damage caused by human neutrophils to Aspergillus fumigatus hyphae by granulocyte colony-stimulating factor and γ interferon. Infect. Immun. 61, 1185–1193 (1993). 124. The International chronic granulomatous disease cooperative study group. A controlled trial of interferon γ to prevent infection in chronic granulomatous disease. New Engl. J. Med. 324, 509–516 (1991). 125. Fietta, A., Sacchi, F., Mangiarotti, P., Manara, G. & Gialdroni-Grassi, G. Defective phagocyte Aspergillus killing associated with recurrent pulmonary Aspergillus infections. Infection 12, 10–13 (1984). 126. Ma, H. R., Mu, S. C., Yang, Y. H., Chen, C. M. & Chiang, B. L. Therapeutic effect of interferon-γ for prevention of severe infection in X-linked chronic granulomatous disease. J. Formos. Med. Assoc. 102, 189–192 (2003). 127. Johnson, C. P. et al. A murine model of invasive aspergillosis: variable benefit of interferon-γ administration under in vitro and in vivo conditions. Surg. Infect. (Larchmt) 6, 397–407 (2005). 128. Shao, C. et al. Transient over expression of γ interferon promotes Aspergillus clearance in invasive pulmonary aspergillosis. Clin. Exp. Immunol. 142, 233–241 (2005). 129. Caesar-TonThat, C. E. & Cutler, J. E. A monoclonal antibody to Candida albicans enhances mouse neutrophil candicidal activity. Infect. Immun. 65, 5354–5357 (1997). 130. Calera, J. A., Zhao, X-J. & Calderone, R. A. Defective hyphal development and avirulence caused by a deletion of the SSK1 response regulator gene in Candida albicans. Infect. Immun. 68, 518–525 (2000). 131. Spellberg, B. J. et al. A phagocytic cell line markedly improves survival of neutropenic infected mice. J. Leukoc. Biol. 78, 338–344 (2005). 132. Mei-Yeh, J., Deschenes, R. J., & Fassler, J. S. Saccharomyces cerevisiae histidine phosphotransferase Ypd1p shuttles between the nucleus and cytoplasm for SLN1-dependent phosphorylation of Ssk1p and Skn7p. Euk. Cell 2, 1304–1314 (2003).
Acknowledgements Some of the data included were supported by NIH-NIAID grants to R.C. and by an NIH-Fogarty International award to R.C. and J.-P.
Competing interests statement The authors declare no competing financial interests.
DATABASES The following terms in this article are linked online to: Entrez Gene: http://www.ncbi.nlm.nih.gov/entrez/query. fcgi?db=gene GPX2 | GSH1 | SLN1 | TRX2 | YPD1 Entrez Genome Project: http://www.ncbi.nlm.nih.gov/ entrez/query.fcgi?db=genomeprj Aspergillus fumigatus | Candida albicans | Neurospora crassa | Saccharomyces cerevisiae UniProtKB: http://ca.expasy.org/sprot Als1p | Cap1p | Cat1p | Chk1p | Cta1p | Hog1p | IFN-γ | IL-15 | Nik1p | Pbs2p | Sho1p | Skn7p | Sln1p | Sod1p | Sod2p | Ssk1p | Ssk2p | Yap1p | Ypd1p
FURTHER INFORMATION Richard Calderone’s homepage: http://microbiology.georgetown.edu/faculty/calderone.html Aspergillus unit, Institut Pasteur: http://www.pasteur.fr/recherche/unites/aspergillus/ index-en.html Access to this links box is available online.
www.nature.com/reviews/micro © 2006 Nature Publishing Group
REVIEWS
The interaction of bacterial pathogens with platelets J. Ross Fitzgerald*, Timothy J. Foster‡ and Dermot Cox§
Abstract | In recent years, the frequency of serious cardiovascular infections such as endocarditis has increased, particularly in association with nosocomially acquired antibioticresistant pathogens. Growing evidence suggests a crucial role for the interaction of bacteria with human platelets in the pathogenesis of cardiovascular infections. Here, we review the nature of the interactions between platelets and bacteria, and the role of these interactions in the pathogenesis of endocarditis and other cardiovascular diseases. Infective endocarditis The formation of an infected thrombus on one of the coronary valves.
Disseminated intravascular coagulation Activation of coagulation in a diffuse manner, often in response to infection. Instead of leading to clot formation it results in consumption of coagulation factors, which can lead to bleeding problems.
Immune thrombocytopenia purpura The formation of anti-platelet antibodies. Platelets with bound antibody are cleared from the circulation, resulting in a decreasing number of circulating platelets. *Centre for Infectious Diseases, The Chancellor’s Building, New Royal Infirmary, University of Edinburgh, Edinburgh EH16 4SB, Scotland, UK. ‡ Department of Microbiology, Moyne Institute of Preventive Medicine, Trinity College, Dublin 2, Ireland. § Molecular and Cellular Therapeutics, School of Pharmacy, Royal College of Surgeons in Ireland, 123 St Stephen’s Green, Dublin 2, Ireland. Correspondence to T.J.F. e-mail:
[email protected] doi:10.1038/nrmicro1425
Opportunistic pathogenic bacteria can occasionally gain entry to the human circulatory system, resulting in a transient bacteraemia. Serious complications of bacteraemia include life-threatening infective endocarditis (IE)1, disseminated intravascular coagulation (DIC)2, immune thrombocytopenia purpura (ITP)3 and even increased risk of myocardial infarction (MI)4,5 or stroke6. All of these conditions are characterized by abnormal platelet function (see below), possibly mediated by a direct or indirect interaction with the invading pathogen. Bacteria–platelet interactions are characterized by the binding of bacteria to platelets either directly through a bacterial surface protein7 or indirectly by a plasmabridging molecule that links bacterial and platelet surface receptors8. Platelet activation is a necessary step in thrombus formation, and bacteria can induce activation either by a direct interaction7 or indirectly through bridging molecules8 or secreted bacterial products9. Adhesion and activation can be independent processes mediated by different bacterial components and different platelet receptors. As a result, bacteria can have an activating phenotype, an adhesive phenotype, both or neither, and different phenotypes might be specific for different diseases. For example, bacteria with an adhesive phenotype might be crucial for colonization of a cardiac valve in IE, whereas bacteria with an activating phenotype might be necessary for the enhanced platelet consumption seen in DIC and thrombocytopenia. The interaction of bacteria with platelets can result in platelet activation. When this occurs in a localized manner, the result is the formation of a thrombus (FIG. 1). It is our contention that endocarditis can occur on undamaged valves by small circulating thrombi caused by bacteria in the bloodstream. Also, sterile thrombi that form on damaged heart valves can become infected by circulating bacteria. In both cases, the development of an infected
NATURE REVIEWS | MICROBIOLOGY
thrombus involving bacterial activation of platelets on the surface of a heart valve can result in heart failure or the formation of a septic embolism. If the interaction of bacteria with platelets occurs in a more diffuse manner, it can result in consumption of the activated platelets, leading to thrombocytopenia; if there is associated activation of coagulation, it can result in DIC. Platelet activation in the laboratory is commonly studied using light transmission aggregometry10. This technique is limited in its sensitivity and, to obtain a maximal response, bacterial cells are used at a concentration that is much higher than the concentration found in bacteraemia. As such, aggregation studied in an aggregometer is an artificial situation. However, the process of platelet activation in vivo can occur at very low densities of bacteria and can lead to severe thrombocytopenia owing to amplification of the activation process. An example of this is Helicobacter pylori-mediated thrombocytopenia, in which there is no evidence of bacteraemia3. Platelet activation causes the release of thromboxane A2, ADP and serotonin, all of which are potent platelet activators that lead to localized platelet activation and thrombus formation and, if they are released to a sufficient extent, can cause thrombocytopenia. This is characteristic of some bacterial infections, such as those caused by H. pylori11,12 and haemolytic uraemic syndrome caused by Escherichia coli13. Thrombocytopenia associated with bacterial infections is thought to occur secondarily to DIC, a generalized activation of the coagulation cascade during bacteraemia that results in the rapid consumption of platelets14. However, thrombocytopenia can also occur in the absence of activation of the coagulation cascade15. IE can be difficult to treat with antibiotics even if susceptible organisms are involved, owing to poor penetration of the thrombus by some antibiotics16. Also, bacteria
VOLUME 4 | JUNE 2006 | 445 © 2006 Nature Publishing Group
REVIEWS a Adhesion
b Activation and signal transduction
c Aggregation
Bacterium
Myocardial infarction The occlusion of one of the coronary vessels resulting in localized tissue damage, owing to lack of oxygen, and a heart attack.
Thrombus The formation of a clot in a blood vessel, usually in response to injury. Thrombi can be either fibrin-rich, such as those that form in veins, or platelet-rich, which usually occur in the arterial system.
Thrombocytopenia A drop in platelet count below the normal range, usually involving large decreases that result in bleeding problems.
Embolism A fragment of a thrombus that breaks away and is carried to a distant blood vessel, which is then occluded.
Light transmission aggregometry The standard method for measuring platelet activity, which is based on the principle that the absorption of a suspension of platelets is dependent on the number of particles. Platelet-rich plasma is stirred and an agonist is added. Platelet activation results in aggregate formation with a decrease in the number of particles and an increase in light transmission.
Atherosclerosis An inflammation of the blood vessels usually associated with an influx of macrophages, deposition of cholesterol and formation of plaque. It results in the narrowing of blood vessels in key organs such as the heart and brain. If the plaque ruptures it triggers clot formation, resulting in a myocardial infarction or stroke.
Acute coronary syndrome A syndrome characterized by decreased blood supply to the heart caused by the rupture of an atherosclerotic plaque. ACS covers the spectrum of unstable angina (chest pain at rest) to myocardial infarction.
Resting platelets
Figure 1 | Platelet activation and aggregation. The sequence of events that occurs when bacterial cells bind to platelets and stimulate activation. a | Bacteria bind to a platelet receptor, either directly or through bridging ligand(s) (not shown). b | A signal-transduction cascade is stimulated. This leads to activation of the platelet integrin GPIIb/IIIa and the expression of more molecules on the platelet surface. c | The activated GPIIb/IIIa binds to soluble fibrinogen in plasma, which causes aggregation.
in a thrombus are isolated from the immune system as thrombi are devoid of neutrophils. Platelets are an important part of the host defence system because of their ability to secrete potent antimicrobial peptides (see below)17. The importance of bacteria–platelet interactions in disease pathogenesis is not yet fully understood. In recent years, important investigations have been carried out into the mechanisms that different bacterial species use to adhere to and activate platelets. These studies are providing insights into bacterial pathogenesis and allowing identification of targets for the development of novel therapeutics against vascular infection. In this review, we describe the interactions that occur between bacteria and platelets, the bacterial proteins involved, their platelet receptors and the mediators of these interactions. We also discuss how bacteria use these interactions to facilitate survival in the harsh environment of the vascular system and how these interactions might be targets for therapy or prophylaxis.
Bacterial infections of the cardiovascular system The frequency of bacteraemic infections has greatly increased in recent years, in part owing to increased use of interventionary procedures including angioplasty and the deployment of stents, catheters, pacemakers and prostheses18,19. Furthermore, the increase in antibiotic resistance has contributed to the increase in bacterial infections that are refractory to treatment20. Bacteraemic infection can lead to serious vascular complications, including life-threatening endocarditis1, DIC2 and thrombocytopenia14. In addition, there is evidence suggesting a role for infection in conditions such as atherosclerosis21,22, MI23 and stroke6,24. A wide range of bacteria can cause IE25 (TABLE 1). Despite improvements in healthcare, there has been a progressive change in risk factors for IE26. New at-risk groups include intravenous-drug users, elderly people with valve sclerosis, patients with intravascular prostheses, those exposed to nosocomial disease and haemodialysis patients. Staphylococcus aureus has now surpassed viridans-group streptococci as the most common cause of IE in the developed world27 and, taken together, S. aureus,
446 | JUNE 2006 | VOLUME 4
streptococci and enterococci are responsible for more than 80% of cases1,26. Native-valve IE is classically associated with congenital heart disease and chronic rheumatic heart disease, indicating that prior valvular damage predisposes to infection. Prosthetic-valve endocarditis (PVE) occurring shortly after surgery is often caused by Staphylococcus epidermidis or S. aureus, whereas the later development of PVE is often due to streptococci and Gram-negative bacteria such as Haemophilus spp. and Actinobacillus spp. 26 HIV-1-positive patients sometimes present with IE caused by unusual organisms, including Bartonella spp., Salmonella serovars and Listeria spp.26 Periodontal disease, dental procedures and oral hygiene practices can lead to bacteraemia caused by oral commensals and, occasionally, to serious cardiovascular disease such as IE or MI28. Nosocomial endocarditis is a rapidly growing category, with many patients having no cardiac predisposing factors. The predominant pathogens are staphylococci and enterococci — organisms that are frequently associated with nosocomial sepsis. Importantly, nosocomial endocarditis has a case-fatality rate greater than 50% (REFS 1,26). Infection has also been shown to have a role in coronary artery disease. Patients who have received antibiotics in the previous year have a reduced risk of MI29–31,whereas those who have had a bacteraemia have an increased risk4. However, recent studies have shown no decrease in the incidence of MI with antibiotic use32,33. Although specific pathogens, especially H. pylori and Chlamydia pneumoniae (Chlamydophila pneumoniae), have been associated with increased risk of MI, more recent evidence suggests that the pathogen burden is more relevant to an increased risk of MI22,23,34,35. In addition, animal studies have shown that exposure to Porphyromonas gingivalis is as effective as a highcholesterol diet at inducing atherosclerosis36–38. Therefore, there is evidence to support a role for infection in cardiovascular and cerebrovascular disease, and to support the fact that the burden of infection might be more important than exposure to a specific agent. However, the role of infection in the pathogenesis of atherosclerosis
www.nature.com/reviews/micro © 2006 Nature Publishing Group
REVIEWS Table 1 | Bacterial agents of vascular infections and reported interactions with platelets Bacterial pathogen
Vascular disease association
Platelet adherence/ aggregation
Bacterial factors
Host factors
Staphylococcus aureus
IE
+
ClfA8,88
Fibrinogen, GPIIb/IIIa, FcγRIIa, IgG88,89
8
IgG88,89
ClfB
89
FnBPA, FnBPB
Fibrinogen, fibronectin, GPIIb/IIIa, FcγRIIa, IgG88,89
SpA8,62
vWF
SdrE8
Complement, IgG
98 ,
151
SraP Efb
Not known
α-Toxin
Not known ND
9
Staphylococcus aureus
Delayed wound healing
Inhibition
Efb
Staphylococcus epidermidis
IE
+ (REF. 170)
SdrG (unpublished data) Fibrinogen, GPIIb/IIIa, FcγRIIa, IgG
Staphylococcus capitis
IE
+ (REF. 170)
ND
ND 112
Streptococcus sanguis
IE
+ (REF. 117)
SrpA, PAAP
GPIb, FcγRIIa112
Streptococcus agalactiae
IE
+ (REF. 132)
FbsA132
Fibrinogen, IgG132 133
Streptococcus pyogenes
IE
+ (REF. 102)
M protein
FcγRIIa133
Streptococcus gordonii
IE
+ (REF. 121)
GspB118,119, HsA120
GPIb118,119
Streptococcus pneumoniae
IE
+ (REF. 171)
ND
ND
Streptococcus mitis
IE
+ (REF. 129)
PblA, PblB, Sm-hPAF
ND
Enterococcus spp.
IE
+ (REF. 172)
ND
ND
Neisseria gonnorrhoeae
IE
ND
ND
ND 163
Pseudomonas aeruginosa
IE
+ (REF. 163)
PLC
ND
Porphyromonas gingivalis
IE
+ (REF. 162)
Gingipains173
FcγRIIa173
Hgp135
GPIbα?
Helicobacter pylori
MI
+ (REF. 139)
ND
vWF, GPIb, IgG139
Borrelia burgdorferi
Lyme disease
+ (REF. 143)
p66144
GPIIb/IIIa143
Borrelia hermsii
Thrombocytopenia
+ (REF. 146)
ND
GPIIb/IIIa146
Haemophilus influenzae
MI
ND
ND
ND
Actinobacillus
IE
+ (REF. 174)
ND
ND
Candida spp.
IE
+ (REF. 175)
ND
ND
Chlamydia pneumoniae
Atherosclerosis
+ (REF. 176)
ND
ND
Listeria monocytogenes
IE
+ (REF. 177)
ND
ND
Bacillus anthracis
Haemorrhage
Inhibition
Oedema toxin153 154
Lethal factor
PKA MAPK pathway
ClfA,B, clumping factor A,B; Efb, extracellular fibrinogen-binding protein; FnBPA,B, fibronectin-binding protein A,B; IE, infective endocarditis; IgG, immunoglobulin G; MAPK, mitogen-activated protein kinase; MI, myocardial infarction; ND, not determined; PAAP, platelet-aggregation-associated protein; PKA, protein kinase A; PLC, phospholipase C; SdrE,G, serine-aspartate repeat protein E,G; SrpA, serine-rich protein A; vWF, von Willebrand factor.
Megakaryocytes The bone marrow cells responsible for the production of platelets.
Fibrinogen A soluble protein in plasma that is essential for clotting of blood. Soluble fibrinogen is converted to fibrin by the action of thrombin, and then polymerizes to form a clot.
remains to be elucidated. Platelets have an important role in atherosclerosis39–41. Platelet interactions with the atherosclerotic lesion are part of the inflammatory process and lead to progression of the lesion. Plaque rupture results in the formation of a thrombus, leading to an acute coronary syndrome (ACS); if the thrombus persists and fully occludes the vessel, the result is an MI. Infection might have a role in several different stages of atherosclerosis: triggering vascular inflammation, which eventually leads to atherosclerosis; accelerating inflammation of an existing lesion by recruiting leukocytes and platelets to the lesion; destabilizing an existing lesion; and enhancing thrombus formation on a ruptured lesion.
NATURE REVIEWS | MICROBIOLOGY
Biology and function of platelets Platelets are anucleate derivatives of megakaryocytes and are the smallest and most numerous (normal range: 3 × 108 to 4 × 108 ml–1) corpuscular component of blood42. The typical shape of resting platelets is discoid and, upon activation, they undergo a shape change to a globular form with pseudopodia. Although they are energetically active, platelets lack nuclei and can perform only limited translation from stable megakaryocyte mRNA43. Platelets contain granules for storage of important mediators such as ADP, serotonin and calcium, and proteins such as fibrinogen, vitronectin and von Willebrand factor (vWF)44. Granules also contain platelet microbicidal proteins VOLUME 4 | JUNE 2006 | 447
© 2006 Nature Publishing Group
REVIEWS a
b
c Fibrinogen
IgG1 FcγRIIa LAT
Receptor Active PLCγ2
Lyn
PTP1B
GPIIb/IIIa inactive
Syk
PI3 kinase
TxA2 Ca2+ PKC
SHIP-1
Actin assembly DAG
SHP-1
TXS
IP3 DAG
IP3
PGH2
PLC PKC activation
Ca2+ release
Phosphorylated tyrosine ITAM domain
Downstream activation
TxA2
Agonist Receptor
COX
PLA2 AA
Platelet
Figure 2 | Signalling pathways in platelet activation. a | Upon binding the Fc region of immunoglobulin G (IgG), the FcγRIIa receptor dimerizes. In addition to homodimers, the receptor can can also form heterodimers with GPIb, the von Willebrand factor receptor (vWF) or complement receptors. After dimerization, tyrosine residues in the ITAM (immunoreceptor tyrosine-based activation motif) domain are phosphorylated by an Src kinase, probably Lyn. The tyrosine kinase Syk then binds to the phosphotyrosine residues and phosphorylates tyrosine residues in the adaptor protein LAT. This results in the recruitment and activation of phospholipase Cγ2 (PLCγ2). PLCγ2 produces inositol triphosphate (IP3), which binds to a calcium-ion channel, allowing an influx of calcium from internal stores. Diacylglycerol (DAG) is also produced and activates protein kinase C (PKC), which triggers downstream activation events. Phosphoinositide-3 (PI3) kinase binds to the dimerized receptor and has an important role in actin assembly, which is essential for platelet aggregation. b | The signal is terminated by the action of several phosphatases: PTP1B dephosphorylates LAT, SHP-1 is recruited to the phosphorylated receptor and dephosphorylates Syk, and the inositol phosphatase SHIP-1 inhibits PI3 kinase activity. c | Platelets can be activated by many agents, including ADP, thrombin, collagen and adrenaline. Activation of receptors by these agonists initiates two distinct signalling pathways. The first is the mobilization of phospholipase A2 which cleaves arachidonic acid (AA) from the platelet membrane, which is then converted to prostaglandin H2 (PGH2) by cyclooxygenase (COX). PGH2 is then converted to thromboxane A2 (TxA2) by thromboxane synthase (TXS). TxA2 is secreted by the platelet and can activate other platelets. It is also a potent vasoconstrictor. In the second pathway, phospholipase C (PLC) releases IP3 and DAG from the platelet membrane. IP3 binds to its receptor on the endoplasmic reticulum, triggering the release of Ca2+ from internal stores. DAG activates PKC. The actions of Ca2+ influx and PKC activate GPIIb/IIIa, which can then bind its ligand fibrinogen. As a divalent molecule, fibrinogen can bind to two GPIIb/IIIa molecules simultaneously. If these are on different platelets, the result is the formation of a platelet aggregate. Based on data from REFS 167–169.
von Willebrand factor Also known as Factor VIIIrelated antigen. Important for platelet adhesion under shear at sites of endothelial damage. In blood, von Willebrand factor is multimeric with 10 or more monomers linked together; each monomer is a 220-kDa protein comprising four distinct domains.
Teichoic acid A phosphate-rich, anionic polysaccharide attached to the peptidoglycan of Gram-positive bacteria.
(PMPs) and peptides; PMPs and thrombin-inducible PMPs (tPMPs) are present in rabbit platelet extracts and thrombin-induced supernatants17. Platelets have a crucial role in thrombus formation. First, single platelets bind to a site of damaged endothelium where the underlying extracellular matrix has been exposed. This involves platelet surface receptors for collagen (α2β1 and GPVI), fibrinogen (GPIIb/IIIa) and vWF (GPIb/IX/V) and results in the capture of platelets and their activation45. Upon platelet activation, the fibrinogen receptor GPIIb/IIIa undergoes a conformational change that allows it to bind soluble fibrinogen46. At the same time, cytoplasmic granule contents are released, triggering activation of other platelets which are then recruited into the growing thrombus, plugging the endothelial damage. FIGURE 2 shows the intracellular pathways that occur during platelet activation. The ability to activate platelets has been exploited by some endovascular pathogens to promote colonization and avoidance of the host immune response. However, platelets exert anti-bacterial effects through release of microbicidal tPMPs upon activation47, which act against a broad
448 | JUNE 2006 | VOLUME 4
range of pathogens, including some strains of S. aureus, viridans-group streptococci and Candida albicans17,48,49. tPMPs released from platelets include the classical chemokines (kinocidins) platelet factor 4, CTAP-3, RANTES, fibrinopeptide B and thymosin β-450–52. Strains of S. aureus that are susceptible to tPMPs in vitro are less likely to cause endovascular infections in humans and are more rapidly cleared from the bloodstream than bacteria that are resistant to tPMPs53–55. In experimental endocarditis, S. aureus cells that are resistant to tPMPs in vitro are more virulent56. Resistance to tPMPs in S. aureus is due to a combination of several factors, including enhanced membrane fluidity57, reduced membrane potential58,59, a high proportion of positively charged groups in the cell membrane and the cell wall owing to lysinylation of phosphatidyl glycerol and d-alanylation of teichoic acid60, and a plasmid-encoded efflux pump61. There is evidence that platelets are capable of engulfing S. aureus cells after activation has occurred62,63 possibly avoiding immune defences63. It is probable that S. aureus can survive within platelets, as the bacterium possesses many mechanisms for avoidance of innate
www.nature.com/reviews/micro © 2006 Nature Publishing Group
REVIEWS Complement A part of the innate immune system comprising serum proteins that can protect against infection.
Fc receptor Surface molecules on various cells that bind to the Fc regions of immunoglobulins, thereby initiating antibody effector functions.
Fibronectin A high-molecular-weight glycoprotein present in soluble form in plasma that binds to integrins and the extracellular matrix.
immune responses64,65. Clearly, the interaction of bacteria with human platelets can have different outcomes, and this might depend on the strain of bacteria causing the infection.
Bacteria–platelet interactions The association of coagulation with infection has been well established in nature. The best-characterized example is that of the amoebocytes of Limulus polyphemus (horseshoe crab). These are the only circulating cells in the crab’s blood and, although nucleate, they are very similar in morphology to mammalian platelets. Upon exposure to bacteria they aggregate and release their granule contents, and therefore the Limulus amoebocytes mediate host defence and haemostasis. In mammals, the platelet has retained the amoebocyte functions despite the development of dedicated immune cells66. Early studies on the interaction of bacteria with platelets focused on the effects of endotoxin on platelet activation67. A study in 1959 showed platelets interacting with mycobacteria68, and a series of studies by Clawson and co-workers during the 1970s examined the response of platelets to whole bacteria69–74. However, the development of a method to measure platelet aggregation10 and the availability of commercial platelet aggregometers was to lead to further studies of platelet interactions with bacteria. Platelets have also been shown to respond to exposure to an antigen–antibody complex75 and complement76, and to secrete bactericidal agents77 in response to activation, indicating a role for platelets in the immune response to infection. However, these studies were limited by a paucity of information on platelet function, and they provided little information on the platelet proteins involved in the interaction and how the interaction resulted in platelet aggregation. Below, we describe some of the most important and best-characterized bacteria that are known to interact with platelets.
a Rapid activation
The interactions of S. aureus with platelets Probably the best-characterized interactions between bacterial cells and platelets are those involving the main IE pathogen, S. aureus. Although the initial studies focused on secreted mediators of platelet activation such as α-toxin78, Hawiger and co-workers showed that staphylococci can trigger platelet aggregation involving protein A, IgG and the Fc receptor on platelets79. Subsequently, clumping factor was shown to be an important mediator of S. aureus adhesion to fibrinogen and to be a virulence factor for endocarditis80 and this interaction was suggested to be independent of GPIIb/ IIIa, the platelet fibrinogen receptor81. In fact, S. aureus has several different surface-associated proteins that can bind to platelets and, in some cases, stimulate platelet activation (TABLE 1). The proteins typically have signal sequences at their N termini, which direct Sec-dependent secretion through the cytoplasmic membrane, and a sorting signal at their C termini for anchoring the protein to the cell wall82. Clumping factor A (ClfA) contains an ~500-residue fibrinogen-binding domain that is projected from the cell surface by an unfolded serine and aspartate dipeptide repeat region (FIG. 3). It is the prototype of a family of related proteins with structurally similar N-terminal ligand-binding domains. The fibronectin-binding proteins FnBPA and FnBPB have an N-terminal A domain that has sequence similarity to ClfA and which interacts with the same region of fibrinogen as ClfA. The A domains of the FnBPs are linked to the cell surface by a series of tandemly repeated, unfolded, fibronectin-binding determinants83,84. In the case of FnBPA, two molecules of fibronectin can interact with each fibronectin-binding region83. Early studies showed that S. aureus could adhere to platelets in a manner that was enhanced by the presence of fibrinogen85. A mutant of S. aureus that lacked the ability to bind platelets had a transposon insertion in the clfA gene, implying that a fibrinogen bridge occurred between ClfA on the bacterial surface and a platelet
b Rapid activation
c Slow activation
S. aureus
Fibronectin bridge
FnBPA/ClfA Fibrinogen bridge
FnBP
ClfA-PY/SdrE Complement
FcγRIIa
IgG
Platelet
GPIIb/IIIa
CR
Signal transduction
Signal transduction
Signal transduction
Platelet activation
Platelet activation
Platelet activation
Figure 3 | The mechanism of Staphylococcus aureus-mediated platelet adherence and activation. a | Rapid activation involving a fibrinogen bridge between a bacterial binding protein (clumping factor A (ClfA) or fibronectinbinding protein A (FnBPA)) and low-affinity resting GPIIb/IIIa, and specific immunoglobulin G (IgG) which engages FcγRIIa. b | Rapid activation involving a fibronectin bridge between FnBP and the platelet integrin GPIIb/IIIa and IgG specific for neo-epitopes created when FnBP binds fibronectin, which engages FcγRIIa. c | Slow activation involving antibody recognizing a surface protein that stimulates complement fixation. Activation involves IgG engaging FcγRIIa and complement being recognized by a platelet complement receptor (CR).
NATURE REVIEWS | MICROBIOLOGY
VOLUME 4 | JUNE 2006 | 449 © 2006 Nature Publishing Group
REVIEWS fibrinogen-binding protein, probably the integrin GPIIb/ IIIa86. It was also suggested that ClfA could bind directly to an as-yet-unidentified protein of 118 kDa on the platelet surface87. By studying strains with one or more mutations affecting surface proteins and by expressing surface proteins separately in a heterologous host (for example, Lactococccus lactis), clumping factors ClfA and ClfB, FnBPA and the serine aspartate repeat protein SdrE were all shown individually to be able to bind platelets and to stimulate platelet aggregation8. More recently, our groups have analysed in detail the molecular basis of platelet activation by extending this approach to different host strains and by studying cells in different phases of growth. First, the major determinants of the ability of S. aureus to promote rapid activation of platelets differ according to the phase of growth of the culture. With cells from stationary phase, ClfA is the dominant pro-aggregatory surface protein88, whereas with cells from exponential growth phase, the fibronectin-binding proteins are most important89. This correlates with the regulation of expression of the genes that encode the proteins — the fnbA and fnbB genes are only transcribed in exponential growth90, whereas clfA is transcribed weakly in exponential phase and is stimulated in a SigB-dependent fashion in stationary phase91. By using a regulatable promoter to control expression of surface proteins in the surrogate host L. lactis, it was shown that a minimum number of ClfA and FnBPA molecules are required on the bacterial surface before activation of platelets can occur88. Once the threshold is achieved, aggregation occurrs rapidly. By studying a mutant of ClfA that did not bind fibrinogen, by using inhibitors specific for platelet receptors and by analysing the soluble plasma proteins required in a reconstituted plasma-free system with washed platelets, it was possible to identify the plasma and platelet components involved88. For rapid activation to occur, the bacteria must express sufficient ClfA molecules on their surface and bind fibrinogen molecules in plasma. GPIIb/ IIIa on resting platelets has a low affinity for fibrinogen but can still bind bacterial-bound fibrinogen. However, this is not sufficient to stimulate activation. ClfA-specific immunoglobulin must also be present to act as a bridge between the bacterial surface protein and the platelet immunoglobulin Fc receptor, FcγRIIa88 (FIG. 3a). Similar observations were made for FnBPA, which possesses two different but related mechanisms of engaging and activating resting platelets89. The fibrinogenbinding A domain behaves in a similar fashion to ClfA and activates platelets through a fibrinogen bridge to GPIIb/IIIa and an IgG bridge to FcγRIIa. In addition, the fibronectin-binding region BCD can independently activate platelets. This requires fibronectin N-terminal type I domains to bind to the bacterial protein by the tandem β-zipper mechanism83 and to GPIIb/IIIa through the integrin-binding RGD domain89. As with ClfA and the FnBPA A domain, a specific antibody (in this case recognizing the neo-epitopes created when FnBPs bind fibronectin92) is also required to engage FcγRIIa89 (FIG. 3b). FnBPB seems to be less effective at inducing platelet aggregation93, although our results contradict this89.
450 | JUNE 2006 | VOLUME 4
Soluble fibrin, which is derived from fibrinogen by the actions of thrombin, is more potent than fibrinogen in promoting platelet activation and aggregation by S. aureus85,94. This suggests that S. aureus might have a high affinity for an existing lesion containing a fibrin clot and could have a role in colonizing a damaged valve. If fibrin is generated, thrombin must be present, which, as a potent platelet activator, would lead to thrombus formation. However, it does not explain how S. aureus can colonize undamaged valves to cause native-valve endocarditis. In fact, S. aureus can directly activate platelets independently of thrombin generation. This is supported by our observation that the presence of a thrombin inhibitor only marginally decreased the potency of S. aureus-promoted aggregation89. In conclusion, rapid platelet activation mediated by ClfA and FnBPs requires two mechanisms of bacterial binding to resting platelets: a fibrinogen or fibronectin bridge to the low-activity platelet integrin, and an immunoglobulin bridge to the FcγRIIa receptor. It is probable that clustering of FcγRIIa initiates the signal transduction cascade that causes the subsequent activation and aggregation of platelets (FIG. 2). Activation therefore depends on the presence of specific antibodies that are present in the sera of most, if not all, individuals95. Although ClfA and FnBPA/B seem to be the dominant surface proteins mediating platelet aggregation in stationary and exponential phase, respectively, additional factors including ClfB, SdrE and SpA have also been shown to have a role8. SpA had previously been reported to bind platelets through gC1qR/p33, a platelet complement receptor96. However, gC1qR/p33 is an intracellular protein on resting platelets and is only exposed when platelets are activated. SpA binds with high affinity to the A1 domain of vWF, a large multimeric glycoprotein which mediates platelet adhesion at sites of endothelial damage (REF. 97; M. O’Seaghdha, unpublished observations). Recent evidence suggests that SpA plays an important part in mediating interactions with platelets under conditions of high shear62 (see the section on bacteria–platelet interactions under high shear conditions). Recently, a high-molecular-mass protein called SraP — which is homologous to the GPIb-binding protein GspB of Streptococcus gordonii (see below) — was shown to promote binding of S. aureus cells to platelets and to enhance virulence in a rabbit model of endocarditis98. SraP did not seem to bind to GPIb, and it was not clear if its interaction with platelets stimulated activation. S. aureus can also activate platelets by a slower mechanism that is independent of GPIIb/IIIa. A nonfibrinogen-binding mutant of ClfA (ClfA-PY) caused slow activation of platelets88, as did SdrE, a surface protein which was previously shown to activate platelets when expressed in L. lactis8, but which does not bind fibrinogen. ClfA-PY- and SdrE-expressing bacteria stimulated activation of platelets after a lag phase lasting between 8 and 20 minutes depending on the plasma donor and the level of protein expression on the bacterial cell surface. In both cases, IgG specific for the surface protein was required to engage FcγRIIa and also to
www.nature.com/reviews/micro © 2006 Nature Publishing Group
REVIEWS provide the stimulus for activation of complement by the classical pathway. Assembled complement components on the bacterial cell surface then bound to complement receptors on the platelet surface. Both complement and antibody are required for activation to occur (FIG. 3c). Some strains of Streptococcus sanguis (S. sanguinis) can also activate platelets in a complement-dependent fashion99 (see below). Indeed, this mechanism might be widespread among bacterial IE pathogens that do not express potent activating molecules such as ClfA and FnBPA. In general, any bacterium should be capable of activating platelets when it expresses a surface protein or proteins at a sufficiently high level provided antibodies to these proteins are present. Antibodies are likely to be present because most IE pathogens are commensals, and most individuals have been exposed and have antibodies to the bacterial surface proteins in their sera95.
Interaction of streptococci with platelets The viridans group of streptococci are usually commensals of the human oral cavity and might contribute to tooth decay. Very occasionally they gain access to the vascular system and cause IE. Until recently S. sanguis was the most common cause of IE but has now been superceded by S. aureus27. Studies using animal models of endocarditis have shown that the ability of S. sanguis to adhere to, and activate, platelets correlates with increased severity of disease100. Initial studies identified M protein isolated from group A streptococci as mediating platelet aggregation in an antibody- and complement-dependent manner101. Subsequently Streptococcus pyogenes102 and S. sanguis were shown to trigger platelet aggregation directly103. A collagen-like component of S. sanguis that was implicated in platelet aggregation104 was subsequently identified as platelet-aggregation-associated protein (PAAP)104,105. PAAP is a rhamnose-rich glycoprotein of 115 kDa that contains a collagen-like platelet-interactive domain comprising residues Pro-Gly-Glu-Gln-Gly-ProLys105. Gong and colleagues reported that PAAP interacts with platelet membrane proteins of 175 kDa and 230 kDa
a
to mediate platelet binding and aggregation106. However, it has recently been shown that the role of PAAP in platelet aggregation is donor specific107. S. sanguis was found to require IgG interacting with the platelet FcγRIIa receptor to mediate platelet aggregation108. However, there is no clear association between IgG levels, FcγRIIa polymorphisms and donor variability109. A role for complement and IgG was also proposed99,110, and anti-GPIb antibodies were found to be inhibitory7,111. An important problem with these studies is that there is considerable variation in the ability of different S. sanguis strains to stimulate platelet aggregation. Kerrigan et al. identified three different phenotypes for platelet interactions7. Type I strains showed strong adhesion and rapid aggregation of platelets, type II strains had longer lag times to platelet activation and did not support bacterial adhesion to resting platelets, and type III strains did not adhere to or activate platelets7. Presumably, there is potential for a type IV strain that supports platelet adhesion but does not induce aggregation. In fact, the data suggest that the ability to support platelet adhesion is distinct from the ability to induce platelet aggregation, and both adhesion and aggregation are probably mediated by different bacterial proteins and different platelet receptors. Platelet activation by type I strains is mediated by a direct interaction between the bacterial cell and GPIb, the platelet vWF receptor (FIG. 4). The interaction with GPIb was localized to the N-terminal 1–225 residues of GPIbα (REF. 7). The S. sanguis protein that interacts with GPIb is the glycoprotein serine-rich protein A (SrpA), a homologue of S. gordonii Hsa and GspB proteins112. Deletion of this protein does not abolish aggregation but it does prolong the lag time112, in effect converting a type I strain into a type III strain. For type III strains with a long lag time, activation required specific antibody and complement assembly similar to the slow activation promoted by S. aureus that was discussed above108,110,113. The signalling pathway stimulated by a S. sanguis strain that induced slow (~12 min) aggregation was shown to be that illustrated in FIG 2a
b
S. aureus
N-linked carbohydrate
S. sanguis S. gordonii Protein A
S-S
S-S S -S
vWF multimer
S-S
GPIbα
S S-
GPIb
FcγRIIa
Glycocalicin GPIb
S-S
S-S
FcγRIIa
S-S
N-terminal globular region
SrpA/GspB/Hsa
S-S
S-S S-
S
O-linked carbohydrate
Membraneproximal mucin-like core
GPIbβ SS
Signal transduction Platelet
Platelet
Platelet activation
Plasma membrane
Cytoplasm
Figure 4 | Bacterial interactions involving the platelet glycoprotein GPIb. a | A model for the interaction of Staphylococcus aureus with platelets involving a von Willebrand factor (vWF) bridge between protein A and GPIb. b | Direct binding to GPIb mediated by streptococcal surface proteins.
NATURE REVIEWS | MICROBIOLOGY
VOLUME 4 | JUNE 2006 | 451 © 2006 Nature Publishing Group
REVIEWS (REF. 114). After rapid phosphorylation of FcγRIIa to initi-
ate the pathway, the proteins become dephosphorylated (FIG. 2b) before aggregation and are rephosphorylated once aggregation occurs. Despite inhibition of fast S. sanguis-induced platelet aggregation by antibodies to FcγRIIa, there was no evidence for a direct role for IgG as aggregation occurred in an antibody-free system with only fibrinogen present7. However, it must be noted that commercial fibrinogen is contaminated with IgG88 and there might have been sufficient IgG present to support an antibody-mediated response. Alternatively, as FcγRIIa and GPIb co-localize, it is possible that the anti-FcγRIIa antibody might sterically block GPIb115. Another member of the viridans group of streptococci that causes IE is S. gordonii116. An early report suggested that S. gordonii strains could not stimulate platelet aggregation117. However, subsequent studies showed that S. gordonii can bind to platelet GPIb through GspB118,119 and can stimulate platelet activation120. It is now well established that S. gordonii can adhere to platelets, and this interaction might contribute to the establishment of endovascular vegetations characteristic of IE120,121. GspB is a 286 kDa surface-anchored protein that mediates adherence to platelets by interacting with GPIb. GspB is translocated from the cytoplasm to the cell wall by a dedicated transport system comprising, at least in part, the SecA2 and SecY2 proteins120,121. Several genes encoding glycosyl transferases lie adjacent to the gspB, secA2 and secY2 genes, suggesting that they are involved in glycosylation of GspB118,122,123. A homologue of GspB known as Hsa, made by a different strain of S. gordonii, is nearly identical to GspB at its N terminus, suggesting it might have a related function119 and also the ability to bind to GPIIb/IIIa124. Hsa is a 203-kDa sialic-acid-binding glycoprotein that mediates binding to platelets and platelet aggregation120. Both GspB and Hsa mediate binding to sialic-acid residues on the platelet receptor GPIb, suggesting that GPIb is an important receptor for S. gordonii–platelet interactions119. Hsa binds specifically to the N-linked oligosaccharides on the globular N-terminal domain of GPIbα, whereas GspB binds to oligosaccharides in this region and in the membrane-proximal mucin core125. The primary function of Has/GspB is to support bacterial adhesion to salivary proteins126,127, and the ability to interact with GPIb seems to be a novel function. Mutation of Hsa has been shown to reduce endocarditis in an animal model128. Streptococcus mitis has been shown to bind to platelets129and mediates adherence through the surface proteins PblA and PblB but does not seem to activate platelets directly117. S. pyogenes has also been shown to induce platelet aggregation102 and to contain a collagenlike protein130, and Streptococcus pneumoniae induces aggregation in an antibody-dependent manner131. Finally, Streptococcus agalactiae is an important human pathogen that causes pneumonia, sepsis and meningitis in neonates. It also poses a serious threat to immunocompromised adult patients, causing arthritis, urinary-tract infections and endocarditis. In a similar manner to S. aureus ClfA, a fibrinogen-binding protein produced by S. agalactiae (FbsA) mediates attachment
452 | JUNE 2006 | VOLUME 4
to, and aggregation of, platelets in a fibrinogen- and specific IgG-dependent manner132.
High shear bacteria–platelet interactions Most of the experiments analysing bacteria–platelet interactions have been carried out under static or low shear conditions. However the vascular system presents conditions of high shear, especially in arterioles, and so it is important that the molecular mechanisms of activation are tested under high shear to simulate better in vivo conditions. It is probable that hydrodynamic shear will affect the molecular interactions between platelets and S. aureus in the vascular system. Pawar and colleagues demonstrated a key role for SpA in mediating platelet activation at high shear rates62. It should be noted that SpA binds to vWF with a Kd in the low-nM range97. We have recently shown that each of the five SpA domains (A–E) can individually bind to the A1 domain of vWF with similar high affinity (M. O’Seaghdha, unpublished observations). A monoclonal antibody directed against vWF partially inhibited platelet activation under these conditions, and an antibody specific for the platelet receptor for vWF (GPIb) also partially inhibited SpAmediated activation, suggesting that an SpA–vWF crossbridge links SpA on the bacterial surface with GPIb (FIG. 4b). It is possible that plasma IgG molecules bound to SpA compete with vWF by occluding the vWF binding site on SpA62. However, the affinity of the vWF–SpA interaction is sufficiently high to withstand this. The fibrinogen-binding protein ClfA can mediate platelet adhesion through the GPIIb/IIIa receptor under both low and high shear rates. Hydrodynamic shear affected the receptor specificity of activation-dependent platelet binding to S. aureus cells, as evidenced by the transition from an SpA-independent/ClfA-dependent process at low shear to an SpA-dependent process at high shear62. When passed over immobilized S. sanguis under high shear conditions, platelets first rolled and then adhered, subsequently forming a large thrombus. This is characteristic of the interaction of platelets with vWF bound to exposed collagen at damaged endothelial surfaces, a mechanism that occurs during blood-clot formation. Although the rolling mediated by vWF only occurred under high shear conditions, platelet rolling over immobilized S. sanguis could also occur under low shear conditions. This interaction was shown to be mediated by GPIb on the platelet and SrpA on the bacteria112. Thrombus formation by S. pyogenes has also been studied under conditions of high shear133. S. pyogenes M protein promoted thrombus formation with fibrinogen- and M-protein-specific IgG being essential, indicating that S. pyogenes and S. aureus share common mechanisms of rapid platelet activation133. Interaction of P. gingivalis with platelets The oral pathogen P. gingivalis has been shown to accelerate atherosclerosis36–38 and to secrete a plateletactivating protease134. P. gingivalis has also been shown to activate platelets directly135. This depends on expression of a surface adhesin, Hgp44, which is processed from haemagglutinin A by surface-associated cysteine
www.nature.com/reviews/micro © 2006 Nature Publishing Group
REVIEWS proteases. Activation requires antibodies in serum to bind to antigens on the bacterial cell surface to interact with FcγRIIa on platelets. A role for GPIbα was indicated by monoclonal antibodies recognizing various domains of GPIbα causing partial or complete inhibition of aggregation, and by proteolytic cleavage of the N-terminal domain inhibiting activation. However, it is not clear whether Hgp44 binds directly to GPIbα or whether the involvement of the platelet glycoprotein is due to its close proximity to FcγRIIa.
Interaction of H. pylori with platelets H. pylori causes gastritis, peptic ulcers, gastric carcinoma, mucosa-associated lymphoid tissue (MALT) lymphoma and pernicious anaemia136. Recent evidence also suggests that H. pylori might have pathogenic effects outside of the gastrointestinal tract. Infection with H. pylori seems to increase the risks of atherosclerosis137,138 and ITP3,11,12. ITP is caused by destruction of platelets, probably by auto-antibodies. H. pylori infections have been identified in patients with ITP, and platelet counts increased after elimination of the infection3. Furthermore, some H. pylori strains could induce platelet activation and aggregation, which seems to be mediated through initial interaction of vWF with bacteria in the blood followed by binding to GPIb, the platelet vWF receptor. Function-blocking antibodies specific for either vWF or GPIb disrupted H. pylori-mediated platelet aggregation139. Specific anti-H. pylori antibodies were also required, as was an interaction with FcγRIIa. The requirement for vWF in H. pylori-induced aggregation suggests a different mechanism compared with S. sanguis, which also requires GPIb but not vWF7, although binding of bacterial-specific IgG to the platelet FcγRIIa receptor is common to both bacteria. Campylobacter fetus infection has been associated with thrombosis in several case reports140. There is no information on a potential interaction with platelets, although Campylobacter rectus has been shown not to induce platelet activation141. Borrelia–platelet interactions Borrelia burgdorferi is a tick-borne spirochaete that is the causative agent of Lyme disease, a chronic, multisystemic infection affecting humans in specific geographical regions in the US and Europe. After several days of localized skin infections at the site of the tick bite, the spirochaete disseminates to multiple tissues, including the synovium, the central nervous system and the heart142. B. burgdorferi binds to activated platelets through the GPIIb/IIIa platelet receptor143. This interaction might occur through the B. burgdorferi surface protein p66 (REF. 144). However, it is not clear whether binding occurs directly or through a plasma protein. A prominent feature of Lyme disease is vascular injury. This might result in localized platelet activation and aggregation, which could then lead to B. burgdorferi adherence to platelets. This interaction could contribute to colonization and/or dissemination to other regions of the vascular system. Borrelia hermsii is a tick-borne spirochaete that is the causative agent of relapsing fever. B. hermsii grows to high density in blood, and platelet–spirochaete complexes NATURE REVIEWS | MICROBIOLOGY
have been observed in blood smears of relapsing-fever patients145. Similarly to B. burgdorferi, B. hermsii was shown to bind to platelets through the GPIIb/IIIa receptor of both activated and resting platelets146. By contrast, B. hermsii induced platelet activation in the absence of added fibrinogen146. B. hermsii infection of mice resulted in severe thrombocytopenia and spirochaete complexes were detected in the blood of infected mice; however, there was no evidence of platelet activation, and binding of the spirochaete to the platelets was independent of GPIIb/IIIa147. This suggests that the platelet–bacteria complexes are cleared from the circulation and that the activation process in mice platelets might differ from that of human platelets.
A general mechanism for platelet activation Increasing evidence indicates that many bacterial pathogens use a similar basic mechanism to mediate platelet activation. Initial adhesion of bacteria to platelets occurs either directly or, more commonly, through a plasma– protein bridge with a platelet receptor, usually GPIIb/IIIa or GPIb. This interaction in itself is insufficient for platelet activation, but requires a circulating antibody specific for the bacterial surface protein to engage the platelet FcγRIIa receptor. This leads to platelet activation and results in platelet aggregation and thrombus formation. Many of the pathogens that cause IE are normal human commensals to which the host has been exposed so that low levels of specific antibodies are present in blood95. Bacteria seem to use these antibodies to subvert the immune response to promote pathogenesis. High levels of antibodies will actually prevent activation by blocking the ligand-binding activity of the surface protein. In this case, a slower complement-dependent mechanism of platelet activation can be used by the pathogen. With H. pylori139, only the sera of individuals who have been infected have antibodies that would be capable of promoting activation. Involvement of bacterial extracellular factors The α-toxin of S. aureus can potently induce aggregation of platelets and contribute to coagulation of blood9, even at concentrations below those which cause cytolysis. α-Toxin initiates prothrombinase-complex formation on platelets148. The process resembles that of ionophore agonists by creating ion channels that stimulate Ca2+ flux (FIG. 2c). Paradoxically, elevated expression of α-toxin by S. aureus leads to reduced virulence in a rabbit model or endocarditis149. This was probably due to excessive damage to platelets causing release of PMPs rather than reduced clearance during the bacteraemic phase or reduced adherence of bacteria to thrombi. S. aureus can produce at least one protein that blocks activation of platelets. The extracellular fibrinogen- binding protein (Efb) is produced during post-exponential growth. It binds fibrinogen and can also bind directly to a receptor on activated platelets150,151. It enhances fibrinogen binding to platelets independently of GPIIb/IIIa (FIG. 5) but blocks platelet aggregation, presumably by altering the conformation of GPIIb/IIIa-bound fibrinogen so it cannot crosslink with another platelet (FIG. 5). An S. aureus efb VOLUME 4 | JUNE 2006 | 453
© 2006 Nature Publishing Group
REVIEWS a
b D Efb
D
Fibrinogen E
E
D
D
c Prothrombinase
Efb GPIIb/IIIa
α-toxin
Blocks aggregation
X
Promotes fibrinogen binding
Ca2+
Figure 5 | Platelet interactions of the staphylococcal secreted proteins Efb and α-toxin. The secreted protein Efb can bind to GPIIb/IIIa and fibrinogen, resulting in both inhibition of platelet aggregation (a) and enhanced binding to fibrinogen (b). α-Toxin stimulates platelet activation through prothrombinase-complex formation leading to the formation of ion channels, which stimulate Ca2+ flux (c).
mutant showed decreased virulence in a model of staphylococcal wound infection, and the presence of Efb contributed to a delay in wound healing150,151. Enterotoxin B secreted from S. aureus has also been shown to inhibit thrombin-induced platelet aggregation152. This suggests that activation of platelets by S. aureus might be disadvantageous in certain disease situations and that S. aureus has evolved mechanisms involving Efb and enterotoxin B to avoid this outcome. The S. aureus efb mutant was not attenuated in an endocarditis model151. Bacillus anthracis expresses two toxins that potently inhibit aggregation of platelets by natural agonists. The oedema toxin is a calmodulin-dependent adenylate cyclase that activates cyclic AMP (cAMP)-dependent protein kinase A inside cells and platelets, which is likely to initiate events leading to failure to aggregate153. The lethal toxin has a completely different mode of action that involves suppressing the p42/44 and p38 mitogen-activated protein kinase pathways154. The action of these two toxins suggests that the platelet is an important target in the pathogenesis of anthrax and helps to explain the haemorrhages that are a feature of the infection. Haemorrhages also occur during whooping cough caused by Bordetella pertussis adenylate cyclase toxin, which is attributed to inhibition of platelet function probably by a similar mechanism to the B. anthracis oedema toxin155.
1.
2.
3.
4.
Petti, C. A. & Fowler, V. G. Jr. Staphylococcus aureus bacteremia and endocarditis. Cardiol. Clin. 21, 219–233 (2003). Levi, M., Keller, T. T., van Gorp, E. & ten Cate, H. Infection and inflammation and the coagulation system. Cardiovasc. Res. 60, 26–39 (2003). Franchini, M. & Veneri, D. Helicobacter pylori infection and immune thrombocytopenic purpura: an update. Helicobacter 9, 342–346 (2004). Ishani, A., Collins, A. J., Herzog, C. A. & Foley, R. N. Septicemia, access and cardiovascular disease in
5.
6.
Periodontitis is a chronic infectious disease in which proteolytic activity contributes to the destruction of tooth-supporting tissue28,156. Several studies have linked periodontal disease to heart disease28,156–158. P. gingivalis is one of the principal causes of adult periodontitis159. Gingipains, a family of cysteine proteases produced by P. gingivalis, contribute to the pathogenesis of periodontitis160,161. Gingipains can directly activate platelets, resulting in platelet aggregation, which requires the proteolytic activity of the gingipains and involves stimulation of protease-activated receptors (PAR-1 and PAR-4) on the platelet surface162. This mechanism is distinct from the activation promoted by surface adhesins on P. gingivalis cells135 described above. Pseudomonas aeruginosa occasionally causes IE, particularly in intravenous-drug users26. P. aeruginosa produces a phospholipase C enzyme that stimulates platelet aggregation in a manner that is dependent on the enzymatic activity of phospholipase C163. Water-soluble extracts of H. pylori have been shown to induce platelet aggregation164. S. mitis has been shown to secrete a platelet-activating factor165 and S. mutans-induced platelet aggregation is mediated by a soluble polysaccharide in an antibody-dependent manner166.
Outlook for novel therapeutics The discovery of a common mechanism of platelet activation by several important pathogens provides increased hope for the development of novel therapeutics to treat serious vascular infections such as IE. In particular, the identification of the crucial role of the FcγRIIa receptor in bacteria-induced platelet activation represents an interesting target for further investigation. Additional analysis will determine if this is a truly universal mechanism. The identification of several host factors that are important for bacteria–platelet interactions indicates that some individuals might be more susceptible to or, conversely, more resistant to, serious complications of bacteraemia. In addition, several bacterial factors that promote interaction with platelets and contribute to pathogenesis have now been identified. Development of specific monoclonal-antibody therapies against these proteins could prove to be useful in treating or preventing serious cardiovascular infections. Ideally, small-molecule inhibitors of the interaction between platelets and bacteria could be developed that would have the advantage of bypassing antibiotic-resistance mechanisms. Accordingly, they would be ideal for prophylaxis of infection in susceptible individuals and would allow restriction of antibiotic therapy to cases of overt infection.
dialysis patients: The USRDS wave 2 study. Kidney Int. 68, 311 (2005). Foley, R. N., Guo, H., Snyder, J. J., Gilbertson, D. T. & Collins, A. J. Septicemia in the United States dialysis population, 1991 to 1999. J. Am. Soc. Nephrol. 15, 1038–1045 (2004). Corrado, E. et al. Markers of inflammation and infection influence the outcome of patients with baseline asymptomatic carotid lesions: a 5-year follow-up study. Stroke 37, 482–486 (2006).
454 | JUNE 2006 | VOLUME 4
7.
8.
Kerrigan, S. W. et al. A role for glycoprotein Ib in Streptococcus sanguis-induced platelet aggregation. Blood 100, 509–516 (2002). Demonstrates the role of GPIb in S. sanguismediated platelet aggregation. O’Brien, L. et al. Multiple mechanisms for the activation of human platelet aggregation by Staphylococcus aureus: roles for the clumping factors ClfA and ClfB, the serine-aspartate repeat protein SdrE and protein A. Mol. Microbiol. 44, 1033–1044 (2002).
www.nature.com/reviews/micro © 2006 Nature Publishing Group
REVIEWS
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24. 25.
26. 27.
28.
29.
Identifies ClfA as a key mediator of S. aureusinduced aggregation. Bhakdi, S. et al. Staphylococcal α-toxin promotes blood coagulation via attack on human platelets. J. Exp. Med. 168, 527–542 (1988). Born, G. The aggregation of blood platelets by adenosine diphosphate and its reversal. Nature 194, 927–929 (1962). Emilia, G. et al. Helicobacter pylori eradication can induce platelet recovery in idiopathic thrombocytopenic purpura. Blood 97, 812–814 (2001). Shows an association between H. pylori infection and thrombocytopenia. Veneri, D., Krampera, M. & Franchini, M. High prevalence of sustained remission of idiopathic thrombocytopenic purpura after Helicobacter pylori eradication: a long-term follow-up study. Platelets 16, 117–119 (2005). Guessous, F. et al. Shiga toxin 2 and lipopolysaccharide cause monocytic THP-1 cells to release factors which activate platelet function. Thromb. Haemost. 94, 1019–1027 (2005). Hambleton, J., Leung, L. L. & Levi, M. Coagulation: consultative hemostasis. Hematology 2002, 335–352 (2002). Proulx, F., Seidman, E. G. & Karpman, D. Pathogenesis of Shiga toxin-associated hemolytic uremic syndrome. Pediatr. Res. 50, 163–171 (2001). Schierholz, J., Beuth, J. & Pulverer, G. ‘‘Difficult to treat infections’’ pharmacokinetic and pharmacodynamic factors—a review. Acta Microbiol. Immunol. Hung. 47, 1–8 (2000). Yeaman, M. & Bayer, A. Antimicrobial host defense. In Platelets (ed. Michelson, A. D.) 469–490 (Academic Press, London, 2002). Wisplinghoff, H. et al. Nosocomial bloodstream infections in US hospitals: analysis of 24,179 cases from a prospective nationwide surveillance study. Clin. Infect. Dis. 39, 309–317 (2004). Petti, C. A. & Fowler, V. G. Jr. Staphylococcus aureus bacteremia and endocarditis. Infect. Dis. Clin. North Am. 16, 413–435, x–xi (2002). Lowy, F. D. Antimicrobial resistance: the example of Staphylococcus aureus. J. Clin. Invest. 111, 1265–1273 (2003). Muhlestein, J. B. & Anderson, J. L. Chronic infection and coronary artery disease. Cardiol. Clin. 21, 333–362 (2003). Ott, S. J. et al. Detection of diverse bacterial signatures in atherosclerotic lesions of patients with coronary heart disease. Circulation 113, 929–937 (2006). Shows the presence of many pathogens in atherosclerotic plaques. Zhu, J. et al. Prospective study of pathogen burden and risk of myocardial infarction or death. Circulation 103, 45–51 (2001). Suggests that the number of different infections that a patient has had is a predictor of cardiovascular outcome. Elkind, M. & Cole, J. Do common infections cause stroke? Semin. Neurol. 26, 88–99 (2006). Baddour, L. M. et al. Infective endocarditis: diagnosis, antimicrobial therapy, and management of complications: a statement for healthcare professionals from the Committee on Rheumatic Fever, Endocarditis, and Kawasaki Disease, Council on Cardiovascular Disease in the Young, and the Councils on Clinical Cardiology, Stroke, and Cardiovascular Surgery and Anesthesia, American Heart Association: endorsed by the Infectious Diseases Society of America. Circulation 111, e394–e434 (2005). Important guideline document on infective endocarditis from the American Heart Association. Moreillon, P. & Que, Y. A. Infective endocarditis. Lancet 363, 139–149 (2004). Fowler, V. G. Jr et al. Staphylococcus aureus endocarditis: a consequence of medical progress. JAMA 293, 3012–3021 (2005). Shows that S. aureus is now the most common cause of IE in the developed world. Seymour, R. A., Preshaw, P. M., Thomason, J. M., Ellis, J. S. & Steele, J. G. Cardiovascular diseases and periodontology. J. Clin. Periodontol. 30, 279–292 (2003). Gupta, S. et al. Elevated Chlamydia pneumoniae antibodies, cardiovascular events, and azithromycin in male survivors of myocardial infarction. Circulation 96, 404–407 (1997).
30. Gurfinkel, E., Bozovich, G., Daroca, A., Beck, E. & Mautner, B. Randomised trial of roxithromycin in nonQ-wave coronary syndromes: ROXIS pilot study. Lancet 350, 404 (1997). 31. Meier, C. R., Derby, L. E., Jick, S. S., Vasilakis, C. & Jick, H. Antibiotics and risk of subsequent first-time acute myocardial infarction. JAMA 281, 427–431 (1999). 32. Jespersen, C. M. et al. Randomised placebo controlled multicentre trial to assess short term clarithromycin for patients with stable coronary heart disease: CLARICOR trial. BMJ 332, 22–27 (2006). 33. Andraws, R., Berger, J. & Brown, D. Effects of antibiotic therapy on outcomes of patients with coronary artery disease: a meta-analysis of randomized controlled trials. JAMA 293, 2641–2647 (2005). 34. Espinola-Klein, C. et al. Impact of infectious burden on extent and long-term prognosis of atherosclerosis. Circulation 105, 15–21 (2002). 35. Rupprecht, H. J. et al. Impact of viral and bacterial infectious burden on long-term prognosis in patients with coronary artery disease. Circulation 104, 25–31 (2001). 36. Brodala, N. et al. Porphyromonas gingivalis bacteremia induces coronary and aortic atherosclerosis in normocholesterolemic and hypercholesterolemic pigs. Arterioscler. Thromb. Vasc. Biol. 25, 1446–1451 (2005). 37. Lalla, E. et al. Oral infection with a periodontal pathogen accelerates early atherosclerosis in apolipoprotein E-null mice. Arterioscler. Thromb. Vasc. Biol. 23, 1405–1411 (2003). Shows that normal exposure to an oral pathogen enhances atherosclerosis without bacteraemia. 38. Li, L., Messas, E., Batista, E. L. Jr, Levine, R. A. & Amar, S. Porphyromonas gingivalis infection accelerates the progression of atherosclerosis in a heterozygous apolipoprotein E-deficient murine model. Circulation 105, 861–867 (2002). Shows that exposure to bacteria enhances atherosclerosis in a mouse model. 39. Gawaz, M. Platelets in the onset of atherosclerosis. Blood Cells Mol. Dis. 36, 206–210 (2006). 40. Shoji, T. et al. Platelet-monocyte aggregates are independently associated with occurrence of carotid plaques in type 2 diabetic patients. J. Atheroscler. Thromb. 12, 344–352 (2005). 41. Shoji, T. et al. Platelet activation is associated with hypoadiponectinemia and carotid atherosclerosis. Atherosclerosis 28 Nov 2005 (doi:10.1016/j.atherosc lerosis.2005.10.034). 42. Hartwig, J. H. Platelet structure. In Platelets (ed. Michelson, A. D.) 37–52 (Academic Press, London. 2002). 43. Macaulay, I. C. et al. Platelet genomics and proteomics in human health and disease. J. Clin. Invest. 115, 3370–3377 (2005). 44. Reed, G. Platelet secretion. In Platelets (ed. Michelson, A. D.) 181–196 (Academic Press, London, 2002). 45. Furie, B. & Furie, B. C. Thrombus formation in vivo. J. Clin. Invest. 115, 3355–3362 (2005). 46. Shattil, S. J. & Newman, P. J. Integrins: dynamic scaffolds for adhesion and signaling in platelets. Blood 104, 1606–1615 (2004). 47. Yeaman, M. R. & Bayer, A. S. Antimicrobial peptides from platelets. Drug. Resist. Updat. 2, 116–126 (1999). 48. Yeaman, M. R., Tang, Y. Q., Shen, A. J., Bayer, A. S. & Selsted, M. E. Purification and in vitro activities of rabbit platelet microbicidal proteins. Infect. Immun. 65, 1023–1031 (1997). 49. Yeaman, M. R. The role of platelets in antimicrobial host defense. Clin. Infect. Dis. 25, 951–968 (1997). 50. Cole, A. M. et al. Cutting edge: IFN-inducible ELR-CXC chemokines display defensin-like antimicrobial activity. J. Immunol. 167, 623–627 (2001). 51. Krijgsveld, J. et al. Thrombocidins, microbicidal proteins from human blood platelets, are C-terminal deletion products of CXC chemokines. J. Biol. Chem. 275, 20374–20381 (2000). 52. Tang, Y. Q., Yeaman, M. R. & Selsted, M. E. Antimicrobial peptides from human platelets. Infect. Immun. 70, 6524–6533 (2002). 53. Bayer, A. S. et al. In vitro resistance to thrombininduced platelet microbicidal protein among clinical bacteremic isolates of Staphylococcus aureus correlates with an endovascular infectious source. Antimicrob. Agents Chemother. 42, 3169–3172 (1998).
NATURE REVIEWS | MICROBIOLOGY
54. Fowler, V. G. et al. In vitro resistance to thrombininduced platelet microbicidal protein in isolates of Staphylococcus aureus from endocarditis patients correlates with an intravascular device source. J. Infect. Dis. 182, 1251–1254 (2000). 55. Fowler, V. G. et al. Persistent bacteremia due to methicillin-resistant Staphylococcus aureus infection is associated with agr dysfunction and low-level in vitro resistance to thrombin-induced platelet microbicidal protein. J. Infect. Dis. 190, 1140–1149 (2004). 56. Dhawan, V. K. et al. Phenotypic resistance to thrombin-induced platelet microbicidal protein in vitro is correlated with enhanced virulence in experimental endocarditis due to Staphylococcus aureus. Infect. Immun. 65, 3293–3299 (1997). Suggests that strains which are more resistant to PMPs might also be more virulent. 57. Bayer, A. S. et al. In vitro resistance of Staphylococcus aureus to thrombin-induced platelet microbicidal protein is associated with alterations in cytoplasmic membrane fluidity. Infect. Immun. 68, 3548–3553 (2000). 58. Bayer, A. S. et al. Transposon disruption of the complex I NADH oxidoreductase gene (snoD) in Staphylococcus aureus is associated with reduced susceptibility to the microbicidal activity of thrombininduced platelet microbicidal protein 1. J. Bacteriol. 188, 211–222 (2006). 59. Koo, S. P., Bayer, A. S., Kagan, B. L. & Yeaman, M. R. Membrane permeabilization by thrombin-induced platelet microbicidal protein 1 is modulated by transmembrane voltage polarity and magnitude. Infect. Immun. 67, 2475–2481 (1999). 60. Weidenmaier, C. et al. DltABCD- and MprF-mediated cell envelope modifications of Staphylococcus aureus confer resistance to platelet microbicidal proteins and contribute to virulence in a rabbit endocarditis model. Infect. Immun. 73, 8033–8038 (2005). 61. Kupferwasser, L. I. et al. Plasmid-mediated resistance to thrombin-induced platelet microbicidal protein in staphylococci: role of the qacA locus. Antimicrob. Agents Chemother. 43, 2395–2399 (1999). 62. Pawar, P., Shin, P. K., Mousa, S. A., Ross, J. M. & Konstantopoulos, K. Fluid shear regulates the kinetics and receptor specificity of Staphylococcus aureus binding to activated platelets. J. Immunol. 173, 1258–1265 (2004). 63. Youssefian, T., Drouin, A., Masse, J. M., Guichard, J. & Cramer, E. M. Host defense role of platelets: engulfment of HIV and Staphylococcus aureus occurs in a specific subcellular compartment and is enhanced by platelet activation. Blood 99, 4021–4029 (2002). 64. Rooijakkers, S. H., van Kessel, K. P. & van Strijp, J. A. Staphylococcal innate immune evasion. Trends Microbiol. 13, 596–601 (2005). 65. Foster, T. J. Immune evasion by staphylococci. Nature Rev. Microbiol. 3, 948–958 (2005). 66. Levin, J. The evolution of mammalian platelets. In Platelets (ed. Michelson, A. D.) 3–20 (Academic Press, London, 2002). 67. Des Prez, R. , Horowitz, H. I. & Hook, E. W. Effects of bacterial endotoxin on rabbit platelets: I. Platelet aggregation and release of platelet factors in vitro. J. Exp. Med. 114, 857–874 (1961). 68. Copley, A., Maupin, B. & Balea, T. The agglutinant and adhesive behaviour of isolated human and rabbit platelets in contact with various strains of mycobacteria. Acta Tuberc. Scand. 37, 151–161 (1959). 69. Clawson, C. Platelet interaction with bacteria. III. Ultrastructure. Am. J. Pathol. 70, 449–471 (1973). 70. Clawson, C., Rao, G. & White, J. G. Platelet interaction with bacteria. IV. Stimulation of the release reaction. Am. J. Pathol. 81, 411–420 (1975). 71. Clawson, C. & White, J. G. Platelet interaction with bacteria. V. Ultrastructure of congenital afibrinogenemic platelets. Am. J. Pathol. 98, 197–211 (1980). 72. Clawson, C., White, J. G. & Herzberg, M. Platelet interaction with bacteria. VI. Contrasting the role of fibrinogen and fibronectin. Am. J. Hematol. 9, 45–53 (1980). 73. Clawson, C. & White, J. G. Platelet interaction with bacteria. I. Reaction phases and effects of inhibitors. Am. J. Pathol. 65, 367–380 (1971). 74. Clawson, C. & White, J. G. Platelet interaction with bacteria. II. Fate of the bacteria. Am. J. Pathol. 65, 381–397 (1971). 75. Humphrey, J. H. & Jaques, R. The release of histamine and 5-hydroxytryptamine (serotonin) from platelets by antigen–antibody reactions (in vitro). J. Physiol. 128, 9 (1955).
VOLUME 4 | JUNE 2006 | 455 © 2006 Nature Publishing Group
REVIEWS 76. Hawiger, J., Marney, S. R. Jr, Colley, D. G. & Des Prez, R. M. Complement-dependent platelet injury by staphylococcal protein A. J. Exp. Med. 136, 68–80 (1972). 77. Weksler, B. B. & Nachman, R. Rabbit platelet bactericidal protein. J. Exp. Med. 134, 1114–1130 (1971). 78. Manohar, M., Maheswaran, S. K., Frommes, S. P. & Lindorfer, R. K. Platelet damaging factor, a fifth activity of staphylococcal α-toxin. J. Bacteriol. 94, 224–231 (1967). 79. Hawiger, J. et al. Staphylococci-induced human platelet injury mediated by protein A and immunoglobulin G Fc fragment receptor. J. Clin. Invest. 64, 931–937 (1979). 80. Moreillon, P. et al. Role of Staphylococcus aureus coagulase and clumping factor in pathogenesis of experimental endocarditis. Infect. Immun. 63, 4738–4743 (1995). 81. Bayer, A. S. et al. Staphylococcus aureus induces platelet aggregation via a fibrinogen-dependent mechanism which is independent of principal platelet glycoprotein IIb/IIIa fibrinogen-binding domains. Infect. Immun. 63, 3634–3641 (1995). 82. Foster, T. J. & Hook, M. Surface protein adhesins of Staphylococcus aureus. Trends Microbiol. 6, 484–488 (1998). 83. Schwarz-Linek, U. et al. Pathogenic bacteria attach to human fibronectin through a tandem β-zipper. Nature 423, 177–181 (2003). 84. Massey, R. C. et al. Fibronectin-binding protein A of Staphylococcus aureus has multiple, substituting, binding regions that mediate adherence to fibronectin and invasion of endothelial cells. Cell. Microbiol. 3, 839–851 (2001). 85. Herrmann, M., Lai, Q. J., Albrecht, R. M., Mosher, D. F. & Proctor, R. A. Adhesion of Staphylococcus aureus to surface-bound platelets: role of fibrinogen/fibrin and platelet integrins. J. Infect. Dis. 167, 312–322 (1993). 86. Sullam, P. M., Bayer, A. S., Foss, W. M. & Cheung, A. L. Diminished platelet binding in vitro by Staphylococcus aureus is associated with reduced virulence in a rabbit model of infective endocarditis. Infect. Immun. 64, 4915–4921 (1996). Suggests a key role for bacteria–platelet interactions in IE pathogenesis. 87. Siboo, I. R., Cheung, A. L., Bayer, A. S. & Sullam, P. M. Clumping factor A mediates binding of Staphylococcus aureus to human platelets. Infect. Immun. 69, 3120–3127 (2001). 88. Loughman, A. et al. Roles for fibrinogen, immunoglobulin and complement in platelet activation promoted by Staphylococcus aureus clumping factor A. Mol. Microbiol. 57, 804–818 (2005). Shows the role for antibody and complement in ClfA-mediated aggregation. 89. Fitzgerald, J. R. et al. Fibronectin-binding proteins of Staphylococcus aureus mediate activation of human platelets via fibrinogen and fibronectin bridges to integrin GPIIb/IIIa and IgG binding to the FcγRIIa receptor. Mol. Microbiol. 59, 212–230 (2006). Describes the molecular mechanisms of FnBPmediated platelet aggregation. 90. Saravia-Otten, P., Muller, H. P. & Arvidson, S. Transcription of Staphylococcus aureus fibronectin binding protein genes is negatively regulated by agr and an agr-independent mechanism. J. Bacteriol. 179, 5259–5263 (1997). 91. Que, Y. A. et al. Fibrinogen and fibronectin binding cooperate for valve infection and invasion in Staphylococcus aureus experimental endocarditis. J. Exp. Med. 201, 1627–1635 (2005). Animal data on the roles of ClfA and FnBp in IE. 92. Casolini, F. et al. Antibody response to fibronectinbinding adhesin FnbpA in patients with Staphylococcus aureus infections. Infect. Immun. 66, 5433–5442 (1998). 93. Heilmann, C. et al. Staphylococcus aureus fibronectinbinding protein (FnBP)-mediated adherence to platelets, and aggregation of platelets induced by FnBPA but not by FnBPB. J. Infect. Dis. 190, 321–329 (2004). 94. Niemann, S. et al. Soluble fibrin is the main mediator of Staphylococcus aureus adhesion to platelets. Circulation 110, 193–200 (2004). 95. Dryla, A. et al. Comparison of antibody repertoires against Staphylococcus aureus in healthy individuals and in acutely infected patients. Clin. Diagn. Lab. Immunol. 12, 387–398 (2005).
96. Nguyen, T., Ghebrehiwet, B. & Peerschke, E. I. Staphylococcus aureus protein A recognizes platelet gC1qR/p33: a novel mechanism for staphylococcal interactions with platelets. Infect. Immun. 68, 2061–2068 (2000). 97. Hartleib, J. et al. Protein A is the von Willebrand factor binding protein on Staphylococcus aureus. Blood 96, 2149–2156 (2000). 98. Siboo, I. R., Chambers, H. F. & Sullam, P. M. Role of SraP, a serine-rich surface protein of Staphylococcus aureus, in binding to human platelets. Infect. Immun. 73, 2273–2280 (2005). 99. Ford, I., Douglas, C. W., Heath, J., Rees, C. & Preston, F. E. Evidence for the involvement of complement proteins in platelet aggregation by Streptococcus sanguis NCTC 7863. Br. J. Haematol. 94, 729–739 (1996). Identifies a role for complement in S. sanguisinduced platelet aggregation. 100. Herzberg, M. C. et al. The platelet interactivity phenotype of Streptococcus sanguis influences the course of experimental endocarditis. Infect. Immun. 60, 4809–4818 (1992). 101. Beachey, E. H. & Stollerman, G. H. Toxic effects of streptococcal M protein on platelets and polymorphonuclear leukocytes in human blood. J. Exp. Med. 134, 351–365 (1971). 102. Kurpiewski, G. E., Forrester, L. J., Campbell, B. J. & Barrett, J. T. Platelet aggregation by Streptococcus pyogenes. Infect. Immun. 39, 704–708 (1983). 103. Herzberg, M. C., Brintzenhofe, K. L. & Clawson, C. C. Aggregation of human platelets and adhesion of Streptococcus sanguis. Infect. Immun. 39, 1457–1469 (1983). 104. Erickson, P. R. & Herzberg, M. C. A collagen-like immunodeterminant on the surface of Streptococcus sanguis induces platelet aggregation. J. Immunol. 138, 3360–3366 (1987). 105. Erickson, P. R. & Herzberg, M. C. The Streptococcus sanguis platelet aggregation-associated protein. Identification and characterization of the minimal platelet-interactive domain. J. Biol. Chem. 268, 1646–1649 (1993). 106. Gong, K., Wen, D. Y., Ouyang, T., Rao, A. T. & Herzberg, M. C. Platelet receptors for the Streptococcus sanguis adhesin and aggregationassociated antigens are distinguished by antiidiotypical monoclonal antibodies. Infect. Immun. 63, 3628–3633 (1995). 107. Herzberg, M. C. et al. Oral streptococci and cardiovascular disease: searching for the platelet aggregation-associated protein gene and mechanisms of Streptococcus sanguis-induced thrombosis. J. Periodontol. 76, 2101–2105 (2005). 108. Sullam, P. M., Jarvis, G. A. & Valone, F. H. Role of immunoglobulin G in platelet aggregation by viridans group streptococci. Infect. Immun. 56, 2907–2911 (1988). Early paper showing a role for IgG in streptococciinduced platelet aggregation. 109. McNicol, A. et al. A role for immunoglobulin G in donor-specific Streptococcus sanguis-induced platelet aggregation. Thromb. Haemost. 95, 288–293 (2006). 110. Ford, I. et al. The role of immunoglobulin G and fibrinogen in platelet aggregation by Streptococcus sanguis. Br. J. Haematol. 97, 737–746 (1997). 111. Ford, I., Douglas, C. W., Preston, F. E., Lawless, A. & Hampton, K. K. Mechanisms of platelet aggregation by Streptococcus sanguis, a causative organism in infective endocarditis. Br. J. Haematol. 84, 95–100 (1993). 112. Plummer, C. et al. A serine-rich glycoprotein of Streptococcus sanguis mediates adhesion to platelets via GPIb. Br. J. Haematol. 129, 101–109 (2005). Identifies SrpA as the S. sanguis protein that interacts with GPIb. 113. Sullam, P. M., Valone, F. H. & Mills, J. Mechanisms of platelet aggregation by viridans group streptococci. Infect. Immun. 55, 1743–1750 (1987). 114. Pampolina, C. & McNicol, A. Streptococcus sanguisinduced platelet activation involves two waves of tyrosine phosphorylation mediated by FcγRIIA and αIIbβ3. Thromb. Haemost. 93, 932–939 (2005). 115. Sullam, P. M. et al. Physical proximity and functional interplay of the glycoprotein Ib-IX-V-complex and the Fc receptor Fcγ RIIA on the platelet plasma membrane. J. Biol. Chem. 273, 5331–5336 (1998). Identifies an association between GPIb and FcγRIIa on the platelet surface.
456 | JUNE 2006 | VOLUME 4
116. Douglas, C. W., Heath, J., Hampton, K. K. & Preston, F. E. Identity of viridans streptococci isolated from cases of infective endocarditis. J. Med. Microbiol. 39, 179–182 (1993). 117. Douglas, C. W., Brown, P. R. & Preston, F. E. Platelet aggregation by oral streptococci. FEMS Microbiol. Lett. 60, 63–67 (1990). 118. Bensing, B. A., Gibson, B. W. & Sullam, P. M. The Streptococcus gordonii platelet binding protein GspB undergoes glycosylation independently of export. J. Bacteriol. 186, 638–645 (2004). 119. Bensing, B. A., Lopez, J. A. & Sullam, P. M. The Streptococcus gordonii surface proteins GspB and Hsa mediate binding to sialylated carbohydrate epitopes on the platelet membrane glycoprotein Ibα. Infect. Immun. 72, 6528–6537 (2004). Identifies S. gordonii proteins that interact with GPIb. 120. Takahashi, Y., Yajima, A., Cisar, J. O. & Konishi, K. Functional analysis of the Streptococcus gordonii DL1 sialic acid-binding adhesin and its essential role in bacterial binding to platelets. Infect. Immun. 72, 3876–3882 (2004). 121. Bensing, B. A. & Sullam, P. M. An accessory sec locus of Streptococcus gordonii is required for export of the surface protein GspB and for normal levels of binding to human platelets. Mol. Microbiol. 44, 1081–1094 (2002). 122. Takamatsu, D., Bensing, B. A. & Sullam, P. M. Genes in the accessory sec locus of Streptococcus gordonii have three functionally distinct effects on the expression of the platelet-binding protein GspB. Mol. Microbiol. 52, 189–203 (2004). 123. Takamatsu, D., Bensing, B. A. & Sullam, P. M. Four proteins encoded in the gspB-secY2A2 operon of Streptococcus gordonii mediate the intracellular glycosylation of the platelet-binding protein GspB. J. Bacteriol. 186, 7100–7111 (2004). 124. Yajima, A., Takahashi, Y. & Konishi, K. Identification of platelet receptors for the Streptococcus gordonii DL1 sialic acid-binding adhesin. Microbiol. Immunol. 49, 795–800 (2005). 125. Takamatsu, D. et al. Binding of the Streptococcus gordonii surface glycoproteins GspB and Hsa to specific carbohydrate structures on platelet membrane glycoprotein Ibα. Mol. Microbiol. 58, 380–392 (2005). 126. Jakubovics, N. S. et al. Functions of cell surfaceanchored antigen I/II family and Hsa polypeptides in interactions of Streptococcus gordonii with host receptors. Infect. Immun. 73, 6629–6638 (2005). 127. Takamatsu, D., Bensing, B. A., Prakobphol, A., Fisher, S. J. & Sullam, P. M. Binding of the streptococcal surface glycoproteins GspB and Hsa to human salivary proteins. Infect. Immun. 74, 1933–1940 (2006). 128. Takahashi, Y. et al. Contribution of sialic acid-binding adhesin to pathogenesis of experimental endocarditis caused by Streptococcus gordonii DL1. Infect. Immun. 74, 740–743 (2006). 129. Bensing, B. A., Rubens, C. E. & Sullam, P. M. Genetic loci of Streptococcus mitis that mediate binding to human platelets. Infect. Immun. 69, 1373–1380 (2001). 130. Rasmussen, M., Eden, A. & Bjorck, L. SclA, a novel collagen-like surface protein of Streptococcus pyogenes. Infect. Immun. 68, 6370–6377 (2000). 131. Zimmerman, T. S. & Spiegelberg, H. L. Pneumococcus-induced serotonin release from human platelets. Identification of the participating plasma/ serum factor as immunoglobulin. J. Clin. Invest. 56, 828–834 (1975). 132. Pietrocola, G. et al. FbsA, a fibrinogen-binding protein from Streptococcus agalactiae, mediates platelet aggregation. Blood 105, 1052–1059 (2005). 133. Sjobring, U., Ringdahl, U. & Ruggeri, Z. M. Induction of platelet thrombi by bacteria and antibodies. Blood 100, 4470–4477 (2002). Describes the interaction of bacteria and platelets under physiological shear conditions. 134. Curtis, M. A., Macey, M., Slaney, J. M. & Howells, G. L. Platelet activation by Protease I of Porphyromonas gingivalis W83. FEMS Microbiol. Lett. 110, 167–173 (1993). 135. Naito, M. et al. Porphyromonas gingivalis-induced platelet aggregation in plasma depends on Hgp44 adhesin but not Rgp proteinase. Mol. Microbiol. 59, 152–167 (2006). 136. Blaser, M. J. & Atherton, J. C. Helicobacter pylori persistence: biology and disease. J. Clin. Invest. 113, 321–333 (2004).
www.nature.com/reviews/micro © 2006 Nature Publishing Group
REVIEWS 137. Danesh, J. et al. Helicobacter pylori infection and early onset myocardial infarction: case-control and sibling pairs study. BMJ 319, 1157–1162 (1999). 138. Pellicano, R., Fagoonee, S., Rizzetto, M. & Ponzetto, A. Helicobacter pylori and coronary heart disease: which directions for future studies? Crit. Rev. Microbiol. 29, 351–359 (2003). 139. Byrne, M. F. et al. Helicobacter pylori binds von Willebrand factor and interacts with GPIb to induce platelet aggregation. Gastroenterology 124, 1846–1854 (2003). 140. Morton, A. R. et al. Campylobacter induced thrombotic thrombocytopenic purpura. Lancet 326, 1133–1134 (1985). 141. Sharma, A. et al. Porphyromonas gingivalis platelet aggregation activity: outer membrane vesicles are potent activators of murine platelets. Oral Microbiol. Immunol. 15, 393–396 (2000). 142. Stanek, G. & Strle, F. Lyme borreliosis. Lancet 362, 1639–1647 (2003). 143. Coburn, J., Leong, J. M. & Erban, J. K. Integrin αIIbβ3 mediates binding of the Lyme disease agent Borrelia burgdorferi to human platelets. Proc. Natl Acad. Sci. USA 90, 7059–7063 (1993). 144. Defoe, G. & Coburn, J. Delineation of Borrelia burgdorferi p66 sequences required for integrin α(IIb)β(3) recognition. Infect. Immun. 69, 3455–3459 (2001). 145. Perine, P. L., Parry, E. H., Vukotich, D., Warrell, D. A. & Bryceson, A. D. Bleeding in louse-borne relapsing fever. I. Clinical studies in 37 patients. Trans. R. Soc. Trop. Med. Hyg. 65, 776–781 (1971). 146. Alugupalli, K. R. et al. Platelet activation by a relapsing fever spirochaete results in enhanced bacterium–platelet interaction via integrin αIIbβ3 activation. Mol. Microbiol. 39, 330–340 (2001). 147. Alugupalli, K. R. et al. Spirochete–platelet attachment and thrombocytopenia in murine relapsing fever borreliosis. Blood 102, 2843–2850 (2003). 148. Arvand, M., Bhakdi, S., Dahlback, B. & Preissner, K. T. Staphylococcus aureus α-toxin attack on human platelets promotes assembly of the prothrombinase complex. J. Biol. Chem. 265, 14377–14381 (1990). 149. Bayer, A. S. et al. Hyperproduction of α-toxin by Staphylococcus aureus results in paradoxically reduced virulence in experimental endocarditis: a host defense role for platelet microbicidal proteins. Infect. Immun. 65, 4652–4660 (1997). 150. Palma, M., Shannon, O., Quezada, H. C., Berg, A. & Flock, J. I. Extracellular fibrinogen-binding protein, Efb, from Staphylococcus aureus blocks platelet aggregation due to its binding to the α-chain. J. Biol. Chem. 276, 31691–31697 (2001). 151. Shannon, O. & Flock, J. I. Extracellular fibrinogen binding protein, Efb, from Staphylococcus aureus binds to platelets and inhibits platelet aggregation. Thromb. Haemost. 91, 779–789 (2004). 152. Tran, U., Boyle, T., Shupp, J. W., Hammamieh, R. & Jett, M. Staphylococcal enterotoxin B initiates protein kinase C translocation and eicosanoid metabolism while inhibiting thrombin-induced aggregation in human platelets. Mol. Cell. Biochem. 21 Mar 2006 (doi:10.1007/s11010-006-9134-6). 153. Alam, S., Gupta, M. & Bhatnagar, R. Inhibition of platelet aggregation by anthrax edema toxin. Biochem. Biophys. Res. Commun. 339, 107–114 (2006). 154. Kau, J. H. et al. Antiplatelet activities of anthrax lethal toxin are associated with suppressed p42/44 and p38
mitogen-activated protein kinase pathways in the platelets. J. Infect. Dis. 192, 1465–1474 (2005). 155. Iwaki, M., Kamachi, K., Heveker, N. & Konda, T. Suppression of platelet aggregation by Bordetella pertussis adenylate cyclase toxin. Infect. Immun. 67, 2763–2768 (1999). 156. Beck, J., Garcia, R., Heiss, G., Vokonas, P. S. & Offenbacher, S. Periodontal disease and cardiovascular disease. J. Periodontol. 67, 1123–1137 (1996). 157. Pihlstrom, B. L., Michalowicz, B. S. & Johnson, N. W. Periodontal diseases. Lancet 366, 1809–1820 (2005). 158. Spahr, A. et al. Periodontal infections and coronary heart disease: role of periodontal bacteria and importance of total pathogen burden in the Coronary Event and Periodontal Disease (CORODONT) study. Arch. Intern. Med. 166, 554–559 (2006). Shows an association between periodontitis, oral pathogen burden and cardiovascular disease. 159. Haffajee, A. D. & Socransky, S. S. Microbial etiological agents of destructive periodontal diseases. Periodontol. 2000 5, 78–111 (1994). 160. Potempa, J. & Travis, J. Porphyromonas gingivalis proteinases in periodontitis, a review. Acta. Biochim. Pol. 43, 455–465 (1996). 161. Potempa, J., Banbula, A. & Travis, J. Role of bacterial proteinases in matrix destruction and modulation of host responses. Periodontol. 24, 153–192 (2000). 162. Lourbakos, A. et al. Activation of protease-activated receptors by gingipains from Porphyromonas gingivalis leads to platelet aggregation: a new trait in microbial pathogenicity. Blood 97, 3790–3797 (2001). Identifies a novel mechanism of bacteria-induced platelet activation involving a family of secreted proteases. 163. Coutinho, I. R., Berk, R. S. & Mammen, E. Platelet aggregation by a phospholipase C from Pseudomonas aeruginosa. Thromb. Res. 51, 495–505 (1988). 164. Kalia, N. et al. Studies on the gastric mucosal microcirculation. 2. Helicobacter pylori water soluble extracts induce platelet aggregation in the gastric mucosal microcirculation in vivo. Gut 41, 748–752 (1997). 165. Ohkuni, H. et al. Purification and partial characterization of a novel human platelet aggregation factor in the extracellular products of Streptococcus mitis, strain Nm-65. FEMS Immunol. Med. Microbiol. 17, 121–129 (1997). 166. Chia, J. S., Lin, Y. L., Lien, H. T. & Chen, J. Y. Platelet aggregation induced by serotype polysaccharides from Streptococcus mutans. Infect. Immun. 72, 2605–2617 (2004). 167. Barkalow, K. L. et al. Role for phosphoinositide 3-kinase in FcγRIIA-induced platelet shape change. Am. J. Physiol. Cell Physiol. 285, C797–C805 (2003). 168. Ragab, A. et al. The tyrosine phosphatase 1B regulates linker for activation of T-cell phosphorylation and platelet aggregation upon FcγRIIa cross-linking. J. Biol. Chem. 278, 40923–40932 (2003). 169. Cooney, D. S., Phee, H., Jacob, A. & Coggeshall, K. M. Signal transduction by human-restricted Fcγ RIIa involves three distinct cytoplasmic kinase families
NATURE REVIEWS | MICROBIOLOGY
leading to phagocytosis. J. Immunol. 167, 844–854 (2001). 170. Usui, Y. et al. Platelet aggregation induced by strains of various species of coagulase-negative staphylococci. Microbiol. Immunol. 35, 15–26 (1991). 171. Guckian, J. C. Effect of pneumococci on blood clotting, platelets, and polymorphonuclear leukocytes. Infect. Immun. 12, 910–918 (1975). 172. Usui, Y., Ichiman, Y., Suganuma, M. & Yoshida, K. Platelet aggregation by strains of enterococci. Microbiol. Immunol. 35, 933–942 (1991). 173. Takii, R., Kadowaki, T., Baba, A., Tsukuba, T. & Yamamoto, K. A functional virulence complex composed of gingipains, adhesins, and lipopolysaccharide shows high affinity to host cells and matrix proteins and escapes recognition by host immune systems. Infect. Immun. 73, 883–893 (2005). 174. Kiley, P. & Holt, S. C. Characterization of the lipopolysaccharide from Actinobacillus actinomycetemcomitans Y4 and N27. Infect. Immun. 30, 862–873 (1980). 175. Willcox, M. D., Webb, B. C., Thakur, A. & Harty, D. W. Interactions between Candida species and platelets. J. Med. Microbiol. 47, 103–110 (1998). 176. Kalvegren, H., Majeed, M. & Bengtsson, T. Chlamydia pneumoniae binds to platelets and triggers P-selectin expression and aggregation: a causal role in cardiovascular disease? Arterioscler. Thromb. Vasc. Biol. 23, 1677–1683 (2003). 177. Czuprynski, C. J. & Balish, E. Interaction of rat platelets with Listeria monocytogenes. Infect. Immun. 33, 103–108 (1981).
Acknowledgements We wish to acknowledge the Health Research Board of Ireland for supporting our research on bacteria–platelet interactions.
Competing interests statement The authors declare no competing financial interests.
DATABASES The following terms in this article are linked online to: Entrez Gene: http://www.ncbi.nlm.nih.gov/entrez/query. fcgi?db=gene ClfA | fnbA | fnbB Entrez Genome Project: http://www.ncbi.nlm.nih.gov/ entrez/query.fcgi?db=genomeprj Bacillus anthracis | Bordetella pertussis | Borrelia burgdorferi | Campylobacter fetus | Candida albicans | Chlamydophila pneumoniae | Escherichia coli | Helicobacter pylori | Lactococcus lactis | Porphyromonas gingivalis | Pseudomonas aeruginosa | Staphylococcus aureus | Staphylococcus epidermidis | Streptococcus agalactiae | Streptococcus gordonii | Streptococcus mitis | Streptococcus pneumoniae | Streptococcus pyogenes | Streptococcus sanguinis UniProtKB: http://ca.expasy.org/sprot α-toxin | clumping factor A | clumping factor B | CTAP-3 | Efb | enterotoxin B | FbsA | fibrinopeptide B | FnBPA | FnBPB | GspB | Hsa | platelet factor 4 | RANTES | SdrE | SecA2 | SecY2 | thymosin β-4 | vWF
FURTHER INFORMATION Timothy J. Foster’s homepage: http://www.tcd.ie/Microbiology/page223.html J. Ross Fitzgerald’s homepage: http://www.bacteriology.ed.ac.uk/groups/lbep Royal College of Surgeons in Ireland: http://www.rcsi.ie Access to this links box is available online.
VOLUME 4 | JUNE 2006 | 457 © 2006 Nature Publishing Group
REVIEWS
The versatile ε-proteobacteria: key players in sulphidic habitats Barbara J. Campbell*§, Annette Summers Engel ‡§, Megan L. Porter ¶ and Ken Takai ||
Abstract | The ε-proteobacteria have recently been recognized as globally ubiquitous in modern marine and terrestrial ecosystems, and have had a significant role in biogeochemical and geological processes throughout Earth’s history. To place this newly expanded group, which consists mainly of uncultured representatives, in an evolutionary context, we present an overview of the taxonomic classification for the class, review ecological and metabolic data in key sulphidic habitats and consider the ecological and geological potential of the ε-proteobacteria in modern and ancient systems. These integrated perspectives provide a framework for future culture- and genomic-based studies.
*College of Marine Studies, University of Delaware, Lewes, Delaware 19958, USA. ‡ Department of Geology and Geophysics, Louisiana State University, Baton Rouge, Louisiana 70803, USA. ¶ Department of Biological Sciences, University of Maryland Baltimore County, Baltimore, Maryland 21250, USA. || Subground Animalcule Retrieval (SUGAR) Program, Extremobiosphere Research Center, Japan Agency for Marine-Earth Science & Technology, 2-15 Natsushima-cho, Yokosuka 237-0061, Japan. § The authors contributed equally to this work. Correspondence to B.J.C. and A.S.E. e-mails:
[email protected] and
[email protected] doi:10.1038/nrmicro1414 Published online 2 May 2006
Although pathogenic species such as Helicobacter pylori have been well studied, the ε-proteobacteria, to which H. pylori is affiliated, is the most poorly characterized class within the Proteobacteria1–3. In 2002, the International Committee on Systematics of Prokaryotes Subcommittee on the taxonomy of Campylobacter and related bacteria4 recognized the increasing number of unclassified and unaffiliated ε-proteobacterial 16S ribosomal RNA (rRNA) sequences deposited into the public databases and recommended that future investigations should deal with this growing problem. Despite recent culture-based investigations and descriptions for novel ε-proteobacterial groups, most lineages are still without cultured representatives or are known only from environmentally retrieved 16S rRNA gene sequences from PCR-based studies of anaerobic to microaerophilic, sulphur-rich marine and terrestrial aquatic environments, or from symbioses with metazoans. Many of these habitats are deemed ‘extreme’ environments — from the hydrothermal fluids of deep-sea vents to the cold darkness of sulphidic caves. A taxonomic framework for the ε-proteobacteria is still lacking. For lineages without cultured representatives, this has made it difficult to fully assess the importance of any newly discovered bacteria. In this review, we evaluate class taxonomic structure as a frame of reference for placing new ε-proteobacterial sequences derived from 16S rRNA gene analyses in an evolutionary context. With this perspective, and to offer recommendations for future research directions, we explore major habitats and highlight ecophysiological diversity patterns based on phylogeny and current metabolic and genomic properties of cultured representatives.
458 | JUNE 2006 | VOLUME 4
Phylogenetic and ecophysiological diversity Ideally, taxonomic classification should be performed through a polyphasic approach using more than one molecular marker and phenotypic information derived from cultured representatives5,6. However, because of the widespread and almost exclusive use of the 16S rRNA gene for phylogenetic studies, and the dearth of cultures, we compiled 1,037 16S rRNA gene sequences (>1,200 bp) from public databases (RDPII, GenBank, EMBL and DDBJ) up to May 2005 and from published reports of clones, strains or sequences described as ‘uncultured bacterium’ with previously determined phylogenetic affinity to the ε-proteobacteria. To construct a phylogenetic foundation for more detailed analyses, a Neighbour Joining (NJ) tree was constructed in PAUP* (REF.7), calculating distances under the general time-reversible model incorporating invariable sites and rate heterogeneity. The analyses revealed that a few previously affiliated ε-proteobacterial 16S rRNA gene sequences were chimeric or misidentified (see Supplementary information S1 (table)). The four clades that contain environmental sequences were then subjected to more rigorous maximum likelihood analyses using PHYML8 with the same model chosen for the NJ analysis. To estimate nodal supports, 100 bootstrap replicates were performed. The ε-proteobacterial sequences currently belong to two valid orders, the Nautiliales (genera Nautilia, Caminibacter and Lebetimonas) 2,9–11 and the Campylobacterales (families Campylobacteraceae, Helicobacteraceae and Hydrogenimonaceae)12,13. Excluding clinical systems (such as infectious associations with humans) affiliated with the Campylobacter and Helicobacter genera, the remaining ε-proteobacterial sequences are diagnosed www.nature.com/reviews/micro
© 2006 Nature Publishing Group
REVIEWS Thermophile An organism that grows optimally at high temperatures, usually above 45°C.
Autotroph An organism that can use carbon dioxide as the sole source of carbon for growth.
Heterotroph An organism that uses organic compounds as nutrients to produce energy for growth.
Chemocline A chemical gradient from high to low concentrations, often consisting of a relatively small stratum where the concentration changes rapidly between the two endpoints.
Mesophile An organism that grows optimally at moderate temperatures, ranging between 20°C and 45°C.
into four robust phylogenetic clusters — classified here as the Nautiliales, Arcobacter, Sulfurospirillum and environmental sequence clusters — that consist of sequences retrieved from various marine systems (for example, deep-sea hydrothermal vents, vent fauna and deep-sea marine subsurfaces) and terrestrial systems (for example, groundwater, caves and springs) (FIG. 1). With few exceptions, ε-proteobacterial sequence affinities strongly correlate with ecotype for each of the phylogenetic clusters (denoted as coloured lines in FIG. 2 and coloured text in Supplementary information S2 (figure)) and metabolic capabilities (denoted as coloured symbols in Supplementary information S2 (figure)). Within the deeply branching group of the Nautiliales, sequences have been retrieved exclusively from hydrothermal systems, and cultured representatives of the family are thermophilic, autotrophic and can reduce elemental sulphur with molecular hydrogen (see Supplementary information S2 (figure), part a). Even within the Sulfurospirillum (FIG. 1; see Supplementary information S2 (figure), part b) and Arcobacter (FIG. 1; see Supplementary information S2 (figure), part c) clusters, nearly all of the sequences are grouped based on environmental setting and metabolism. For instance, although all characterized Sulfurospirillum spp. ferment
Ar co ba c
ic o ba cte r
lla line Wo
Campylobacter
te r
Candidatus A. sulfidicus Oilfield 'FWKO B'
Hel
m illu pir s o fur Sul sp. Am-N Hydrogenimonas Thioreductor Nitratiruptor Caminibacter Lebetimonas Nautilia Nautiliales
Sulfurovum
Nitratifractor
Thiovulum Thiomicrospira sp. CVO Thiomicrospira denitrificans
tal men Environ
Sulfuricurvum
Sulfurimonas
Figure 1 | Phylogeny of 1,037 near full-length (>1,200 bp) ε-proteobacterial sequences collected from public databases and published research. Sequences were aligned using Muscle v3.52 (REF. 116) followed by removal of highly divergent and ambiguous regions using Gblocks v0.91b (REF. 117). The phylogeny was reconstructed using Neighbour Joining under a general time-reversible model of evolution. Major taxonomic divisions, and all of the currently recognized genera, are indicated. Branches for environmental sequences are coloured to represent either marine (blue) or terrestrial (green) habitats.
NATURE REVIEWS | MICROBIOLOGY
using fumarate and can reduce nitrate to ammonia, with the exception of Sulfurospirillum multivorans14, one feature that phylogenetically distinguishes the cultured sulfurospirilla is their ability to respire using alternative electron acceptors under heterotrophic conditions15 (TABLE 1; see Supplementary information S2 (figure), part b). Sequences from different strains that respire using similar elements are more closely related to each other compared with other species within the family, despite strains originating from different geographical locations (for example, Sulfurospirillum carboxydovorans, Sulfurospirillum arcachonense and Sulfurospirillum sp. Am-N). Although arcobacters have been implicated in human and animal enteric diseases 16, few studies have combined isolation and molecular techniques to examine their habitat range17. The type species of the genus Arcobacter nitrofigilis was isolated from a salt-marsh plant root18, but there is still significant diversity among the arcobacters. Similar to the sulfurospirilla, Arcobacter sequences retrieved from marine and terrestrial habitats group together (FIG. 1) and with ecotype (see Supplementary information S2 (figure), part c). Although the metabolic capabilities of most arcobacters have not been studied in detail, many of the cultured representatives originate from marine environments with a well defined geochemical interface between dissolved oxygen and sulphide concentrations17. For example, ‘Candidatus Arcobacter sulfidicus’ was isolated from coastal marine sediments with an oxygen–sulphide chemocline19. This bacterium undergoes mesophilic, chemolithoautotrophic growth, and produces filamentous sulphur with sulphide and oxygen as the electron donor and acceptor, respectively. Based on radio- and stable-isotopic experiments of carbonfixation processes, Candidatus A. sulfidicus was the first ε-proteobacterium thought to assimilate inorganic carbon sources, not through the Calvin–Benson pathway, but by means of the reductive TCA cycle (rTCA cycle)19. The phylogenetic assignment of the remaining sequences is problematic. Based on the bootstrap supported phylogenetic topology, there is a large group that is distinct from the other major clusters (FIGS 1,2; see Supplementary information S2 (figure), part d). This cluster represents the largest increase in 16S rRNA genesequence diversity throughout the ε-proteobacteria and includes several recently described genera. Currently, this sequence cluster has no hierarchical taxonomic classification and future taxonomic revision is required to elucidate the possibility that the cluster might represent more than one hierarchical group. We have provisionally named this clade Thiovulgaceae fam. nov. for ease of reference throughout this review20–22. Thiovulgaceae is derived from thio meaning ‘sulphur’ and vulgar meaning ‘of, pertaining to, common’, forming Thiovulga meaning ‘pertaining to sulphur’;-aceae represents the ending to denote a family. The cultured genera that belong to the family are Gram-negative bacteria that have rod-, vibrio- or filamentous-shaped non-spore-forming cells. Organisms are found in mesophilic conditions, and cultured representatives are
VOLUME 4 | JUNE 2006 | 459 © 2006 Nature Publishing Group
REVIEWS gro
up I
ro up
I
Mari ne
Te rr es tr ia
lg
Sulfurovum
Nitratifractor
n ou Gr
Lower Kane Cave group VI
Sulfuricurvum
I up gro
Lower Kane Cave group I Lower Kane Cave group IV
r wate und Gro
Lower Kane Cave group II/V
group
II
Marine: deep-sea vents, sediments Marine: basinal sediments Marine: deep-sea vent metazoans Terrestrial: water, contaminated water Terrestrial: acid mine drainage, lakes, springs Terrestrial: hydrocarbon groundwater Terrestrial: cave microbial mats Marine: pelagic Marine: whale bone
r ate dw
Termite gut
Sulfurimonas Marine group II
Thiomicrospira
Thiovulum
Figure 2 | The provisional Thiovulgaceae fam. nov. clade. This figure is expanded from FIG. 1. Branches are coloured to represent ecotype. Based on additional, more rigorous maximum likelihood analysis, terrestrial group I is placed outside of the Thiovulgaceae fam. nov. as shown in Supplementary information S2 (figure). Chemolithoautotroph An organism that obtains energy from inorganic compounds and carbon from CO2.
Calvin–Benson pathway Also known as the Calvin– Benson cycle. A series of biochemical, enzyme-mediated reactions in which CO2 is reduced and incorporated into organic molecules.
Reductive TCA cycle (rTCA cycle). The TCA cycle in reverse, leading to the fixation of CO2. Represents a putatively ancient metabolic pathway in which autotrophic carbon fixation occurs under anaerobic conditions.
chemolithoautotrophic and can use molecular hydrogen and/or reduced sulphur compounds as electron donors. Members of the family have been isolated from both marine and freshwater habitats. The Thiovulgaceae fam. nov. family is a member of the order Campylobacterales and comprises the genera Thiovulum 23, Nitratifractor 24, Sulfurovum 25, Sulfuricurvum26, Thiomicrospira27,28 and Sulfurimonas29, which cluster into four main sequences groups, within which are two discrete ecological units — marine group (MG) and groundwater group (GG) (FIG. 2; see Supplementary information S2 (figure), part d). Ecotype groups are closely related to each other (for example MG I and MG II), but relatedness is not supported by bootstrap values, which indicates that significant diversity has yet to be uncovered within the cluster. Unlike the phylogenetic and ecotype patterns within the Sulfurospirillum and Arcobacter clusters, there are little recognizable
460 | JUNE 2006 | VOLUME 4
fine-scale ecotype clade associations. Both MG I and MG II are composed of sequences retrieved from deep-sea vents and sediments, or are associated with vent fauna; however, MG II contains a slightly broader ecotype diversity than MG I, as MG II contains a clade of terrestrial wastewater (sludge) organisms and Thiomicrospira spp. The newly described genera Sulfurovum 25 and Nitratifractor24 are affiliated with MG I, and the sulphuroxidizing genera Sulfurimonas 29, Thiovulum 23 and Thiomicrospira27,28 are affiliated with MG II. A large group of sequences retrieved from groundwater is separated into two clusters, GG I and GG II. Whereas GG I includes the genus Sulfuricurvum26, Lower Kane Cave groups I and IV30, and sequences isolated from wastewater, sludge or groundwater contaminated with petroleum, uranium or tricholoroethene, GG II consists of sequences only from Lower Kane Cave (FIG. 2; see Supplementary information S2 (figure), part d).
www.nature.com/reviews/micro © 2006 Nature Publishing Group
REVIEWS Table 1 | Physiological characteristics of ε-proteobacteria from deep-sea hydrothermal habitats and other selected environments Isolate/phylogenetic Isolation association site
Growth Carbon Electron temperature metabolism donor
Electron acceptor
Sulphur/nitrate reduction to:
Ref.
Order Nautiliales, Family Nautiliaceae Nautilia lithotrophica
Alvinella pompejana 53 ˚C tube, 13˚N EPR
Mixotroph
H2, formate
Sulphite, elemental sulphur
H2S
10
Nautilia sp. str. Am-H
Alvinella pompejana 45 ˚C tube, 13˚N EPR
Mixotroph
H2, formate
Elemental sulphur
H2S
44
Caminibacter hydrogeniphilus
Alvinella pompejana 60 ˚C tube, 13˚N EPR
Mixotroph
H2, complex organic compounds
Nitrate, elemental sulphur
H2S/ NH3
9
Caminibacter profundus
Vent cap, Rainbow Field, MAR
55 ˚C
Autotroph
H2
Nitrate, oxygen (microaerobic) elemental sulphur
H2S/ NH3
2
Caminibacter mediatlanticus
Chimney, Rainbow Field, MAR
55 ˚C
Autotroph
H2
Nitrate, elemental sulphur
H2S/ NH3
46
Lebetimonas acidiphila
In situ colonization system, TOTO, MA
50 ˚C
Autotroph
H2
Elemental sulphur
H2S
11
55 ˚C
Autotroph
H2
Nitrate, oxygen (microaerobic), elemental sulphur
H2S/ NH3
13
55 ˚C
Autotroph
H2
Nitrate, oxygen (microaerobic), elemental sulphur
H2S /N2
24
32 ˚C
Autotroph
H2
Nitrate, elemental sulphur
H2S/ NH3
45
Order uncertain, Family Hydrogenimonaceae Hydrogenimonas thermophila
Chimney, Kairei Field, CIR
Order uncertain, Family Nitratiruptoraceae Nitratiruptor tergarcus
Chimney, Iheya North Field, OT
Order uncertain, Family Thioreductoraceae Thioreductor micantisoli
Sediment, Iheya North Field, OT
Order Campylobacterales, Family Campylobacteraceae Sulfurospirillum sp. str. Am-N
Alvinella pompejana, 13˚N EPR
41 ˚C
Heterotroph
Formate, fumarate
Elemental sulphur
H2S
44
Arcobacter sp. str. FWKO B
Production water, Coleville oil field
30 ˚C
Autotroph
H2, formate, sulphide
Nitrate, oxygen (microaerobic), elemental sulphur
H2S/NO2–
28
Order uncertain, Family Thiovulgaceae Sulfurovum lithotrophicum
Sediment, Iheya North Field, OT
30 ˚C
Autotroph
Elemental sulphur, thiosulphate
Nitrate, oxygen (microaerobic)
N2
25
Nitratifractor salsuginis
Chimney, Iheya North Field, OT
37 ˚C
Autotroph
H2
Nitrate, oxygen (microaerobic)
N2
24
Sulfurimonas autotrophica
Sediment, Hatoma Knoll, OT
25 ˚C
Autotroph
Elemental sulphur, thiosulphate
Oxygen (microaerobic)
Sulfuricurvum kujiense
Groundwater, Japan oil storage cavity
25 ˚C
Autotroph
H2, sulphide, thiosulphate, elemental sulphur
Nitrate, oxygen (microaerobic)
NO2–
26
Thiomicrospira sp. str. CVO
Production water, Coleville oil field
30 ˚C
Mixotroph
Sulphide, elemental sulphur
Oxygen (microaerobic), nitrate, nitrite
N2, N2O
28
29
Location abbreviations: CIR, Central Indian Ridge; EPR, East Pacific Rise; MA, Mariana Volcanic Arc; MAR, Mid-Atlantic Ridge; OT, Okinawa Trough; TOTO, TOTO caldera deep-sea hydrothermal field. Chemical abbreviations: H2S, hydrogen sulphide; NH3, ammonia; N2O, nitrous oxide; NO2–, nitric oxide.
An additional sequence cluster, representing other terrestrial ecotypes (TG), is placed outside of the Thiovulgaceae fam. nov. and other families within the order Campylobacterales (see Supplementary information S2 (figure), part d). The TG I sequences from acid mine drainage, Lower Kane Cave, contaminated groundwater and termite guts might represent greater
NATURE REVIEWS | MICROBIOLOGY
taxonomic diversity than previously hypothesized. Moreover, other than these few termite-gut sequences, virtually nothing is known about the potential of ε-proteobacterial symbioses with terrestrial organisms. Future work in these poorly investigated or unexplored terrestrial systems should increase the known diversity of ε-proteobacterial groups. VOLUME 4 | JUNE 2006 | 461
© 2006 Nature Publishing Group
REVIEWS Box 1 | Integrating ecology and biogeochemistry: hydrothermal vents as a case study The ε-proteobacteria have been found at, and sometimes dominate, four main deep-sea hydrothermal vent-specific habitats: mats on the surfaces of rocks, chimneys and animal surfaces (a in the figure); discharged vent fluids and subseafloor (b); within the hydrothermal vent plume (c); and symbiotic associations with vent animals such as Alvinella pompejana, Alviniconcha aff. hessleri, and Rimicaris spp.103 (d). The metabolically versatile ε-proteobacteria are uniquely suited to thrive in deep-sea habitats and other extreme settings. These hydrothermal-vent habitats are all dynamic suboxic to anaerobic environments, and the ε-proteobacteria use many metabolic processes, including sulphur oxidation, sulphur/ sulphite reduction, nitrate (NO3– ) or nitric oxide (NO2–) reduction to ammonium or nitrogen, and hydrogen and formate oxidation (see figure). In the figure, known electron donors are shown in the yellow shaded blocks and electron acceptors in the green shaded blocks. Some of the ε-proteobacteria might use complex sulphur species for sulphur oxidation or they might biotically influence the formation of iron-sulphur minerals, such as pyrite, at vents104–106. Carbon monoxide might also be used as an electron donor and metal(oid)s (iron, manganese, arsenic and selenium) might be used as electron acceptors, although these metabolisms have not been fully examined in vent habitats. These energy pathways can either be coupled with autotrophy (probably through the reductive TCA cycle), mixotrophy or heterotrophy (TABLE 1). The ε-proteobacterial groups establish themselves as the primary (and perhaps the first) colonizers in the dynamic diffuse flow vent environment because of their metabolic flexibility ( TABLE 1), specialized gene assemblages87, the possibility of special modes for attachment to surfaces43, rapid colonization at O2–H2S interfaces (possibly by formation of filamentous sulphur from hydrogen sulphide39) and phylotypic diversification over time as new habitats are colonized following eruptions or with titanium ring for Alvinella colonization (TRAC) deployment34,43,48,70. Ecological principles indicate that there is a tendency for the most productive species in an ecosystem to be the most dominant in a habitat, thereby pushing other species to comparatively lower densities. It is not surprising that the metabolically versatile ε-proteobacteria colonize extensive areas that are warmer (20–60oC) and have higher concentrations of sulphur species than locations where typical chemolithoautotrophic γ-proteobacteria are found. Hydrothermal fluids: Anaerobic, >350°C Substrates: H2, CO, CO2, H2S, As, FeSx, Mn, Se, organic acids
Seawater: Aerobic, 4°C Substrates: O2, NO3–
c Hydrothermal vent plume Caminibacter profundus Lebetimonas acidiphila
H2 S0, NO3–, O2 H2 S0
a Chimney structures Caminibacter mediatlanticus Hydrogenimonas thermophila Nitratifractor salsuginis Nitratiruptor tergarcus
H2 H2 H2 H2
NO3–, S0 NO3–, O2, S0 NO3–, O2 NO3–, O2, S0
–
–
a Vent fauna associations Caminibacter hydrogeniphilus Nautilia lithotrophica Nautilia sp. Am-H Sulfurospirillium sp. Am-N
H2, organics H2, formate H2, formate Formate, fumarate
NO3–, S0 S0, SO32– S0, SO32– S0
d Symbiotic associations No cultured representatives
b Hydrothermal sediments/subsurface Sulfurimonas autotrophica S0, S2O32– O2 Sulfurovum lithotrophicum S0, S2O32– NO3–, O2 NO3–, S0 Thioreductor micantisoli H2
Sea water seeping through crust and back to the vent system Heat from magma below
In the figure, the cultured species from each hydrothermal-vent-specific habitat are shown. Coloured arrows indicate the flow of either hot hydrothermal fluids (red) or cold sea water (blue). S0, elemental sulphur; SO32– sulphite; S2O3, thiosulphate.
Mixotroph An organism that can use both heterotrophic and autotrophic metabolic processes.
ε-Proteobacteria from marine systems Hydrothermal vents and vent-associated subsurfaces. To interpret the possible ecological and geological significance of the uncultured ε-proteobacterial groups, we look to the exemplary hydrothermal vent system (BOX 1). Since the discovery of hydrothermal vents in 1977, the importance of microorganisms as the prevailing biological feature in these environments has been clearly established. Whereas early studies focused on the endosymbiotic microbial assemblages of the vent tube worm, Riftia pachyptila31, the accumulated
462 | JUNE 2006 | VOLUME 4
knowledge about the microbial diversity of vent sites now reveals that ε-proteobacteria are probably key players in the cycling of (at least) carbon, nitrogen and sulphur, and have important roles in symbiotic associations with vent metazoans (BOX 2). Moreover, deep-sea hydrothermal environments can be regarded as one of the largest reservoirs of diverse environmental ε-proteobacteria on Earth, ranging from the deeplybranching Nautiliales and Nitratiruptor groups to the Sulfurospirillum, Arcobacter, and the MG I and MG II of the Thiovulgaceae fam. nov.
www.nature.com/reviews/micro © 2006 Nature Publishing Group
REVIEWS Although PCR biases and differences in library construction and screening32 might skew interpretations, most 16S rRNA gene-based studies indicate an overwhelming dominance of ε-proteobacteria in the free-living populations in vent fluids or on, or near, the natural surfaces of vent chimney structures where the most intensive hydrogeochemical mixing occurs between ambient (1–4oC), oxygenated bottom sea water and high-temperature, anoxic and sulphide-enriched vent fluid diffusing from the interior portions of the chimney 33–37. The full-cycle rRNA approach (which includes 16S rRNA gene clone library construction and fluorescence in situ hybridization (FISH)) has unequivocally shown that up to 90% of the microbial communities found in these hydrothermal sites are composed of ε-proteobacteria, predominately associated with the Nautilia and Sulfurimonas genera33,37. Other lines of evidence also point to the importance of ε-proteobacteria at vents. Taylor and Wirsen38 showed that the flocculent discharge emanating from diffuse flow vents was similar to the filamentous sulphur production by chemolithoautotrophic sulphur-oxidizing bacteria, which were later shown to belong to the genus Arcobacter19. Filamentous sulphur mats composed of both vibrioid and filamentous sulphide-oxidizers, many of which were probably ε-proteobacteria, have also been found on titanium devices deployed at vents39. Several other research groups have retrieved ε-proteobacterial 16S rRNA gene sequences from vent caps or other in situ colonization devices34,40–43. For instance, ε-proteobacteria comprised ~81% of the total microbial community from an in situ colonization device deployed into diffuse flow vent emissions for 4 days in the Mid-Okinawa Trough34. In this study, water samples collected from 2 m and 10 m away from the chimney structure had dramatically different percentages of ε-proteobacteria, from 83% to 17.6%, respectively, indicating that the ε-proteobacteria tolerate the immediate and proximal vent conditions, probably owing to the increased availability of energy sources compared with more distal habitats or cold sea water (BOX 1).
Deep-sea hydrothermal-vent habitats have also been important for obtaining pure cultures of diverse phylogenetic groups. The first successful isolations were of hydrogen-oxidizing, sulphur-reducing, thermophilic chemolithoautotrophs from Alvinella pompejana symbiont-associated biomass and tube samples; these isolates belong to the Nautiliales10,44. Recently, several previously uncultivated, phylogenetically diverse ε-proteobacterial groups were isolated from various geologically and geographically distinct deepsea hydrothermal fields, all with a diverse range of physiological characteristics and utilization of electron donors (for example, hydrogen and sulphur) and acceptors (for example, sulphur and nitrogen) coupled to carbon fixation2,11,13,24,29,34,45–47 (FIG. 1; TABLE 1; see Supplementary information S2 (figure)). A sub-seafloor model proposed by Huber et al. indicates that ε-proteobacteria thrive in diffuse flow areas surrounding vent chimneys and heated crustal fields and sediments, where sea water mixes with hydrothermal fluids; ε-proteobacterial cells are discharged when eruptions at mid-ocean-ridge axes flush fluids and resident microorganisms within the seafloor crust and sediment to the seafloor surface and into sea water. Indeed, eruption plumes contain diverse microbial communities, but ε-proteobacteria make up ~20–60% of 16S rRNA gene clone libraries from these plumes, with more intergroup diversity occurring from particle-attached libraries than from free-living populations 48. Eventually, the particles and microorganisms fall to the ocean floor and again become part of the marine sediment and subsurface habitats. These hydrogeological processes have the potential to link microorganisms in all of the marine habitats, and might result in a homogenized genetic pool of microorganisms over time. This could be one explanation why members of ε-proteobacteria from the deep-sea hydrothermal vent and sediment systems are closely related to each other.
Box 2 | Episymbionts of Alvinella pompejana
Phylotype A group of sequences that show some threshold of sequence similarity, usually >97%, and that also form a monophyletic clade.
Epibiont An organism that lives attached to a host organism without apparent consequence (benefit or detriment) to the host.
At least two endemic hydrothermal vent fauna, Alvinella pompejana (East Pacific Rise) and Rimicaris exoculata (Mid-Atlantic Ridge), contain ε-proteobacterial episymbionts107,108. A. pompejana, also known as the Pompeii worm because of its heat tolerance, builds paper-like tube colonies attached to hydrothermal-vent chimneys along the East Pacific Rise. The hydrothermal vent shrimp, R. exoculata, forms large clusters on the warmer sections of vents along the Mid-Atlantic Ridge. A. pompejana, the biology of which has been extensively reviewed109, contains two closely related ε-proteobacterial phylotypes that comprise over 65% of a 16S rRNA gene library107,110 derived from episymbiont biomass. These two groups, both within Marine Group I, are filamentous and distinctly separated horizontally on individual dorsal expansions of A. pompejana, indicating niche specialization (M. T. Cottrell and S. C. Cary, unpublished data and REF. 110). More recently, the first endosymbiotic ε-proteobacterium, represented by a single ε-proteobacterial phylotype, was discovered in a deep-sea hydrothermal-vent-endemic gastropod Alvinoconcha spp.111,112 There have been few clues as to the role of the epibionts of A. pompejana because, despite many attempts, the filamentous ε-proteobacterial symbionts have not yet been cultured. Surveys of geochemical conditions within A. pompejana tubes revealed high temperatures (~20– 80oC) and anoxia, exceeding that of any known metazoan habitat, surprisingly low or trace free hydrogen sulphide (<0.2–46.53 µM), pH values between 5.3 and 6.4, high concentrations of potential electron acceptors (sulphate, nitrate and iron), and potentially lethal levels of heavy metals (zinc, nickel, vanadium, copper, lead, cadmium, cobalt and silver)113–115. Preliminary analysis of a metagenomic library of the A. pompejana symbionts supports the hypothesis that at least a portion of the symbiotic ε-proteobacteria detoxify sulphide by rendering it biologically unavailable through metal-transport and sulphide-oxidation processes so that A. pompejana can thrive in this extreme microhabitat (S. C. Cary et al., unpublished data). Also, sequence analysis of a fosmid library revealed the potential for these symbionts to use the reductive TCA cycle, owing to the presence of a key gene in the pathway, aclBA (ATP citrate lyase)87.
NATURE REVIEWS | MICROBIOLOGY
VOLUME 4 | JUNE 2006 | 463 © 2006 Nature Publishing Group
REVIEWS Marine and subsurface environments. Several recent studies indicate that ε-proteobacteria, specifically those affiliated with the Thiovulgaceae fam. nov., have important roles in nutrient cycling and ecosystem function at other marine interfaces and the overall marine subsurface habitat49. PCR-based studies reveal that ε-proteobacteria occur in high abundance at oxic– anoxic interfaces, such as the chemocline-transition zone where hydrogen sulphide from the sediment meets oxygenated sea water. This is particularly true in both the Black Sea and Cariaco Basin (southern Caribbean Sea), two of the largest anoxic basins on the planet50,51. The ε-proteobacteria are also prevalent in deep-sea sediment push cores, the hydrocarbon-rich Guaymas Basin sediment cores, methane cold seeps, gas hydrates and deltaic mud52–56. In many of these habitats, there is also a relatively high abundance of δ-proteobacteria. In one study of cold-seep sediments, δ- and ε-proteobacteria comprised >78% of the metabolically active fraction (RNA), with the ε-proteobacteria dominating the lower, sulphide-rich sediment fractions56. The strong relationship between δ- and ε-proteobacteria could be due to their respective roles in the sulphur cycle (that is, sulphate reduction for the δ-proteobacteria and sulphur oxidation for the ε-proteobacteria).
Push cores Soft sediment collected using a hollow plastic collection tube that is pushed into the sediment, after which the ends are closed.
Methane cold seeps Areas of the deep ocean floor where oil and methane gas bubble up from under seasediment layers at ambient temperatures, providing an energy source that can sustain deep-sea microbial communities.
ε-Proteobacteria in terrestrial systems According to some researchers, an immense subsurface microbial biosphere might exist that is not just associated with marine sediments or deep-sea hydrothermal vent systems57,58. On the basis of recent PCR-based investigations of terrestrial ecotypes, including naturally sulphur-rich environments such as oil-field brines15,27,28,59, hydrocarbon-contaminated groundwater15,59–61, uncontaminated groundwater62, sulphidic springs63–65 and limestone caves30,66–68, we are beginning to discover the importance of ε-proteobacteria in these habitats. Although there are terrestrial sequences belonging to the Sulfurospirillium and Arcobacter clusters, generally, most of the recently acquired ε-proteobacterial sequences from natural, uncontaminated terrestrial habitats are affiliated with the Thiovulgaceae fam. nov. (FIG. 2; see Supplementary information S2 (figure)). Several chemolithoautotrophic, nitrate-reducing, sulphur-oxidizing, microaerophilic ε-proteobacteria have been isolated from oil-field brines and oil-contaminated groundwater, including Arcobacter sp. strain FWKO B28, Thiomicrospira sp. strain CVO27,28 and Sulfuricurvum kujiense26,61 of GG I (FIGS 1,2; see Supplementary information S2 (figure)). So far, S. kujiense (isolated from hydrocarbon-contaminated Japanese groundwater60) is the only cultured terrestrial representative in the Thiovulgaceae fam. nov. Sulphidic caves (limestone caves with discharging hydrogen-sulphide-rich groundwater) allow easy access to the subsurface and are currently one of the beststudied natural terrestrial sites for ε-proteobacteria30,63,66–68. Lower Kane Cave serves as an ideal model system for understanding terrestrial ε-proteobacteria, particularly the terrestrial Thiovulgaceae fam. nov., because all of the sequences retrieved so far from Lower Kane Cave are affiliated with this evolutionary lineage (FIG. 2; see
464 | JUNE 2006 | VOLUME 4
Supplementary information S2 (figure))30,68. Subsurface terrestrial habitats are geographically isolated from one another owing to geological structures, hydrostratigraphic connectivity and plate tectonics, and terrestrial organisms should have limited dispersal mechanisms and are not capable of atmospheric dispersal processes68,69. Therefore, high sequence similarity for the GG I sequences retrieved from around the world indicates that the ancestral population giving rise to the modern group might have originated from one geographical location. However, because GG II consists of only Lower-KaneCave-derived sequences, this clade might be endemic only to Lower Kane Cave. More research is needed to validate these relationships.
Ecological significance of ε-proteobacteria The ε-proteobacteria have significant roles in the habitats in which they thrive, as primary colonizers, primary producers or in symbiotic associations. Lopez-Garcia et al.70 suggest that, in deep-sea habitats, the ε-proteobacteria maximize overall ecosystem function owing to their high biomass and growth rates, rapid adaptations to changing geochemical conditions and metabolic versatility. These factors all facilitate the colonization of new substrates and habitats39,43. For Candidatus A. sulfidicus, the formation of filamentous sulphur might also stimulate colonization of surfaces in marine habitats35. Because nearly all of the microorganisms isolated from deep-sea hydrothermal-vent or marine-sediment ecotypes are chemolithoautotrophs (TABLE 1), these colonizers also serve as one of the crucial sources for organic carbon to the ecosystems (BOXES 1,2), especially at oxic–anoxic interfaces50,71,72. To assess the importance of chemolithoautotrophy in marine and terrestrial settings, however, in situ rates of carbon production or substrate use are needed. Carbon-fixation rates estimated for Candidatus A. sulfidicus were equal to, or exceeded, those of known sulphur-oxidizing bacteria that use the Calvin cycle19. Furthermore, based on previous work that describes the organic carbon-stable-isotope compositions for some marine-vent organisms73,74 and corresponding carbon-isotope fractionation patterns, organic carbon is probably supplied from primary producers that use the rTCA cycle74. In Lower Kane Cave, M. L. P. estimated that the rate of chemolithoautotrophic primary productivity by H14CO3 assimilation was 96.5 ± 6.0 mg carbon gram dry weight per hour for the ε-proteobacterial-dominated microbial mats24, which is comparable to rates of other autotrophic organisms75. Stable-carbon-isotope analyses in Lower Kane Cave corroborate that chemolithoautotrophically produced carbon supports the otherwise nutrient-poor system30. Although the autotrophic carbonfixation pathways were not evident from the study, the Calvin–Benson cycle was implicated in carbon fixation. In addition to cycling carbon, we know that ε-proteobacteria metabolically convert various forms of reduced and oxidized sulphur and nitrogen compounds (TABLE 1, which has important bearing on the speciation of sulphur/nitrogen within a habitat and on global sulphur/nitrogen cycling, as well as on geochemical and geological processes. Few studies have measured rates
www.nature.com/reviews/micro © 2006 Nature Publishing Group
REVIEWS NTPs, dNTPs
Wood–Ljungdahl pathway) or the rTCA (or Arnon) cycle,
Gluconeogenesis/glycolysis Amino acids PEP CO2
Pyruvate PS
Fdox 2[H]
CO2
LPS
Oxaloacetate
2[H] Fdred
Malate
Acetyl-CoA
Fumarate 2[H]
ATP-CL
ATP, CoA
ADP + P
Frd Succinate
Lipids
Citrate
ATP + CoASH
Isocitrate Succinyl CoA KGS
2-Oxoglutarate (α-Ketoglutarate)
CO2
CO2, 2[H]
Fdred 2[H]
Fdox
Ammonium assimilation
Figure 3 | The reductive or reverse TCA (rTCA) cycle of carbon fixation. The two ferredoxin-linked (Fd) CO2-fixation reactions (green) are oxygen sensitive; therefore, this cycle is generally found in anaerobic to microaerophilic microorganisms. The net product of the cycle is one molecule of acetyl-coenzyme A (CoA) synthesized from two molecules of CO2. Acetyl-CoA can be converted to pyruvate and phosphoenolpyruvate (PEP), which can either regenerate the intermediates of the cycle or be used for gluconeogenesis. Many rTCA-cycle intermediates are used in the generation of other cellular components, as indicated by the green arrows. Key enzymes ATP citrate lyase (ATP-CL), pyruvate synthase (PS, also known as pyruvate:ferredoxin oxidoreductase), ketoglutarate synthase (KGS, also known as 2-oxoglutarate:ferredoxin oxidoreductase) and fumarate reductase (Frd) are shown in blue ovals. ATP-CL, KGS and Frd allow the TCA cycle to operate in reverse (red arrow indicates reverse direction). A shared feature of the Calvin–Benson and rTCA cycles is their bidirectionality; in the presence of small organic compounds, microorganisms can use the rTCA cycle in the forward, oxidizing direction.
of sulphur/nitrogen oxidation/reduction in cultured organisms, much less in the environments in which they dominate. In Lower Kane Cave, an assessment of sulphide-consumption rates in the cave revealed that, under microaerophilic conditions, the chemolithoautotrophic ε-proteobacterial Lower Kane Cave group II (GG II of the Thiovulgaceae fam. nov.) consumed sulphide more rapidly than abiotic hydrogen sulphide loss mechanisms, and were consequently found to be responsible for sulphuric-acid dissolution of the cave host limestone76. These results not only linked the biogeochemical carbon and sulphur cycles but also provided evidence for the geological importance of the ε-proteobacteria to processes such as cave development76. Wood–Ljungdahl pathway Also known as the acetylcoenzyme A pathway. An ancient carbon-fixation pathway found in bacteria and archaea in which CO2 is converted to acetate; the key enzyme is acetyl-coenzyme A synthase/CO dehydrogenase.
ε-Proteobacteria and the rTCA cycle Many of the ε-proteobacteria studied so far are chemolithoautotrophs, and it is relevant to the evolutionary history of this group that chemolithoautotrophy is thought to be the first type of metabolic pathway to have evolved77,78. One of two extant autotrophic pathways, the acetyl-coenzyme A (CoA) pathway (also called the
NATURE REVIEWS | MICROBIOLOGY
most closely resembles the first known autotrophic pathway79–83 (FIG. 3). Until the latest studies on chemolithoautotrophic ε-proteobacteria, the rTCA cycle had been described in only a few microorganisms, including the green sulphur bacterium Chlorobium limicola (Chlorobiaceae), a few members of the δ-proteobacteria (for example Desulfobacter hydrogenophilus) and some members of the thermophilic Aquificales and archaeal Thermoproteaceae groups84–86. Within the ε-proteobacteria, the rTCA cycle was initially thought to be a potential CO2 fixation pathway in Candidatus A. sulfidicus 19 . Subsequently, two fosmids were sequenced from fosmid libraries linked to the dominant ε-proteobacterial episymbionts of A. pompejana87. Both fosmids contained the key indicator gene in the rTCA cycle, ATP citrate lyase (aclBA). Evidence for the potential presence and significance of the rTCA cycle for autotrophic carbon fixation at deep-sea vents has accumulated from phylogenetic analysis of rTCA genes amplified directly from hydrothermal vent chimney samples, from enzymatic expression analyses of aclB88–90, and from genetic analyses of the cultures of Candidatus A. sulfidicus and the chemolithoautotrophic Nautilia sp. strain AmH19,87. Phylogenetic evidence points to the close evolutionary relatedness among the acl gene-encoded ATP citrate lyases (ATP-CLs) of Persephonella marina (Aquificales), the ε-proteobacteria, and plants and animals 87,90. The plant and bacterial ATP-CLs are encoded by two subunits, aclB (the small subunit) and aclA (the large subunit). Sections of the acl subunits have significant homologies to the large subunit of succinyl-CoA synthetase and the small subunit of succinyl-CoA synthetase and citrate synthetase91. The acl gene of C. limicola is more distantly related and might be an ancestral form92. The phylogenetic relationships point to the possibility of transfer of the acl gene to the eukaryotic population somewhere between the Chlorobium and Aquificales split. However, there is evidence of two types of citrate-cleaving systems within the Aquificales themselves, with the citrate-cleaving citryl-CoA lyase/synthetase enzymes found in both Hydrogenobacter thermophilus and Aquifex aeolicus, and ATP citrate lyase in P. marina87,91.
ε-Proteobacteria throughout Earth’s history Although more data are needed to resolve the evolution of the citrate-cleaving system in relation to the evolution of the rTCA cycle and the early evolution of life, the presence and use of ATP-CL in some ε-proteobacteria clearly incite some interesting questions with respect to the overall evolutionary history of the ε-proteobacteria. How old is the subdivision? What role did these organisms have in Early Earth habitats? Because marine deep-sea vents are thought to be some of the most ancient colonized habitats on Earth93, it is not far-reaching to hypothesize that, as the ε-proteobacteria are dominant and important organisms in modern vents and similar extreme habitats, this group has been significant to ecological and biogeochemical processes throughout much of Earth’s history. VOLUME 4 | JUNE 2006 | 465
© 2006 Nature Publishing Group
REVIEWS Aphotic Receiving no light or energy from the sun.
1.
2.
Research to reconstruct the evolutionary relationships among prokaryotic genomes using comparative 16S rRNA phylogeny and protein-sequence analyses has resulted in mixed information regarding the origin of non-photosynthetic sulphur-oxidizing bacteria and the ε-proteobacteria94–98. However, there is sufficient evidence to indicate that the ε-proteobacteria, along with the δ-proteobacteria, have closer genetic ancestry with the Chlorobiaceae and Aquificales than with the other Proteobacteria94,97. Sheridan et al. estimated that the divergence of 16S rRNA gene sequences of ε-proteobacteria dated to 1.37 billion years ago (Ga), which also corresponds to work by Brocks et al., who detected the presence of the Chlorobiaceae, along with the purple sulphur bacteria (Chromatiaceae) in rocks from 1.64 Ga based on unique hydrocarbon biomarkers (molecular fossils). However, much remains to be explored about the evolution of the ε-proteobacteria because, unlike the Chlorobiaceae and Chromatiaceae99, no molecular fossils have yet been identified in the rock record for the ε-proteobacteria. On the basis of biological and isotopic evidence98–102, the time period when ε-proteobacteria might have arisen on Early Earth, as far back as ~2 Ga, is marked by a shift from a reducing to an oxidizing ocean and global atmosphere owing to cyanobacterial photosynthesis98,100. Although considerable debate surrounds the nature of the geochemical conditions on Early Earth, and the relative timing of this transition is hotly deliberated101, recent sulphur-isotope data from ancient mineral deposits indicate that free oxygen was present in the atmosphere and surface environments as early as ~2.2 Ga102, which has implications for the evolution of oxygen-dependent sulphur-oxidation pathways. For present-day sulphuroxidizing bacteria that require aerobic or even microaerophilic conditions (for instance, those affiliated with the γ-proteobacteria), the availability of free oxygen would have been crucial for growth. However, oxygen is not essential for many of the isolated sulphur-reducing or sulphur-oxidizing ε-proteobacteria, especially for the deeply-branching Nautiliales, which are obligate anaerobes, or for other ε-proteobacteria that use various alternative electron acceptors (TABLE 1). Because the metabolic characteristics and ecotype preferences of the modern ε-proteobacteria, including thermophilic growth, anaerobic metabolism and autotrophy through the rTCA cycle, are similar to the Chlorobiaceae and Aquificales, the evolution and significance of the ε-proteobacteria throughout Earth’s history are provocative avenues to pursue in future research.
Garrity, G. M., Bell, J. A. & Lilburn, T. In Bergey’s Manual of Systematic Bacteriology Vol. 2 (eds Brenner, D. J., Krieg, N. R., Staley, J. T. & Garrity, G. M.) 1145 Part C (Springer, New York, 2005). Miroshnichenko, M. L. et al. Caminibacter profundus sp. nov., a novel thermophile of Nautiliales ord. nov within the class ‘ε-proteobacteria’, isolated from a deep-sea hydrothermal vent. Int. J. Syst. Evol. Microbiol. 54, 41–45 (2004). This paper erected the order Nautiliales, (containing the family Nautiliacea and the genera Nautilia and Caminibacter), which is only the second order to be described from the ε-proteobacteria.
3.
4.
5.
Future prospects The ε-proteobacteria represent a unique assemblage of microorganisms that, despite the attention given to the pathogenic members, have had little defined taxonomic or ecological consideration. Our taxonomic considerations of the ε-proteobacteria reveal that there is a large group of environmentally relevant ε-proteobacteria known mainly from phylogenetic studies of 16S rRNA genes, which we have provisionally termed Thiovulgaceae fam. nov. Our phylogenetic analysis indicates that there are several clades of ε-proteobacteria (for example, Thiovulgaceae fam. nov., Thioreductor and Nitratiruptor) that will most likely reveal more diversity in the future. In many sulphidic habitats, especially at oxic– anoxic interfaces, ε-proteobacteria are not only present in the microbial communities, but might be the dominant microorganisms involved in the cycling and recycling of carbon, nitrogen and sulphur compounds. Quantitative measurements of ε-proteobacteria (for example, using FISH) and biogeochemical cycling (by measuring uptake or consumption rates) are few, and more studies are needed to correlate the roles of ε-proteobacteria with their dominance in aphotic, sulphur-rich environments. Certainly, cave and terrestrial spring environments are more suited to these types of measurements than deep-sea hydrothermal vents or the deep subsurface because these are readily accessible sites where phototrophic productivity can be eliminated. Molecular methods, including FISH, metabolic gene presence/expression quantification and genome sequencing, hold promise for understanding the biogeochemical roles of ε-proteobacteria in remote extreme environments. At present, genomes from at least six environmentally relevant chemolithoautotrophic ε-proteobacteria from the Nautilia, Caminibacter, Arcobacter, Thiomicrospira, Sulfurovum and Nitratiruptor genera are being sequenced. More projects are needed, however, to understand the metabolic flexibility of this group, and to better characterize the pathogenic ε-proteobacteria. Genome projects will also further our understanding of ε-proteobacterial phylogenetic diversity and ecophysiology, and will undoubtedly allow for the identification of molecular markers to elucidate the evolutionary history of the entire class, including the Thiovulgaceae fam. nov. Another major challenge is to integrate this extensive molecular information and in situ biogeochemical culture-based strategies to improve our ability to isolate diverse metabolic groups.
Olsen, G. J., Woese, C. R. & Overbeek, R. The winds of (evolutionary) change: breathing new life into microbiology. J. Bacteriol. 176, 1–6 (1994). On, S. L. W. International Committee on Systematics of Prokaryotes Subcommittee on the taxonomy of Campylobacter and related bacteria. Int. J. Syst. Evol. Microbiol. 54, 291–292 (2004). Vandamme, P. et al. Polyphasic taxonomy, a consensus approach to bacterial systematics. Microbiol. Rev. 60, 407–438 (1996). An important overview of the plethora of methods used to classify bacteria, including illustrations of
466 | JUNE 2006 | VOLUME 4
6.
7.
8.
how to integrate these diverse methods into a unified taxonomic scheme. Gevers, D. et al. Re-evaluating prokaryotic species. Nature Rev. Microbiol. 3, 733–739 (2005). Swofford, D. PAUP*: Phylogenetic Analysis Using Parsimony (* and Other Methods) 4th edn (Sinauer Associates, Sunderland, 2003). Guindon, S., Lethiec, F., Duroux, P. & Gascuel, O. PHYML Online — a web server for fast maximum likelihood-based phylogenetic inference. Nucleic Acids Res. 33, W557–W559 (2005).
www.nature.com/reviews/micro © 2006 Nature Publishing Group
REVIEWS 9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
Alain, K. et al. Caminibacter hydrogeniphilus gen. nov., sp. nov., a novel thermophilic, hydrogen-oxidizing bacterium isolated from an East Pacific Rise hydrothermal vent. Int. J. Syst. Evol. Microbiol. 52, 1317–1323 (2002). Miroshnichenko, M. L. et al. Nautilia lithotrophica gen. nov., sp. nov., a thermophilic sulfur-reducing ε-proteobacterium isolated from a deep-sea hydrothermal vent. Int. J. Syst. Evol. Microbiol. 52, 1299–1304 (2002). Takai, K. et al. Lebetimonas acidiphila gen. nov., sp. nov., a novel thermophilic, acidophilic, hydrogen-oxidizing chemolithoautotroph within the ε-proteobacteria, isolated from a deep-sea hydrothermal fumarole in the Mariana Arc. Int. J. Syst. Evol. Microbiol. 55, 183–189 (2005). Meinersmann, R. J., Patton, C. M., Evins, G. M., Wachsmuth, I. K. & Fields, P. I. Genetic diversity and relationships of Campylobacter species and subspecies. Int. J. Syst. Evol. Microbiol. 52, 1789–1797 (2002). Takai, K., Nealson, K. H. & Horikoshi, K. Hydrogenimonas thermophila gen. nov., sp. nov., a novel thermophilic, hydrogen-oxidizing chemolithoautotroph within the ε-proteobacteria, isolated from a black smoker in a Central Indian Ridge hydrothermal field. Int. J. Syst. Evol. Microbiol. 54, 25–32 (2004). Luijten, M. L. G. C. et al. Description of Sulfurospirillum halorespirans sp. nov., an anaerobic, tetrachloroethene-respiring bacterium, and transfer of Dehalospirillum multivorans to the genus Sulfurospirillum as Sulfurospirillum multivorans comb. nov. Int. J. Syst. Evol. Microbiol. 53, 787–793 (2003). Stolz, J. F. et al. Sulfurospirillum barnesii sp. nov. and Sulfurospirillum arsenophilum sp. nov., new members of the Sulfurospirillum clade of the ε-proteobacteria. Int. J. Syst. Bacteriol. 49, 1177–1180 (1999). Snelling, W. J., Matsuda, M., Moore, J. E. & Dooley, J. S. G. Under the microscope: Arcobacter. Lett. Appl. Microbiol. 42, 7–14 (2006). Maugeri, T. L. et al. Detection and enumeration of Arcobacter spp. in the coastal environment of the Straits of Messina (Italy). New Microbiol. 28, 177–182 (2005). Mcclung, C. R., Patriquin, D. G. & Davis, R. E. Campylobacter nitrofigilis sp. nov., a nitrogen-fixing bacterium associated with roots of Spartina alterniflora Loisel. Int. J. Syst. Bacteriol. 33, 605–612 (1983). Wirsen, C. O. et al. Characterization of an autotrophic sulfide-oxidizing marine Arcobacter sp. that produces filamentous sulfur. Appl. Environ. Microbiol. 68, 316–325 (2002). The process of filamentous sulphur formation is attributed to the isolate Candidatus A. sulfidicus that was later found to use the rTCA pathway for CO2 fixation. De Vos, P. & Trüper, H. G. Judicial Commission of the International Committee on Systematic Bacteriology. IXth International (IUMS) Congress of Bacteriology and Applied Microbiology. Minutes of the meetings, 14, 15 and 18 August 1999, Sydney, Australia. Int. J. Syst. Evol. Microbiol. 50, 2239–2244 (2000). Euzeby, J. P. et al. ‘List of Changes in Taxonomic Opinion’: making use of the new lists. Int. J. Syst. Evol. Microbiol. 54, 1429–1430 (2004). Euzeby, J. P. Notification of changes in taxonomic opinion previously published outside the IJSEM. Int. J. Syst. Evol. Microbiol. 56, 11 (2006). Wirsen, C. O. & Jannasch, H. W. Physiological and morphological observations on Thiovulum sp. J. Bacteriol. 136, 765–774 (1978). Nakagawa, S., Takai, K., Inagaki, F., Horikoshi, K. & Sako, Y. Nitratiruptor tergarcus gen. nov., sp. nov. and Nitratifractor salsuginis gen. nov., sp. nov., nitratereducing chemolithoautotrophs of the ε-proteobacteria isolated from a deep-sea hydrothermal system in the Mid-Okinawa Trough. Int. J. Syst. Evol. Microbiol. 55, 925–933 (2005). Inagaki, F., Takai, K., Nealson, K. H. & Horikoshi, K. Sulfurovum lithotrophicum gen. nov., sp. nov., a novel sulfur-oxidizing chemolithoautotroph within the ε-proteobacteria isolated from Okinawa Trough hydrothermal sediments. Int. J. Syst. Evol. Microbiol. 54, 1477–1482 (2004).
26. Kodama, Y. & Watanabe, K. Sulfuricurvum kujiense gen. nov., sp. nov., a facultatively anaerobic, chemolithoautotrophic, sulfur-oxidizing bacterium isolated from an underground crude oil storage cavity. Int. J. Syst. Evol. Microbiol. 54, 2297–2300 (2004). 27. Voordouw, G. et al. Characterization of 16S rRNA genes from oil field microbial communities indicates the presence of a variety of sulfate-reducing, fermentative, and sulfide-oxidizing bacteria. Appl. Environ. Microbiol. 62, 1623–1629 (1996). 28. Gevertz, D., Telang, A. J., Voordouw, G. & Jenneman, G. E. Isolation and characterization of strains CVO and FWKOB, two novel nitrate-reducing, sulfide-oxidizing bacteria isolated from oil field brine. Appl. Environ. Microbiol. 66, 2491–2501 (2000). This study is the first description of chemoautotrophic isolates of ε-proteobacteria from any environment. 29. Inagaki, F., Takai, K., Hideki, K. I., Nealson, K. H. & Horikishi, K. Sulfurimonas autotrophica gen. nov., sp. nov., a novel sulfur-oxidizing ε-proteobacterium isolated from hydrothermal sediments in the MidOkinawa Trough. Int. J. Syst. Evol. Microbiol. 53, 1801–1805 (2003). 30. Engel, A. S., Porter, M. L., Stern, L. A., Quinlan, S. & Bennett, P. C. Bacterial diversity and ecosystem function of filamentous microbial mats from aphotic (cave) sulfidic springs dominated by chemolithoautotrophic ‘ε-proteobacteria’. FEMS Microbiol. Ecol. 51, 31–53 (2004). 31. Cavanaugh, C. M., Jones, M. L., Jannasch, H. W. & Waterbury, J. B. Prokaryotic cells in the hydrothermal vent tube worm Riftia pachyptila Jones: possible chemoautotrophic symbionts. Science 213, 340–342 (1981). 32. Wintzingerode, F. V., Göbel, U. B. & Stackebrandt, E. Determination of microbial diversity in environmental samples: pitfalls of PCR-based rRNA analysis. FEMS Microbiol. Rev. 21, 213–229 (1997). 33. Takai, K. et al. Spatial distribution of marine Crenarchaeota group I in the vicinity of deep-sea hydrothermal systems. Appl. Environ. Microbiol. 70, 2404–2413 (2004). 34. Nakagawa, S. et al. Distribution, phylogenetic diversity and physiological characteristics of ε-proteobacteria in a deep-sea hydrothermal field. Environ. Microbiol. 7, 1619–1632 (2005). 35. Moyer, C. L., Dobbs, F. C. & Karl, D. M. Phylogenetic diversity of the bacterial community from a microbial mat at an active, hydrothermal vent system, Loihi Seamount, Hawaii. Appl. Environ. Microbiol. 61, 1555–1562 (1995). This study is the first to focus on characterizing the genetic diversity of 16S rRNA gene sequences of a hydrothermal vent microbial community, finding that ε-proteobacteria dominate the samples. 36. Longnecker, K. & Reysenbach, A. L. Expansion of the geographic distribution of a novel lineage of ε-proteobacteria to a hydrothermal vent site on the Southern East Pacific Rise. FEMS Microbiol. Ecol. 35, 287–293 (2001). 37. Nakagawa, T. et al. Geomicrobiological exploration and characterization of a novel deep-sea hydrothermal system at the TOTO caldera in the Mariana Volcanic Arc. Environ. Microbiol. 8, 37–49 (2006). 38. Taylor, C. D. & Wirsen, C. O. Microbiology and ecology of filamentous sulfur formation. Science 277, 1483–1485 (1997). This study links the widespread occurrence of filamentous sulphur formation in marine habitats to laboratory sulphur produced by enrichment cultures of a chemolithoautotrophic sulphur-oxidizing bacterium isolated from coastal marine seawater. 39. Taylor, C. D., Wirsen, C. O. & Gaill, F. Rapid microbial production of filamentous sulfur mats at hydrothermal vents. Appl. Environ. Microbiol. 65, 2253–2255 (1999). 40. Corre, E., Reysenbach, A. L. & Prieur, D. ε-proteobacterial diversity from a deep-sea hydrothermal vent on the Mid-Atlantic Ridge. FEMS Microbiol. Lett. 205, 329–335 (2001). 41. Reysenbach, A. L., Longnecker, K. & Kirshtein, J. Novel bacterial and archaeal lineages from an in situ growth chamber deployed at a Mid-Atlantic Ridge hydrothermal vent. Appl. Environ. Microbiol. 66, 3798–3806 (2000). 42. Higashi, Y. et al. Microbial diversity in hydrothermal surface to subsurface environments of Suiyo Seamount, Izu-Bonin Arc, using a catheter-type in situ growth chamber. FEMS Microbiol. Ecol. 47, 327–336 (2004).
NATURE REVIEWS | MICROBIOLOGY
43. Alain, K. et al. Early steps in microbial colonization processes at deep-sea hydrothermal vents. Environ. Microbiol. 6, 227–241 (2004). 44. Campbell, B. J., Jeanthon, C., Kostka, J. E., Luther, G. W. & Cary, S. C. Growth and phylogenetic properties of novel bacteria belonging to the epsilon subdivision of the proteobacteria enriched from Alvinella pompejana and deep-sea hydrothermal vents. Appl. Environ. Microbiol. 67, 4566–4572 (2001). The first study to describe characterized enrichment cultures of four ε-proteobacteria from A. pompejana symbionts and vent chimney samples. 45. Nakagawa, S., Inagaki, F., Takai, K., Horikoshi, K. & Sako, Y. Thioreductor micantisoli gen. nov., sp. nov., a novel mesophilic, sulfur-reducing chemolithoautotroph within the ε-proteobacteria isolated from hydrothermal sediments in the Mid-Okinawa Trough. Int. J. Syst. Evol. Microbiol. 55, 599–605 (2005). 46. Voordeckers, J. W., Starovoytov, V. & Vetriani, C. Caminibacter mediatlanticus sp. nov., a thermophilic, chemolithoautotrophic, nitrate-ammonifying bacterium isolated from a deep-sea hydrothermal vent on the Mid-Atlantic Ridge. Int. J. Syst. Evol. Microbiol. 55, 773–779 (2005). 47. Takai, K. et al. Isolation and phylogenetic diversity of members of previously uncultivated ε-proteobacteria in deep-sea hydrothermal fields. FEMS Microbiol. Lett. 218, 167–174 (2003). This study describes the isolation of metabolically diverse marine ε-proteobacteria, predominately from vent sites, whereby many have recently been described as new taxa. 48. Huber, J. A., Butterfield, D. A. & Baross, J. A. Bacterial diversity in a subseafloor habitat following a deep-sea volcanic eruption. FEMS Microbiol. Ecol. 43, 393–409 (2003). 49. Kormas, K. A., Smith, D. C., Edgcomb, V. & Teske, A. Molecular analysis of deep subsurface microbial communities in Nankai Trough sediments (ODP Leg 190, Site 1176). FEMS Microbiol. Ecol. 45, 115–125 (2003). 50. Madrid, V. M., Taylor, G. T., Scranton, M. I. & Chistoserdov, A. Y. Phylogenetic diversity of bacterial and archaeal communities in the anoxic zone of the Cariaco Basin. Appl. Environ. Microbiol. 67, 1663–1674 (2001). 51. Vetriani, C., Tran, H. V. & Kerkhof, L. J. Fingerprinting microbial assemblages from the oxic/anoxic chemocline of the Black Sea. Appl. Environ. Microbiol. 69, 6481–6488 (2003). 52. Li, L., Kato, C. & Horikoshi, K. Microbial diversity in sediments collected from the deepest cold-seep area, the Japan Trench. Marine Biotechnol. 1, 391–400 (1999). 53. Mills, H. J., Hodges, C., Wilson, K., MacDonald, I. R. & Sobecky, P. A. Microbial diversity in sediments associated with surface-breaching gas hydrate mounds in the Gulf of Mexico. FEMS Microbiol. Ecol. 46, 39–52 (2003). 54. Todorov, J. R., Chistoserdov, A. Y. & Aller, J. Y. Molecular analysis of microbial communities in mobile deltaic muds of Southeastern Papua New Guinea. FEMS Microbiol. Ecol. 33, 147–155 (2000). 55. Teske, A. et al. Microbial diversity of hydrothermal sediments in the Guaymas Basin: evidence for anaerobic methanotrophic communities. Appl. Environ. Microbiol. 68, 1994–2007 (2002). 56. Inagaki, F., Sakihama, Y., Inoue, A., Kato, C. & Horikoshi, K. Molecular phylogenetic analyses of reverse-transcribed bacterial rRNA obtained from deep-sea cold seep sediments. Environ. Microbiol. 4, 277–286 (2002). 57. Whitman, W. B., Coleman, D. C. & Wiebe, W. J. Prokaryotes: The unseen majority. Proc. Natl Acad. Sci. USA 95, 6578–6583 (1998). 58. Pedersen, K. Exploration of deep intraterrestrial microbial life: current perspectives. FEMS Microbiol. Lett. 185, 9–16 (2001). 59. Schumacher, W., Kroneck, P. M. H. & Pfennig, N. Comparative systematic study on Spirillum-5175, Campylobacter and Wolinella species — description of Spirillum-5175 as Sulfurospirillum deleyianum gen. nov., sp. nov. Arch. Microbiol. 158, 287–293 (1992). 60. Watanabe, K., Kodama, Y., Syutsubo, K. & Harayama, S. Molecular characterization of bacterial populations in petroleum-contaminated groundwater discharged from underground crude oil storage cavities. Appl. Environ. Microbiol. 66, 4803–4809 (2000).
VOLUME 4 | JUNE 2006 | 467 © 2006 Nature Publishing Group
REVIEWS 61. Kodama, Y. & Watanabe, K. Isolation and characterization of a sulfur-oxidizing chemolithotroph growing on crude oil under anaerobic conditions. Appl. Environ. Microbiol. 69, 107–112 (2003). 62. Pedersen, K., Nilsson, E., Arlinger, J., Hallbeck, L. & O’Neill, A. Distribution, diversity and activity of microorganisms in the hyper-alkaline spring waters of Maqarin in Jordan. Extremophiles 8, 151–164 (2004). 63. Barton, H. A. & Luiszer, F. Microbial metabolic structure in a sulfidic cave hot spring: potential mechanisms of biospeleogenesis. J. Cave Karst Stud. 67, 28–38 (2005). 64. Elshahed, M. S. et al. Bacterial diversity and sulfur cycling in a mesophilic sulfide-rich spring. Appl. Environ. Microbiol. 69, 5609–5621 (2003). 65. Rudolph, C., Moissl, C., Henneberger, R. & Huber, R. Ecology and microbial structures of archaeal/bacterial strings-of-pearls communities and archaeal relatives thriving in cold sulfidic springs. FEMS Microbiol. Ecol. 50, 1–11 (2004). 66. Angert, E. R. et al. Molecular phylogenetic analysis of a bacterial community in Sulphur River, Parker Cave, Kentucky. Amer. Mineral. 83, 1583–1592 (1998). 67. Engel, A. S., Porter, M. L., Kinkle, B. K. & Kane, T. C. Ecological assessment and geological significance of microbial communities from Cesspool Cave, Virginia. Geomicrobiol. J. 18, 259–274 (2001). 68. Engel, A. S. et al. Filamentous ‘ε-proteobacteria’ dominate microbial mats from sulfidic cave springs. Appl. Environ. Microbiol. 69, 5503–5511 (2003). The first study to show that filamentous ε-proteobacteria are abundant and of metabolic significance in a natural, terrestrial, subterranean system. 69. Finlay, B. J. Global dispersal of free-living microbial eukaryote species. Science 296, 1061–1063 (2002). 70. Lopez-Garcia, P. et al. Bacterial diversity in hydrothermal sediment and ε-proteobacterial dominance in experimental microcolonizers at the Mid-Atlantic Ridge. Environ. Microbiol. 5, 961–976 (2003). 71. Casamayor, E. O., Garcia-Cantizano, J., Mas, J. & Pedros-Alio, C. Primary production in estuarine oxic/ anoxic interfaces: contribution of microbial dark CO2 fixation in the Ebro River Salt Wedge Estuary. Mar. Ecol. Prog. Ser. 215, 49–56 (2001). 72. Taylor, G. T. et al. Chemoautotrophy in the redox transition zone of the Cariaco Basin: a significant midwater source of organic carbon production. Limnol. Oceanogr. 46, 148–163 (2001). 73. Van Dover, C. L. & Fry, B. Stable isotopic compositions of hydrothermal vent organisms. Mar. Biol. 102, 257–263 (1989). 74. Preuss, A., Schauder, R., Fuchs, G. & Stichler, W. Carbon isotope fractionation by autotrophic bacteria with 3 different CO2 fixation pathways. Z. Naturforsch. [C] 44, 397–402 (1989). 75. Porter, M. L. Ecosystem Energetics of Sulfidic Karst. Thesis, Univ. Cincinnati (1999). 76. Engel, A. S., Stern, L. A. & Bennett, P. C. Microbial contributions to cave formation: new insights into sulfuric acid speleogenesis. Geology 32, 369–372 (2004). 77. Russell, M. J. & Hall, A. J. The emergence of life from iron monosulphide bubbles at a submarine hydrothermal redox and pH front. J. Geol. Soc. London 154, 377–402 (1997). 78. Wachtershauser, G. The case for the chemoautotrophic origin of life in an iron-sulfur world. Orig. Life Evol. Biosph. 20, 173–176 (1990). 79. Wachtershauser, G. Evolution of the 1st metabolic cycles. Proc. Natl Acad. Sci. USA 87, 200–204 (1990). This paper presents the hypothesis that a version of the rTCA cycle was the first metabolic cycle to have evolved. 80. Smith, E. & Morowitz, H. J. Universality in intermediary metabolism. Proc. Natl Acad. Sci. USA 101, 13168–13173 (2004). 81. Russell, M. J. & Martin, W. The rocky roots of the acetyl-CoA pathway. Trends Biochem. Sci. 29, 358–363 (2004). 82. Pereto, J. G., Velasco, A. M., Becerra, A. & Lazcano, A. Comparative biochemistry of CO2 fixation and the evolution of autotrophy. Int. Microbiol. 2, 3–10 (1999). 83. Lindahl, P. A. & Chang, B. The evolution of acetyl-CoA synthase. Orig. Life Evol. Biosphere 31, 403–434 (2001).
84. Evans, M. C. W., Buchanan, B. B. & Arnon, D. I. A new ferredoxin-dependent carbon reduction cycle in a photosynthetic bacterium. Proc. Natl Acad. Sci. USA 55, 928–934 (1966). 85. Fuchs, G., Stupperich, E. & Eden, G. Autotrophic CO2 fixation in Chlorobium limicola — evidence for the operation of a reductive tricarboxylic acid cycle in growing cells. Arch. Microbiol. 128, 64–71 (1980). 86. Hugler, M., Huber, H., Stetter, K. O. & Fuchs, G. Autotrophic CO2 fixation pathways in archaea (Crenarchaeota). Arch. Microbiol. 179, 160–173 (2003). 87. Campbell, B. J., Stein, J. L. & Cary, S. C. Evidence of chemolithoautotrophy in the bacterial community associated with Alvinella pompejana, a hydrothermal vent polychaete. Appl. Environ. Microbiol. 69, 5070–5078 (2003). 88. Campbell, B. J. & Cary, S. C. Abundance of reverse tricarboxylic acid cycle genes in free-living microorganisms at deep-sea hydrothermal vents. Appl. Environ. Microbiol. 70, 6282–6289 (2004). The abundance and expression of rTCA genes in microbial communities from deep-sea hydrothermal vents establishes the ubiquity of ε-proteobacteria and chemolithoautotrophy at vents. 89. Hugler, M., Wirsen, C. O., Fuchs, G., Taylor, C. D. & Sievert, S. M. Evidence for autotrophic CO2 fixation via the reductive tricarboxylic acid cycle by members of the ε subdivision of proteobacteria. J. Bacteriol. 187, 3020–3027 (2005). 90. Takai, K. et al. Enzymatic and genetic characterization of carbon and energy metabolisms by deep-sea hydrothermal chemolithoautotrophic isolates of ε-proteobacteria. Appl. Environ. Microbiol. 71, 7310–7320 (2005). 91. Aoshima, M., Ishii, M. & Igarashi, Y. A novel enzyme, citryl-CoA lyase, catalysing the second step of the citrate cleavage reaction in Hydrogenobacter thermophilus TK-6. Mol. Microbiol. 52, 763–770 (2004). 92. Fatland, B. L. et al. Molecular characterization of a heteromeric ATP-citrate lyase that generates cytosolic acetyl-coenzyme A in Arabidopsis. Plant Physiol. 130, 740–756 (2002). 93. Reysenbach, A. L., Banta, A. B., Boone, D. R., Cary, S. C. & Luther, G. W. Microbial essentials at hydrothermal vents. Nature 404, 835–835 (2000). 94. Gupta, R. S. The phylogeny of proteobacteria: relationships to other eubacterial phyla and eukaryotes. FEMS Microbiol. Rev. 24, 367–402 (2000). 95. Wolf, Y. I., Rogozin, I. B., Grishin, N. V., Tatusov, R. L. & Koonin, E. V. Genome trees constructed using five different approaches suggest new major bacterial clades. BMC Evol. Biol. 1, 8 (2001). 96. Gupta, R. S. & Griffiths, E. Critical issues in bacterial phylogeny. Theor. Popul. Biol. 61, 423–434 (2002). 97. Sheridan, P. P., Freeman, K. H. & Brenchley, J. E. Estimated minimal divergence times of the major bacterial and archaeal phyla. Geomicrobiol. J. 20, 1–14 (2003). 98. Canfield, D. E. & Teske, A. Late Proterozoic rise in atmospheric oxygen concentration inferred from phylogenetic and sulphur-isotope studies. Nature 382, 127–132 (1996). 99. Brocks, J. J. et al. Biomarker evidence for green and purple sulphur bacteria in a stratified Palaeoproterozoic sea. Nature 437, 866–870 (2005). 100. Johnson, C. M. & Beard, B. L. Geochemistry. Biogeochemical cycling of iron isotopes. Science 309, 1025–1027 (2005). 101. Des Marais, D. J. Palaeobiology: sea change in sediments. Nature 437, 826–827 (2005). 102. Farquhar, J. & Wing, B. Multiple sulfur isotopes and the evolution of the atmosphere. Earth Planet. Sci. Lett. 213, 1–13 (2003). 103. Karl, D. M. The microbiology of deep-sea hydrothermal vents. (CRC Press, Boca Raton, 1995). 104. Wirsen, C. O., Jannasch, H. W. & Molyneaux, S. J. Chemosynthetic microbial activity at Mid-Atlantic Ridge hydrothermal vent sites. J. Geophys. Res. 98, 9693–9703 (1993). 105. Zbinden, M., Martinez, I., Guyot, F., CambonBonavita, M. A. & Gaill, F. Zinc-iron sulphide mineralization in tubes of hydrothermal vent worms. Eur. J. Mineral. 13, 653–658 (2001).
468 | JUNE 2006 | VOLUME 4
106. Goffredi, S. K., Waren, A., Orphan, V. J., Van Dover, C. L. & Vrijenhoek, R. C. Novel forms of structural integration between microbes and a hydrothermal vent gastropod from the Indian Ocean. Appl. Environ. Microbiol. 70, 3082–3090 (2004). 107. Haddad, A., Camacho, F., Durand, P. & Cary, S. Phylogenetic characterization of the epibiotic bacteria associated with the hydrothermal vent polychaete Alvinella pompejana. Appl. Environ. Microbiol. 61, 1679–1687 (1995). The first paper showing the dominance of ε-proteobacterial 16S rRNA gene sequences in the episymbionts of A. pompejana. 108. Polz, M. F. & Cavanaugh, C. M. Dominance of one bacterial phylotype at a Mid-Atlantic Ridge hydrothermal vent site. Proc. Natl Acad. Sci. USA 92, 7232–7236 (1995). 109. Desbruyéres, D. et al. Biology and ecology of the ‘Pompeii worm’ (Alvinella pompejana ), a normal dweller of an extreme deep-sea environment: a synthesis of current knowledge and recent developments. Deep Sea Res. II 45, 383–422 (1998). 110. Cary, S. C., Cottrell, M. T., Stein, J. L., Camacho, F. & Desbruyeres, D. Molecular identification and localization of filamentous symbiotic bacteria associated with the hydrothermal vent annelid Alvinella pompejana. Appl. Environ. Microbiol. 63, 1124–1130 (1997). 111. Urakawa, H. et al. Hydrothermal vent gastropods from the same family (Provannidae) harbour ε- and γ-proteobacterial endosymbionts. Environ. Microbiol. 7, 750–754 (2005). 112. Suzuki, Y. et al. Novel chemoautotrophic endosymbiosis between a member of the ε-proteobacteria and the hydrothermal vent gastropod Alviniconcha aff. hessleri (Gastropoda: Provannidae) from the Indian Ocean. Appl. Environ. Microbiol. 71, 5440–5450 (2005). 113. Cary, S. C., Shank, T. & Stein, J. Worm basks in extreme temperatures. Nature 391, 545 (1998). 114. Di Meo-Savoie, C. A., Luther, G. W. & Cary, S. C. Physicochemical characterization of the microhabitat of the epibionts associated with Alvinella pompejana, a hydrothermal vent annelid. Geochim. Cosmochim. Acta 68, 2055–2066 (2004). 115. Luther, G. W. et al. Chemical speciation drives hydrothermal vent ecology. Nature 410, 813–816 (2001). 116. Edgar, R. C. MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Res. 32, 1792–1797 (2004). 117. Castresana, J. Selection of conserved blocks from multiple alignments for their use in phylogenetic analysis. Mol. Biol. Evol. 17, 540–552 (2000).
Acknowledgements B.J.C was partially supported by the National Science Foundation. A.S.E was partially supported by the College of Basic Sciences at Louisiana State University, USA. The authors thank T.E. Hanson for his valuable input in to this review.
Competing interests statement The authors declare no competing financial interests.
DATABASES The following terms in this article are linked online to: Entrez Genome Project: http://www.ncbi.nlm.nih.gov/ entrez/query.fcgi?db=genomeprj Aquifex aeolicus | Chlorobium limicola | Helicobacter pylori | Nautilia sp. strain AmH | Persephonella marina
FURTHER INFORMATION Barbara J. Campbell’s homepage: http://www.ocean.udel.edu/cms/bcampbell Annette Summers Engel’s laboratory: http://www.geol.lsu.edu/Faculty/Engel/profile.html A metagenome of Alvinella pompejana symbiont database: http://ocean.dbi.udel.edu/index.php DDBJ: http://www.ddbj.nig.acjp EMBL: http://www.ebi.ac.uk/embl ε-Proteobacteria sequence data files and alignments: http://geol.lsu.edu/Faculty/Engel/epsilon.htm GenBank: http://www.ncbi.nlm.nih.gov/Genbank RDPII: http://wdcm.nig.ac.jp/RDP/html/index.html
SUPPLEMENTARY INFORMATION See online article: S1 (table) | S2 (figure) Access to this links box is available online.
www.nature.com/reviews/micro © 2006 Nature Publishing Group
PERSPECTIVES SCIENCE AND SOCIETY
Advances in tuberculosis vaccine strategies Yasir A. W. Skeiky and Jerald C. Sadoff
Abstract | Tuberculosis (TB), an ancient human scourge, is a growing health problem in the developing world. Approximately two million deaths each year are caused by TB, which is the leading cause of death in HIV-infected individuals. Clearly, an improved TB vaccine is desperately needed. Heterologous prime–boost regimens probably represent the best hope for an improved vaccine regimen to prevent TB. This first generation of new vaccines might also complement drug treatment regimens and be effective against reactivation of TB from the latent state, which would significantly enhance their usefulness. Mycobacterium tuberculosis, the causative agent of tuberculosis (TB), is the second leading infectious cause of mortality worldwide — M. tuberculosis has infected approximately two billion individuals in total, which is one third of the world’s population1. Currently, TB is the leading cause of death in HIV-infected individuals owing to the high incidence of dual infections and the accompanying inhibition of the immune system by both HIV/ AIDS and TB. Ninety percent of the estimated deaths from TB and 95% of the estimated eight million new cases of TB each year occur in developing countries, which comprise 85% of the world’s population2. Incidence is increasing fastest in African countries that are affected by HIV, followed by Eastern Europe and the former Soviet Union, both of which are plagued by multi-drug resistant strains of TB3,4 that present an ominous global threat. The TB pandemic has continued despite widespread use of the only available licensed TB vaccine bacille Calmette–Guérin (BCG)5 under the WHO expanded program on immunization (EPI) and despite the use of directly observed chemotherapy programmes (DOTS) for those diagnosed with active disease. Failure or delay in diagnosis of active TB cases, relapse rates of up to 5% after DOTS6,7, and the lengthy treatment required to cure TB, which often results in poor compliance and the consequent emergence of multi-drug resistant M. tuberculosis strains8–10, present
serious challenges for the DOTS strategy. Although BCG seems to provide protection against disseminated disease in newborns and children, its efficacy against pulmonary TB is poor, which highlights the need for a better vaccine regimen. There are two potential vaccination strategies against TB; those given before (pre-exposure, prophylactic) or after (postexposure, therapeutic) exposure to the pathogen. Therapeutic vaccines can also be used in combination with drug therapy. Specific vaccines designed to prevent latent TB or reactivation from the latent state are in the early stages of development. Here, we review the development of crucial new prophylactic vaccines, the primary goal of which is to prevent infection itself or disease that occurs after infection. Such vaccines include recombinant BCG (rBCG), attenuated strains of M. tuberculosis, subunit vaccine approaches and live, non-replicating viral vector-based delivery systems used alone or in prime–boost regimens. A formidable challenge facing the development of these new vaccines is their safety, especially in populations afflicted with both HIV and TB. Infection with Mycobacterium tuberculosis The development of a delayed-type hypersensitivity response to tuberculin antigen (TST) is considered an indicator of M. tuberculosis infection. Only 20–50% of individuals
NATURE REVIEWS | MICROBIOLOGY
that are exposed to M. tuberculosis become infected, as evidenced by a positive TST in single case contact studies11. Around 5% of infected individuals develop pulmonary disease or some other clinical form of TB over 2–5 years, whereas the other 95% of infected TST positive individuals develop latent TB infection (LTBI). Individuals with LTBI have about a 5% lifetime chance of reactivating their latent infection, resulting in the development of clinical TB11. M. tuberculosis is transmitted by the aerosol route of infection. After phagocytosis, the organism successfully hides from immune-mediated destruction inside the endosome of the macrophage using various mechanisms, which include changing the endosomal pH, inhibiting apoptosis and destroying toxic superoxides that are secreted by immune cells. Infected alveolar macrophages and dendritic cells migrate to adjacent lymph nodes where mycobacterial antigens are presented and a T-helper-1 (TH1)-type response is initiated. This complex immune response ultimately results in the formation of the granuloma around infected macrophages. IFN-γ, which synergizes with tumour-necrosis factor-α (TNF-α), is central to the activation of macrophages and the isolation of M. tuberculosis inside the granuloma. The granuloma is initially composed of CD4+ and CD8+ T cells, but a complex array of T cells, including CD4+, CD8+, γδ T cells and CD1-restricted αβ T cells are also invoked to orchestrate immune responses and to contain the infection12–14. At later stages, the granuloma is surrounded by a fibrotic wall and lymphoid follicular structures. The granuloma can persist for decades and can contain the tubercle infection in a state of dormancy by depriving mycobacteria of oxygen and nutrients. However, failure to contain the infection can result in release of organisms, and caseation and necrosis with active clinical disease and transmission (FIG. 1). Requirements for TB vaccine development Scientific rationale. Most humans have complete or partial resistance to tuberculous disease following exposure or infection with M. tuberculosis. Further indication that resistance to TB is mediated in part by the
VOLUME 4 | JUNE 2006 | 469 © 2006 Nature Publishing Group
PERSPECTIVES a
TNF-α
d
CD8+ T cell
IFN-γ
Class I
TH1
Fibroblast
MHC class II
e M. tuberculosisinfected macrophage
Macrophage –
CD4+ T cell
γδ T cell
Nucleus
Perforin, granzyme
Early phagosome
c Cross-priming (weak)
UreC, SOD, nlaA gene product
Granuloma, containment TH1
Dendritic cell
b
CD1
Apoptosis
f rBCG–pfo + ∆SOD ∆nlaA rBCG∆UreC:Hly+
Cross-priming (strong)
Reactivation, dissemination
Figure 1 | The immune response to Mycobacterium tuberculosis infection or vaccination with BCG or recombinant modified BCG. a | M. tuberculosis subverts macrophage phagosome maturation through the production of UreC, which inhibits the acidification of the early phagosome. M. tuberculosis also produces factors such as superoxide dismutase (SOD) and the nlA gene product, which might inhibit host defences by interfering with host cell apoptosis. In the phagosome, M. tuberculosis antigens have access to the MHC class II machinery resulting in a predominant CD4+ T-cell response. However, apoptosis of infected macrophages produces extracellular vesicles carrying glycolipids and protein antigens. These vesicles are taken up by dendritic cells as a result of cross-priming and presented in the context of MHC class I (CD8+ T-cell responses), MHC class II (CD4+ T-cell responses) and CD1 molecules. b | Enhancement of MHC class I antigen presentation and CD8+ T-cell responses by modified BCG constructs. rBCG vaccines such as rBCG–pfo (with deletion of UreC and expression of the pH-independent perfringolysin) or rBCGΔUreC:Hly+ (with deletion of UreC and expression
immune system stems from the findings that certain groups of immunocompromised people have a high susceptibility to tuberculous disease. These include people with Mendelian inherited defects in immune mediators such as INF-γ 15,16, patients treated with TNF-α-antagonist drugs for rheumatoid arthritis17, and AIDS patients with decreased numbers of CD4+ T cells that are also functionally impaired18. Results from clinical trials with BCG, although not consistent, have nonetheless provided proofof-principle that vaccination can be effective
+
of the acidic pH-dependent Hly (listeriolysin)) facilitate the formation of pores in the membrane of the early phagosome, thereby allowing leakage of BCG mycobacteria and antigens into the cytoplasm. This leads to potent MHC class I presentation, enhancement of apoptosis and increased cross-priming and uptake by dendritic cells (c), resulting in enhanced CD8+ and CD4+ T-cell responses as well as γδ T-cell and CD1-restricted αβ T-cell activation. rBCG constructs with diminished SOD activity or deletion of nlaA can further enhance apoptosis. d | CD8+ T cells produce effector cytokines (IFN-γ and TNF-α) and lyse infected cells through the release of granules (perforin and granzyme) or by Fas-mediated lysis, whereas activated CD4+ TH1 cells produce IFN-γ, which activates macrophages to kill M. tuberculosis. e | Infected macrophages can also be contained within granulomas in a state of mycobacterial dormancy for decades. f | However, M. tuberculosis can reactivate as a result of a weakened immune system, with release of organisms from the granuloma and progression to active clinical disease.
against the consequences of infection with M. tuberculosis. Therefore, although resistance to TB is in part a property of the hostmediated immune response, in people not naturally resistant, vaccination could confer an additional degree of protection. Although animal-challenge models have not been proven to correlate with protection in humans, new rBCG vaccines or heterologous prime–boost regimens (TABLE 1) show better protection than BCG in such models. Such prime–boost regimens consist of a prime vaccine using BCG or rBCG,
470 | JUNE 2006 | VOLUME 4
followed by a heterologous booster vaccine that contains antigens shared with the prime. Immunological correlates of protection. Rapid vaccine development and improvement requires reliable correlates of vaccineinduced human protection. In mouse TB challenge studies, IFN-γ-secreting CD4+ T cells are important mediators of protection19,20, and attempts to induce TB-antigenspecific IFN-γ-secreting CD4+ T cells has been the dominant theme of most TB vaccine research.
www.nature.com/reviews/micro © 2006 Nature Publishing Group
PERSPECTIVES Table 1 | Summary of new-generation tuberculosis vaccine work in progress
Vaccine
Description
Development stage
Researchers/sponsors
Live mycobacterial vaccines: modified BCG rBCG30
BCG Tice engineered to overexpress Ag85B
Phase I trial in US, completed in 2004; no M. Horwitz, UCLA; D. Hoft, St Louis serious adverse events University, US; T. Littlejohn, Winston, Salem, North Carolina, US Sponsor: Sequella, Aeras*
rBCG-Aeras 403
BCG Danish with endosome escape and overexpression of several proteins including Ag85A, Ag85B and TB10.4
Ongoing pre-clinical studies and GMP production; Phase I clinical trial planned for 2006
R. Sun, D. Hone, M. Stone, J. Sadoff , Aeras*
rBCGΔUre:CHly+
BCG Pasteur with endosome escape
Ongoing pre-clinical studies and GMP production; Phase I clinical trial planned for 2006
S. Kaufmann, Max Planck Institute for Infection Biology, Berlin, Germany
BCG::RDI
BCG Pasteur with reintroduction of RD-1 locus which contains protective Ags
Ongoing pre-clinical studies
S. Cole, Institute Pasteur, Paris, France
Pro-apoptotic BCG
BCG Tice with diminished superoxide dismutase activity
Ongoing pre-clinical studies
D. Kernodle, Vanderbilt, US Sponsor: NIH, VA, Aeras*
Live mycobacterial vaccines: modified Mycobacterium tuberculosis M. tuberculosis PhoP
Deletion of virulence associated gene phoP from Mtb MT103 strain
Ongoing pre-clinical studies and GLP production
B. Martin, University of Zaragoza, Spain; B. Gicquel, Institute Pasteur, Paris, France
M. tuberculosis mc26030
Mtb H37Rv with deletion of panCD and RD-1 locus
Ongoing pre-clinical studies and GMP production; Phase I clinical testing planned for 2006
W. Jacobs, Albert Einstein College of Medicine, New York Sponsor: NIH
M. tuberculosis mc26020
Mtb H37Rv with deletion of the lysA and the panCD locus
Ongoing pre-clinical studies and GMP production; Phase I clinical testing planned for 2006
W. Jacobs, Albert Einstein College of Medicine Sponsor: NIH
Mtb72F
Recombinant fusion protein (Mtb39 and Mtb32) in AS02A and AS01B adjuvants
Phase I trials completed in US and Europe; no serious adverse events; additional trials ongoing in Europe
Y. Skeiky, S. Reed, Corixa Corp. Seattle, US Sponsor: Glaxo Smith Kline, Aeras*
Hybrid-1 (85B–ESAT6)
Recombinant fusion of Ag85B–ESAT-6 in IC31 adjuvant
Phase I clinical trials commenced in 2005
P. Andersen, SSI, Denmark Sponsor: SSI, TBVAC
HyVac-4 (Ag85B–TB10.4)
Recombinant fusion of Ag85B–TB10.4 in IC31 adjuvant
Ongoing pre-clinical studies and GLP production; Phase I clinical trial planned for 2007
P. Andersen, SSI, Denmark Sponsor: SSI, Aeras*
Heatshock protein
Nascent BCG proteins associated with purified heat-shock proteins
Ongoing pre-clinical studies
C. Colaco, ImmunoBiology, UK Sponsor: ImmunoBiology, Aeras*
Subunit vaccines
Naked DNA and viral-vectored vaccines Hsp65 (GroEL) DNA
Conserved antigen from Mycobacterium leprae for immunotherapy
Ongoing pre-clinical studies
D. Lowrie, National Institute for Medical Research, London UK. Sponsor: Sequella Inc.
MVA85A
Attenuated strain of vaccinia (modified vaccinia virus) expressing Ag85A
Completed and ongoing Phase I clinical trials; immunogenic, no serious adverse events reported
A. Hill, H. McShane, Oxford University, UK Sponsor: Wellcome Trust
Aeras 402 (Ad35.TB-S)
Non-replicating Ad35 expressing multiple TB proteins including Ag85A, Ag85B and TB10.4
Ongoing pre-clinical studies; GMP production for Phase I clinical trials planned for 2006
Aeras*, Crucell Sponsor: Bill and Melinda Gates Foundation
Ongoing pre-clinical studies
J. Fulkerson, D. Hone, M. Stone, D. Onyabe, J. Sadoff Sponsor: Aeras*
Double-stranded RNA capsids Double-stranded RNA capsids
Double-stranded RNA capsids encoding TB antigens for oral delivery
*Supported by the Bill and Melinda Gates Foundation. Phase I clinical trial: trial that assesses the safety and immunogenicity of a drug or a vaccine in healthy volunteers. BCG, bacille Calmette–Guérin; GLP, good laboratory practice; GMP, good manufacturing practice; NIH, National Institutes of Health; SSI, Statens Serum Institute; TB, tuberculosis; VA, Medical Research Service, Department of Veterans Affairs.
MHC class I restricted, IFN-γ-secreting CD8+ T cells are also important in TB challenge models19,21 and experiments in knockout mice have reported that vaccines can induce significant protection in the
absence of CD4+ T cells22. CD8+ T cells might also be important in the control of LTBI23. The potential role of both vaccineinduced CD4+ T cells and CD8+ T cells is
NATURE REVIEWS | MICROBIOLOGY
not surprising as the granuloma is composed of CD4+ T cells with an outer ring of CD8+ T cells24,25. Potent new TB vaccines will probably need to induce a balanced antigen-specific CD4+ and CD8+ T-cell
VOLUME 4 | JUNE 2006 | 471 © 2006 Nature Publishing Group
PERSPECTIVES response, but the detailed phenotypes and numbers of such cells required for protection is unknown. Furthermore, antibody production against mycobacterial cell wall components might also be important26. At present, there are no proven human immunological correlates of vaccine-induced protection. The detailed phenotypic properties and levels of vaccine-induced T cells in protected versus non-protected volunteers in Phase IIb and Phase III efficacy studies are needed to generate such correlates. In an ongoing trial conducted by the South African TB Vaccine Initiative (SATVI) group (G. Hussey, T. Hawkridge, L. Geiter and W. Hanekom), cells from over 7,000 BCG-vaccinated 10-week-old infants have been collected and banked27,105 for the detailed analysis of vaccine-induced T cells. One correlate of protective immunity that could be explored using this approach is the activation of antigen-specific memory CD4+ and CD8+ T cells that secrete IFN-γ, IL-2 and TNF-α28. Combinations of these and other phenotypic markers in individual T cells — such as CD69, MIP1β, CD45RA, CD27, memory markers such as CD127 (IL-7 receptor)28,29, cytotoxic markers such as FAS ligand, CD107a/b and CD3/4/8 phenotype28 — show potential as correlates of immune protection. γδ T cells that recognize phospholigands, and CD1-restricted T cells that recognize the abundant glycolipids of the mycobacterial cell wall through CD1 molecules (reviewed in REFS 12–14), might also require examination for their potential as immune correlates. Strategies to improve BCG The current BCG vaccine. BCG, the attenuated strain of Mycobacterium bovis, was derived from the parental strain by passage on potato-derived media. Almost three billion doses have been used in various forms since 1921. Approximately 115 million doses are distributed each year30,31 providing almost 80% coverage of infants worldwide32. Neonatal vaccination with BCG has been reported as effective in reducing the incidence of childhood TB in endemic areas33 and in reducing the rate of complications from disseminated TB. However, the ongoing SATVI trial of over 11,000 newborns vaccinated with BCG showed an incidence of culture or smear-proven TB of around 3% by the age of 3 years27,105. The protective effect of BCG might wane over time resulting in a variable outcome that is insufficient to control TB in most endemic areas34. Although several older
Antigen overexpression. Antigen 85B (Ag85B) is a secretory and immunogenic protein of M. tuberculosis and BCG with mycolyl-transferase activity, which is required for mycobacterial cell wall synthesis. M. Horwitz’s group (UCLA) have produced a TB vaccine candidate, rBCG30 that overexpresses Ag85B and that induces increased protection compared with its parental BCG strain in animal challenge studies40,41. There were no significant safety issues in a Phase I clinical trial of rBCG30 in over 30 healthy adult volunteers. Increased immunogenicity compared with the parent BCG strain was reported42, although the study did not meet its predefined immunogenicity endpoints.
apoptosis-mediated CD8+ T-cell responses. Listeria monocytogenes Hly (listeriolysin) is a sulphydryl-activated cytolysin that forms pores in the membrane of the early phagosome. S. Kaufmann’s group (Max Planck Institute) have produced an rBCG vaccine (rBCGΔUreC:Hly+) that secretes Hly, which punches holes in the endosome and allows some leakage of BCG from the phagosome into the cytoplasm of infected host cells43, also resulting in apoptosis of these cells (FIG. 1). Because the activity of Hly is optimal in an acidic environment, the BCG urease gene (ureC), which enables mycobacteria to block the acidification of the early macrophage phagosome, was also deleted. Enhanced presence of BCG in the cytoplasm increases MHC class I presentation of BCG antigens to CD8+ T cells, thereby increasing the CD8+ T-cell immune response. Furthermore, translocation of BCG into the cytoplasm triggers apoptosis, which in turn kills the bacilli and releases antigens, a powerful signal for induction of cellular immunity44. The antigens released as apoptotic blebs (vesicles that carry mycobacterial antigens) are subsequently taken up and processed by dendritic cells, which activate T cells by a mechanism referred to as ‘cross-priming’ (reviewed in REF. 45). In preclinical mouse-protection studies, the rBCGΔUreC:Hly+ vaccine induced better efficacy than the BCG strain and was safer in immunocompromised SCID mice46. Also, unlike the parental BCG strain the ΔUreC:Hly+ rBCG was shown to protect significantly against the virulent M. tuberculosis strain, Beijing/W. This vaccine is scheduled for Phase I trials in early 2006 (REFS 47,48). R. Sun, M. Stone and D. Hone’s group at Aeras have constructed rBCG strains that escape from the endosome using the pH-independent perfringolysin (rBCG–pfo, (FIG. 1)), instead of Hly49, with overexpression of several immunodominant antigens including Ag85A, Ag85B and TB10.4. This strain combines the increased protection seen with rBCG30 with the increased protection and safety of the endosome-escape mutants. This vaccine is scheduled for Phase I trials in 2006 (REF. 27).
Endosome escape. Because protective immunity against M. tuberculosis infection seems to involve both CD4+ and CD8+ T cells, antigens must be presented by both MHC class I and II molecules. BCG that resides in the phagosome gains access to MHC class II molecules, thereby primarily activating CD4+ T cells. To be presented by MHC class I molecules and to activate CD8+ T cells, BCG must either escape from the endosome and gain access to the cytoplasm of the infected cells or induce
Reducing BCG interference with immunity. Usually, host superoxides produced by macrophages kill M. tuberculosis and can amplify host T-cell-recruiting signals. D. Kernodle’s group (Vanderbilt) has shown that secretion of superoxide dismutase (SOD) and other compounds that inactivate superoxides by M. tuberculosis promotes granuloma formation in which M. tuberculosis is secluded, and therefore contributes to inhibition of host defences and persistence of the organism50.
studies demonstrated efficacy as high as 80% in children, adolescents and young adults, with effects lasting up to 60 years in a study of native Americans35–37, and a recent study from Turkey claims that vaccination with two doses of BCG confers some protection in children against infection and disease38, a large-scale study from India over a wide age range did not show any overall efficacy of two different BCG vaccines compared with placebo39. The reasons for the varying efficacy of BCG in protection against pulmonary TB are not completely understood. Potential explanations that have been suggested (reviewed in REFS 30,32) include: interference with the immune response to BCG caused by previous exposure to environmental mycobacteria; differences among BCG vaccine sub-strains; and phenotypic changes in the vaccine during passage from the original cultures and during the manufacturing process; the deletion of protective antigens from BCG; failure of BCG to stimulate adequate, balanced anti-mycobacterial CD4+ and CD8+ T-cell responses; variability in dose, route of administration, age of administration and genetic differences among vaccinees; and lyophilization of the vaccine. Several strategies are being used to improve the efficacy of BCG and are described below.
472 | JUNE 2006 | VOLUME 4
www.nature.com/reviews/micro © 2006 Nature Publishing Group
PERSPECTIVES Attenuated M. tuberculosis and rBCG with reduced secretion of functional SOD are more potent vaccines than their parental strains51. M. tuberculosis and BCG, similar to other successful human pathogens, also produce various other gene products that inhibit the human immune system. W. Jacobs (Albert Einstein College of Medicine) recently suggested that promoting a pro-apoptotic vaccine by specific deletion of genes such as nlaA (Rv3238c), which inhibit apoptosis of infected cells, might also lead to a superior vaccine52. Cells infected with rBCG or recombinant M. tuberculosis vaccines that cannot inhibit caspase-8-mediated apoptosis that usually accompanies infection consequently undergo apoptosis. This leads to cross priming and powerful cellular immune responses against these live vaccines, as caspase-8-mediated apoptosis is one of the most powerful known inducers of cellular immunity53,54. Reintroduction of deleted genes. S. Cole’s group (Institute Pasteur) have reintroduced selected genes into BCG that were deleted during BCG attenuation55. Examples include the RD1 locus56 (see TubercuList in Online links box), which encodes the immunodominant and protective ESAT-6 and CFP-10 proteins — reintroduction of this locus enhanced the protective efficacy of the recombinant BCG::RD1-2F9 vaccine compared with BCG57. However, this approach, although increasing the potency of BCG, might also increase its virulence58 and pose problems for human use. New attenuated live TB vaccines An alternative to BCG modification is to use molecularly attenuated M. tuberculosis that contains over 120 M. tuberculosis genes not found in BCG59–61. A desirable feature of effective and safe live M. tuberculosis vaccine strains is their ability to persist in the host for a limited duration, greater protective efficacy, as well as an improved safety profile compared with the currently used BCG vaccine. The obstacles facing live TB vaccines with respect to safety concerns, public perception and regulatory hurdles were discussed at a recent meeting, and the danger of reversion of live TB vaccines to a virulent state was considered a crucial issue62. At the minimum, two non-reverting independent mutations as well as proven safety in immunocompromised SCID mice are recommended for attenuated M. tuberculosis vaccines. Also, the potential release of antibiotic-resistance gene markers from live vaccines into the environment, with subsequent transfer to other organisms, is considered an important concern.
Inactivation of PhoP, which regulates expression of many M. tuberculosis proteins, by C. Soto’s group (University of Zaragoza, Spain)63 attenuated M. tuberculosis virulence while maintaining the pattern of immune responses associated with the parental M. tuberculosis strain64. This mutant has shown promising early animal protection but will probably require a second, independent, nonreverting mutation to satisfy human safety concerns62. Auxotrophic mutant vaccine candidates (requiring the addition of a nutrient(s) for survival) have been produced by transposon mutagenesis. These mutants maintain their infective ability, have a limited duration of replication in the host65,66 and protect against virulent TB challenge in the guinea pig model67. The extent of the attenuation and safety of these vaccines remains to be defined in immunocompromised SCID mice. The Jacobs and Bloom groups have targeted the deletion of proteins involved in M. tuberculosis lipid metabolism by disrupting both the panC and panD genes (which are involved in pantothenic acid synthesis). Further deletion of the lysA gene produced vaccine mc26020 (panC panD lysA), which does not replicate in vivo. Deletion of the RD1 region (implicated in the attenuation of BCG) from the panC, panD mutant produced vaccine mc26030 (panC panD ∆RD1)68–70. Both of these highly attenuated mutants were safer than BCG in immunocompromised animals and provided similar levels of protection in mouse challenge studies. The mc26030 is scheduled for Phase I trials in 2006 (REF. 52). Subunit vaccine strategies Vaccine antigen identification. Antigens of M. tuberculosis that activate T cells in previously infected humans and animals are considered good vaccine candidates. Biochemical approaches initially identified several of the most abundant proteins (Ag85 complex, MPT32, PhoS, DnaK, GroES, MPT46, MPT53, MPT63 and the 19-kDa lipoprotein) secreted by M. tuberculosis71–73. Subsequently, antigens in the low molecular mass range of 6–10 kDa (ESAT-6 family) and 26–34 kDa (Ag85B and MPT59)74,75 were defined. T-cell and serological expression cloning approaches76–79 identified several important dominant immunogenic antigens including Mtb32 (a secreted serine protease) and Mtb39 (a member of the PPE (proline-prolineglumatic acid) family of proteins presumably localized on the cell wall)80,81. Proteomic approaches82,83 have also identified several promising antigens such as Rv3407 (REF. 84).
NATURE REVIEWS | MICROBIOLOGY
Antigen vaccines have advanced towards clinical development using animal protection models and assessment of their potency in stimulating immune cells derived from peripheral blood mononuclear cells of healthy, previously infected donors to produce IFN-γ. Antigen vaccine candidates therfore include Ag85A and Ag85B (REFS 85,86), ESAT-6, TB10.4 and Mtb9.9 (REFS 75,79,87) and the PPE family of proteins, Mtb39a–e and Mtb41 (REFS 76,78,81). Recombinant fusion proteins. Y. Skeiky, S. Reed and P. Andersen’s groups have used a genetic fusion approach to generate hybrid protein vaccines consisting of Mtb39 and Mtb32 (Mtb72F)88, or Ag85B and ESAT-6 (Hybrid-1)89, respectively. These candidate vaccines were selected on the basis of their strong recognition by immune cells in infected individuals and experimentally vaccinated or infected animals. The generation of recombinant protein subunit vaccines comprising multiple open reading frames is both cost effective and simple, and the fusion vaccines can be more immunogenic than the individual components. Both fusion recombinant vaccine constructs (Mtb72F and Hybrid-1) in selected adjuvants protect as well as BCG in the mouse and guinea pig TB challenge models88,90–92. Multiple Phase I clinical trials of Mtb72F have been completed in the United States and Europe (results pending) and additional trials are being conducted in Europe. Phase I trials of Hybrid-I started in Europe in November 2005. ESAT-6 (the component of Hybrid-I that is absent from all BCG, but present in M. tuberculosis strains) is the basis of a new generation of diagnostic tests for M. tuberculosis infection. Therefore, if used as a vaccine, Hybrid-I could potentially compromise the diagnostic utility of ESAT-6 in distinguishing between infection and vaccination93,94. Therefore, P. Andersen’s group in collaboration with Aeras has developed a new fusion construct, HyVac-4, which consists of Ag85B–TB10.4, and which avoids diagnostic interference and should also boost BCG-primed individuals because TB10.4 is immunogenic in BCG-vaccinated subjects87,95. Prime–boost vaccine strategies. The attraction of the heterologous prime–boost approach is that it preferentially expands TB-specific pre-existing memory T cells against antigenic epitopes shared by both the prime and booster vaccines. This approach also minimizes the potential of
VOLUME 4 | JUNE 2006 | 473 © 2006 Nature Publishing Group
PERSPECTIVES generating antibody and T-cell neutralizing effects that would impact on the ‘take’ and effectiveness of a non-replicating viralvector boost (FIG. 2). An early study using a BCG prime–boost approach showed that mice boosted with an adjuvanted Ag85A protein when they had lost their BCG-induced capacity to resist an aerosol TB challenge were protected better than non-BCG-primed animals given adjuvanted protein96. This suggests that adjuvanted proteins including Mtb72F, Hybrid-1 and HyVac-4 could induce superior protection when given as boosters in primary regimens, or in adolescence. Such recombinant protein regimens, however, might induce primarily antigen-specific CD4+ T cells instead of the desired balanced CD4+ and CD8+ T-cell responses. A recent study in mice of a BCG prime followed by a booster with an M. tuberculosis DNA vaccine encoding the gene product Rv3407 (a 10-kDa protein of unknown function) also showed superior protection compared with BCG alone84. Protection that was superior to BCG was also demonstrated in guinea pigs immunized with a BCG prime–Mtb72F DNA boost regimen followed by aerosol challenge with virulent M. tuberculosis. Remarkably, histological examination of the lungs of these guinea pigs revealed minimal granulomatous lesions with evidence of lesion healing and airway remodelling at a time when animals immunized with BCG and then challenged were dying90. Whereas, so far, the success of human DNA vaccination in general has been disappointing, non-replicating recombinant poxviruses and adenoviruses are efficient at boosting previously primed T-cell responses and show great potential for use in heterologous prime–boost immunization strategies. Priming with BCG and boosting with a recombinant Modified Vaccinia Ankara virus (MVA) that expresses Ag85A induced increased levels of CD4+ and CD8+ T cells and enhanced protection in animals compared with a vaccination regimen using either BCG or MVA85A alone97. In a Phase I clinical trial recently reported by H. McShane and A. Hill’s group (Oxford University, UK), immunization of volunteers from 6 months to 38 years after BCG vaccination with MVA resulted in persistent, high levels of IFN-γ secreting CD4+ T cells, which were substantially higher than seen following vaccination with BCG or MVA85A alone, which only induced transient responses106. Adenoviral vectors are potent vaccinedelivery vehicles. A recent study showed that vaccination of mice by the intranasal
a Newborn
b 14 weeks
c 24 weeks and adolescence Recombinant protein with adjuvant
rBCG or BCG
Recombinant protein with adjuvant
Dendritic cell Viral vector
Nucleocaspid in bacteria
CD4+ T cell
CD8+ T cell
Expansion of Ag-specific T cells
Memory T cells
Amplification of Ag-specific T cells
Figure 2 | Prime-boost vaccination strategies. Boosting or augmenting BCG or a recombinant, modified BCG (rBCG), rather than trying to completely replace it, is a practical approach for future immunization against TB. Heterologous prime–boost regimens begin with vaccination of newborns using BCG or modified rBCG as the prime (a). This is followed by a booster vaccination in infancy (b) and later in childhood and adolescence (c) using different antigen-delivery systems (heterologous boosting) composed of selected antigens in common with the prime. The booster can include viral vectors such as recombinant adenovirus or poxvirus engineered to express immunogenic and protective mycobacterial candidate antigens (for example, Ag85A, Ag85B and TB10.4) or recombinant proteins (for example, HyVac-4, Hybrid-1 and Mtb72F) delivered with adjuvants (such as AS02A, AS01B, IC31 and LipoVac). Priming with BCG or rBCG followed by a booster with adenovirus and a subsequent boost with recombinant protein in adjuvant formulation is also being contemplated. Alternatively, Shigella spp. nucleocapsid vectors offer a possible alternative oral vaccine boost (b). Shigella vectors deliver double stranded RNA capsids into the cytoplasm where they produce mRNA that directs cellular production of M. tuberculosis antigens. The heterologous prime–boost approach preferentially expands BCG or TB-specific pre-existing memory T cells against antigenic epitopes shared by both the prime and booster vaccines. Administration of the booster vaccination in an alternative vector minimizes the generation of anti-vector immunity that is typically seen with repeated administration using the same vaccine (homologous boosting), thereby inducing a powerful synergistic immune response. Unlike BCG (which induces a predominant CD4+ T-cell response), rBCG that has access to the cytoplasm activates a broad repertoire of antigen-specific CD4+ and CD8+ T cells (a). Subsequent boosting results in the expansion of selected antigen-specific memory CD4+ and CD8+ T cells and the selective enrichment of high avidity T cells to combat exposure to the pathogen (b,c).
mucosal route with a recombinant adenoviral (Ad5 based) construct expressing the M. tuberculosis antigen Ag85A (AdAg85A) conferred protection (mediated by both CD4+ and CD8+ T cells) that was superior to that seen with BCG vaccination against a pulmonary M. tuberculosis challenge98. Aeras, in collaboration with Crucell, has constructed non-replicating Ad35 vectors that are less prevalent in the environment than Ad5 (REFS 99 and 100), and that express several
474 | JUNE 2006 | VOLUME 4
M. tuberculosis proteins including Ag85A, Ag85B and TB10.4 (REFS 27,100). Vaccination of mice with a single dose of the rAd35–TB construct conferred protection against M. tuberculosis challenge107. BCG prime followed by a booster with a single dose of these adenovirus recombinant vaccines in mice have yielded significantly increased antigenspecific CD4+ and CD8+ T-cell responses compared with adenovectored TB vaccines or BCG administered alone. Aeras has
www.nature.com/reviews/micro © 2006 Nature Publishing Group
PERSPECTIVES constructed rBCG vaccines with over 50% escape from the endosome that overexpress these same antigens to function as a prime for this adenovector boost27 (R. Sun et al., unpublished observations). J. Fulkerson, D. Hone, D. Onyabe and M. Stone at Aeras have also constructed double stranded RNA capsids which can be orally delivered using engineered Shigella spp. vectors. Shigella delivers these capsids into the cytoplasm where they produce mRNA that directs cellular production of M. tuberculosis antigens101 (J. Fulkerson et al., unpublished observations) and induces cellular immunity. Such an oral vaccine boost to an rBCG prime might have the low cost and ease of deliverability desirable in a TB vaccine for the developing world. Conclusions A safe and effective heterologous prime– boost regimen, which boosts or augments BCG or rBCG, is perhaps the most realistic strategy for future TB control through immunization. Ideally the rBCG prime in such a regimen would include the overexpression of important antigens from different stages of the pathogen’s life cycle and would induce cross-priming and increased CD4+ and CD8+ T-cell responses, as well as increased safety in immunocompromised individuals. The ideal booster would comprise recombinant proteins delivered with adjuvants, or delivered by non-replicating viral or capsid vectors. Although there is a possibility that the current generation of new vaccines might protect against reactivation from the latent state, these vaccines are not generally focused on the ‘latency’ dosR regulated antigens recently identified by microarray and other techniques102,103. Furthermore, animal models of latency for preclinical vaccine evaluation are just being developed104. The best hope for these vaccines probably lies with overexpression of dosR regulated proteins by rBCG followed by a multivalent booster vaccine. Based on current clinical development timelines, first generation vaccine regimens for protection against infection and disease could be licensed and available in the next 7–9 years with vaccines against latency and reactivation 2–3 years later. Because of its great promise and practical usefulness, rBCG prime– heterologous boost strategies for malaria and HIV are also currently being explored. Yasir A. W. Skeiky and Jerald C. Sadoff are at the Aeras Global TB Vaccine Foundation, 1405 Research Blvd, Rockville, Maryland 20850, USA. e-mails:
[email protected];
[email protected] doi10.1038/nrmicro1419
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14. 15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
Kochi, A. The global tuberculosis situation and the new control strategy of the World Health Organization 1991. Bull. World Health Organ.79, 71–75 (2001). Hopewell, P. C. Impact of human immunodeficiency virus infection on the epidemiology, clinical features, management and control of tuberculosis. Clin. Infect. Dis.15, 540–547 (1992). Raviglione, M. C. et al. Tuberculosis trends in eastern Europe and the former USSR. Tuber. Lung Dis. 75, 400–416 (1994). Maher, D. et al. Tuberculosis deaths in countries with high HIV prevalence: what is their use as an indicator in tuberculosis programme monitoring and epidemiological surveillance? Int. J. Tuberc. Lung Dis. 9, 123–127 (2005). Bloom, B. R. & Fine, P. E. In Tuberculosis: Pathogenesis, Protection, and Control (ed. Bloom, B. R.) 531–557 (American Society for Microbiology, Washington DC, 1994). Summary from the 6th Annual Conference of the International Union Against Tuberculosis and Lung Disease. (Chicago, 2001). Ormerod, L. P., Horsfield, N. & Green, R. M. Tuberculosis treatment outcome monitoring: Blackburn 1988–2000. Int. J. Tuberc. Lung Dis. 6, 662–665 (2002). Sudre, P., ten Dam, G. & Kochi, A. Tuberculosis: a global overview of the situation today. Bull. World Health Organ.70, 149–159 (1992). Dolin, P. J., Raviglione, M. C. & Kochi, A. Global tuberculosis incidence and mortality during 1990– 2000. Bull. World Health Organ. 72, 213–220 (1994). Hopewell, P. et al. Evaluation of new anti-infective drugs for the treatment and prevention of infections caused by the Mycobacterium avium complex. Infectious Diseases Society of America and the Food and Drug Administration. Clin. Infect. Dis. 15 (Suppl. 1), 296–306 (1992). Smith, P. G. & Moss, A. R. In Tuberculosis: Pathogenesis, Protection, and Control. (ed Bloom, B. R.) 47–59 (American Society for Microbiology, Washington DC, 1994). Kaufmann, S. H. & McMichael, A. J. Annulling a dangerous liaison: vaccination strategies against AIDS and tuberculosis. Nature Med.11 (Suppl. 4), 33–44 (2005). Lewinsohn, D. M., Lewinsohn, D. A. & Grotzke, J. E. TB vaccines at the turn of the century: insights into immunity to M. tuberculosis and modern approaches for prevention of an ancient disease. Semin. Respir. Infect. 18, 320–338 (2003). Flynn, J. L. & Chan, J. Immunology of tuberculosis. Annu. Rev. Immunol. 19, 93–129 (2001). Jouanguy, E. et al. Interferon- γ-receptor deficiency in an infant with fatal bacille Calmette-Guérin infection. N. Engl. J. Med. 335, 1956–1961 (1996). Jouanguy, E. et al. A human IFNGR1 small deletion hotspot associated with dominant susceptibility to mycobacterial infection. Nature Genet. 21, 370–378 (1999). Toussirot, E. & Wendling, D. The use of TNF-α blocking agents in rheumatoid arthritis: an overview. Expert Opin. Pharmacother. 5, 581–594 (2004). Havlir, D. V. & Barnes, P. F. Tuberculosis in patients with human immunodeficiency virus infection. N. Engl. J. Med. 340, 367–373 (1999). Cooper, A. M. et al. Disseminated tuberculosis in IFN-γ gene-disrupted mice. J. Exp. Med.178, 2243–2247 (1993). Flynn, J. L. et al. An essential role for IFN-γ in resistance to Mycobacterium tuberculosis infection. J. Exp. Med. 178, 2249–2254 (1993). Flynn, J. L. et al. Major histocompatibility complex class I-restricted T cells are required for resistance to Mycobacterium tuberculosis infection. Proc. Natl Acad. Sci. USA 89, 12013–12017 (1992). Derrick, S. C. et al. Immunization with a DNA vaccine cocktail protects mice lacking CD4 cells against an aerogenic infection with Mycobacterium tuberculosis. Infect. Immun. 72, 1685–1692 (2004). van Pinxteren, L. A. et al. Control of latent Mycobacterium tuberculosis infection is dependent on CD8 T cells. Eur. J. Immunol. 30, 3689–3698 (2000). Gonzalez-Juarrero, M. et al. Temporal and spatial arrangement of lymphocytes within lung granulomas induced by aerosol infection with Mycobacterium tuberculosis. Infect. Immun. 69, 1722–1728 (2001).
NATURE REVIEWS | MICROBIOLOGY
25. Ulrichs, T. et al. Human tuberculous granulomas induce peripheral lymphoid follicle-like structures to orchestrate local host defence in the lung. J. Pathol. 204, 217–228 (2004). 26. Teitelbaum, R. et al. A mAb recognizing a surface antigen of Mycobacterium tuberculosis enhances host survival. Proc. Natl Acad. Sci. USA 95, 15688–15693 (1998). 27. Sadoff, J. Public private partnership approach to vaccine development. Presented at New Approaches to Vaccine Development. (Sept. 8–10, Berlin, Germany, 2005). 28. Bouneaud, C. et al. Lineage relationships, homeostasis, and recall capacities of central- and effector-memory CD8 T cells in vivo. J. Exp. Med. 201, 579–590 (2005). 29. Seder, R. A. & Ahmed, R. Similarities and differences in CD4+ and CD8+ effector and memory T cell generation. Nature Immunol. 4, 835–842 (2003). 30. Brewer, T. F. & Colditz, G. A. Relationship between bacille Calmette-Guérin (BCG) strains and the efficacy of BCG vaccine in the prevention of tuberculosis. Clin. Infect. Dis. 20, 126–135 (1995). 31. Colditz, G. A. et al. Efficacy of BCG vaccine in the prevention of tuberculosis. Meta-analysis of the published literature. JAMA 271, 698–702 (1994). 32. Fine, P. E. M., Carneiro, I. A. M., Milstien, J. B. & Clements, C. J. Issues relating to the use of BCG in immunization programmes (World Health Organization, Geneva, 1999). 33. Colditz, G. A. et al. The efficacy of bacillus CalmetteGuérin vaccination of newborns and infants in the prevention of tuberculosis: meta-analyses of the published literature. Pediatrics 96, 29–35 (1995). 34. Sterne, J. A., Rodrigues, L. C. & Guedes, I. N. Does the efficacy of BCG decline with time since vaccination? Int. J. Tuberc. Lung Dis. 2, 200–207 (1998). 35. Aronson, N. E. et al. Long-term efficacy of BCG vaccine in American Indians and Alaska Natives: A 60-year follow-up study. JAMA 291, 2086–2091 (2004). 36. Fine, P. E. The BCG story: lessons from the past and implications for the future. Rev. Infect. Dis. 11 (Suppl. 2), 353–359 (1989). 37. Comstock, G. W. Field trials of tuberculosis vaccines: how could we have done them better? Control Clin. Trials 15, 247–276 (1994). 38. Soysal, A. et al. Effect of BCG vaccination on risk of Mycobacterium tuberculosis infection in children with household tuberculosis contact: a prospective community-based study. Lancet 366, 1443–1451 (2005). 39. Tuberculosis Research Centre (ICMR), Chennai. Fifteen year follow up of trial of BCG vaccines in south India for tuberculosis prevention. Indian J. Med. Res. 110, 56–69 (1999). 40. Horwitz, M. A. & Harth, G. A new vaccine against tuberculosis affords greater survival after challenge than the current vaccine in the guinea pig model of pulmonary tuberculosis. Infect. Immun.71, 1672–1679 (2003). 41. Horwitz, M. A. et al. Recombinant bacillus CalmetteGuérin (BCG) vaccines expressing the Mycobacterium tuberculosis 30-kDa major secretory protein induce greater protective immunity against tuberculosis than conventional BCG vaccines in a highly susceptible animal model. Proc. Natl Acad. Sci. USA 97, 13853–13858 (2000). 42. Hoft, D. F. Results of the 1st Phase I trial of a recombinant BCG TB vaccine (rBCG30). Presented at US–Japan Cooperative Medical Science Program; 40th Tuberculosis and Leprosy Research Conference. (Seattle, Washington, 2005). 43. Hess, J. et al. Mycobacterium bovis bacille CalmetteGuérin strains secreting listeriolysin of Listeria monocytogenes. Proc. Natl Acad. Sci. USA 95, 5299–5304 (1998). 44. Albert, M. L., Sauter, B. & Bhardwaj, N. Dendritic cells acquire antigen from apoptotic cells and induce class I-restricted CTLs. Nature 392, 86–89 (1998). 45. Nasser Eddine, A. & Kaufmann, S. H. E. Improved protection by recombinant BCG. Microbes Infect. 7, 939–946 (2005). 46. Grode, L. et al. Increased vaccine efficacy against tuberculosis of recombinant Mycobacterium bovis bacille Calmette-Guérin mutants that secrete listeriolysin. J. Clin. Invest.115, 2472–2479 (2005). 47. Kaufmann, S. H. E. Rational vaccine design against tuberculosis. Presented at New Approaches to Vaccine Development. (Sept. 8–10, Berlin, Germany, 2005).
VOLUME 4 | JUNE 2006 | 475 © 2006 Nature Publishing Group
PERSPECTIVES 48. Laufer, A. Managing the transition from academic vaccine research to license oriented development. Presented at New Approaches to Vaccine Development. (Sept. 8–10, Berlin, Germany, 2005). 49. Portnoy, D. A. et al. Capacity of listeriolysin O, streptolysin O, and perfringolysin O to mediate growth of Bacillus subtilis within mammalian cells. Infect. Immun. 60, 2710–2717 (1992). 50. Edwards, K. M. et al. Iron-cofactored superoxide dismutase inhibits host responses to Mycobacterium tuberculosis. Am. J. Respir. Crit. Care Med. 164, 2213–2219 (2001). 51. Kernodle, D. S. Pro-apoptotic BCG vaccine that enhances cellular immune responses and reduces tissue damage. Aeras Global TB Vaccine Foundation Scott Thaler Lecture Series. (Bethesda, Maryland, 2004). 52. Jacobs, B. Live attenuated vaccines for TB: Mining the immune evasion function of the tubercle bacillus. Presented at New Approaches to Vaccine Development. (Sept. 8–10,Berlin, Germany, 2005). 53. Chattergoon, M. A. et al. Targeted antigen delivery to antigen-presenting cells including dendritic cells by engineered Fas-mediated apoptosis. Nature Biotechnol. 18, 974–979 (2000). 54. Sheridan, J. P. et al. Control of TRAIL-induced apoptosis by a family of signaling and decoy receptors. Science 277, 818–821 (1997). 55. Brosch, R. et al. Comparative genomics of the mycobacteria. Int. J. Med. Microbiol. 290, 143–152 (2000). 56. Mahairas, G. G. et al. Molecular analysis of genetic differences between Mycobacterium bovis BCG and virulent M. bovis. J. Bacteriol. 178, 1274–1282 (1996). 57. Pym, A. S. et al. Recombinant BCG exporting ESAT-6 confers enhanced protection against tuberculosis. Nature Med. 9, 533–539 (2003). 58. Pym, A. S. et al. Loss of RD1 contributed to the attenuation of the live tuberculosis vaccines Mycobacterium bovis BCG and Mycobacterium microti. Mol. Microbiol. 46, 709–717 (2002). 59. Behr, M. A. et al. Comparative genomics of BCG vaccines by whole-genome DNA microarray. Science 284, 1520–1523 (1999). 60. Gordon, S. V. et al. Identification of variable regions in the genomes of tubercle bacilli using bacterial artificial chromosome arrays. Mol. Microbiol. 32, 643–645 (1999). 61. Salamon, H. et al. Detection of deleted genomic DNA using a semiautomated computational analysis of GeneChip data. Genome Res. 10, 2044–2054 (2000). 62. Kamath, A. T. et al. In New Live Mycobacterial Vaccines: Defining Essential Steps Towards Clinical Development. (WHO, Geneva, Switzerland, 2004). 63. Soto, C. Y. et al. IS6110 mediates increased transcription of the phoP virulence gene in a multidrug-resistant clinical isolate responsible for tuberculosis outbreaks. J. Clin. Microbiol. 42, 212–219 (2004). 64. Perez, E. et al. An essential role for phoP in Mycobacterium tuberculosis virulence. Mol. Microbiol. 41, 179–187 (2001). 65. Guleria, I. et al. Auxotrophic vaccines for tuberculosis. Nature Med. 2, 334–337 (1996). 66. Smith, D. A. et al. Characterization of auxotrophic mutants of Mycobacterium tuberculosis and their potential as vaccine candidates. Infect. Immun. 69, 1142–1150 (2001). 67. Jackson, M. et al. Persistence and protective efficacy of a Mycobacterium tuberculosis auxotroph vaccine. Infect. Immun. 67, 2867–2873 (1999). 68. Sambandamurthy, V. K. et al. Long-term protection against tuberculosis following vaccination with a severely attenuated double lysine and pantothenate auxotroph of Mycobacterium tuberculosis. Infect. Immun. 73, 1196–1203 (2005). 69. Sambandamurthy, V. K. et al. A pantothenate auxotroph of Mycobacterium tuberculosis is highly attenuated and protects mice against tuberculosis. Nature Med. 8, 1171–1174 (2002). 70. Sambandamurthy, V. K. & Jacobs, W. R. Live attenuated mutants of Mycobacterium tuberculosis as candidate vaccines against tuberculosis. Microbes Infect. 7, 955–961(2005). 71. Andersen, A. B. & Hansen, E. B. Structure and mapping of antigenic domains of protein antigen b, a 38,000-molecular-weight protein of Mycobacterium tuberculosis. Infect. Immun. 57, 2481–2488 (1989).
72. Nagai, S. et al. Isolation and partial characterization of major protein antigens in the culture fluid of Mycobacterium tuberculosis. Infect. Immun. 59, 372–382 (1991). 73. Young, D. B. & Garbe, T. R. Lipoprotein antigens of Mycobacterium tuberculosis. Res. Microbiol. 142, 55–65 (1991). 74. Andersen, P. et al. Recall of long-lived immunity to Mycobacterium tuberculosis infection in mice. J. Immunol. 154, 3359–3372 (1995). 75. Skjot, R. L. et al. Comparative evaluation of lowmolecular-mass proteins from Mycobacterium tuberculosis identifies members of the ESAT-6 family as immunodominant T-cell antigens. Infect. Immun. 68, 214–220 (2000). 76. Alderson, M. R. et al. Expression cloning of an immunodominant family of Mycobacterium tuberculosis antigens using human CD4+ T cells. J. Exp. Med. 191, 551–560 (2000). 77. Dillon, D. C. et al. Molecular and immunological characterization of Mycobacterium tuberculosis CFP-10, an immunodiagnostic antigen missing in Mycobacterium bovis BCG. J. Clin. Microbiol. 38, 3285–3290 (2000). 78. Skeiky, Y. A. et al. T-cell expression cloning of a Mycobacterium tuberculosis gene encoding a protective antigen associated with the early control of infection. J. Immunol. 165, 7140–7149 (2000). 79. Laal, S. & Skeiky, Y. A. W. In Tuberculosis and the Tubercle Bacillus. (eds Cole, S. T. et al.) 71–83 (American Society for Microbiology Press, Washington DC, 2005). 80. Skeiky, Y. A. et al. Cloning, expression, and immunological evaluation of two putative secreted serine protease antigens of Mycobacterium tuberculosis. Infect. Immun. 67, 3998–4007 (1999). 81. Dillon, D. C. et al. Molecular characterization and human T-cell responses to a member of a novel Mycobacterium tuberculosis mtb39 gene family. Infect. Immun. 67, 2941–2950 (1999). 82. Covert, B. A. et al. The application of proteomics in defining the T-cell antigens of Mycobacterium tuberculosis. Proteomics 1, 574–586 (2001). 83. Mattow, J. et al. Comparative proteome analysis of culture supernatant proteins from virulent Mycobacterium tuberculosis H37Rv and attenuated M. bovis BCG Copenhagen. Electrophoresis 24, 3405–3420 (2003). 84. Mollenkopf, H. J. et al. Application of mycobacterial proteomics to vaccine design: improved protection by Mycobacterium bovis BCG prime–Rv3407 DNA boost vaccination against tuberculosis. Infect. Immun. 72, 6471–6479 (2004). 85. Huygen, K. et al. Immunogenicity and protective efficacy of a tuberculosis DNA vaccine. Nature Med. 2, 893–898 (1996). 86. Baldwin, S. L. et al. Evaluation of new vaccines in the mouse and guinea pig model of tuberculosis. Infect. Immun. 66, 2951–2959 (1998). 87. Skjot, R. L. et al. Epitope mapping of the immunodominant antigen TB10.4 and the two homologous proteins TB10.3 and TB12.9, which constitute a subfamily of the esat-6 gene family. Infect. Immun. 70, 5446–5453 (2002). 88. Skeiky, Y. A. et al. Differential immune responses and protective efficacy induced by components of a tuberculosis polyprotein vaccine, Mtb72F, delivered as naked DNA or recombinant protein. J. Immunol. 172, 7618–7628 (2004). 89. Weinrich Olsen, A. et al. Protection of mice with a tuberculosis subunit vaccine based on a fusion protein of antigen 85b and ESAT-6. Infect. Immun. 69, 2773–2778 (2001). 90. Brandt, L. et al. The protective effect of the Mycobacterium bovis BCG vaccine is increased by coadministration with the Mycobacterium tuberculosis 72-kilodalton fusion polyprotein Mtb72F in M. tuberculosis-infected guinea pigs. Infect. Immun. 72, 6622–6632 (2004). 91. Olsen, A. W. et al. Efficient protection against Mycobacterium tuberculosis by vaccination with a single subdominant epitope from the ESAT-6 antigen. Eur. J. Immunol. 30, 1724–1732 (2000). 92. Olsen, A. W. et al. Protective effect of a tuberculosis subunit vaccine based on a fusion of antigen 85B and ESAT-6 in the aerosol guinea pig model. Infect. Immun. 72, 6148–6150 (2004). 93. Mazurek, G. H. et al. Comparison of a whole-blood IFN-γ assay with tuberculin skin testing for detecting latent Mycobacterium tuberculosis infection. JAMA 286, 1740–1747 (2001).
476 | JUNE 2006 | VOLUME 4
94. Mazurek, G. H. & Villarino, M. E. Guidelines for using the QuantiFERON-TB test for diagnosing latent Mycobacterium tuberculosis infection. Centers for Disease Control and Prevention. MMWR Recomm. Rep. 52, 15–18 (2003). 95. Dietrich, J. et al. Exchanging ESAT-6 with TB10. 4 in an Ag85B fusion molecule-based tuberculosis subunit vaccine: efficient protection and ESAT-6-based sensitive monitoring of vaccine efficacy. J. Immunol. 174, 6332–6339 (2005). 96. Brooks, J. V. et al. Boosting vaccine for tuberculosis. Infect. Immun. 69, 2714–2717 (2001). 97. Goonetilleke, N. P. et al. Enhanced immunogenicity and protective efficacy against Mycobacterium tuberculosis of bacille CalmetteGuérin vaccine using mucosal administration and boosting with a recombinant modified vaccinia virus Ankara. J. Immunol. 171, 1602–1609 (2003). 98. Wang, J. et al. Single mucosal, but not parenteral, immunization with recombinant adenoviral-based vaccine provides potent protection from pulmonary tuberculosis. J. Immunol. 173, 6357–6365 (2004). 99. Vogels, R. et al. Replication-deficient human adenovirus type 35 vectors for gene transfer and vaccination: efficient human cell infection and bypass of pre-existing adenovirus immunity. J. Virol. 77, 8263–8271 (2003). 100. Havenga et al. Novel replication-incompetent adenoviral B-group vectors. J. Virol. (In the press). 101. Hone, D. Optimization of nucleic acid vaccine delivery by bacterial vectors. Presented at New Approaches to Vaccine Development. (Sept. 8–10, Berlin, Germany, 2005). 102. Schnappinger, D. et al. Transcriptional adaptation of Mycobacterium tuberculosis within macrophages: insights into the phagosomal environment. J. Exp. Med. 198, 693–704 (2003). 103. Voskuil, M. I. et al. Inhibition of respiration by nitric oxide induces a Mycobacterium tuberculosis dormancy program. J. Exp. Med. 198, 705–713 (2003). 104. Capuano, S. V. et al. Experimental Mycobacterium tuberculosis infection of cynomolgus macaques closely resembles the various manifestations of human M. tuberculosis infection. Infect. Immun. 71, 5831–5844 (2003). 105. Hussey, G., Hawkridge, T., Geiter, L. & Hanekom, W. Presented at TB vaccines for the world (April 19–21, Vienna, Austria, 2006). 106. McShane, H. et al. Recombinant MVA85A boosts BCG-primed and naturally acquired antimycobacterial immunity in humans. Nature Med. 10, 1240–1244 (2004). 107. Radosevic, K. et al. Presented at TB vaccines for the world (April 19–21, Vienna, Austria, 2006).
Acknowledgements We are grateful to Dr. L. Barker for his input and critical reading of this article. The Aeras Global TB Vaccine Foundation is supported by a major grant from the Bill and Melinda Gates Foundation.
Competing interests statement The authors declare no competing financial interests.
DATABASES The following terms in this article are linked online to: Entrez Gene: http://www.ncbi.nlm.nih.gov/entrez/query. fcgi?db=gene lysA | panC | panD | Rv3238c | ureC Entrez Genome Project: http://www.ncbi.nlm.nih.gov/ entrez/query.fcgi?db=genomeprj Listeria monocytogenes | Mycobacterium bovis | Mycobacterium tuberculosis UniProtKB: http://ca.expasy.org/sprot Ag85A | Ag85B | DnaK | GroES | Hly | IFN-γ | MPT32 | MPT46 | MPT53 | MPT63 | Mtb9.9 | Mtb32 | Phop | TB10.4
FURTHER INFORMATION Author’s homepage: www.aeras.org CDC HIV/AIDS Prevention: http://www.cdc.gov/hiv/dhap.htm South African TB Vaccine Initiative: http://www.satvi.uct.ac.za TubercuList: http://genolist.pasteur.fr./TubercuList Access to this links box is available online.
www.nature.com/reviews/micro © 2006 Nature Publishing Group
PERSPECTIVES OPINION
The trypanolytic factor of human serum Etienne Pays, Benoit Vanhollebeke, Luc Vanhamme, Françoise Paturiaux-Hanocq, Derek P. Nolan and David Pérez-Morga
Abstract | African trypanosomes (the prototype of which is Trypanosoma brucei brucei) are protozoan parasites that infect a wide range of mammals. Human blood, unlike the blood of other mammals, has efficient trypanolytic activity, and this needs to be counteracted by these parasites. Resistance to this activity has arisen in two subspecies of Trypanosoma brucei — Trypanosoma brucei rhodesiense and Trypanosoma brucei gambiense — allowing these parasites to infect humans, and this results in sleeping sickness in East Africa and West Africa, respectively. Study of the mechanism by which T. b. rhodesiense escapes lysis by human serum led to the identification of an ionic-pore-forming apolipoprotein — known as apolipoprotein L1 — that is associated with high-density-lipoprotein particles in human blood. In this Opinion article, we argue that apolipoprotein L1 is the factor that is responsible for the trypanolytic activity of human serum. Similar to other species of African trypanosome (such as Trypanosoma congolense and Trypanosoma vivax), Trypanosoma brucei brucei, Trypanosoma brucei rhodesiense and Trypanosoma brucei gambiense are protozoan parasites that are transmitted to various mammals by tsetse flies (Glossina palpalis and Glossina morsitans) (FIG. 1). During a blood meal by the tsetse fly, infective trypanosome forms, known as metacyclic forms, that are present in the salivary glands of the fly are inoculated into mammalian hosts, such as humans, cattle, antelopes, buffaloes and lions. This allows the transformation of the parasites into long, slender bloodstream forms, which actively divide and colonize the blood until a quorum-sensing signal triggers their differentiation into short, stumpy forms. These short, stumpy forms do not divide and are competent for transformation into procyclic forms as soon as they are ingested by a tsetse fly. The procyclic forms then proliferate in the mid-gut of the fly and eventually differentiate into metacyclic forms after a complex journey in the insect vector. The metacyclic forms are quiescent and are ready to be transferred back to mammals. Because they reside extracellularly in their mammalian hosts, the bloodstream forms of trypanosomes need to resist all components of host defence. The main protective feature of these parasites is a thick and dense surface coat, which covers the entire plasma membrane. This coat consists
of a monolayer of ∼107 molecules of a single glycoprotein known as variant surface glycoprotein (VSG). The VSG molecules are attached to the plasma membrane by glycosylphosphatidylinositol (GPI) anchors, and they form homodimers1. Close interactions between adjacent VSG homodimers prevent antibodies from reaching the inner antigenic determinants of the coat, and the elongated shape of the VSGs keeps toxins away from the plasma membrane. In addition to these protective functions, VSGs undergo continuous variation, which leads to the periodic change of the VSG loops that are exposed at the surface of the parasite. These loops contain the only antigenic determinants of living intact trypanosomes that are recognized by the immune system. Therefore, such variation is known as antigenic variation, and it allows the parasite to evade antibody-mediated clearance2–5. Only one VSG variant is produced at any one time, because only one VSG allele at a time can be transcribed from the repertoire of VSG genes, which consists of more than 1,000 sequences. This transcription occurs at one of several VSG gene expression sites (ESs), and antigenic variation can result from two distinct mechanisms: transcriptional switching between VSG ESs, and homologous recombination between the active VSG gene and another VSG gene from the repertoire. In addition to the defences that are usually encountered in mammals, African
NATURE REVIEWS | MICROBIOLOGY
trypanosomes need to defy a novel innate immune mechanism that evolved in humans and in some non-human primates — an efficient trypanolytic factor that is present in serum. Here, we summarize current knowledge of the trypanolytic activity of human serum and of how trypanosomes that infect humans resist this activity. Two serum proteins — haptoglobin-related protein (HPR) and apolipoprotein L1 (APOL1) — have been proposed as candidates for providing the trypanolytic activity. In this Opinion article, we argue that the characterization of the mechanism of resistance of T. b. rhodesiense, as well as the study of the phenotype of lysis induced by APOL1, indicates that APOL1 is the sole factor that is responsible for trypanolysis. Studies of trypanolytic activity Between 1902 and 1912, Laveran and Mesnil6 reported that sera from humans and other primates — such as various Papio (baboon), Cercocebus (mangabey) and
Procyclic form (mid-gut)
Metacyclic form (salivary glands)
Proliferative
Quiescent Insect host
Mammalian host Quiescent
Proliferative
Short, stumpy form (blood)
Long, slender form (blood)
Figure 1 | Life cycle of Trypanosoma brucei. The life cycle of T. brucei alternates between the insect host (the tsetse fly) and a mammalian host (such as humans, cattle, antelopes, buffaloes and lions). In both the insect vector and the mammalian host, colonization occurs through the proliferation of rapidly dividing trypanosome forms. These forms eventually transform into resting (quiescent) parasite cells that are preprogrammed for cellular differentiation after changes in the environment (that is, transfer from one host to the other). The long, slender bloodstream forms also adapt to their environment while present in the mammalian host, and this adaptation involves continuous variation of the main surface antigen, which is known as variant surface glycoprotein.
VOLUME 4 | JUNE 2006 | 477 © 2006 Nature Publishing Group
PERSPECTIVES Mandrillus (forest baboon) species — kill trypanosomes, although they did note differences in activity between these sera. Human serum did not affect the T. brucei subspecies that infect humans: that is, T. b. rhodesiense and T. b. gambiense. By contrast, sera from various Papio species lysed T. b. rhodesiense and showed 10–25-fold higher activity than human serum against human-serumsensitive trypanosomes6. Sera from Cercocebus and Mandrillus species were much less active against human-serumsensitive trypanosomes, and serum from Pan troglodytes (chimpanzees) was devoid of activity, despite the close evolutionary link between chimpanzees and humans6. These findings were confirmed in subsequent studies, which also showed that serum from Gorilla gorilla (gorillas) has trypanolytic activity7–9. Trypanolysis occurs through high-densitylipoprotein-particle-mediated endocytosis. Mainly as a result of the pioneering work of Mary Rifkin10,11, the trypanolytic activity was shown to be associated with high-density lipoprotein (HDL) particles (BOX 1), in particular with the densest HDL subfraction, HDL3 (REFS 12–15). Further studies indicated that receptor-mediated endocytosis of these HDL particles by the trypanosome was involved in lysis16–22. In this case, internalization of the lytic particles would involve their delivery to increasingly acidic compartments of the endocytic pathway. That these processes occur is supported by the findings that either neutralization of the pH of endosomes (with a membrane-permeable weak base, such as chloroquine)16,17 or inhibition of endocytosis itself (by RNA-interference-mediated knockdown of actin mRNA)22 results in inhibition of lysis. The identity of the trypanosome cell-surface receptor for HDL particles remains elusive, but it is probably a lipoprotein scavenger receptor, because HDL particles and low-density lipoprotein (LDL) particles have both been shown to compete with binding of the trypanolytic factor to
the trypanosome surface23. However, the lipoprotein scavenger receptor of trypanosomes must considerably differ from that of other eukaryotes, because sequence analysis did not uncover any candidate genes in the trypanosome genome. HPR as the trypanolytic factor. Purification of the trypanolytic HDL particles in the HDL3 subfraction yielded two types of complex, which differed in their lipid content but contained several of the same proteins, including APOA1 and the serine-proteaselike protein HPR24–27. The involvement of APOA1 in trypanolysis was quickly discounted because of its wide distribution among different types of HDL complex and because of its failure to show trypanolytic activity28–30. Several lines of evidence, however, suggested that HPR has a role in trypanolysis. First, haptoglobin-specific antibodies, which also recognize HPR, inhibited trypanolysis when added to lytic HDL particles24. Second, haptoglobin, which is similar to HPR, also prevented trypanolysis when added to these assays of trypanolysis, possibly as a result of competition30,31 (although a differential effect was observed for the two types of trypanolytic complex that had been purified32–34). Third, HPR is not expressed in chimpanzees, the serum of which lacks trypanolytic activity (as indicated earlier)6– 9,35. Last, HPR was also thought to be the specific component of the trypanolytic HDL particles that is recognized by the parasite cell-surface receptor36. The trypanolytic effect of the HPRcontaining HDL particles was thought to be a consequence of damage to the lysosomal membrane. According to Hajduk and colleagues16,24,37, HDL particles caused lipid peroxidation of the lysosomal membrane of the parasite, leading to disruption of this membrane, release of proteolytic enzymes into the cytoplasm and auto-digestion of the parasite cell. The peroxidase activity of HDL particles was initially ascribed to a putative HPR–haemoglobin dimer24 and then to the
Box 1 | High-density lipoprotein (HDL) particles In the blood, most lipids are contained in soluble complexes known as lipoproteins. HDL particles are spherical particles that comprise a hydrophobic lipid core (which mainly consists of triglycerides and cholesteryl esters) surrounded by a hydrophilic layer (which consists of phospholipids, unesterified cholesterol and several proteins that are collectively known as apolipoproteins). In terms of protein content, HDL particles mainly contain apolipoprotein A1 (APOA1), which specifically captures and solubilizes free cholesterol, thereby enabling HDL particles to function as cholesterol scavengers. Several HDL-particle subfractions can be separated on the basis of density. The subfraction known as HDL3, which contains both APOL1 and haptoglobin-related protein (HPR), is the densest. The high protein content of HDL particles renders them denser than other lipoprotein particles, including low-density lipoprotein (LDL) particles.
478 | JUNE 2006 | VOLUME 4
Fenton reaction between hydrogen peroxide and iron (an iron-salt-dependent decomposition of hydrogen peroxide, which generates highly reactive hydroxyl radicals)37. Findings supporting these views included the inhibition of lysis by antioxidants and protease inhibitors (factors that were proposed to block lipid peroxidation and cellular autodigestion, respectively)16,24,37, the detection of end-products of lipid peroxidation during trypanolysis37 and the observation (using electron microscopy) of discontinuities in the lysosomal membrane of cells undergoing lysis16. By contrast, Raper, Tomlinson and colleagues27,38 could not detect binding of haemoglobin to HPR, found no evidence of lipid peroxidation, observed no effect of antioxidants and protease inhibitors on trypanolysis, and found no reactive oxygen intermediates associated with trypanolysis. As detailed later, one explanation for these conflicting findings could be that both the proposed identity and the proposed mode of action of the trypanolytic factor were incorrect. Studies of resistance to trypanolysis In addition to studies of the trypanolytic activity, research on the resistance of certain trypanosome subspecies to lysis provided another angle from which to approach identification of the trypanolytic factor. In contrast to T. b. brucei, the subspecies T. b. gambiense and T. b. rhodesiense escape the trypanolytic activity of human serum and cause sleeping sickness, a lethal disease in humans. T. b. gambiense is permanently resistant to human serum. By contrast, T. b. rhodesiense loses resistance after being isolated from humans and transferred to other animals: for example, after 30–70 passages in mice6,39. Following the injection of these mice with human serum, T. b. rhodesiense regains resistance to lysis, and this acquisition of resistance was shown to be associated with antigenic variation39. Specifically, for a given isolate of T. b. rhodesiense, such resistance arising on selection in human serum involved a change in expression of the VSG to a particular VSG variant known as ETat 1.10. This VSG, however, was not itself responsible for resistance to lysis, because other humanserum-resistant T. b. rhodesiense clones were shown to express different VSGs, including VSGs that are expressed by clones that are sensitive to lysis39. This paradox was resolved by the discovery that, in these trypanosomes, selection in human serum involved transcriptional switching to a particular VSG ES, which was named R-ES40. Although
www.nature.com/reviews/micro © 2006 Nature Publishing Group
PERSPECTIVES R-ES initially contained the gene encoding the VSG variant ETat 1.10, replacement of this gene by gene conversion resulted in conservation of the resistance phenotype, indicating that a component of R-ES, but not the VSG gene itself, conferred resistance. All VSG ESs have polycistronic transcription units that contain several genes (known as expression-site-associated genes, ESAGs) in addition to the VSG gene41, and these are under monoallelic control of expression, similar to the VSG genes. R-ES was found to contain a gene known as serum-resistance associated (SRA) (FIG. 2), which had previously been identified as being transcribed only in resistant clones42. The presence of this gene in a VSG ES that is systematically selected for transcription in human serum provided a straightforward explanation for its exclusive expression in human-serumresistant clones. Transfection of SRA alone into T. b. brucei (which is sensitive to lysis by human serum) conferred full resistance, thereby identifying SRA as necessary and sufficient for resistance of T. b. rhodesiense40. The finding that SRA is an ESAG present in a particular VSG ES explained the link between resistance to lysis by human serum and antigenic variation. SRA encodes a truncated VSG that is devoid of surface-exposed loops, owing to an in-frame deletion43–45. Together with the localization of this gene in a VSG ES, this finding indicates that SRA was generated by antigenic-variation-associated homologous recombination of VSG genes. Interestingly, another component that is involved in adaptation of the parasite to the host, the parasite surface receptor for host transferrin, was also found to originate from a VSG gene, and the genes that encode this heterodimeric receptor are also located in VSG ESs41. So, the VSG ESs are loci that result in the generation of various VSG-derived proteins that help the parasite to adapt to the host. SRA and R-ES seem to be conserved in most isolates of T. b. rhodesiense46–50. So, these genetic elements are clearly important for the propagation of this trypanosome subspecies, for which they are also excellent specific markers51. Given the high probability that SRA evolved only once, T. b. rhodesiense is likely to have arisen as a clone of T. b. brucei that differs mainly or solely by its ability to express SRA on selection in human serum. The genetic variability that is observed between T. b. rhodesiense isolates is probably a result of spreading of R-ES into various backgrounds, through genetic exchange48. Using reverse-transcription PCR, it has also been shown that R-ES
a
VSG repertoire
ES1
7
6
5
4
8
8
7
6
R-ES
6
7
5
4 SRA
5
ES transcription
b
VSGn
VSG
VSG
VSG Homologous recombination: gene conversion VSG 3 2 11 1
Homologous recombination: reciprocal exchange
Transcriptional switching ES2
VSG
8
8
3
2
11
1
VSG
VSG 70-Base-pair repeats
Telomeric repeats
Trypanosoma brucei rhodesiense metacyclic form
VSG variant expressed VSG
VSG
SRA
VSG
Human serum
Transcriptional switching Homologous recombination
Non-human mammalian serum ?
Figure 2 | Antigenic variation and resistance to human serum of Trypanosoma brucei rhodesiense. a | Genetic mechanisms of antigenic variation. The Trypanosoma brucei genome contains more than 1,000 variant surface glycoprotein (VSG) genes (shown as coloured boxes). Only one of these VSG variants is expressed at any one time. Gene expression occurs in a telomeric expression site (ES), where the VSG gene is co-transcribed with expression-site-associated genes (ESAGs; shown as numbered boxes).The T. brucei genome contains a set of 15–20 similar (but not identical) ESs, 3 examples of which are depicted. In T. b. rhodesiense, one particular ES, known as R-ES, contains the serum-resistance associated (SRA) gene, which confers resistance to lysis by human serum40. Antigenic variation (that is, the expression of a different VSG variant) can result from two mechanisms (indicated in red): transcriptional switching and homologous recombination. Transcriptional switching occurs by a process known as in situ activation, in which expression of the active ES is turned off, and expression of a previously silent ES is turned on. This process does not involve DNA rearrangement. By contrast, homologous recombination involves replacement of the active VSG gene. This can occur by one of two mechanisms: gene conversion, which involves replacement with a copy of a VSG gene from the repertoire; or reciprocal exchange, which involves replacement with a VSG gene from another ES (and thereby exchange of a VSG gene between two ESs). These replacements occur through recombination between homologous regions, such as 70-base-pair repeats. b | Gain and loss of resistance to human serum. Tsetse flies inject the mammalian host with T. b. rhodesiense in the metacyclic form, and these trypanosomes then transform into long, slender bloodstream forms (FIG. 1), in which different VSG ESs can be activated. Only trypanosomes in which the R-ES is active can resist lysis by human serum. Under these conditions, in situ activation (transcriptional switching) of other VSG ESs is counter-selected (indicated by a small red cross), owing to the requirement for SRA expression to resist lysis. Therefore, antigenic variation only occurs through homologous recombination targeted to the active VSG gene. In non-human hosts, expression of the R-ES seems to be counter-selected, although in situ inactivation of the R-ES does require many passages in mice6,39. This counter-selection might result from the absence of most ESAGs from the R-ES, because these ESAGs might not be completely dispensable.
NATURE REVIEWS | MICROBIOLOGY
VOLUME 4 | JUNE 2006 | 479 © 2006 Nature Publishing Group
PERSPECTIVES a APOL1 Met60
Trp235
Ala238
Pro304
Ala339
Leu398 C terminus
Pore-forming domain
N terminus
Membrane-addressing domain
b
c
d
pH ~7 N
C C
Asp249
SRA-interacting domain
N SRA (amino acids 31–79)
Arg287 Glu253 Glu256
Arg284 Ile58, Leu61 and Ile62
Glu260 APOL1 (amino acids 248–291)
C
APOL1 (amino acids 340–392)
N APOL1 (amino acids 90–235)
C
pH ~5 N
N
C
N
C N Colicin A (pore-forming domain)
Figure 3 | Models of the three domains of apolipoprotein L1 (APOL1). A diagrammatic representation of APOL1 and the location of each of its domains is shown in a. The structure of each domain is also shown in a ribbon-diagram format in b,c,d, where the protein backbone is shown as a coloured ribbon and the amino-acid side chains are shown in black (uncharged atoms), blue (positively charged atoms) and red (negatively charged atoms). a | APOL1 contains three domains: a pore-forming domain, a membrane-addressing domain and an SRA (serum-resistance associated)-interacting domain. b | The region of APOL1 that spans the methionine residue at position 60 and the tryptophan residue at position 235 contains a domain that is structurally and functionally similar to the ionic-pore-forming domain of bacterial colicins such as colicin A. Analogous helices of the two proteins are shown in the same colour. Colicin A diagram modified with permission from REF. 64 © (2005) American Association for the Advancement of Science. c | The region of APOL1 that spans the alanine residue at position 238 and the proline residue at position 304 contains a membrane-addressing domain. Computer modelling of this region predicts a pH-dependent structure. At neutral pH, such as in the blood, the two α-helices of the membrane-
is not active when the parasite is present in non-human sera52. This finding indicates that R-ES, which lacks several ESAGs that are usually found in other VSG ESs40 (FIG. 2a), is counter-selected when the parasite is not exposed to human serum (FIG. 2b). The data also confirm and explain the early observations of Laveran and Mesnil6, as well
addressing domain interact through two salt bridges (indicated by pink squares), which are formed by the side chains of the amino acids indicated. This hairpin structure shows a segregation between hydrophobic (orange) and hydrophilic (green) surfaces, as indicated by space-filling models. At acidic pH, such as in the lysosome (pH 5.3), the salt bridges are predicted to dissociate as a result of neutralization of the negatively charged residues, and this would lead to loss of the large hydrophobic surface. Experimental evidence indicates that this surface is required for the association of APOL1 with high-density lipoprotein (HDL) particles and that treatment with acid results in dissociation of APOL1 from HDL particles (B.V. and E.P., unpublished observations). Figure modified with permission from REF. 64 © (2005) American Association for the Advancement of Science. d | The region of APOL1 that spans the alanine residue at position 339 and the leucine residue at position 398 forms a long α-helix that strongly and specifically interacts with the N-terminal α-helix of the Trypanosoma brucei rhodesiense protein SRA. The residues of SRA that are important for this interaction, as determined by studies involving point mutations, are indicated. Figure modified with permission from Nature REF. 54 © (2003) Macmillan Publishers Ltd.
as more recent observations in the field47, that T. b. rhodesiense is sensitive to human serum when grown in non-primate animals. Therefore, when living in non-human hosts, T. b. rhodesiense probably does not differ from T. b. brucei, but the difference arises when T. b. rhodesiense is exposed to human serum, which triggers selection of R-ES
480 | JUNE 2006 | VOLUME 4
and transcription of SRA. Interestingly, T. b. gambiense does not carry SRA, despite its constitutive resistance to human serum53. Therefore, the mechanism ensuring resistance of this subspecies to lysis must differ from that of T. b. rhodesiense, and this mechanism is under investigation at present.
www.nature.com/reviews/micro © 2006 Nature Publishing Group
PERSPECTIVES APOL1 Although SRA has a signal peptide, it is not targeted to the cell surface of the parasite (unlike the VSGs from which it is derived) but accumulates in the lysosome54. This could result from the absence of a GPI anchor (REFS 55,56, and D.P.N., F.P.H. and E.P., unpublished observations). Similar to VSGs, SRA is characterized by a long N-terminal hairpin that contains two amphipathic α-helices1,44. The analysis of various mutated forms of SRA expressed in T. b. brucei (which itself does not express SRA) showed that only the N-terminal amphipathic α-helix helix A is required to provide resistance54. In VSGs, this helix is usually involved in coil-coiling interactions with the other amphipathic α-helix, helix B, and with helices of adjacent VSGs, to form a dimer1. So, by analogy, it was envisaged that SRA interacts with the trypanolytic factor of human serum by coil-coiling. Chromatographic separation of human serum on SRA affinity columns showed that the protein APOL1 bound strongly and specifically to SRA. This binding was found to result from hydrophobic coil-coiling interactions between helix A of SRA and the C-terminal α-helix of APOL1 (REF. 54) (FIG. 3a,d). In addition, point mutations that disrupt helix A of SRA were shown to result in loss of binding to APOL1 (REF. 54). APOL1 is associated with the HDL3 subfraction of human serum57. Trypanosomes internalize APOL1containing particles, and APOL1 is then routed to the lysosome, where it colocalizes with SRA54. Under conditions that dissociate serum lipoprotein complexes (that is, in the presence of detergent), separation of human serum by affinity chromatography on either APOL1-specific antibodies or SRA selectively removed APOL1, and the only other detectable components that were subsequently eluted with APOL1 were serum albumin and antibodies54. This chromatographic step, followed by dialysis to remove the detergent, resulted in total loss of the trypanolytic activity of the serum54. Applying the same protocol but using affinity chromatography on an SRA molecule with a disrupted helix A did not result in the removal of APOL1 or in the loss of trypanolytic activity, showing that the depletion of APOL1 (and not the treatment with detergent) was responsible for the loss of trypanolytic activity54. Furthermore, addition of purified native or recombinant APOL1 to the depleted serum or to fetal calf serum (which naturally lacks APOL1) restored or conferred, respectively,
the ability to lyse human-serum-sensitive trypanosomes but not human-serumresistant trypanosomes54. The resistance of trypanosomes to lysis was associated with the C-terminal helix of APOL1 (which interacts with helix A of SRA), because removal of this helix from APOL1 conferred the ability to lyse both humanserum-sensitive and human-serumresistant trypanosomes54. Finally, the natural phenotype of this system can be entirely reconstituted with recombinant proteins: recombinant-SRA-expressing T. b. brucei were found to be resistant to lysis following exposure to recombinant APOL1, whereas wild-type T. b. brucei were readily lysed by APOL1 (REF. 54). This finding indicates that APOL1 mimics the natural killing activity of human serum, and it also proves that SRA blocks the activity of APOL1 on trypanosomes. From these studies, it was concluded that APOL1 is the trypanolytic factor of human serum and that, in T. b. rhodesiense, SRA neutralizes APOL1 through coil-coiling interactions with the C-terminal helix of APOL1. Accordingly, APOL1 was found to be absent from chimpanzee serum, which is non-trypanolytic (as discussed earlier)6,8,9, and sequencing of the chimpanzee genome provided a straightforward explanation for this finding: the APOL1 gene is absent from these animals58. Resistance to APOL1 during the T. b. rhodesiense life cycle. In long, slender bloodstream forms of T. b. rhodesiense (FIG. 1), the expression of SRA allows neutralization of APOL1 in the lysosome54. This mechanism of resistance considerably differs from the previously proposed mechanism, the selective inhibition of endocytosis of the trypanolytic factor19. Recent data obtained using T. b. brucei transfected with an epitope-tagged version of SRA indicate that neutralization of the trypanolytic factor by SRA can occur in endosomes, before the lytic factor reaches the lysosome59. These findings are consistent with the observation that the interaction between SRA and APOL1 can occur under both neutral and acidic conditions (that is, between pH 7.5 and 5.0)54. In accordance with the apparent lability of SRA40,46,59, the distribution of SRA between the endosomes and the lysosome might depend on the relative level of endocytic-protease activity, as indicated by studies in which the endo-lysosomal protease trypanopain was inhibited (L.V. and E.P., unpublished observations). Therefore, it can be speculated that SRA is routed to the
NATURE REVIEWS | MICROBIOLOGY
lysosome through the endocytic pathway and that the interaction of SRA with APOL1 can occur at different steps of this pathway and is followed by the proteolytic cleavage of this complex. Degradation of APOL1, however, is not required for the trypanosome to be resistant to lysis, because the presence of both APOL1 and SRA in the lysosome (in contrast to APOL1 alone) has been shown to prevent lysis from occurring54. That is, interaction with SRA seems to be sufficient for inactivation of APOL1. In procyclic forms of T. b. rhodesiense (FIG. 1), which are present in the tsetse fly, R-ES is inactive, so SRA is not expressed. However, procyclic forms can resist lysis by human serum in vitro60, and this is probably a consequence of the low level of endocytosis by trypanosomes at this stage22. Metacyclic forms of T. b. rhodesiense (FIG. 1), which are transferred from the tsetse fly to the mammalian host, are the first to be confronted with mammalian blood. Because R-ES is probably inactive in these forms of the parasite3, it is not clear how they resist APOL1 before their transformation into bloodstream forms. In vitro, metacyclic forms of T. b. rhodesiense were found to be mostly sensitive to lysis by human serum61. In the same assay, metacyclic forms of T. b. brucei and T. b. gambiense were always sensitive and resistant, respectively, to lysis by human serum. Therefore, metacyclic forms of T. b. rhodesiense are not constitutively resistant to human serum. After inoculation by the tsetse fly, trypanosomes can remain for several days in tissue spaces or lymphatic vessels, so it is possible that some metacyclic forms escape exposure to APOL1 before they begin to express SRA. Proposed mechanism of APOL1-mediated trypanolysis. The physiological function of APOL1 in humans remains elusive, although there is indirect evidence that it is involved in lipid metabolism62, as well as evidence of a link between overexpression in the brain and schizophrenia63. For this reason, the way in which APOL1 kills trypanosomes was not apparent. A weak similarity was, however, found between the N-terminal region of APOL1 and the pore-forming domain of colicins, which are bacterial toxins that kill competing bacteria by forming ionic pores in the inner cell membrane. This finding led to experiments that uncovered the capacity of APOL1 to generate ionic pores in biological membranes, both in vivo and in vitro64. Briefly, the region of APOL1 that spans amino acids 60 to 235, which was predicted by computer modelling to form a structure
VOLUME 4 | JUNE 2006 | 481 © 2006 Nature Publishing Group
PERSPECTIVES A Recombinant APOL1
Normal human serum
B a
0h
1h
2h
3h
4h
c
b
5h
Flagellar pocket
Fixation artefacts Lysosome
Lysosome
Nucleus
1 µm
1 µm
1 µm
Figure 4 | Phenotype of trypanosome lysis by human serum or by recombinant apolipoprotein L1 (APOL1). A | Micrographs show the expansion of the lysosome of Trypanosoma brucei brucei over time in the presence of 1 µg per ml recombinant APOL1 or 10% normal human serum. Lower panels are reproduced with permission from REF. 64 © (2005) American Association for the Advancement of Science. B | Transmission electron micrographs show fixed T. b. brucei that either were not exposed to human serum (a) or were exposed to human serum for 1.5 hours (b,c). In control cells (a), the lysosome is small and dispersed, and cannot be seen clearly. In treated cells (b,c), the swollen lysosome is indicated. Fixation artefacts, which have led to the suggestion that the lysosomal membrane is disrupted during lysis16, are also shown (c). Quantitative measurements of lysosomal surface area and cytoplasmic surface area over time are shown in Supplementary information S1 (figure)). Both the swelling of the lysosome and the absence of general cellular swelling can be observed in Supplementary information S2 (movie).
that resembles the pore-forming domain of colicins (FIG. 3a,b), showed bactericidal activity with a marked similarity to that of colicins and generated anionic pores in synthetic lipid-bilayer membranes. This pore-forming domain of APOL1 was found to be adjacent to a pH-sensitive membraneaddressing domain, which was predicted to bind HDL particles at neutral pH only (that is, in the blood but not in the late endosomes or the lysosome)64 (FIG. 3a,c). Both domains were required for the toxic activity of APOL1 in vivo, against bacteria and against trypanosomes. By contrast, the C-terminal α-helix of APOL1 (which interacts with SRA and is required for resistance to lysis) was dispensable for toxic activity in both cases. In trypanosomes, recombinant APOL1 was found to mimic completely the lytic activity of normal human serum, causing
depolarization of the lysosomal membrane (that is, loss of the differential charge between the two faces of the membrane), followed by continuous swelling of the lysosome until the parasite cell lysed64 (FIG. 4). In addition, APOL1 was detected at the periphery of the lysosome but not in the lumen64. Given the predicted structure of its membraneaddressing domain (FIG. 3c), APOL1 is thought to undergo a conformational change that allows its dissociation from HDL particles at low pH (such as in the late endosomes and the lysosome) and its subsequent insertion into the lysosomal membrane, where it forms a pore. On exposure to recombinant APOL1, similar to normal human serum, lysosomal swelling was found to be the first detectable morphological change, and as this process continued, there were no marked alterations in other intracellular structures
482 | JUNE 2006 | VOLUME 4
(FIG. 4; see Supplementary information S1,S2 (figure and movie)), in particular no general vacuolization and cellular swelling (although this is in contrast to findings reported in REF. 65; discussed later). Lysosomal swelling was found to be associated with the influx and intracellular accumulation of chloride ions (Cl–). Both of these processes were blocked by either depletion of extracellular Cl– or addition of the anionchannel blocker DIDS (4,4′-diisothiocyano2,2′-disulphonate stilbene)64. Although DIDS completely inhibited the influx of Cl– into trypanosomes and the subsequent lysis, it had no direct effect on the pore-forming activity of APOL1 in vitro, indicating that additional Cl– channels are involved in trypanolysis. So, it was concluded that APOL1 triggers the lysosomal influx of Cl– from the cytoplasm, where the concentration of this anion is high (106 mM)66. This movement would then reduce the cytoplasmic Cl– concentration, activating the compensatory entry of extracellular Cl– through DIDS-sensitive channels in the plasma membrane64. This model of the trypanolytic process is shown in FIG. 5. The cascade of Cl– movement would be accompanied by the influx of water into the lysosome and osmotic swelling of this compartment. This accounts for the observed inhibition of trypanolysis by osmotically active molecules such as sucrose (the presence of which in the cytosol reduces the influx of water into the lysosome)67. The internal pressure resulting from the continuous enlargement of the lysosome would be responsible for the lysis of the parasite, because this pressure is expected to irreversibly damage the plasma membrane of the parasite. This model also explains the observed fraying of the parasite surface coat and the leakage of ions that occur before lysis67. The finding that trypanosomes with considerably swollen lysosomes can survive for a period of time clearly conflicts with the idea that lysis involves disruption of the lysosomal membrane following lipid peroxidation, a mechanism that is still proposed by some researchers16,20,24. Indeed, the extreme swelling of the lysosome seems to be incompatible with a damaged lysosomal membrane. The main argument for lysosomal-membrane disruption is that discontinuities can be detected in the membrane, using electron microscopy16; however, these are probably the result of common fixation artefacts68 (FIG. 4Bc). A recent study proposed an alternative mechanism for trypanolysis: that lytic HDL particles generate cation-selective pores that are active at the plasma membrane after they have been recycled from the lysosome69.
www.nature.com/reviews/micro © 2006 Nature Publishing Group
PERSPECTIVES Given the cation concentration gradients at the plasma membrane (that is, extracellular to cytoplasmic concentration ratios of 140 mM/14 mM and 6 mM/116 mM for sodium ions (Na+) and potassium ions (K+), respectively 66), these pores would allow an influx of Na+ into the cytoplasm. However, in this study 69, the electrophysiological measurements were carried out on a crude population of lipoproteins that contained various pore-forming proteins and peptides, and this population generated complex ion currents, possibly owing to the formation of irrelevant (non-physiological) pores or the influx of ions through nonspecific channels. In addition, although pore-forming activity was monitored under conditions that mimic the interface between the lysosomal and cytoplasmic compartments, its involvement in trypanolysis was entirely attributed to ionic fluxes produced at the plasma membrane, and we think that this conclusion is questionable for two main reasons. First, it is unlikely that plasma-membrane fluxes are relevant to the trypanolytic process that is induced by normal human serum, because we have found that this process is characterized by swelling of the lysosome but not of the cytoplasm64 (FIG. 4; see Supplementary information S1, S2 (figure and movie)). This phenotype is in contrast to that observed using purified lytic HDL particles65, indicating that the lysis that occurs in the presence of purified lytic HDL particles is likely to involve nonspecific toxicity. Second, the specificity of the pores for particular cations was not measured in the study reported in REF. 69. However, if the pores conduct Na+ to a similar extent to which they seem to conduct K+, then (because the Na+ and K+ gradients across the plasma membrane are approximately equal in magnitude but opposite in direction66) the most likely initial outcome would be the electroneutral exchange of the two cations, which would have no osmotic consequences. Therefore, in our opinion, the hypothesis that the trypanolytic factor generates cationic pores is not proven by the data in REF. 69, and pore formation at the plasma membrane does not account for the phenotype of trypanolysis by normal human serum. Identification of the trypanolytic factor of human serum: APOL1 or HPR? Initially, HPR was considered to be the trypanolytic factor of human serum, and recent papers present evidence for why this protein, in addition to APOL1, is a trypanolytic factor9,69,70. This view is based, in part, on the observation that serum from baboons is
Blood pH ~7 HDL particle
DIDS Flagellum
APOL1
Lysosome Cl–
Flagellar pocket
Cl– and H2O
SRA pH 5.3 Swollen lysosome pH ~7 Late endosome or lysosome Plasma membrane Early endosome
Cytosol
Figure 5 | Model of the mechanism of trypanolysis by apolipoprotein L1 (APOL1). APOL1 contains a pore-forming domain (red) and a membrane-addressing domain (blue). It is associated with highdensity lipoprotein (HDL) particles (green) that are present in the blood. These particles are internalized by bloodstream-form trypanosomes through HDL-receptor-mediated endocytosis that is initiated in the flagellar pocket. Following internalization, HDL particles progress through the endocytic pathway, from early endosomes (which have neutral pH) to late endosomes (which are acidic) and then to the lysosome (which is also acidic). We think that most data indicate that trypanolysis occurs by the following process. The lysosomal pH induces a conformational change in the membrane-addressing domain of APOL1, leading to dissociation from the HDL particle and binding to the lysosomal membrane. The pore-forming domain of APOL1 enables it to form a pore in the membrane, and this leads to the flux of chloride ions (Cl–) from the cytoplasm to the lumen of the lysosome. Cytoplasmic Cl– deprivation is compensated by the activity of a DIDS (4,4′-diisothiocyano-2,2′-disulphonate stilbene)sensitive Cl– transporter in the plasma membrane. This ionic flux is presumably accompanied by secondary cationic movements64,67,69, and it triggers the movement of water (H2O) into the lysosome and osmotic swelling of this compartment. The resultant uncontrolled swelling of the lysosome increases intracellular pressure, probably leading to damage to the plasma membrane (dashed line), which is ultimately the cause of cell death. The Trypanosoma brucei rhodesiense protein serum-resistance associated (SRA) interacts with APOL1 in late endosomes and the lysosome, and this prevents APOL1 from forming pores, possibly through sequestration of APOL1 followed by proteolytic degradation.
trypanolytic and contains HPR but not APOL1 (REFS 8,9). However, serum from these primates kills both human-serumsensitive and human-serum-resistant T. b. rhodesiense clones, and it seems to have considerably more lytic activity than does human serum6,8,9. In our opinion, these observations indicate that the trypanolytic activity of baboon serum differs from the trypanolytic activity of human serum and that HPR cannot be the common trypanolytic factor in both cases, because HPR does not lyse human-serum-resistant T. b. rhodesiense. It is also argued that HPR is a trypanolytic factor because HPR-specific antibodies inhibit the activity of lytic HDL particles9,24,70. These experiments, however, involved lipoprotein complexes that were not dissociated with detergent. In addition, it has been reported that antibodies specific
NATURE REVIEWS | MICROBIOLOGY
for APOA1 (a protein that does not have trypanolytic activity28–30) inhibit lysis70. So, one possible explanation for the inhibitory activity of HPR-specific antibodies is that antibodies specific for any exposed constituent of HDL particles can inhibit lysis through a nonspecific mechanism, perhaps by aggregation of HDL particles or by interference in the endocytic process. In their recent study, Hajduk and colleagues70 propose that both HPR and APOL1 are trypanolytic but that these proteins need to be associated in the same HDL particle for maximal activity. This proposal is based on trypanolytic assays carried out using purified HPR and APOL1 or using human HDL particles fractionated into particles that contain both HPR and APOL1, either protein alone or neither protein. In the first case (using purified proteins), toxicity was
VOLUME 4 | JUNE 2006 | 483 © 2006 Nature Publishing Group
PERSPECTIVES Box 2 | The function of apolipoproteins (APOLs) In evolutionary terms, APOL1 appeared recently, as it has been identified only in humans and gorillas8,75. It belongs to a multigene family that has six members76,77, and homologous sequences are present in other eukaryotes (particularly in other mammals). Among APOL1-like proteins, the sequence of the pore-forming domain is similar, whereas that of the membrane-addressing domain is more variable; the C-terminal helix is the most conserved region (see Supplementary information S3 (figure)). Unlike other APOL1-family members, APOL1 is secreted (probably because of the presence of a signal peptide), but sequestration in high-density lipoprotein (HDL) particles presumably neutralizes its activity in serum. Other APOL-family members are most probably intracellular proteins. Given the pore-forming activity of APOL1 (REF. 64), intracellular APOLs could be ion channels in organelles, such as the lysosome, and might thereby control the volume of these organelles. The potential of the C-terminal helix of APOL1 to interact with the Trypanosoma brucei rhodesiense protein serum -resistance associated (SRA) and thereby neutralize APOL1 activity, together with the evidence that this helix is dispensable for trypanolytic activity despite its high level of sequence conservation54,64, raises the interesting possibility that this region controls the ionic-pore-forming activity of APOLs through interaction with proteins that have SRA-like helices. It is tempting to propose that the generation of a secreted version of these pore-forming proteins in the great apes of Africa was linked to the presence of trypanosomes in this region and that it was selected because of its trypanolytic activity. However, the conservation of APOL1 in humans (despite the wide geographical dispersion of humans) does not support this idea but indicates that APOL1 fulfils additional functions. In this regard, it is interesting that the expression of APOL1, APOL2 and APOL3 is strongly induced by the pro-inflammatory cytokine tumour-necrosis factor75,78 and that overexpression of APOL6 triggers apoptosis, presumably through a BH3 domain (B-cell lymphoma 2 (BCL-2)-homology domain 3)79. So, APOLs might be involved in various cellular responses to danger signals.
Pore-forming proteins and cell death The ability of pore-forming proteins to trigger cell death can be illustrated by several well-known examples — bacterial colicins, diphtheria toxin and apoptotic proteins from the BCL-2 family — although different mechanisms are involved in the cell death that is mediated by these three types of protein (reviewed in REF. 80). Despite a low level of sequence homology, these proteins all have poreforming domains with a similar structure: that is, several α-helices organized around a conserved hydrophobic hairpin. Computer modelling and experimental data indicate that the pore-forming domain of APOL1 also belongs to this category64 (FIG. 3). As is the case for BCL-2-family members, the pore-forming domain of APOL1 might have been inherited from bacteria. HDL particles and innate immunity The association of APOL1 with HDL particles is likely to occur through hydrophobic interactions that are mediated by the membrane-addressing domain at neutral pH64 (FIG. 3). The presence of pore-forming microbicidal components in HDL particles is not an unprecedented finding. Various APOLs that are associated with HDL particles can show bactericidal activity81. In addition, defensins and cathelicidins, which are pore-forming antibacterial peptides, also seem to be associated with lipoproteins in serum82. The anchoring of microbicidal proteins in lipoprotein particles would ensure the sequestration of these proteins when their activity is not required, thereby preventing cytotoxic effects while maintaining high serum concentrations. Therefore, lipoproteins could have a role as carriers for components of the innate immune system21,70.
detected following direct incubation of trypanosomes with high concentrations of either protein. However, we argue that these data cannot be fully interpreted without information relating to the phenotype of lysis. For example, when the antimicrobial peptides cathelicidins are purified from lipoprotein complexes, they can kill trypanosomes in vitro through disruption of the plasma membrane71; however, there is no evidence that these cathelicidins are active against trypanosomes in vivo. In the second case (using fractionated HDL particles), almost 100% of the trypanolytic activity was recovered in the HDL-particle fraction that contained both HPR and APOL1, whereas less than 0.4% of the trypanolytic activity was detected in fractions that contained
either APOL1 or HPR. We think that the small trypanolytic effect of these HDL particles that contain undefined amounts of either HPR or APOL1 could be a result of nonspecific toxicity conferred by the experimental treatment. Indeed, fractionation does seem to confer non-physiological toxicity, because the phenotype of trypanolysis by fractionated HDLs differed from that of human serum and involved generalized vacuolization and cellular swelling65. Furthermore, fractionation resulted in a higher trypanolytic activity than that naturally present in serum31,33, and it rendered chimpanzee HDL particles trypanolytic9. One possibility is that HDLparticle fractionation activates microbicidal peptides, including cathelicidins71. A definitive assessment of this idea could be made by
484 | JUNE 2006 | VOLUME 4
determining the effect of fractionated HDL particles on trypanosomes that are naturally resistant to human serum. In summary, we outline the following five arguments in support of our contention that APOL1 is the only trypanolytic factor of human serum. First, the addition of recombinant APOL1 to fetal calf serum, which lacks HPR, rendered the serum trypanolytic, and the phenotype of lysis was identical to that observed for human serum54,64. It should be noted that free APOL1 (that is, APOL1 not associated with HDL particles), whether native or recombinant protein, seems to have less trypanolytic activity than human serum64,70, and this finding could be interpreted in two ways. The reduced activity of free APOL1 might result from a lack of synergy with HPR69,70: that is, both APOL1 and HPR are required for maximal trypanolytic activity. Alternatively, it is our contention that the reduced activity of free APOL1 reflects a requirement for APOL1 to be associated with HDL particles for efficient binding and uptake by the HDL receptor of the parasite11,23. Indeed, kinetic studies indicated that free recombinant APOL1 seemed to be as effective as human serum at lysis of trypanosomes, except that lysis was delayed64. Accordingly, trypanolysis by recombinant APOL1 was strongly accelerated on reconstitution of APOL1 in lipoproteins54,64. Therefore, we think that the activity of APOL1 does not require synergy with HPR. APOL1-transgenic mice would be useful for evaluating the in vivo efficiency of trypanosome killing by APOL1 alone, but such mice are not available at present. Injection of recombinant APOL1 into wildtype mice, however, was shown to be necessary and sufficient for complete inhibition of infection with T. b. brucei 72. It remains to be determined whether similar experiments with HPR-transgenic mice would show increased parasite killing by APOL1. Second, the presence of HPR in APOL1free serum did not trigger trypanolysis and did not affect trypanosome growth. This was observed in vivo for transgenic mice that expressed HPR at a level similar to that of humans73, as well as for human serum that was specifically depleted of APOL1 (through separation of the proteins present in detergent-dissociated HDL particles by affinity chromatography on SRA)54,64. Third, depletion of APOL1 from human serum without removal of HPR (as above) led to a complete loss of trypanolytic activity54,64. The converse experiment — selective depletion of HPR without removal of APOL1— has not yet been carried out.
www.nature.com/reviews/micro © 2006 Nature Publishing Group
PERSPECTIVES Fourth, trypanolysis did not occur when recombinant APOL1 that was mutated in either the pore-forming domain or the membrane-addressing domain was added to serum specifically depleted of APOL1, even if HPR was present64. Conversely, recombinant APOL1 that lacked the C-terminal (SRA-interacting) helix was necessary and sufficient to kill T. b. rhodesiense independent of HPR54,72. Fifth, expression of SRA alone conferred complete resistance to either human serum or recombinant APOL1. This finding can be explained by the interaction between SRA and APOL1 (REF. 54), and there is no evidence that SRA could interfere with HPR. For example, no trace of HPR has been found in the APOL1-containing serum fraction that binds SRA affinity columns in the presence of detergent (discussed earlier)54. We propose, therefore, that APOL1 is responsible for most, if not all, of the trypanolytic activity that is associated with human serum. In our opinion, the idea that APOL1 operates synergistically with HPR for maximal trypanolytic activity in the context of a subclass of HDL particles that contains APOL1, HPR and APOA1 is not supported by the available evidence. Conclusions and future directions Despite considerable progress, many questions that relate to the trypanolytic activity of human serum remain to be answered. In addition, the physiological function of APOL1 and other APOLs remains unclear (BOX 2; see Supplementary information S3 (figure)). At present, the most important points that require further investigation are the identity of the trypanosome receptor for the trypanolytic HDL particles, the basis of resistance of T. b. gambiense to APOL1 and the putative involvement of HPR in the mechanism of trypanosome killing by APOL1. To resolve the last issue, a crucial experiment will be to compare the trypanosome-infection characteristics of mice that are transgenic for HPR, APOL1 or both genes. So far, however, attempts to generate APOL1-transgenic mice have not been successful. Recent advances in this field of research include a breakthrough in the diagnosis of sleeping sickness, for which the presence of SRA has proved to be a reliable marker of infection with T. b. rhodesiense47–51, and the use of APOL1 as a treatment that might cure the disease. As discussed in this article, when APOL1 is not bound to HDL particles, it kills trypanosomes but with delayed kinetics compared with human serum64. So, to convert APOL1 into a drug that is effective against
trypanosomes that infect humans, two modifications have been carried out. First, the C-terminal region, which is recognized by SRA, has been removed. Second, this truncated APOL1 molecule has been fused to an antibody module (known as a nanobody) that is derived from single-chain camel antibodies, thereby targeting the toxin (APOL1) to invariant determinants (high-mannose side chains) of the trypanosome surface74. The efficacy of trypanolysis by this protein construct was shown in mice: injection of the modified APOL1 resulted in inhibition of parasitaemia caused by either T. b. brucei or T. b. rhodesiense72. APOL1 might also be useful for solving the problem of nagana, a lethal disease that is caused by infection of cattle with T. b. brucei. Transgenic cattle expressing APOL1 derivatives such as the truncated-APOL1–nanobody conjugate would be expected to resist infection by T. b. brucei and T. b. rhodesiense, because this truncated trypanolytic factor cannot be neutralized by the T. b. rhodesiense ‘antidote’ SRA. As well as allowing healthy animal production in areas in which T. b. rhodesiense is endemic, the use of such ‘trypanolytic cattle’ should lead to a marked reduction in the main reservoir of T. b. rhodesiense, thereby contributing to the prevention of epidemics of sleeping sickness. Etienne Pays, Benoit Vanhollebeke, Luc Vanhamme, Françoise Paturiaux-Hanocq and David Pérez-Morga are at the Laboratory of Molecular Parasitology, Institute of Molecular Biology and Medicine (IBMM), Université Libre de Bruxelles, 12 rue des Professeurs Jeener et Brachet, B-6041 Gosselies, Belgium.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
Derek P. Nolan is at the Department of Biochemistry, Trinity College, Dublin 2, Ireland.
1.
2. 3.
4. 5.
6.
7.
8.
Correspondence to E.P. e-mail:
[email protected]
24.
doi:10.1038/nrmicro1428
25.
Blum, M. L. et al. A structural motif in the variant surface glycoproteins of Trypanosoma brucei. Nature 362, 603–609 (1993). Cross, G. A. M. Antigenic variation in trypanosomes. Proc. R. Soc. Lond. B 202, 55–72 (1978). Barry, J. D. The relative significance of mechanisms of antigenic variation in African trypanosomes. Parasitol. Today 13, 212–218 (1997). Borst, P. Antigenic variation and allelic exclusion. Cell 109, 5–8 (2002). Pays, E., Vanhamme, L. & Pérez-Morga, D. Antigenic variation in Trypanosoma brucei: facts, challenges and mysteries. Curr. Opin. Microbiol. 7, 369–374 (2004). Laveran, A. & Mesnil, F. in Trypanosomes et Trypanosomiases 126–183 (Libraires de l’Académie de Médecine, Paris, 1912) (in French). Seed, J. R., Sechelski, J. B. & Loomis, M. R. A survey for a trypanocidal factor in primate sera. J. Protozool. 37, 393–400 (1990). Poelvoorde, P., Vanhamme, L., Van Den Abbeele, J., Switzer, W. M. & Pays, E. Distribution of apolipoprotein L-I and trypanosome lytic activity among primate sera. Mol. Biochem. Parasitol. 134, 155–157 (2004).
NATURE REVIEWS | MICROBIOLOGY
26.
27.
28.
29.
30.
Lugli, E. B., Pouliot, M., Portela, M. P., Loomis, M. R. & Raper, J. Characterization of primate trypanosome lytic factors. Mol. Biochem. Parasitol. 138, 9–20 (2004). Rifkin, M. R. Identification of the trypanocidal factor in normal human serum: high density lipoprotein. Proc. Natl Acad. Sci. USA 75, 3450–3454 (1978). Rifkin, M. R. Role of phospholipids in the cytotoxic action of high density lipoprotein on trypanosomes. J. Lipid Res. 32, 639–647 (1991). Hajduk, S. L. et al. Lysis of Trypanosoma brucei by a toxic subspecies of human high-density lipoprotein. J. Biol. Chem. 264, 5210–5217 (1989). Gillett, M. P. & Owen, J. S. Trypanosoma brucei brucei: differences in the trypanocidal activity of human plasma and its relationship to the level of high density lipoproteins. Trans. R. Soc. Trop. Med. Hyg. 85, 612–616 (1991). Gillett, M. P. & Owen, J. S. Comparison of the cytolytic effects in vitro on Trypanosoma brucei brucei of plasma, high density lipoproteins, and apolipoprotein A-I from hosts both susceptible (cattle and sheep) and resistant (human and baboon) to infection. J. Lipid Res. 33, 513–523 (1992). Lorenz, P., Owen, J. S. & Hassall, D. G. Human serum resistant Trypanosoma brucei rhodesiense accumulates similar amounts of fluorescently-labelled trypanolytic human HDL3 particles as human serum sensitive T. b. brucei. Mol. Biochem. Parasitol. 74, 113–118 (1995). Hager, K. M. et al. Endocytosis of a cytotoxic human high density lipoprotein results in disruption of acidic intracellular vesicles and subsequent killing of African trypanosomes. J. Cell Biol. 126, 155–167 (1994). Lorenz, P., Barth, P. E., Rudin, W. & Betschart, B. Importance of acidic intracellular compartments in the lysis of Trypanosoma brucei brucei by normal human serum. Trans. R. Soc. Trop. Med. Hyg. 88, 487–488 (1994). Ortiz-Ordonez, J. C., Sechelski, J. B. & Seed, J. R. Mechanism of lysis of Trypanosoma brucei gambiense by human serum. J. Parasitol. 80, 924–930 (1994). Hager, K. M. & Hajduk, S. L. Mechanism of resistance of African trypanosomes to cytotoxic human HDL. Nature 385, 823–826 (1997). Shimamura, M., Hager, K. M. & Hajduk, S. L. The lysosomal targeting and intracellular metabolism of trypanosome lytic factor by Trypanosoma brucei brucei. Mol. Biochem. Parasitol. 115, 227–237 (2001). Vanhamme, L. & Pays, E. The trypanosome lytic factor of human serum and the molecular basis of sleeping sickness. Int. J. Parasitol. 34, 887–898 (2004). Garcia-Salcedo, J. A. et al. A differential role for actin during the life cycle of Trypanosoma brucei. EMBO J. 23, 780–789 (2004). Green, H. P., del Pilar Molina Portela, M., St Jean, E. N., Lugli, E. B. & Raper, J. Evidence for a Trypanosoma brucei lipoprotein scavenger receptor. J. Biol. Chem. 278, 422–427 (2003). Smith, A. B, Esko, J. D. & Hajduk, S. L. Killing of trypanosomes by the human haptoglobin-related protein. Science 268, 284–286 (1995). Raper, J., Nussenzweig, V. & Tomlinson, S. The main lytic factor of Trypanosoma brucei brucei in normal human serum is not high density lipoprotein. J. Exp. Med. 183, 1023–1029 (1996). Tomlinson, S., Muranjan, M., Nussenzweig, V. & Raper, J. Haptoglobin-related protein and apolipoprotein AI are components of the two trypanolytic factors in human serum. Mol. Biochem. Parasitol. 86, 117–120 (1997). Muranjan, M., Nussenzweig, V. & Tomlinson, S. Characterization of the human serum trypanosome toxin, haptoglobin-related protein. J. Biol. Chem. 273, 3884–3887 (1998). Rifkin, M. R. Trypanosoma brucei: cytotoxicity of host high-density lipoprotein is not mediated by apolipoprotein A-I. Exp. Parasitol. 72, 216–218 (1991). Owen, J. S., Gillett, M. P. & Hughes, T. E. Transgenic mice expressing human apolipoprotein A-I have sera with modest trypanolytic activity in vitro but remain susceptible to infection by Trypanosoma brucei brucei. J. Lipid Res. 33, 1639–1646 (1992). Tomlinson, S. et al. High-density-lipoproteinindependent killing of Trypanosoma brucei by human serum. Mol. Biochem. Parasitol. 70, 131–138 (1995).
VOLUME 4 | JUNE 2006 | 485 © 2006 Nature Publishing Group
PERSPECTIVES 31. Smith, A. B. & Hajduk, S. L. Identification of haptoglobin as a natural inhibitor of trypanocidal activity in human serum. Proc. Natl Acad. Sci. USA 92, 10262–10266 (1995). 32. Raper, J., Nussenzweig, V. & Tomlinson, S. Lack of correlation between haptoglobin concentration and trypanolytic activity of normal human serum. Mol. Biochem. Parasitol. 76, 337–338 (1996). 33. Raper, J., Fung, R., Ghiso, J., Nussenzweig, V. & Tomlinson, S. Characterization of a novel trypanosome lytic factor from human serum. Infect. Immun. 67, 1910–1916 (1999). 34. Barker, C., Barbour, K. W., Berger, F. G. & Hajduk, S. L. Activity of human trypanosome lytic factor in mice. Mol. Biochem. Parasitol. 117, 129–136 (2001). 35. McEvoy, S. M. & Maeda, N. Complex events in the evolution of the haptoglobin gene cluster in primates. J. Biol. Chem. 263, 15740–15747 (1988). 36. Drain, J., Bishop, J. R. & Hajduk, S. L. Haptoglobinrelated protein mediates trypanosome lytic factor binding to trypanosomes. J. Biol. Chem. 276, 30254–30260 (2001). 37. Bishop, J. R., Shimamura, M. & Hajduk, S. L. Insight into the mechanism of trypanosome lytic factor-1 killing of Trypanosoma brucei brucei. Mol. Biochem. Parasitol. 118, 33–40 (2001). 38. Molina Portela, M. P., Raper, J. & Tomlinson, S. An investigation into the mechanism of trypanosome lysis by human serum factors. Mol. Biochem. Parasitol. 110, 273–282 (2000). 39. Van Meirvenne, N., Magnus, E. & Janssens, P. G. The effect of normal human serum on trypanosomes of distinct antigenic type (ETat 1 to 12) isolated from a strain of Trypanosoma brucei rhodesiense. Ann. Soc. Belg. Med. Trop. 56, 55–63 (1976). 40. Xong, H. V. et al. A VSG expression site-associated gene confers resistance to human serum in Trypanosoma rhodesiense. Cell 95, 839–846 (1998). 41. Pays, E., Lips, S., Nolan, D., Vanhamme, L. & Pérez-Morga, D. The VSG expression sites of Trypanosoma brucei: multipurpose tools for the adaptation of the parasite to mammalian hosts. Mol. Biochem. Parasitol. 114, 1–16 (2001). 42. De Greef, C., Imberechts, H., Matthyssens, G., Van Meirvenne, N. & Hamers, R. A gene expressed only in serum-resistant variants of Trypanosoma brucei rhodesiense. Mol. Biochem. Parasitol. 36, 169–176 (1989). 43. De Greef, C. & Hamers, R. The serum resistanceassociated (SRA) gene of Trypanosoma brucei rhodesiense encodes a variant surface glycoprotein-like protein. Mol. Biochem. Parasitol. 68, 277–284 (1994). 44. Campillo, N. & Carrington, M. The origin of the serum resistance associated (SRA) gene and a model of the structure of the SRA polypeptide from Trypanosoma brucei rhodesiense. Mol. Biochem. Parasitol. 127, 79–84 (2003). 45. Vanhamme, L. et al. The Trypanosoma brucei reference strain TREU927/4 contains T. b. rhodesiense-specific SRA sequences, but displays a distinct phenotype of relative resistance to human serum. Mol. Biochem. Parasitol. 135, 39–47 (2004). 46. Milner, J. D. & Hajduk, S. L. Expression and localization of serum resistance associated protein in Trypanosoma brucei rhodesiense. Mol. Biochem. Parasitol. 104, 271–283 (1999). 47. Welburn, S. C. et al. Identification of human-infective trypanosomes in animal reservoir of sleeping sickness in Uganda by means of serum-resistanceassociated (SRA) gene. Lancet 358, 2017–2019 (2001). 48. Gibson, W., Backhouse, T. & Griffiths, A. The human serum resistance associated gene is ubiquitous and conserved in Trypanosoma brucei rhodesiense throughout East Africa. Infect. Genet. Evol. 1, 207–214 (2002). 49. Radwanska, M. et al. The serum resistance-associated gene as a diagnostic tool for the detection of Trypanosoma brucei rhodesiense. Am. J. Trop. Med. Hyg. 67, 684–690 (2002).
50. Gibson, W. & Ferris, V. Conservation of the genomic location of the human serum resistance associated gene in Trypanosoma brucei rhodesiense. Mol. Biochem. Parasitol. 130, 159–162 (2003). 51. Gibson, W. C. The SRA gene: the key to understanding the nature of Trypanosoma brucei rhodesiense. Parasitology 131, 143–150 (2005). 52. Vanhamme, L. et al. Differential RNA elongation controls the variant surface glycoprotein gene expression sites of Trypanosoma brucei. Mol. Microbiol. 36, 328–340 (2000). 53. De Greef, C., Chimfwembe, E., Kihang’a Wabacha, J., Bajyana Songa, E. & Hamers, R. Only the serumresistant bloodstream forms of Trypanosoma brucei rhodesiense express the serum resistance associated (SRA) protein. Ann. Soc. Belg. Med. Trop. 72 (Suppl. 1), 13–21 (1992). 54. Vanhamme, L. et al. Apolipoprotein L-I is the trypanosome lytic factor of human serum. Nature 422, 83–87 (2003). 55. Triggs, V. P. & Bangs, J. D. Glycosylphosphatidylinositoldependent protein trafficking in bloodstream stage Trypanosoma brucei. Eukaryot. Cell 2, 76–83 (2003). 56. Wang, J., Böhme, U. & Cross, G. A. M. Structural features affecting variant surface glycoprotein expression in Trypanosoma brucei. Mol. Biochem. Parasitol. 128, 135–145 (2003). 57. Duchateau, P. N. et al. Apolipoprotein L, a new human high density lipoprotein apolipoprotein expressed by the pancreas. Identification, cloning, characterization, and plasma distribution of apolipoprotein L. J. Biol. Chem. 272, 25576–25582 (1997). 58. Mikkelsen, T. S. et al. Initial sequence of the chimpanzee genome and comparison with the human genome. Nature 437, 69–87 (2005). 59. Oli, M. W., Cotlin, L. F., Shiflett, A. M. & Hajduk, S. L. Serum resistance-associated protein blocks lysosomal targeting of trypanosome lytic factor in Trypanosoma brucei. Eukaryot. Cell 5, 132–139 (2006). 60. Moore, D. R. et al. Developmentally regulated sensitivity of Trypanosoma brucei brucei to the cytotoxic effects of human high-density lipoprotein. Exp. Parasitol. 81, 216–226 (1995). 61. Brun, R. & Jenni, L. Human serum resistance of metacyclic forms of Trypanosoma brucei brucei, T. brucei rhodesiense and T. brucei gambiense. Parasitol. Res. 73, 218–223 (1987). 62. Duchateau, P. N. et al. Plasma apolipoprotein L concentrations correlate with plasma triglycerides and cholesterol levels in normolipidemic, hyperlipidemic, and diabetic subjects. J. Lipid Res. 41, 1231–1236 (2000). 63. Mimmack, M. L. et al. Gene expression analysis in schizophrenia: reproducible up-regulation of several members of the apolipoprotein L family located in a high-susceptibility locus for schizophrenia on chromosome 22. Proc. Natl Acad. Sci. USA 99, 4680–4685 (2002). 64. Pérez-Morga, D. et al. Apolipoprotein L-I promotes trypanosome lysis by forming pores in lysosomal membranes. Science 309, 469–472 (2005). 65. Raper, J., Molina Portela, M. P., Lugli, E., Frevert, U. & Tomlinson, S. Trypanosome lytic factors: novel mediators of innate immunity. Curr. Opin. Microbiol. 4, 402–408 (2001). 66. Nolan, D. P. & Voorheis, H. P. Factors that determine the plasma-membrane potential in bloodstream forms of Trypanosoma brucei. Eur. J. Biochem. 267, 4615–4623 (2000). 67. Rifkin, M. R. Trypanosoma brucei: biochemical and morphological studies of cytotoxicity caused by normal human serum. Exp. Parasitol. 58, 81–93 (1984). 68. Cardoso de Almeida, M. L., Geuskens, M. & Pays, E. Cell lysis induces redistribution of the GPI-anchored variant surface glycoprotein on both faces of the plasma membrane of Trypanosoma brucei. J. Cell Sci. 112, 4461–4473 (1999). 69. del Pilar Molina-Portela, M., Lugli, E. B., Recio-Pinto, E. & Raper, J. Trypanosome lytic factor, a subclass of high-density lipoprotein, forms cation-selective pores in membranes. Mol. Biochem. Parasitol. 144, 218–226 (2005).
486 | JUNE 2006 | VOLUME 4
70. Shiflett, A. M., Bishop, J. R., Pahwa, A. K. & Hajduk, S. L. Human HDLs are platforms for the assembly of multi-component innate immune complexes. J. Biol. Chem. 280, 32578–32585 (2005). 71. McGwire, B. S., Olson, C. L., Tack, B. F. & Engman, D. M. Killing of African trypanosomes by antimicrobial peptides. J. Infect. Dis. 188, 146–152 (2003). 72. Baral, T. N. et al. Experimental therapy of African trypanosomiasis with a nanobody-conjugated human trypanolytic factor. Nature Med. 12, 580–584 (2006). 73. Hatada, S. et al. No trypanosome lytic activity in the sera of mice producing human haptoglobin-related protein. Mol. Biochem. Parasitol. 119, 291–294 (2002). 74. Stijlemans, B. et al. Efficient targeting of conserved cryptic epitopes of infectious agents by single domain antibodies. African trypanosomes as paradigm. J. Biol. Chem. 279, 1256–1261 (2004). 75. Monajemi, H., Fontijn, R. D., Pannekoek, H. & Horrevoets, A. J. The apolipoprotein L gene cluster has emerged recently in evolution and is expressed in human vascular tissue. Genomics 79, 539–546 (2002). 76. Page, N. M., Butlin, D. J., Lomthaisong, K. & Lowry, P. J. The human apolipoprotein L gene cluster: identification, classification, and sites of distribution. Genomics 74, 71–78 (2001). 77. Duchateau, P. N., Pullinger, C. R., Cho, M. H., Eng, C. & Kane, J. P. Apolipoprotein L gene family: tissuespecific expression, splicing, promoter regions; discovery of a new gene. J. Lipid Res. 42, 620–630 (2001). 78. Horrevoets, A. J. et al. Vascular endothelial genes that are responsive to tumor necrosis factor-α in vitro are expressed in atherosclerotic lesions, including inhibitor of apoptosis protein-1, stannin, and two novel genes. Blood 93, 3418–3431 (1999). 79. Liu, Z., Lu, H., Jiang, Z., Pastuszyn, A. & Hu, C. A. Apolipoprotein L6, a novel proapoptotic Bcl-2 homology 3-only protein, induces mitochondriamediated apoptosis in cancer cells. Mol. Cancer Res. 3, 21–31 (2005). 80. Lazebnik, Y. Why do regulators of apoptosis look like bacterial toxins? Curr. Biol. 11, R767–R768 (2001). 81. Concha, M. I. et al. Apolipoproteins A-I and A-II are potentially important effectors of innate immunity in the teleost fish Cyprinus carpio. Eur. J. Biochem. 271, 2984–2990 (2004). 82. Sorensen, O., Bratt, T., Johnsen, A. H., Madsen, M. T. & Borregaard, N. The human antibacterial cathelicidin, hCAP-18, is bound to lipoproteins in plasma. J. Biol. Chem. 274, 22445–22451 (1999).
Acknowledgements This paper is dedicated to the memory of M. Steinert. This work was supported by the Fonds National de la Recherche Scientifique (FNRS), the United Nations Children’s Fund, United Nations Development Programme, World Bank, and World Health Organization Special Programme for Research and Training in Tropical Diseases (TDR), and the Interuniversity Attraction Poles Programme (Belgian Science Policy).
Competing interests statement The authors declare no competing financial interests.
DATABASES The following terms in this article are linked online to: Entrez Genome Project: http://www.ncbi.nlm.nih.gov/ entrez/query.fcgi?db=genomeprj T. brucei UniProtKB: http://ca.expasy.org/sprot APOA1 | APOL1 | HPR
FURTHER INFORMATION Etienne Pays’s laboratory: http://www.ulb.ac.be/ibmm/homeuk_13.html
SUPPLEMENTARY INFORMATION See online article: S1 (figure) | S2 (movie) | S3 (figure) Access to this links box is available online.
www.nature.com/reviews/micro © 2006 Nature Publishing Group
