In this issue
M
ore than 1,400 bacterial, viral and parasitic species are known to cause human disease. Understanding how these pathogens colonize, survive and inflict disease remains one of the key goals of microbiology. This, the first issue
of 2008, features several Reviews that discuss microbial virulence. On page 79, John Boothroyd and Jean-Francois Dubremetz discuss
the specialized rhoptry organelle, which exists in apicomplexan parasites such as Toxoplasma gondii. In T. gondii, recent findings indicate that these organelles contain the molecular machinery that enables the parasite to affect host gene expression and co-opt host functions. Dealing with bacteria, Samuel Miller and colleagues discuss the
▶ cover: ‘A to B’ by George Marshall, inspired by the Review on p28.
mechanisms that enable Salmonellae to interact with, and manipulate, host cells (page 53). Specifically, they focus on the interplay between various bacterial and host proteins that enables the invasion of epithelial cells, stimulation and repression of signalling cascades, sensing of the intracellular environment and establishment of a niche for intracellular replication. So, how can we harness our understanding of microbial virulence to stem
david o’connell
the continuing rise in infectious-disease-related morbidity and mortality?
susan jones
Antibiotic-based strategies, which typically target bacterial viability, have resulted in widespread resistance — meticillin-resistant Staphylococcus aureus infections reached epidemic levels last autumn in some parts of the sheilagh molloy
United States. Perhaps rather than targeting bacteria for eradication, we
sharon Ahmad
should focus instead on de-clawing pathogens by inhibiting the virulence mechanisms that promote infection or that cause disease symptoms. On page 17, Scott Hultgren and colleagues highlight various bacterial virulence mechanisms that could be targeted and consider the recent efforts towards, and the remaining challenges that face, antivirulence-based drug discovery.
Editorial offices
editorial assistants: Laura Firman,
The Macmillan Building, 4 Crinan Street, London N1 9XW, UK Tel: +44 (0)20 7843 3620; Fax: +44 (0)20 7843 3629 chief editor: David O’Connell senior editors: Susan Jones, Sheilagh Molloy associate editor: Sharon Ahmad assistant editor: Asher Mullard copy editor: Gillian Young art editor: George Marshall
web production manager, UK:
london
[email protected]
editorial support manager: Elinor Faulkner
Robert Monk
Alexander Thurrell
marketing managers: Kellie Lane, Amy Maurer
production Director: Yvonne Strong
director, web publishing: Timo Hannay
head of web production: Jeremy Macdonald
Management offices london
[email protected]
The Macmillan Building, 4 Crinan Street, London N1 9XW, UK Tel: +44 (0)20 7833 4000; Fax: +44 (0)20 7843 4596/7 office manager: Sheryl Ocampo publisher: Hugh Blackbourn
new york
[email protected] Nature Publishing Group, 75 Varick Street, 9th floor, New York, NY 10013-1917, USA Tel: +1 212 726 9200; Fax: +1 212 696 9006
tokyo
[email protected]
Chiyoda Building 5F, 2-37-1 Ichigayatamachi, Shinjuku-ku, Tokyo 162-0843, Japan Tel: +81 3 3267 8751; Fax: +81 3 3267 8746 asia–pacific publisher: Antoine E Bocquet manager: Koichi Nakamura
asia–pacific sales director: Kate Yoneyama
senior marketing manager: Peter Yoshihara
marketing/production manager: Takesh Murakami
india 5A/12 Ansari Road, Daryganj,
managing director:
Howard Ratner
New Delhi 110 002, India Tel/Fax: +91 11 2324 4186
Patrick Morgan
editor-in-chief, nature publications:
Greg Suprock
Harpal Singh Gill
Judith Shadwell
associate directors:
Simon Fenwick
editorial production director:
senior Copy editor: Carrie Patis art controller: Susanne Harris senior art editors: Vicky Askew, managing production editor: deputy production editor: production controller: Natalie Smith
chief technology officer:
Steven Inchcoombe
director of new technology:
Philip Campbell
head of web services:
Jenny Henderson, Tony Rudland
naturejobs publisher: Ben Crowe head of nature research & reviews marketing: Sara Girard
James McQuat
Anthony Barrera
nature reviews | microbiology
sales and marketing manager, india: Copyright © 2008 Nature Publishing Group Research Highlight images courtesy of Getty Images unless otherwise credited. Printed in Wales by Cambrian Printers on acid-free paper
volume 6 | january 2008 | © 2008 Nature Publishing Group
Editorial The value of vaccines Our ability to control infectious diseases is continuously being eroded by antimicrobial resistance, the decline in industrial antibiotic development and increasing development timelines from discovery to market. There has never been a better time to rediscover the value of vaccination.
the direct economic value that is associated with vaccines is negligible compared with that of pharmaceutical drugs.
As a preventative strategy in the fight against infectious diseases, vaccination is considered to be the most costeffective medical intervention. Indeed, in many developed countries, measles, mumps, rubella, hepatitis B, Haemophilus influenzae type b, tetanus, pertussis and diphtheria have been controlled or nearly eliminated as chief causes of morbidity and mortality by the use of vaccines. Yet, despite their role in delivering many of the successes that have been achieved in infectious disease control, there seems to be relatively little enthusiasm for vaccine development among those who have the capability for vaccine development and production. There are a number of contributing factors to this lack of enthusiasm, most of which are linked to economics. The simple truth is that the direct economic value that is associated with vaccines is negligible compared with that of pharmaceutical drugs. Globally, potential vaccine sales are approximately US$6 billion per year. Although the vaccines sector is out-performing the rest of the industry in terms of revenue growth, it still represents less than 2% of the worldwide pharmaceutical market. Because of the low return for their investment, high regulatory costs, uncertain market conditions and exposure to legal liability, most pharmaceutical manufacturers do not consider the development and production of vaccines as an attractive business opportunity. For example, over the past 30 years, the number of companies that distribute vaccines in the United States has decreased from 30 to 5, a situation that has directly contributed to serious shortages of influenza, tetanus–diphtheria, measles–mumps–rubella, pneumococcal, meningococcal and other vaccines. Not surprisingly, there is even less incentive for the vaccine industry to develop new vaccines against diseases that are largely limited to the developing world. Governments do have a number of well-documented and under-used tools at their disposal to make vaccines more attractive to industry. These options include: tax breaks or subsidies to reduce research and development expenses; extending patent protection for intellectual property that is related to vaccines of public health importance; working with industry and others to find ways to reduce the costs of meeting regulatory requirements; and allowing tiered pricing whereby vaccines can be sold at
| january 2008 | volume 6
higher prices in developed countries and lower prices in under-developed countries. It should also be possible to decrease liability risks (by measures such as the Vaccine Injury Compensation Program in the United States) and protect manufacturers from lawsuits that are related to the unanticipated adverse effects of a properly manufactured, safe and effective vaccine. Ultimately, however, an economic solution is required. There have been numerous demonstrations of the cost-effectiveness of immunization. Attributing the true economic value to vaccines — both real and intangible — will create a self-sustaining system to ensure the development and adequate supply of vaccines to those that need them most. But how can governments be mobilized into action? If the ever-increasing threat of a public health calamity does not provide the necessary incentive, the promise of success might do the trick. An example that illustrates the effectiveness of a global approach to vaccination is the campaign for the eradication of poliomyelitis. Launched in 1985 for South America, it was taken up by the World Assembly, which in 1988 committed to the global eradication of poliovirus by the year 2000. Progress towards eradication of the virus has been fast. In 1988, 125 countries in 5 continents reported endemic poliovirus but by 2003 only 6 polio-endemic countries were reported and, officially, only 4 remain: India, Pakistan, Afghanistan and Nigeria. So, although the 2000 deadline passed with the incidence of disease stalled at around 2,000 cases per year, and there have been setbacks, including a recent outbreak in Nigeria, the initiative is an unequivocal example of how a globally coordinated effort can, within a short period of time, reduce the incidence of an infectious disease by more than 99.9%. The elimination of the disease from the remaining endemic regions is now achievable, leaving the welcome problem of whether, and how long, vaccination should be continued against a disease that no longer exists. In their efforts to control infectious diseases, governments and non-governmental organizations need look no further than the poliomyelitis eradication campaign, both for their inspiration and justification in taking that first important step — placing vaccination back where it belongs, at the top of the global public health agenda. www.nature.com/reviews/micro
© 2008 Nature Publishing Group
Research highlights
hi v
Infection, superinfection and (lack of) protection Superinfection with HIV-1 occurs when an individual who is already infected with one strain of HIV-1 acquires a second strain from a different partner. Because a re-infection event suggests that the immune response generated against the original infection is not sufficient to protect against later exposures, there are obvious implications for HIV-1 vaccine design. However, despite the potential importance of superinfection in HIV disease and vaccine development, there is uncertainty regarding its incidence and timing. Now, an in-depth, population-level assessment of HIV-1 superinfection, published in PLoS Pathogens, confirms that natural HIV-1 infection does not always elicit a protective immune response, and that this lack of protection is largely independent
of the timing of re-exposure and the relatedness of the virus strains. Prior to this study, approximately twenty cases of HIV-1 superinfection had been reported in the literature, which suggested that natural infection does not always generate a protective immune response. However, if researchers are to accurately assess the impact of superinfection on the success of an HIV-1 vaccine, questions need to be answered about the frequency of superinfection and whether this process is restricted to the times in infection when an HIV‑1-specific immune response has not yet developed. To address these issues, Piantadosi and colleagues screened a cohort of high-risk Kenyan women for HIV-1 superinfection by comparing 2 partial gene sequences (gag and envelope) over a 5-year period, beginning at the time
…this lack of protection is largely independent of the timing of reexposure and the relatedness of the virus strains.
CTL NtAb HIV-1 RNA
0
1
2
3
Time (years post-infection)
4
5
Superinfection with an HIV-1 virus from a second partner (pink) can occur several years after an initial HIV-1 infection (green). At this time (see arrow), cytotoxic T-cell responses (CTL) and neutralizing antibody responses (NTAb) to the first infection have had time to develop. | Microbiology Figure redrawn, with permission, from one kindly provided by Julie Overbaugh, Fred Hutchinson CancerNature ResearchReviews Center, Washington, USA.
nature reviews | microbiology
of primary infection. Analysing more than one region of the HIV-1 genomes allowed the authors to detect cases in which the initial and superinfecting viruses had recombined, a frequent event during HIV-1 replication. By analysing the proviral sequences from 36 women infected with HIV-1, seven cases of superinfection were detected, including 3 cases in which both viruses belonged to the same HIV-1 subtype; this finding indicates that closely related viruses do not confer protection. The authors also showed that in 5 of the 7 cases, infection with the second virus occurred over 1 year after the initial infection, and, in some cases, 5 years after initial infection. In other words, superinfection occurred after the host immune response to the first HIV-1 strain should have had time to develop and mature. This study clearly demonstrates that HIV-positive individuals are at continued risk of acquiring a second HIV infection. Indeed, if the lack of protection against re-infection is shared by at least a proportion of all HIV-infected individuals, the potential value of an HIV-1 vaccine that attempts to mimic the immune response to natural infection is called into question. Follow-up studies to compare the immune responses of those individuals that become superinfected with those that do not, and deciphering the contribution of these responses to immune protection, will be the focus of future research.
David O’Connell
ORIGINAL RESEARCH PAPER Piantadosi, A. et al. Chronic HIV-1 infection frequently fails to protect against superinfection. PLoS Pathog. 3, e177 (2007)
volume 6 | january 2008 © 2008 Nature Publishing Group
Research highlights
MicroBIAL Physiology
How low can you grow? Removal of the greenhouse gas methane by methanotrophic bacteria is an important climatic process. All methanotrophs characterized so far are meso- or thermophilic members of the phylum Proteobacteria. Now, two papers published in Nature report the cultivation, characterization and draft genome analysis of two new methanotrophs from the phylum Verrucomicrobia that are the most acidophilic methanotrophs ever studied. Some environments that are rich in methane are also extremely acidic, so two groups of researchers set out to investigate whether novel methanotrophs thrive in these conditions. PCR primer sets for the conserved pmoA gene, which encodes a subunit of the particulate form of methane monooxygenase, have been used routinely to survey numerous environments for methanotrophs. Pol et al. extracted DNA from a hot, acidic, methaneproducing fumarole in the Solfatara, Italy, and used primers for pmoA to construct an environmental clone library. They found two classes of pmoA sequences in the library — gammaproteobacterial-like sequences and a divergent set of pmoA homologues. Intrigued by these unusual pmoA sequences, they isolated a strain into pure culture that had a divergent pmoA sequence that was similar to those present in the clone library and a 16S rRNA sequence that was typical of Verrucomicrobia, and named it SolV (provisional name Acidimethylosilex fumarolicum).
In a parallel study, Dunfield et al. analysed 16S rRNA genes of bacteria in a sample from Hell’s Gate, an acidic geothermal site in New Zealand that is rich in methane, and identified novel Verrucomicrobia sequence signatures that were unrelated to any cultured bacteria. A strain named Verrucomicrobia isolate V4 (provisional name Methylokorus infernorum) was isolated into pure culture from this sample by supplying methane as the sole carbon and energy source. Both SolV and V4 have one striking phenotype: V4 can grow at pH 1.0 and SolV can grow at pH 0.8. This is the first time that extremely acidophilic methanotrophs have been described. Draft genomes of both new species were assembled and scrutinized for insights into their physiologies; both are remarkable in that they have three pmoCAB operons that are completely different. Most proteobacterial methanotrophs typically encode up
nature reviews | microbiology
to three similar pmoCAB operons. Phylogenetic analyses of the three pmoA alleles encoded by V4 and those encoded by SolV indicate that these acidophilic lineages diverged from Proteobacteria a long time ago, and exclude the possiblity that Verrucomicrobia species have acquired the pmoA gene by a recent horizontal-gene-transfer event. Inspection of the genomes indicated that both species might also use novel methylotrophic pathways. These studies have provided an important step forward in the understanding of the physiology and diversity of methane consumption by bacteria in natural environments.
Susan Jones
ORIGINAL RESEARCH PAPERS Pol, A. et al. Methanotrophy below pH1 by a new Verrucomicrobia species. Nature 14 Nov 2007 (doi 10.1038/nature06222) | Dunfield, P. F. et al. Methane oxidation by an extremely acidophilic bacterium of the phylum Verrucomicrobia. Nature 14 Nov 2007 (doi 10.1038/nature06411)
volume 6 | january 2008 © 2008 Nature Publishing Group
Research highlights
RNA
Extending the network of sRNA control Mechanisms by which the small non-coding RNAs (sRNAs) of Gram-negative bacteria regulate multiple mRNA targets have proved difficult to pin down, owing to the limited complementarity between sRNAs and target mRNAs. In a paper published in Genes & Development, Jörg Vogel and colleagues report that GcvB, a 200-nucleotide sRNA, directly regulates 7 mRNAs that encode periplasmic peptide and aminoacid transport proteins in Salmonella enterica serovar Typhimurium (S. typhimurium). The GcvB sRNA represses translation of the dppA and oppA peptide transporter genes in nutrient-rich conditions. Intriguingly, gcvB mutants are pleiotropic, and computational analyses indicated that GcvB might have numerous target genes, which led Vogel and colleagues to investigate how the GcvB sRNA functions. RNA-structure probing pinpointed a 29-nucleotide G- and U-rich linker region in GcvB, named R1, that interacts with sequences that span the ribosome binding site (RBS) of both peptide transporter genes. Strikingly, deletion of R1 destroyed the ability of GcvB to repress dppA and oppA. Bioinformatic searches revealed that R1 is well-conserved among the gcvB genes of Pasteurella spp., Vibrio spp. and many enterobacteria. By combining scrutiny of the
periplasmic-protein profiles of an S. typhimurium strain that is overexpressing gcvB with an in silico screen for binding partners of R1, Vogel and co-workers identifed five additional candidate GcvB targets that encode periplasmic amino-acid transport proteins — gltI, livJ, livK, argT and STM4351. The predicted GcvB-binding sites in these genes were not conserved, but were rich in C and A residues. Biochemical assays revealed that GcvB functions as a sequence-specific inhibitor of translation, but the details of inhibition vary among the different target mRNAs. GcvB binds to the RBS of dppA and oppA mRNAs to repress
nature reviews | microbiology
translation by occlusion of the ribosome. Surprisingly, GcvB binds to sequences upstream of the RBS of other regulated genes, but still prevents the 30S ribosome from docking onto the mRNA to initiate translation. Fusion of the 5′-untranslated region of gltI, which contains the GcvB-docking site, with the unrelated E. coli ompR gene conferred GcvBmediated repression to ompR, proving that GcvB can regulate any mRNA that contains a cognate Cand A‑rich R1-binding motif. The precise mechanism of this novel type of translational control remains unclear. Other studies have shown that an sRNA (514 nucleotides) named RNAIII functions to repress multiple virulence factors in the Gram-positive pathogen Staphylococcus aureus by complementary base-pairing with the RBS of target-gene mRNAs. Together, these studies show that specific interactions between sRNAs and mRNAs in both Gram-positive and Gram-negative bacteria can regulate multiple genes.
Susan Jones
ORIGINAL RESEARCH PAPER Sharma, C. M., Darfeuille, F., Plantinga, T. H. & Vogel, J. A small RNA regulates multiple ABC transporter mRNAs by targeting C/A-rich elements inside and upstream of ribosome-binding sites. Genes Dev. 21, 2804–2817 (2007)
volume 6 | january 2008 © 2008 Nature Publishing Group
Research highlights
Host response
IgA — peacemaker in the gut
Using a novel gnotobiotic mouse model in which the diverse gut microbiota is reduced to a single bacterial species, and the antibody repertoire to a single monoclonal immunoglobulin A (IgA) antibody that is directed against the bacterium’s capsular polysaccharide, Jeffrey Gordon and colleagues provide evidence that supports a role for IgA in the gut as a mediator of tolerance. The IgA antibody response has a key role in establishing and maintaining a non-inflammatory relationship between the host and microbiota in the gut. Germ-free mice that are colonized with normal gut microbiota develop bacteria-specific IgA antibody responses, but the effects of these responses on the biology of the host and microbiota are not well defined. Peterson et al. developed a gnotobiotic mouse model to study these effects. As their model symbiont, they chose the bacterium Bacteroides
thetaiotaomicron, which is a prominent, obligately anaerobic, Gramnegative member of the human distal intestinal ecosystem that also efficiently colonizes the intestines of adult germ-free C57BL/6J mice. B. thetaiotaomicron was introduced into germ-free, recombinationactivating-gene-1-deficient (Rag1–/–) mice (which lack mature B and T cells) or germ-free Rag1–/– mice that had been injected under their dorsal skin with B. thetaiotaomicron-primed IgA-producing hybridoma cells. An inverse relationship was found between IgA antibody levels and the levels of its B. thetaiotaomicron epitope in the intestinal lumen — that is, infected mice with IgA secreted by the hybridoma cells had lower levels of epitope expression than mice without IgA. In the absence of IgA, B. thetaiotaomicron elicited a more robust innate immune oxidative response, and
nature reviews | microbiology
adapted to this response by inducing genes that are involved in the metabolism of oxidative products of the host response. The presence of IgA, however, reduced intestinal pro-inflammatory signalling (by downregulating signal transducer and activator of transcription 3 and interferon regulatory factor 8, for example), and bacterial epitope expression. Therefore, these results suggest that it is the IgA antibody response that establishes a quiescent relationship between B. thetaiotaomicron and its host. They are also consistent with a model in which a set of adaptations that involve both the symbiont and its host lead to co-evolved homeostasis.
Sharon Ahmad
ORIGINAL RESEARCH PAPER Peterson, D. A. et al. IgA response to symbiotic bacteria as a mediator of gut homeostasis. Cell Host Microbe 2, 328–339 (2007)
volume 6 | january 2008 © 2008 Nature Publishing Group
Research highlights
in brief b a c t e r i a l pat h o g e n e s i s
Invasive and adherent bacterial pathogens co-opt host clathrin for infection Veiga, E. et al. Cell Host & Microbe 2, 340–351 (2007)
Although many viruses enter host cells by clathrin-mediated endocytosis, it was thought that clathrin-coated vesicles were too small to internalize bacteria. So, it was a surprise when Listeria monocytogenes was shown to enter host cells by a clathrin-dependent mechanism. Building on this previously published finding, Veiga et al. investigated whether clathrin is required for the entry of other pathogens. They found that bacteria that enter cells after interactions with specific receptors, such as Staphylococcus aureus, require clathrin for entry, whereas those that inject effectors into host cells to facilitate their own entry by altering the host cytoskeleton, such as Shigella flexneri, do not. Clathrin was also required to assemble pedestals beneath adherent enteropathogenic Escherichia coli, which remain extracellular. molecular ecology
Mutational activation of niche-specific genes provides insight into regulatory networks and bacterial function in a complex environment Giddens, S. R. et al. Proc. Natl Acad. Sci. USA 104, 18247–18252 (2007)
The identification of environment-induced loci (EIL) is hampered by a lack of discernable phenotypes in the laboratory when EIL are mutated. Giddens et al. describe a new, broadly applicable method — SpyVet (suppressor analysis with in vivo expression technology) — to identify regulatory loci that control EIL in the rhizosphere bacterium Pseudomonas fluorescens. Promoters of EIL are fused to dapB (required for lysine biosynthesis), rendering strains prototrophic in the rhizosphere, where EIL are expressed, but auxotrophic in minimal growth media, where EIL are not expressed. Transposon mutagenesis, coupled with a simple screen for prototrophy in the laboratory, can quickly identify genes that regulate EIL–dapB fusions. A secondary screen for other members of the regulatory hierarchy was also feasible, which enabled these researchers to identify a regulatory network of seven regulators. b ac t e r i a l s e c r e t i o n
A minimal Tat system from a Gram-positive organism: a bifunctional TatA subunit participates in discrete TatAC and TatA complexes Barnett, J. P. et al. J. Biol. Chem. 26 Nov 2007 (doi 10.1074/jbc. M708134200)
The Tat (twin-arginine) secretion system transports fully folded proteins across bacterial membranes. Gram-negative bacteria encode three subunits (TatABC) that are essential for secretion. TatABC form a substrate-binding complex that is thought to recruit a separate and heterogeneous TatA complex, which can form a translocation pore. Bacillus subtilis has three TatA and two TatC variants. However, it lacks tatB, as do all Gram-positive bacteria. The B. subtilis tatAd gene complemented a tatB mutant of Escherichia coli, indicating that tatAd can replace TatA and TatB. Importantly, a B. subtilis TatAdCd system secreted a Tat substrate upon expression in E. coli, even though the TatAd complex is discrete, in contrast to the heterogeneous E. coli TatA complex. Direct experiments in B. subtilis are now needed to resolve the current controversy over how Tat secretes proteins.
nature reviews | microbiology
volume 6 | january 2008 © 2008 Nature Publishing Group
Research highlights
b ac t erial d e v elop m en t
Moving in the right direction New data published in a recent issue of Molecular Microbiology might have finally solved the controversy over the function of SpoIIIE during sporulation in Bacillus subtilis. During sporulation, B. subtilis undergoes an asymmetric cell division that generates two cells of unequal size — a small forespore and a larger mother cell — and a chromosome must be accurately segregated into each cell. The asymmetric-division septum is formed before chromosome segregation is completed, pinching one of the replicated chromosomes into a small lobe, which constitutes ~30% of the genome and is present in the forespore, and a large lobe, which constitutes ~70% of the genome and is present in the mother cell. To ensure successful completion of chromosome segregation, the trapped large lobe must be translocated across the septum from the mother cell into the forespore. Chromosome segregation during sporulation requires the doublestranded DNA translocase SpoIIIE, a protein that is related to the FtsK translocase of non-spore-forming bacteria. In recent years, results from different laboratories have supported two models of SpoIIIE/FtsK function. In the first, which is based on single-molecule studies of FtsK, the proteins function as reversible DNA translocases, and the polarity of the translocated DNA is determined by the DNA substrate. In the second, SpoIIIE functions as a DNA exporter, and the polarity of the translocated DNA is determined by the cellspecific assembly of a stable SpoIIIE
translocation complex. In this work, Eric Becker and Kit Pogliano set out to determine which of these models is correct. They began by devising a method that would allow them to tag SpoIIIE specifically in the mother cell and the forespore. To do so, the authors made use of the high-affinity interaction between the leucine zipper domains of cFos and cJun — the leucine zipper domain of cFos was fused to the carboxyl terminus of SpoIIIE and the leucine zipper domain of cJun was fused to the amino terminus of green fluorescent protein (GFP). Cellspecific promoters were used to express this GFP fusion protein in either the mother cell or the forespore. The results showed that wild-type SpoIIIE forms a focus only in the mother cell, supporting
nature reviews | microbiology
the second model discussed above. However, analysis of translocation-defective SpoIIIE showed that, in the absence of DNA translocation, SpoIIIE complexes formed on both sides of the septum, supporting the first model. To resolve this conundrum, Becker and Pogliano moved on to look at the effects of chromosome orientation. In a B. subtilis strain in which the chromosome partitioning proteins Soj and Spo0J had been deleted, the SpoIIIE complex could assemble in, and move DNA out of, the forespore. Finally, using a range of time-lapse fluorescence microscopy studies the authors were able to show that there was a direct correlation between the polarity of SpoIIIE-mediated chromosome segregation and the position of oriC after septation — SpoIIIE moves the chromosome into the cell that contains the replication origin. So, it seems that aspects of both models of SpoIIIE/FtsK function were correct — the assembly of a stable SpoIIIE translocation complex is indeed cell-specific, but this specificity, and the resulting translocation polarity, are determined by the orientation of the chromosome.
Sheilagh Molloy
ORIGINAL RESEARCH PAPER Becker, E. C. & Pogliano, K. Cell-specific SpoIIIE assembly and DNA translocation polarity are dictated by chromosome orientation. Mol. Microbiol. 66, 1066–1079 (2007)
volume 6 | january 2008 © 2008 Nature Publishing Group
Research highlights
V IR U S S T R U C T U RE
One of a kind! Despite the availability of a commercial vaccine, the highly contagious measles virus remains a substantial health risk for which there is currently no specifically targeted treatment or antiviral therapy. The structure of the measles virus haemagglutinin (MVH), which the virus uses to bind to host-cell receptors, has now been revealed in a recent study published in Nature Structural & Molecular Biology. Uniquely, species of Morbillivirus (such as the measles virus) lack neuraminidase activity and, instead of binding to the host using sialic acid, these viruses bind directly to the host-cell receptors SLAM and/or CD46. In this study, Colf and colleagues solved the structure of MVH at a resolution of 2.7 Å. The overall fold was similar to that of a β-propeller, with six interconnected blades (B1–6) — each of which contains four antiparallel β-strands — that surround a large cavity. Four potential N-linked glycosylation sites were detected, and the authors were able to model the glycan electron densities of two of these, Asn200 and Asn215, so supporting the theory that asparagine-linked glycosylation enables MVH folding and stablization. The authors then compared the structure of MVH with those of other viruses and found that it was most similar to the haemagglutinin/neuraminidase fold of the parainfluenza virus (PIV). However,
B1
B2
N-linked glycosylation Asn215 B3
Asn200
B6
B4
B5
C N
Cartoon of the structure of the measles virus haemagglutinin. Figure 1b from Colf, L. A., Juo, Z. S. & Garcia, K. C. Structure of the measles virus hemagglutinin. Nature Struct. Mol. Biol. 18 Nov 2007 (doi:10.1038/ nsmb1342).
large positional differences were detected, which suggests that, despite having similar folds, the haemagglutinin structures of the two viruses have diverged considerably. Finally, mutagenesis and antibody-blocking data were mapped onto the solved MVH structure and it was found that the binding surfaces of SLAM and CD46 are far apart and distinct from the large cavity that is analogous to that used by other viruses for sialic acid binding. This study further demonstrates the unique features of species of Morbillivirus, which retain a
nature reviews | microbiology
neuraminidase structural fold that is similar to other viruses, such as PIV, but share few other structural or sequence similarities. By using the structure of MVH as a template, structure-based drugs could be designed that target the measles virus at its point of entry and bring us one step closer to treating this potentially fatal viral disease.
Gillian Young
Original Research Paper Colf, L. A., Juo, Z. S. & Garcia, K. C. Structure of the measles virus hemagglutinin. Nature Struct. Mol. Biol. 18 Nov 2007 (doi:10.1038/nsmb1342)
volume 6 | january 2008 © 2008 Nature Publishing Group
Research highlights
b ac t erial v ir u lence
The cycle of success for Legionella Two recent papers have provided an insight into the mechanisms that Legionella pneumophila uses to subvert the host-cell vesicular trafficking pathway to create its replicative niche. L. pneumophila must establish a replicative niche within alveolar macrophages. This process involves remodelling of the Legionellacontaining vacuole (LCV) by exploiting the ability of the host protein Rab1 to recruit material from the vesicular transport pathway between the endoplasmic reticulum and Golgi apparatus. Rab proteins achieve this by reversibly associating with lipid membranes through a process that is known as Rab membrane cycling. Active GTP-bound prenylated Rab proteins are present in membranes. Once inactivated by a GTPaseactivating protein (GAP), GDP–Rabs are removed from the membrane and maintained in an inactive form in the cytosol through association with a guanine nucleotide-dissociation inhibitor (GDI). The GDI can be displaced by a GDI-displacement factor (GDF), and the GDP-bound
Rab protein is then recruited back to the membrane and activated by a guanine nucleotide-exchange factor (GEF). After the active Rab has carried out its function, membrane cycling is completed when the Rab is inactivated by a GAP protein. L. pneumophila uses a type IV secretion system to secrete effector proteins into the host-cell cytoplasm. Previous work had shown that one type IV effector, DrrA/SidM, is required for the recruitment of Rab1 to the LCV and has Rab1-specific GEF activity. Machner and Isberg, and Ingmundson et al. were interested in whether DrrA/SidM also has GDF activity. Both groups purified overexpressed, tagged forms of Rab1 and Rab-GDI from eukaryotic cells to ensure that they were prenylated and then showed that DrrA/SidM interferes with Rab1–Rab-GDI complex formation by displacing the Rab-GDI, thus confirming that DrrA/SidM does have GDF activity. Machner and Isberg went on to look at the effects of the presence of liposomal membranes. They found that the
activation of Rab1 by the GEF activity of DrrA/SidM was enhanced in the presence of a lipid bilayer. They were then able to follow the association of the proteins with the membrane during nucleotide exchange, which confirmed that once the GDF activity of DrrA/SidM has displaced the Rab-GDI, the free Rab1 protein is inserted in the membrane and can then be activated by the GEF activity of DrrA/SidM. Both groups were also interested in other L. pneumophila type IV effectors that might function downstream of DrrA. Machner and Isberg showed that L. pneumophila LidA binds GTP–Rab1 that has been activated by DrrA/SidM and supports the accumulation of activated Rab1 on the LCV. Ingmundson et al. found that LepB is delivered into the host-cell cytoplasm shortly after bacterial uptake, is present on the early LCV membrane and can disrupt secretory transport in a mammalian cell line. They completed their work on LepB by demonstrating that this L. pneumophila effector has GAP activity towards Rab1 and propose that this activity completes the membrane cycling of Rab1 that is begun by DrrA/SidM. DrrA/SidM is the first protein to be identified that has both GDF and GEF activity. Both groups point out that this raises the possibility that there might be eukaryotic GEFs that have similar dual functions. Sheilagh Molloy
ORIGINAL RESEARCH PAPERS Machner, M. P. & Isberg, R. I. A bifunctional bacterial protein links GDI displacement to Rab1 activation. Science 318, 974–977 (2007) | Ingmundson, A. et al. Legionella pneumophila proteins that regulate Rab1 membrane cycling. Nature 450, 365–369 (2007)
nature reviews | microbiology
volume 6 | january 2008 © 2008 Nature Publishing Group
news & analysis genome watch
A poultry existence Helena Seth-Smith The two poultry pathogens discussed in this month’s Genome Watch are closely related to well-characterized organisms that infect humans, so scrutinizing their genomes could reveal factors that determine host-specificity. Avian influenza has the potential to ravage the poultry industry and be transmitted to humans. Bacterial pathogens of poultry are also economically important. Bordetella avium causes non-lethal respiratory disease, which can predispose birds (mainly turkeys) to secondary infections. Infection of the respiratory tract by avian pathogenic Escherichia coli (APEC) strains can result in colibacillosis, which can be lethal. The genome of B. avium consists of 3,732,255 base pairs and contains 3,417 coding sequences (CDSs)1. The genome of the related strain Bordetella bronchiseptica, which is responsible for respiratory disease in mammals, was sequenced previously. A comparison of the two could pinpoint those genomic features that are important to host specificity, that is, the infection of either an avian or a mammalian host. These 2 strains share a conserved Bordetella backbone of 2,380 orthologous CDSs that constitutes two thirds of the B. avium genome. The CDSs that are unique to B. avium mainly encode surface-associated proteins that might be important in adhesion and immune-system evasion. The B. avium genome
contains CDSs that encode many novel adhesins that are presumably specific for the avian trachea, including several filamentous haemagglutinins. There are potentially novel biosynthetic pathways for both lipopolysaccharide (LPS) and capsular polysaccharide, which would also modify the bacterial-cell surface. There are eleven CDSs that seem to encode fimbriae, which could aid binding of the bacteria to cells of the respiratory tract. The interaction with the host might also be mediated by 2 novel surface proteins, each of which comprises more than 4,000 amino acids, and 7 autotransporters. Bordetella species often synthesize toxins; B. avium can produce dermonecrotic toxin, which has been shown to be important in causing disease in turkeys, and tracheal cytotoxin. Most E. coli strains are associated with the intestine. Strains that have strayed from this lifestyle are known as ExPEC (extraintestinal pathogenic E. coli) and include APEC and UPEC (uropathogenic) strains, of which the UPEC strains are responsible for urinary-tract infections in humans. Multiple E. coli genomes have been sequenced, including laboratory strains, those that cause food poisoning and other ExPEC strains. It has been proposed that UPEC originated from APEC, and genome comparisons could help us to understand these evolutionary connections. Strain APEC 01 has a chromosome of 5,082,025 base pairs, and comprises 4,467 CDSs2. It also harbours four plasmids: one involved in virulence; one involved in antibiotic resistance;
| january 2008 | volume 6
and two cryptic plasmids. Comparative analyses revealed that 78% of the CDSs from the APEC 01 chromosome are common to all sequenced E. coli strains, which means that these CDSs comprise the minimal E. coli backbone. A further 9% of the CDSs are shared among all the sequenced ExPECs, which might indicate their importance to the survival of the bacteria outside the intestine. These genes include those that code for LPS synthesis, fimbriae, virulence‑associated pili and iron acquisition. The remaining 13% of CDSs are unique to APEC 01 and are predominantly located on 4 APEC‑specific pathogenicity islands (PAIs), which are phage-related and might contain additional virulence factors. The differences between the avian and mammalian Bordetella strains seem to reflect evolution over millions of years through the vertical transfer of genes. The E. coli genomes show more evidence of horizontal gene transfer, and many of the differences between strains are restricted to PAIs and plasmids. Helena Seth-Smith is at the Sanger Institute, Wellcome Trust Genome Campus, Hinxton, Cambridge CB10 1SA, UK. e-mail:
[email protected] doi:10.1038/nrmicro1830 1. Sebaihia, M. et al. Comparison of the genome sequence of the poultry pathogen Bordetella avium with those of B. bronchiseptica, B. pertussis, and B. parapertussis reveals extensive diversity in surface structures associated with host interaction. J. Bacteriol. 188, 6002–6015 (2006). 2. Johnson, T. J. et al. The genome sequence of avian pathogenic Escherichia coli strain O1:K1:H7 shares strong similarities with human extraintestinal pathogenic E. coli genomes. J. Bacteriol. 189, 3228–3236 (2007).
DATABASES Entrez Genome Project: http://www.ncbi.nlm.nih.gov/ entrez/query.fcgi?db=genomeprj APEC 01 | Bordetella avium | Bordetella bronchiseptica All links are active in the online pdf
www.nature.com/reviews/micro © 2008 Nature Publishing Group
disease watch | in the news
Neil Smith
Choose your battles wisely
Animals, like plants, can respond in different ways to parasite infections: they can mount a full-scale immune response or they can tolerate the invader. Andrew Read and colleagues from the University of Edinburgh, UK, introduced rodent malaria into five strains of laboratory mice and monitored parasite load and animal health, as measured by anaemia and body mass. In some strains of mice, parasite load increased and the mice stayed healthy, which indicated that the mice could tolerate the parasite. In other mice strains, however, parasite load decreased, which indicated that the mice were resisting infection. The authors suggest that these findings could have important implications for pathogen evolution: pathogens might not be pushed to evolve increased virulence if they are tolerated rather than destroyed. However, tolerated pathogens are also more likely to spread. In a somewhat related paper, Lynn Martin, from the University of South Florida, and colleagues report that fast-living strains of mice develop high fevers in response to simulated infection, but slow-living strains do not. The authors propose that the ‘live fast, die young’ mice tolerate the harm that fever inflicts on bodily tissues, as they already have short lifespans. Slow-living mice, however, have more to lose and may therefore adopt targeted strategies, such as antibody production. Science/Funct. Ecol.
Dandruff on the collar? The genome of Malassezia globosa — a fungus that causes discomforting and embarrassing dandruff — has been sequenced by an international team led by Thomas Dawson at Procter & Gamble. Malassezia spp.associated dandruff affects more than 50% of humans, and more than UK£3 billion are spent annually on dandruff treatments. Genome sequencing revealed that the M. globosa genome is among the smallest of all the free-living fungi analysed so far, containing as few as 4,285 predicted proteincoding genes. Of these, approximately 50 encode secreted proteins, including many lipid-degrading proteins, that are probably responsible for breaking down hair and scalp, thereby causing the irritation, inflammation and flaking that are hallmarks of dandruff. Not only might this sequenced genome represent a gold mine for developers of anti-dandruff shampoos, they also have agricultural implications as M. globosa is closely related to various common pathogenic plant fungi that affect corn, wheat and other food crops. PNAS/The Telegraph
Novel MRSA virulence factors identified Although community-acquired meticillinresistant Staphylococcus aureus (CA-MRSA) causes the majority of the staphylococcal infections that result in trips to emergency rooms, the basic cause of CA-MRSA virulence remains unidentified. However, a class of secreted cytolytic peptides (so‑called PSMs) that function primarily to destroy leukocytes has now been identified by Michael Otto, from the Rocky Mountain Laboratories, USA, and colleagues. Deletion of PSM genes resulted in a decrease in the severity of CA-MRSA in two mouse abscess and bacteraemia models. Moreover, in vitro studies indicated that PSM genes caused S. aureus to recruit, activate and lyse human neutrophils. Notably, hospital-acquired MRSA strains exhibit a significantly lower expression of PSM genes than do CA-MRSA strains, which are more virulent. Nature Med.
was unexpectedly ended after preliminary results showed that the vaccine was not conferring resistance. Further analysis suggests that the vaccine — which consists of 3 HIV genes and an attenuated adenovirus 5 vector — may have increased HIV susceptibility in recipients. Notably, 21 of 392 male vaccine-recipient patients with pre-existing immunity to adenovirus 5 acquired HIV infections, compared with 9 of 386 from the placebo group. As the trial was halted abruptly, researchers cannot determine whether these results are statistically significant. Among other explanations, vaccination may have increased the production of CD4+ T cells — HIV’s favorite infection target — in adenovirus-5-immune patients. An upcoming trial of a vaccine developed by the National Institutes of Health has been delayed to allow further analysis. Nature/ New York Times
HIV numbers revised The United Nations has lowered their estimate of how many people are infected with HIV globally from 39.5 million to 33.2 million. These revised numbers, however, are attributed to improved methods of data collection rather than to a real decline in HIV infection. The number of HIV carriers is still rising, although the global prevalence of HIV seems to be levelling off. An estimated 2.5 million new infections occurred in the past year (which works out at a staggering 6,800 new infections per day), down from a likely peak in the late 1990s of over 3 million infections per year. Sub-Saharan Africa remains the hardest hit by HIV, accounting for nearly 50% of the new cases of infection and 68% of the total number of cases. WHO
Vaccine may have increased HIV risk In late September, a trial of Merck’s candidate HIV vaccine, which included over 3,000 people from 9 countries,
nature reviews | microbiology
STOCKBYTE
volume 6 | january 2008 | © 2008 Nature Publishing Group
N e w s & a n a ly s i s
Use condoms! Although the global prevalence of HIV is showing signs of levelling off, the transmission rates for sexually transmitted infections (STIs) in the United Kingdom are still on the rise, despite concerted public health efforts to reverse this trend. During 2006, report the Health Protection Agency (HPA), 376,508 STIs were newly diagnosed — a 2.2% increase compared with the number of new STIs in 2005. The number of STIs diagnosed each year in the United Kingdom has increased almost continually since the 1990s. In particular, the sexual health of young adults has worsened, although the HPA also warns of a continuing HIV and STI epidemic among men who have sex with men. An estimated 73,000 adults in the United Kingdom are now HIV-positive, although one-third of these individuals remain unaware that they are infected. More funding and education is needed to tackle these problems, say experts. HPA/BBC
Anti-polio programme gets a booster
Outbreak news
IMAGE SOURCE
Can Africa cope with avian influenza? African nations are unlikely to be able to achieve the five priorities that have been identified by the World Health Organization (WHO) to fight avian influenza, argue Robert Webster and colleagues. So far there have been only 40 cases of avian influenza in Africa, for which there has been a 40% fatality rate. However, the predominant strain of avian influenza in Africa, the Qinghai strain, “has acquired several troubling properties, including respiratory rather than faecal transmission in poultry.” The five priorities of the WHO for the combat of avian flu are: reduce human exposure; strengthen surveillance; intensify rapid containment; enhance response capacity; and coordinate global research. Owing to a number of factors, including inadequate surveillance, the denial of outbreaks and the poor communication of risks to the public, African nations are unlikely to achieve these targets in the opinion of some observers. For example, it took 4 weeks to officially confirm the outbreaks in all of the affected African countries, apart from Egypt. Webster and colleagues call for each African nation to realistically assess its status, conduct regular active surveillance and be more forthcoming with data. Lancet Infect. Dis.
autoinducers. V. cholerae produces two autoinducers, autoinducer 2 (AI‑2) and CAI‑1. The structure of AI-2 has been known for many years, and in this latest paper Bassler et al. characterized and synthesized CAI‑1. To assess the feasibility of controlling V. cholerae by manipulating QS the authors showed that synthetic CAI‑1 can repress production of the V. cholerae virulence factor toxin co-regulated pilus as well as natural CAI-1. “This work”, the authors argue, “provides a demonstration that interference with quorum-sensing processes in general ... has great promise in the clinical setting.” Nature
The campaign to eradicate polio has received a much-needed US$200 million (£97 million) grant from the Bill and Melinda Gates Foundation and Rotary International. An initial $100 million will be distributed through grants to the WHO and UNICEF to support mass immunization campaigns in polio-affected countries, polio-surveillance activities, and community-education and outreach programmes. Although immunization programmes have made huge strides over the past 20 years, polio is still endemic in Nigeria, Pakistan, India and Afghanistan, and approximately 700 cases of polio are reported each year. Despite this generous donation, WHO officials report that the polio-eradication programme is still facing a $650 million shortfall over the next 2 years. BBC
Cracking cholera’s lines of communications The Vibrio cholerae signalling molecule cholera autoinducer-1 (CAI-1) may form the basis of therapeutic intervention against V. cholerae infection, reports Bonnie Bassler and colleagues. V. cholerae, and other bacteria, use quorum sensing (QS) to sense and respond to population density, and this cell-tocell communication is mediated by
10 | january 2008 | volume 6
Rift Valley Fever. An outbreak of Rift Valley Fever has killed 164 people out of a total of 451 who have contracted the disease in Sudan. The majority of these infections are the result of direct or indirect contact with the blood and organs of infected animals, although some cases are also the result of bites from infected mosquitoes. Sudanese health authorities initially sought help to control the outbreak in mid-October. Rift Valley Fever normally has a fatality rate of ~1%, but the fatality rate is ~50% in patients who develop the haemorrhagic form. WHO/Associated Press Encephalitis. A ‘mystery’ pathogen that causes encephalitis has resulted in 21 deaths, mostly children, in a remote region of Bangladesh. An additional 200 people are affected. Bangladesh’s International Epidemiology Research Centre and the Centers for Disease Control and Prevention in the United States are working together to identify the causative pathogen. Reuters Ebola. WHO officials have confirmed that Ebola virus has killed 16 people and affected an additional 50 in western Uganda. The virus belongs to a distinct subtype that has not been detected before, as indicated through tests conducted by national laboratories in Uganda and the US Center for Disease Control. Ebola is fatal in approximately 80% of cases, and as yet there is no known cure. BBC 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.
www.nature.com/reviews/micro © 2008 Nature Publishing Group
Progress The versatility of Shigella effectors Michinaga Ogawa, Yutaka Handa, Hiroshi Ashida, Masato Suzuki and Chihiro Sasakawa
Abstract | When Shigella infect the intestinal epithelium, they deliver several effectors through the type III secretion system (T3SS) into the surrounding space and directly into the host-cell cytoplasm, where they can mimic and usurp host cellular functions or subvert host-cell signalling pathways and the immune response. Although bacterial strategies and mechanisms of infection vary greatly, recent studies of Shigella effectors have revealed that Shigella possess a highly evolved strategy for infection. The intestinal epithelium acts as an intrinsic defensive barrier against microbial invaders. The barrier function has four major elements: the resident commensal bacteria; the integrity of the epithelial barrier; the rapid epithelial-cell turnover; and the innate immune response. Nevertheless, many pathogenic bacteria can circumvent these defences and exploit the intestine as a replicative niche, a shelter from host immune surveillance or a port of entry for dissemination into deeper tissues. Shigella species (abbreviated here to Shigella) are human-adapted pathogens that are capable of colonizing the intestinal epithelium by exploiting epithelial-cell functions and circumventing the host innate immune response. Shigella have neither adherence factors nor flagella. Following ingestion by the faecal–oral route, Shigella move through the small intestine to the colon and rectum, where they cross the epithelial barrier through microfold cells that overlie solitary lymphoid nodules. When the epithelial barrier has been breached, Shigella immediately enter the macrophages that reside within the microfold‑cell pocket. Once within the macrophages, the infecting bacteria disrupt the phagosomal membrane and disseminate from the phagosome into the macrophage cytoplasm, where they multiply and induce rapid cell death1,2. Shigella that are released from dead macrophages can enter the surrounding enterocytes by inducing the production of membrane ruffles at the basolateral surface,
which eventually leads to entry by macro pinocytosis. As soon as a bacterium is surrounded by a membrane vacuole within an epithelial cell, it disrupts the vacuolar membrane and escapes into the cytoplasm. Shigella can multiply in the epithelial-cell cytoplasm and move both intra- and intercellularly, and so infection of the intestinal epithelium by Shigella elicits a strong inflammatory response1,2. To proceed through the series of steps that are involved in infection, Shigella secrete many effector proteins through the type III secretion system (T3SS)3. TABLE 1 lists some of the characterized Shigella effectors and their homologues in other bacterial pathogens. This Progress article highlights recent advances in our understanding of the roles of the Shigella effectors that are delivered into host cells by the T3SS during intestinal infection. Effectors in the early stages of infection The ability to invade, colonize and translocate across mucosal barriers is an important feature of enteric pathogens. Notably, in the case of Shigella, the ability to invade epithelial cells and subsequently spread from cell to cell is pivotal in establishing an intestinal infection. When Shigella come into contact with epithelial cells, they deliver a subset of effectors through the T3SS both around the bacterial surface and directly into host cells. These effectors, which include IpaA, IpaB, IpaC, IpgB1, IpgB2, IpgD and VirA, are involved in
nature reviews | microbiology
promoting bacterial basolateral entry into polarized epithelial cells1,4 (FIG. 1). IpaA binds the amino-terminal head domain of vinculin, which is a key component of focal adhesions, and stimulates actin depolymerization5. IpaA also targets β1-integrin and stimulates the GTPase activity of RhoA, thereby inducing the loss of actin stress fibres6. Thus, IpaA might facilitate recycling of the free actin pool by the destruction of stress fibres and contribute to the production of membrane ruffles. The interaction between IpaB and the CD44 receptor, which is present on the epithelial cell membrane in areas that are rich in lipid rafts, stimulates basolateral invasion7. IpaB is a multifunctional effector2,4,8. As well as being secreted at the tip of the T3SS needle and acting as a T3SS translocator (with IpaC), it is also involved in vacuole disruption, activates caspase‑1 in macrophages and modulates cell-cycle progression in epithelial progenitor cells (discussed below). IpaC is delivered into the epithelial-cell cytoplasm and integrates into the membrane, where it is assumed to have a key role, either directly or indirectly, in inducing actin polymerization4. Although there is no direct evidence as yet, IpaC might be involved in promoting Shigella invasion of epithelial cells by inducing actin foci through the accumulation of the host non-receptor tyrosine kinase Src at the site of bacterial entry4. On entry, the phosphorylation of cortactin — a cortical actin-binding protein that activates the Arp2/3 complex and induces actin polymerization — occurs in a Src-dependent manner and Crk (a Src-related tyrosine kinase) activates the small GTPase Rac1 (Ref. 4).
IpgB1 is assumed to have a major role in producing membrane ruffles by activating Rac1 through ELMO–Dock180, a Rac1 guanine nucleotide-exchange factor9. For example, when IpgB1 is ectopically expressed in HeLa cells, large membrane ruffles are produced. Pulldown assays have identified ELMO as the IpgB1-binding partner; IpgB1 binds to the amino‑terminal region of ELMO, which also binds RhoG. An in vitro binding assay showed that RhoG binding to volume 6 | january 2008 | 11
© 2008 Nature Publishing Group
Progress Table 1 | Activities of Shigella type III secretion system (T3SS) effectors
T3SS effector
Biochemical activity
Target (or targets) Role in infection
Selected homologues
IpaA
Unknown
Vinculin
Bacterial invasion
Salmonella spp. SipA (also called SspA)
IpaB
T3SS translocon
Caspase-1, CD44 and Mad2L2
Macrophage apoptosis and cell-cycle arrest
Salmonella spp. SipB (also called SspB) and Yersinia spp. YopB
6,34,36
IpaC
T3SS translocon
Unknown
Bacterial invasion
Salmonella spp. SipC (also called SspC)
37
IpgB1
RhoG mimic
ELMO
Bacterial invasion
Salmonella spp. SifA and SifB; EHEC, EPEC and Citrobacter rodentium Map; and EHEC EspM1 and EspM2
IpgB2
RhoA mimic
mDia and ROCK
Unknown
IpgD
Inositol phosphate Phosphatidylinositol phosphatase 4,5-bisphosphate
Bacterial invasion and host-cell survival
Salmonella spp. SopB (also called SigD)
VirA
Cysteine protease
Tubulin
Bacterial invasion and intracellular spreading
EHEC, EPEC and C. rodentium EspG and EPEC EspG2 (also called Orf3)
IcsB
Unknown
Shigella VirG (also called IcsA)
Escape from autophagy
Burkholderia spp. BopA
20
OspC1
Unknown
Unknown
Polymorphonuclear transepithelial migration
Shigella OspC2, OspC3 and OspC4 and Vibrio parahaemolyticus OspC2
38
OspE2
Unknown
Unknown
Intercellular spreading
Shigella OspE1; Salmonella spp. EspO1STYM; and EHEC EspO1-1 and EspO1-2
39
OspF
Phosphothreonine lyase
MAP kinases
Suppression of innate immune responses
Salmonella spp. SpvC; Pseudomonas syringae HopAI1; and Chromobacterium violaceum VirA
27,29,38
OspG
Serine/threonine kinase
E2 ubiquitinconjugating enzymes
Suppression of innate immune responses
Yersinia enterocolitica YE2447; C. rodentium NleH; and EHEC NleH1-1 and NleH1-2
26
IpaH9.8
E3 ubiquitin ligase
U2AF35 and yeast Ste7
Suppression of innate immune responses
24,25
IpaH7.8
E3 ubiquitin ligase
Unknown
Escape from endocytic vacuoles of phagocyte
Shigella IpaH4.5; Salmonella spp. SspH1, SspH2 and SlrP; Yersinia pestis YP3416 and YP3418; P. syringae PSPTO1492 and PSPTO4093; and Rhizobium spp. Y4fR
Chromosomal IpaHs
Possibly E3 ubiquitin ligase
Unknown
Suppression of innate immune responses
Refs 9,10
11,12
7,8
13,14,17
25,40 23
EHEC, enterohaemorrhagic Escherichia coli; EPEC, enteropathogenic E. coli, MAP, mitogen-activated protein.
ELMO is competitively inhibited by IpgB1, implying that IpgB1 mimics the function of RhoG in producing membrane ruffles during Shigella invasion9. IpgB2 is an IpgB1 homologue that binds to mDia1, which facilitates actin nucleation, and the Rho kinase ROCK through its GTPasebinding domains. In this way, IpgB2 mimics the activity of RhoA in inducing the formation of stress fibres, although its specific involvement in bacterial invasion remains unclear10. IpgD possesses phosphatidylinositol (4,5) bisphosphate (PI(4,5)P2) phosphatase activity and catalyses the hydrolysis of PI(4,5)P2 to phosphatidylinositol 5‑monophosphate (PI(5)P), thereby promoting local actin polymerization4,11,12. VirA, which is a member of the EspG/ VirA family, shares significant amino-acid homology, as well as functional similarity, with EspG. EspG/VirA family members
are found in enteropathogenic Escherichia coli (EPEC), enterohaemorrhagic E. coli (EHEC) and Shigella. Intriguingly, VirA is delivered into the host-cell cytoplasm near the site of bacterial entry and induces local microtubule (MT) degradation13. As degradation of MTs by EspG results in the release of various MT‑associated proteins, including GEF‑H1, and GEF‑H1 activates RhoA14, VirA activity is assumed to contribute to ruffle formation during Shigella invasion through the cross-talk between RhoA and Rac1. Clearly, these studies indicate that Shigella exploit the interactions between a subset of effector proteins and their host target molecules (or target functions). Such interactions are capable of stimulating several host-cell signalling pathways that are involved in inducing actin polymerization and remodelling the architecture of the host-cell surface (FIG. 1).
12 | january 2008 | volume 6
Intra- and intercellular motility Many pathogenic bacteria can invade host cells, and some species, such as Shigella, Listeria monocytogenes, Mycobacterium marinum, Rickettsia conorii and Burkholderia pseudomallei, are capable of inducing actin nucleation at one pole of the bacterium to gain the propulsive force that is necessary to move through the host cell and into neighbouring host cells. This activity, often called actin-based bacterial motility, is crucial for establishing an infectious foothold as well as renewing replicative niches. Intriguingly, the bacterial proteins that are involved in mediating actin nucleation vary from species to species. Nevertheless, these proteins share the ability to recruit and activate the Arp2/3 complex, which is required for actin polymerization near the bacterial surface8. In the case of Shigella, an outermembrane protein called VirG (also known as IcsA) has a key role, as VirG can directly www.nature.com/reviews/micro
© 2008 Nature Publishing Group
Progress
T3SS ELMO
RhoG
Shigella
IpgB1 T3SS effectors CD44
Dock180
IpaC
IpaB
IpgB2 mDia Rac1
RhoA
IpgB2 Rock
Cdc42
WAVE
IpaC
Microtubule destabilization
Major actin rearrangements Arp2/3
Membrane ruffling
PI(4,5)P2
VirA
IpgD
IpaA Vinculin IpaB IpaC
IpaA
Entry
α5β1 integrin
T3SS effectors
Figure 1 | A simplified model of membrane-ruffle production in response to the stimulation of host cellular signalling by Shigella effectors. Upon contact between Shigella and an epithelial cell membrane, the bacterium delivers several effectors through the type III secretion Nature Reviews | Microbiology system (T3SS) around the bacterial surface and into the host-cell cytoplasm. By interacting with their host binding partners the effectors are eventually able to activate the Rac1–WAVE–Arp2/3 pathway, which leads to the protrusion of membrane ruffles around the bacterial entry point. It should be noted that this model does not rule out any other additional, as yet uncharacterized, signalling pathways that may be involved in mediating actin polymerization during Shigella invasion. See the main text for details.
interact with N‑WASP (neural Wiskott– Aldrich syndrome protein)15. When VirG and Cdc42 interact with N‑WASP, N‑WASP becomes activated, which, in turn, recruits and activates the Arp2/3 complex16. The formation of the VirG–N-WASP–Arp2/3 complex at one pole of the bacterium allows Shigella to induce actin nucleation and elongation, thus gaining propulsive force (FIG. 2). During bacterial movement, some bacteria impinge on the host-cell membrane and cause the membrane to protrude and penetrate into that of neighbouring cells, thereby allowing the bacteria to disseminate into adjacent cells.
The movement of bacteria within host cells is highly variable and depends on their location in the host cell. For example, some motile Shigella suddenly change direction, spin around or stop moving. It has recently been shown that Shigella movement within the host-cell cytoplasm is severely hindered by MTs, but a motile bacterium can destroy surrounding MTs using VirA17 (FIG. 2). As described above, VirA is also used to promote bacterial entry. Thus, VirA has dual roles in both bacterial invasion and intracellular spreading. Characterization of VirA activity indicated that the degradation of MTs by VirA depends on its α‑tubulin-specific
nature reviews | microbiology
cysteine-protease-like activity17. Consistent with this observation, a virA mutant that lacked this activity was found to be less capable of moving smoothly within the host cytoplasm than the wild type and was incapable of maintaining continuous cell–cell spreading17. Although there is no direct evidence as yet, the ability to clear a path through the host-cell MTs might not be unique to Shigella. L. monocytogenes ActA is capable of directly recruiting the Arp2/3 complex in the vicinity of the bacterial surface, and the destruction of MTs is occasionally detected along the path of L. monocytogenes movement. Intriguingly, L. monocytogenes ActA is not a VirA homologue but it indirectly recruits Op18, an MT‑sequestering host protein, near the bacterial surface18. It is tempting to speculate that Op18 facilitates L. monocytogenes movement within epithelial cells. The discovery of a novel bacterial activity that destroys MTs demonstrates that, although MTs are an obstacle to bacterial movement within the host-cell cytoplasm, certain bacteria have evolved an activity to remove this barrier. Escape from autophagy Autophagy is a ubiquitous degradation system in eukaryotic cells that is required not only for the cellular response to starvation and stress, and for the removal of damaged or surplus organelles, but also for removing bacterial pathogens that invade the cytoplasm of host cells. For example, Streptococcus pyogenes (Group A Streptococcus) and Staphylococcus aureus are capable of invading epithelial cells, and both pathogens are targeted by the autophagic machinery and eventually undergo lysosomal degradation. Autophagy is achieved by a series of autophagy-related (Atg) proteins that are highly conserved from yeast to humans19. During multiplication within epithelial cells, Shigella are recognized by components of the autophagic pathway. However, the secretion of IcsB through the T3SS on entry into the cytoplasm allows the bacteria to escape autophagic destruction20. Although an icsB mutant is fully invasive and can escape from the phagocytic vacuole in epithelial cells, it is ultimately enclosed by autophagosomes. Surprisingly, IcsB does not directly inhibit autophagy itself. Instead, the VirG protein — which is required for actin-based bacterial motility (FIG. 2) — is targeted for autophagic recognition by binding to Atg5 (a protein that is involved in the elongation of isolation membranes). In in vitro binding assays, both IcsB and volume 6 | january 2008 | 13
© 2008 Nature Publishing Group
Progress MT network
Capping protein
Profilin
F-actin
VirA
Vinculin Arp2/3 G-actin
N-WASP
VirA
IcsB VirG
Cdc42
Shigella VirA
IcsB VirA
VirA
Figure 2 | Shigella movement within the host-cell cytoplasm requires actin polymerization and microtubule degradation. Asymmetric distribution of VirG Nature (also known IcsA) on the Reviewsas| Microbiology bacterial surface is essential for the polar movement of Shigella in epithelial cells. The accumulation of VirG at one pole of the bacterium recruits and activates the Arp2/3 complex by the interaction between VirG and N‑WASP (neural Wiskott–Aldrich syndrome protein). Motile Shigella secrete VirA through the type III secretion system and destroy local microtubules (MTs), thereby facilitating bacterial movement.
Atg5 can bind to VirG. Intriguingly, IcsB and Atg5 share the same binding region on VirG, and the affinity of IcsB for VirG is much stronger than the affinity of Atg5 for VirG, which suggests that IcsB acts as an anti-Atg5-binding protein. The interaction between IcsB and VirG near the bacterial surface, therefore, seems to provide a mechanism of escape from autophagic recognition1,20. L. monocytogenes can also escape autophagy during multiplication in the host-cell cytoplasm, but the mechanism that underlies this escape seems to differ from that of Shigella21. Although the precise mechanism is still unknown, in contrast to Shigella VirG, ActA, which is present on the surface of L. monocytogenes and mediates actin polymerization at one pole of the bacterium, seems to be involved in the escape of L. monocytogenes from autophagic recognition. The recognition of intracellular pathogens by autophagy is not limited to pathogens that invade the cytoplasm. As long as the innate immune response of the host is intact, intracellular pathogens, such as Mycobacterium tuberculosis, Legionella pneumophila and Coxiella burnetii, which are sequestered in vacuolar compartments, can be also targeted by autophagy. Unless they are able to modify the vesicles in which they are contained or avoid autophagic uptake at an early stage of infection, they are entrapped by a lamellar membranous structure that is associated with LC3, an essential autophagic protein. Although the mechanisms that are involved remain
unclear, recent studies have indicated that these intracellular pathogens are also capable of certain manoeuvres that allow them to circumvent autophagic recognition. Modulation of the innate immune response Major bacterial-cell components, such as lipopolysaccharide (LPS), peptidoglycan (PGN), and, perhaps, nucleic acids (DNA and RNA), are readily released from multiplying and killed bacteria. These components are recognized by Toll-like receptors and Nod-like receptors, which activate the host innate defence systems, stimulate inflammatory signalling cascades and induce cellular and humoral immune responses. Nevertheless, many bacterial pathogens that infect the intestinal mucosa can circumvent the host innate immune response and colonize their replicative niches2. The mechanisms that are used by bacteria to dampen host inflammatory signals have been extensively studied using a range of animal and plant pathogens. Shigella release LPS and PGN into host cells, and these pathogen-associated molecular patterns are the main cause of the strong inflammatory response that is induced by Shigella infection2. For example, the LPS that is released by Shigella as they escape from phagosomes into the macrophage cytoplasm activates caspase‑1, which induces the production of interleukin‑1β and cell death. PGN released from Shigella that are multiplying within epithelial cells activates the Nod1–RICK pathway, which stimulates signalling cascades that involve mitogen-activated protein kinases
14 | january 2008 | volume 6
(MAPKs) and nuclear factor (NF)-κB and results in the production of proinflammatory cytokines22. To circumvent the innate immune response, Shigella deliver more than ten effector proteins, including the IpaH proteins, OspG and OspF, into host cells through the T3SS2,23 (FIG. 3; TABLE 1). IpaH9.8, a member of the IpaH protein family, translocates to the nucleus, where it interacts with U2AF35, an mRNA splicing factor24. Infection of epithelial cells by an ipaH9.8 mutant increased the levels of proinflammatory cytokines, and RNA interference (RNAi)-mediated knockdown of U2AF35 production decreased these levels. In a murine lung-infection model, the ipaH9.8 mutant induced more severe inflammatory responses and greater proinflammatory cytokine production than did wild-type Shigella, which resulted in a 30-fold decrease in bacterial colonization24. It has recently been suggested that the IpaH homologues (including IpaH9.8) that are produced by plant and animal pathogens, such as Yersinia species, Salmonella species, Edwardsiella ictaluri, Bradyrhizobium japonicum, Rhizobium species and some Pseudomonas species, share an E3 ubiquitin ligase activity in the highly conserved carboxy‑terminal region25 (TABLE 1). Although the host target (or targets) of each ubiquitin ligase remains unknown, in a yeast-cell system, IpaH9.8 was shown to have an activity that interferes with the pheromone response by the ubiquitination of the MAPK kinase Ste7, which is then degraded by proteasomes25. These findings suggest that IpaH and IpaH homologues have a central role in dampening host inflammatory-related signals during bacterial infection (FIG. 3). OspG can bind to ubiquitinated E2s (ubiquitin-conjugating enzymes), such as UbcH5b, a component of the Skp1–culin–Fbox protein (SCF)β–TrCP complex that is involved in phospho‑IκBα ubiquitination and its subsequent degradation by the proteasome26. Thus, OspG interferes with IκBα degradation, thereby resulting in a repression of NF‑κB activation. Consistent with this observation, the characterization of the phenotype of an ospG mutant in in vitro and in vivo models of infection has shown that OspG is involved in downregulating the host inflammatory response to Shigella infection26. OspF also translocates to the host-cell nucleus. OspF has a specific phosphatase activity that dephosphorylates and inactivates MAPKs, such as ERK1/2, JNK and p38, www.nature.com/reviews/micro
© 2008 Nature Publishing Group
Progress thereby blocking the phosphorylation of the serine 10 residue of histone H3, which is required for the transcription of a subset of NF‑κB-regulated genes27–29. Intriguingly, in contrast to the OspF activity that is proposed above, Shigella OspF, Salmonella spp. SpvC and Pseudomonas syringae HopAl1 have also been shown to share the ability to dephosphorylate MAPKs through their phosphothreonine lyase activities, thereby enabling these bacterial effectors to interfere with MAPK activity3,29 (TABLE 1). In rabbit ileal loops and a mouse-lung infection model, the ability of a Shigella ospF mutant to induce inflammatory responses was much greater than that of wild-type Shigella. Although there is some controversy regarding the activity of OspF, which might be attributed to the different experimental conditions that are used in each study or the dual activities that are encoded by OspF, these studies have clearly indicated that OspF participates in dampening the inflammatory response27. The reason for the presence of such a variable number of effectors in Shigella and other pathogens remains a matter of speculation. As numerous signal-transduction pathways are engaged in inducing the inflammatory response to bacterial infection, many effectors must be engaged in modulating the various inflammatory signal pathways at different levels and different times during bacterial infection. Slowing of rapid epithelial-cell turnover The intestinal epithelium renews itself every several days, which provides an important intrinsic defence system that limits bacterial colonization. The rapid turnover of intestinal epithelial cells forms a crucial physical, as well as functional, barrier, and renewal is sustained by the vigorous proliferation of epithelial progenitor cells that migrate upwards from the bottom of the intestinal crypts. Nevertheless, many pathogenic bacteria, including Shigella, are capable of colonizing the intestinal epithelium. Recent studies have indicated that a growing family of bacterial toxins, effectors and small compounds, called cyclomodulins, are capable of interfering with the eukaryotic cell cycle30,31. Some of the cyclomodulins — for example, the EPEC protein cycling inhibitor factor (Cif) — inhibit host-cell mitosis after transfer from the bacteria through the T3SS. Cells that have been transformed by Cif accumulate 4n DNA and re-initiate DNA synthesis without dividing, which results in cells that contain 8–16n DNA31.
Shigella T3SS effectors
Nod1/2
OspG
PGN LPS
IKK complex
p65
IpaHs
β-TrCP
IκB MAPK
E2s
Ub UbcH5b
RICK
p50
Ub
IκB
NF-κB
Ub
Unknown
Proteasome OspF P P
IpaH9.8 MAPK
MAPK
U2AF35
Interleukin-8 production
IL-8 Histone H3
p65 p50
Figure 3 | Shigella and the downregulation of the host inflammatory response. During the multiplication of Shigella in epithelial cells, the bacteria shed components, as peptidoglycan Naturesuch Reviews | Microbiology (PGN) and lipopolysaccharide (LPS), into the cytoplasm. The recognition of PGN by Nod1 activates the nuclear factor (NF)-κB and mitogen-activated protein kinase (MAPK)-dependent inflammatory signals. Shigella delivers a set of effectors, including IpaH9.8, IpaHs (IpaH9.8 homologues), OspF and OspG, into the host cell through the type III secretion system (T3SS), which enables them to circumvent the host inflammatory response and inactivate the innate immune system. See the main text for details.
Other cyclomodulins, the cytolethal distending toxins (CDTs), are produced by Shigella dysenteriae, Campylobacter jejuni, E. coli and Salmonella enterica serovar Typhi (S. typhi). One of the CDTs that is produced by C. jejuni possesses a deoxyribonuclease‑I-like activity and causes limited DNA damage when it is delivered into the host-cell nucleus, which leads to the activation of ATM (a PI3 kinase) and eventually results in cell-cycle arrest32. Helicobacter pylori VacA, which induces cellular vacuolation in epithelial cells, is also capable of efficiently blocking the proliferation of T cells by inducing G1/S cell-cycle arrest33. Although the biological importance of each cyclomodulin and its target host cells in bacterial infection remains largely unclear, it is expected that some cell-cycle inhibitors will prolong the pathogen’s presence by interfering with the rapid turnover of epithelial cells.
nature reviews | microbiology
Shigella IpaB is involved in mediating the translocation of effectors through the T3SS and also functions as an invasin that interacts with CD44 (Refs 1,4,7). It has been shown that IpaB is delivered into the epithelial-cell cytoplasm by intracellular bacteria and causes cell-cycle arrest by targeting Mad2L2, an anaphase-promoting complex/cyclosome (APC) inhibitor34. Cell-cycle progression is stringently controlled by cell-cycle-specific proteolysis, which involves the ubiquitination of target proteins by two main types of E3 ligase complexes: the APC complex and the SCF complex35. The APC complex is a multisubunit complex that targets substrates for degradation only during mitosis and the G1 phase; it also targets mitotic cyclin A and cyclin B1, so allowing mitotic progression. Cyclin B1 ubiquitination assays have shown that APC undergoes unscheduled activation in response to volume 6 | january 2008 | 15
© 2008 Nature Publishing Group
Progress IpaB interaction with Mad2L2. Indeed, synchronized HeLa cells infected with Shigella fail to accumulate APC substrates, such as cyclin B1, Cdc20 and Plk1, which causes cell-cycle arrest at the G2/M phase in an IpaB–Mad2L2-dependent manner. IpaB–Mad2L2-dependent cell-cycle arrest by Shigella infection can be visualized in the intestinal crypt progenitors of rabbit ileal loops, and the IpaB-mediated arrest contributes to the efficient colonization of the host cells34. The finding that bacterial activity can retard intestinal epithelial renewal for the purpose of gaining an infectious foothold adds a new facet to bacterial infection. Conclusions In this Progress article, we have discussed recent advances in our understanding of the strategies that are used by Shigella to infect the intestinal epithelium. During basolateral entry into polarized epithelial cells in the early stage of infection, Shigella deliver a subset of effectors, including IpaA, IpaB, IpaC, IpgB1, IpgB2, IpgD and VirA, around the bacterial surface and directly into the host cells through the T3SS (TABLE 1). During multiplication within the epithelium, Shigella secrete another subset of effectors, including IcsB, VirA, OspF, OspG and IpaH family proteins, again through the T3SS, which enable the bacteria to survive intracellularly, promote intraand intercellular movement and modulate the host inflammatory response (TABLE 1). Shigella then seem to directly access the intestinal crypts and invade non-polarized epithelial progenitor cells, in which they deliver IpaB and slow the progression of the cell cycle, thereby prolonging the lifespan of their replicative niche. From the work reviewed here and elsewhere1–5, it is clear that bacterial effectors have pivotal roles in the various aspects of bacterial infection. Importantly, some of the pathogenic functions of the effectors that are involved in infection or in bacterial defence against the host’s immune system are occasionally shared by some other pathogens (TABLE 1). Consequently, studies of the roles of Shigella effectors (and effectors from other bacteria) will contribute to our understanding of host–bacteria interactions as highly dynamic and widely divergent biological events.
Chihiro Sasakawa is also at the Department of Infectious Disease Control, International Research Center for Infectious Diseases, Institute of Medical Science, University of Tokyo, 4‑6‑1, Shirokanedai, Minato-ku, Tokyo 108‑8639, Japan and CREST, Japan Science and Technology Agency (JST), Kawaguchi, 332‑0012, Japan. Correspondence to C.S. e‑mail:
[email protected] doi:10.1038/nrmicro1814 Published online 3 December 2007 1. 2. 3. 4. 5.
6. 7.
8. 9. 10.
11. 12. 13.
14.
15.
16. 17. 18.
19. 20. 21. 22.
Michinaga Ogawa, Yutaka Handa, Hiroshi Ashida, Masato Suzuki and Chihiro Sasakawa are at the Department of Microbiology and Immunology, Institute of Medical Science, University of Tokyo, 4‑6‑1, Shirokanedai, Minato-ku, Tokyo 108‑8639, Japan.
23.
Ogawa, M. & Sasakawa, C. Intracellular survival of Shigella. Cell. Microbiol. 8, 177–184 (2006). Sansonetti, P. J. & Di Santo, J. P. Debugging how bacteria manipulate the immune response. Immunity 26, 149–161 (2007). Mattoo, S., Lee, Y. M. & Dixon, J. E. Interactions of bacterial effector proteins with host proteins. Curr. Opin. Immunol. 19, 392–401 (2007). Cossart, P. & Sansonetti, P. J. Bacterial invasion: the paradigms of enteroinvasive pathogens. Science 304, 242–248 (2004). Nhieu, G. T., Enninga, J., Sansonetti, P. & Grompone, G. Tyrosine kinase signaling and type III effectors orchestrating Shigella invasion. Curr. Opin. Microbiol. 8, 16–20 (2005). Skoudy, A. et al. CD44 binds to the Shigella IpaB protein and participates in bacterial invasion of epithelial cells. Cell. Microbiol. 2, 19–33 (2000). Niebuhr, K. et al. Conversion of PtdIns(4,5)P2 into PtdIns(5)P by the S. flexneri effector IpgD reorganizes host cell morphology. EMBO J. 21, 5069–5078 (2002). Pendaries, C. et al. PtdIns(5)P activates the host cell PI3-kinase/Akt pathway during Shigella flexneri infection. EMBO J. 25, 1024–1034 (2006). Bourdet-Sicard, R. et al. Binding of the Shigella protein IpaA to vinculin induces F‑actin depolymerization. EMBO J. 18, 5853–5862 (1999). Demali, K. A., Jue, A. L. & Burridge, K. IpaA targets β1 integrins and rho to promote actin cytoskeleton rearrangements necessary for Shigella entry. J. Biol. Chem. 281, 39534–39541 (2006). Handa, Y. et al. Shigella IpgB1 promotes bacterial entry through the ELMO–Dock180 machinery. Nature Cell Biol. 9, 121–128 (2007). Alto, N. M. et al. Identification of a bacterial type III effector family with G protein mimicry functions. Cell 124, 133–145 (2006). Yoshida, S. et al. Shigella deliver an effector protein to trigger host microtubule destabilization, which promotes Rac1 activity and efficient bacterial internalization. EMBO J. 21, 2923–2935 (2002). Matsuzawa, T., Kuwae, A., Yoshida, S., Sasakawa, C. & Abe, A. Enteropathogenic Escherichia coli activates the RhoA signaling pathway via the stimulation of GEF‑H1. EMBO J. 23, 3570–3582 (2004). Suzuki, T., Miki, H., Takenawa, T. & Sasakawa, C. Neural Wiskott–Aldrich syndrome protein is implicated in the actin-based motility of Shigella flexneri. EMBO J. 17, 2767–2776 (1998). Pantaloni, D., Le Clainche, C. & Carlier, M. F. Mechanism of actin-based motility. Science 292, 1502–1506 (2001). Yoshida, S. et al. Microtubule-severing activity of Shigella is pivotal for intercellular spreading. Science 314, 985–989 (2006). Pfeuffer, T., Goebel, W., Laubinger, J., Bachmann, M. & Kuhn, M. LaXp180, a mammalian ActA-binding protein, identified with the yeast two-hybrid system, co-localizes with intracellular Listeria monocytogenes. Cell. Microbiol. 2, 101–114 (2000). Mizushima, N. & Klionsky, D. J. Protein turnover via autophagy: implications for metabolism. Annu. Rev. Nutr. 27, 19–40 (2007). Ogawa, M. et al. Escape of intracellular Shigella from autophagy. Science 307, 727–731 (2005). Birmingham, C. L. et al. Listeria monocytogenes evades killing by autophagy during colonization of host cells. Autophagy 3, 442–451 (2007). Fritz, J. H., Ferrero, R. L., Philpott, D. J. & Girardin, S. E. Nod-like proteins in immunity, inflammation and disease. Nature Immunol. 7, 1250–1257 (2006). Ashida, H., Toyotome, T., Nagai, T. & Sasakawa, C. Shigella chromosomal IpaH proteins are secreted via the type III secretion system and act as effectors. Mol. Microbiol. 63, 680–693 (2007).
16 | january 2008 | volume 6
24. Okuda, J. et al. Shigella effector IpaH9.8 binds to a splicing factor U2AF(35) to modulate host immune responses. Biochem. Biophys. Res. Commun. 333, 531–539 (2005). 25. Rohde, J. R., Breitkreutz, A., Chenal, A., Sansonetti, P. J. & Parsot, C. Type III secretion effectors of the IpaH family are E3 ubiquitin ligases. Cell Host Microbe 1, 77–83 (2007). 26. Kim, D. W. et al. The Shigella flexneri effector OspG interferes with innate immune responses by targeting ubiquitin-conjugating enzymes. Proc. Natl Acad. Sci. USA 102, 14046–14051 (2005). 27. Arbibe, L. et al. An injected bacterial effector targets chromatin access for transcription factor NF‑κB to alter transcription of host genes involved in immune responses. Nature Immunol. 8, 47–56 (2007). 28. Kramer, R. W. et al. Yeast functional genomic screens lead to identification of a role for a bacterial effector in innate immunity regulation. PLoS Pathog. 3, 179–190 (2007). 29. Zhang, J. et al. A Pseudomonas syringae effector inactivates MAPKs to suppress PAMP-induced immunity in plants. Cell Host Microbe 1, 175–185 (2007). 30. Nougayrede, J. P., Taieb, F., De Rycke, J. & Oswald, E. Cyclomodulins: bacterial effectors that modulate the eukaryotic cell cycle. Trends Microbiol. 13, 103–110 (2005). 31. Oswald, E., Nougayrede, J. P., Taieb, F. & Sugai, M. Bacterial toxins that modulate host cell-cycle progression. Curr. Opin. Microbiol. 8, 83–91 (2005). 32. Lara-Tejero, M. & Galan, J. E. A bacterial toxin that controls cell cycle progression as a deoxyribonuclease I‑like protein. Science 290, 354–357 (2000). 33. Gebert, B., Fischer, W., Weiss, E., Hoffmann, R. & Haas, R. Helicobacter pylori vacuolating cytotoxin inhibits T lymphocyte activation. Science 301, 1099–1102 (2003). 34. Iwai, H. et al. A bacterial effector targets Mad2L2, an APC inhibitor, to modulate host cell cycling. Cell 130, 611–623 (2007). 35. Vodermaier, H. C. APC/C and SCF: controlling each other and the cell cycle. Curr. Biol. 14, R787–R796 (2004). 36. Chen, Y., Smith, M. R., Thirumalai, K. & Zychlinsky, A. A bacterial invasin induces macrophage apoptosis by binding directly to ICE. EMBO J. 15, 3853–3860 (1996). 37. Nhieu, G. T., Caron, E., Hall, A. & Sansonetti, P. J. IpaC induces actin polymerization and filopodia formation during Shigella entry into epithelial cells. EMBO J. 18, 3249–3262 (1999). 38. Zurawski, D. V., Mitsuhata, C., Mumy, K. L., McCormick, B. A. & Maurelli, A. T. OspF and OspC1 are Shigella flexneri type III secretion system effectors that are required for postinvasion aspects of virulence. Infect. Immun. 74, 5964–5976 (2006). 39. Miura, M. et al. OspE2 of Shigella sonnei is required for the maintenance of cell architecture of bacterium-infected cells. Infect. Immun. 74, 2587–2595 (2006). 40. Fernandez-Prada, C. M. et al. Shigella flexneri IpaH7.8 facilitates escape of virulent bacteria from the endocytic vacuoles of mouse and human macrophages. Infect. Immun. 68, 3608–3619 (2000).
Acknowledgments
This work was supported by a Grant-in-aid for the Scientific Research on Priority Areas, Grant-in-aid-for Scientific Research (C) and Grant-in-aid-for Young Scientists (B) from the Ministry of Education, Culture, Sports, Science and Technology of Japan (MEXT) and the Special Coordination Funds for Promoting Science from Japan Science and Technology Agency (JSTA).
DATABASES Entrez Genome Project: http://www.ncbi.nlm.nih.gov/ entrez/query.fcgi?db=genomeprj Bradyrhizobium japonicum | Burkholderia pseudomallei | Campylobacter jejuni | Coxiella burnetii | Edwardsiella ictaluri | Escherichia coli | Helicobacter pylori | Legionella pneumophila | Listeria monocytogenes | Mycobacterium marinum | Mycobacterium tuberculosis | Pseudomonas syringae | Rickettsia conorii | Salmonella typhi | Shigella dysenteriae | Staphylococcus aureus | Streptococcus pyogenes All links are active in the online pdf
www.nature.com/reviews/micro © 2008 Nature Publishing Group
REVIEWS The biology and future prospects of antivirulence therapies Lynette Cegelski*, Garland R. Marshall‡, Gary R. Eldridge§ and Scott J. Hultgren*
Abstract | The emergence and increasing prevalence of bacterial strains that are resistant to available antibiotics demand the discovery of new therapeutic approaches. Targeting bacterial virulence is an alternative approach to antimicrobial therapy that offers promising opportunities to inhibit pathogenesis and its consequences without placing immediate lifeor-death pressure on the target bacterium. Certain virulence factors have been shown to be potential targets for drug design and therapeutic intervention, whereas new insights are crucial for exploiting others. Targeting virulence represents a new paradigm to empower the clinician to prevent and treat infectious diseases.
Riboswitch An mRNA control element that changes conformation in response to the binding of a metabolite (for example, glycine, lysine and coenzyme B12) and influences gene expression.
Microbiota The entire collection of microorganisms (bacteria, archaea, fungi, sometimes protozoa and viruses) that are resident on or in the host.
*Department of Molecular Microbiology, ‡Center for Computational Biology and Department of Biochemistry and Molecular Biophysics, Washington University, Saint Louis, Missouri 63110, USA. § Sequoia Sciences, Saint Louis, Missouri 63114, USA. Correspondence to S.J.H. e-mail: hultgren@borcim. wustl.edu doi:10.1038/nrmicro1818
Bacteria that are resistant to current antibiotics have wreaked havoc in the clinic and are a primary cause of death in the intensive-care units of our hospitals worldwide1,2. Currently, most infections are caused by important bacterial pathogens, such as Staphylococcus aureus, that are penicillin resistant, and up to 50% are resistant to stronger drugs, such as meticillin3. Meticillin-resistant S. aureus (MRSA) infections have reached epidemic levels this autumn (2007) in some parts of the United States, and are spreading through many sports centres, schools and gymnasiums, affecting predominantly student athletes, but also younger schoolchildren, and have already caused deaths in a matter of weeks. Moreover, as most bacterial strains are becoming resistant to multiple antibiotics, including vancomycin (the current drug of choice for the treatment of MRSA), there is an urgent need for antibiotics that have new modes of action on therapeutic targets4,5. Currently used antibiotics were derived by screening natural products and compound libraries against whole organisms, which identified compounds that have bacteriostatic or bacteriocidal activity. Analogues of these parent drugs, which have improved potency, pharmacological properties and efficacy against resistant strains, have increased the number of antibiotics that are clinically available6. However, the commercialization of new classes of antibiotics over the past 20 years has not met expectations, and current pharmaceutical pipelines lack new, broad-spectrum antibiotics7–9. The antibiotics that are available today are primarily variations on a single theme — bacterial eradication based on different modes of action at the molecular level. Some target cell-wall biosynthesis, whereas others
nature reviews | microbiology
inhibit protein synthesis or DNA replication. More recently, fatty-acid biosynthesis has been proposed as a viable bactericidal target. Lysine analogues have also been identified that target lysine riboswitches and inhibit bacterial growth10. Further study of such new targets is an important element in the development of new drugs. However, all these strategies target bacterial cellular processes that are crucial for microbial survival. In our current battle against infectious diseases, clinicians are limited to the use of antibiotics that stimulate bacterial evolution11,12. New tactics and weapons are needed to combat bacteria that are, owing to evolution and selection, moving targets. Targeting bacterial virulence is an alternative approach to the development of new antimicrobials that can be used to disarm pathogens in the host13–15. The overall strategy is to inhibit specific mechanisms that promote infection and are essential to persistence in a pathogenic cascade (for example, binding, invasion, subversion of host defences and chemical signalling), and/or cause disease symptoms (for example, the secretion of toxins). Stripping microorganisms of their virulence properties without threatening their existence may offer a reduced selection pressure for drug-resistant mutations. Virulence-specific therapeutics would also avoid the undesirable dramatic alterations of the host microbiota that are associated with current antibiotics. Indeed, the microbial cells that comprise the microbiota of a healthy human outnumber human cells by tenfold16; they colonize distinct sites throughout the body and confer numerous advantages to the host, some of which are only beginning to be understood. The balance of bacterial populations in the gut, for example, influences volume 6 | january 2008 | 17
© 2008 Nature Publishing Group
REVIEWS Table 1 | Representative adhesive fibres, fibre classification and disease association Adhesive fibre
Assembly proteins
Adhesin
Organisms
Associated diseases
Fibres that use the chaperone–usher pathway Type 1 pili
FimC and FimD
FimH
Escherichia coli, Klebsiella pneumoniae and Salmonella species
Cystitis
P pili
PapD and PapC
PapG
E. coli
Cystitis and pyelonephritis
Prs pili
PrsD and PrsC
PrsG
E. coli
Cystitis
S pili
SfaE and SfaF
SfaS
E. coli
Urinary-tract infection and newborn meningitis
Hif pili
HifB and HifC
HifE
Haemophilus influenzae
Otitis media and meningitis
Type 2 and 3 pili
FimB and FimC
FimD
Bordetella pertussis
Whooping cough
Pef pili
PefD and PefC
Unknown
Salmonella enterica serovar Typhimurium (S. typhimurium)
Gastroenteritis
Long polar fimbriae
LpfB and LpfC
Unknown
S. typhimurium
Gastroenteritis
MR/K(type 3) pili
MrkB and MrkC
MrkD
K. pneumoniae
Pneumonia
Myf fimbriae
MyfB and MyfC
Unknown
Yersinia enterocolitica
Enterocolitis
CooD
E. coli
Diarrhoea
Alternate chaperone pathway CS1 pili
CooB and CooC
Extracellular nucleation-precipitation pathway Curli
CsgB (nucleator), CsgE and CsgF (assembly) and CsgG (secretion)
CsgA (major subunit)
E. coli
Sepsis
Tafi
AgfB (nucleator)
AgfA (major subunit)
Salmonella enterica serovar Enteritidis
Mouse typhoid
General secretion apparatus
PilC
Neisseria gonorrhoea
Gonorrhoea
Pilin protein
Pseudomonas aeruginosa, Vibrio cholerae, Cholera Mycobacterium bovis and Dichelobacter nodosis
LPXTG-motif-mediated export and covalent polymerization
Unknown
Corynebacterium diphtheriae, Streptococcus mutans and Streptococcus pneumoniae
General secretion pathway Type 4 pili
Gram-positive pili Pili
Pneumoccocal diseases
Tafi, thin aggregative fimbriae.
Pilus A non-flagellar filamentous appendage that is formed on the surface of many bacteria.
Quorum sensing (QS). The process by which bacteria use signalling molecules to monitor bacterial density and coordinate gene expression in a populationdensity-dependent manner.
Adhesin The surface-exposed bacterial molecule that mediates specific binding to a receptor or ligand on a target cell.
energy harvest and caloric intake through complex interbacterial metabolic networks17. The microbiota is dynamic, and shifts in the balance of microorganisms that alter the sizes of different bacterial populations — for example, the use of traditional antibiotic therapy — can lead to the loss of symbiotic benefits and the proliferation of disease-causing opportunistic pathogens. A commitment to develop therapeutics that target virulence requires a serious change in our perspective for treating infectious diseases. Some elements of virulence do seem to be fundamental for many pathogens, and drugs that target these elements should exhibit broaderspectrum activity. However, many antivirulence drugs could be designed to target specific pathogens and the virulence factors that are unique to their pathogenic cascades. In addition, virulence-gene expression is a function of time and space throughout infection, in which factors such as pili could function early to mediate adhesion, and others, such as toxin production and quorum sensing (QS), could operate later, and in a different
18 | january 2008 | volume 6
niche, as bacteria replicate and respond to host defences. As the coordinated regulation of virulence-gene expression is dynamic and location specific in the host, it should be possible to identify and target vital genetic or molecular bottlenecks — that is, to target the Achilles’ heel of a pathogen during infection18–20. Systems biology is engaged in mapping the genetic control networks and molecular correlates of pathogenesis (ideally, using parameters obtained from relevant in vivo models) to drive the discovery of these bottlenecks and identify new drug targets. Vaccination and other immunomodulatory strategies are additional crucial avenues that are being pursued to combat infectious disease and antibiotic resistance, both in immunocompetent and immunologically compromised hosts. The complex relationship between host immunity and microbial pathogenesis, the balance between protective immunity and immunopathology and strategies to exploit the many networks that are involved in antimicrobial strategies have been extensively reviewed21–24 and will not be discussed in detail www.nature.com/reviews/micro
© 2008 Nature Publishing Group
REVIEWS Binding
Superficial facet cell Spread to new cells
Invasion and replication
Underlying transitional cell
Biofilm formation IBC
Biomass dispersion and cell exit
Figure 1 | Multi-step pathogenic cascade of uropathogenic Escherichia coli (UPEC). UPEC coordinate highly organized temporal and spatial events to Reviews colonize| Microbiology the urinary Nature tract. UPEC bind to and invade the superficial facet cells that line the bladder lumen, where they rapidly replicate to form a biofilm-like intracellular bacterial community (IBC). In the IBC, bacteria find safe haven, are resistant to antibiotics and subvert clearance by host innate immune responses. UPEC can persist for months in a quiescent bladder reservoir following acute infection and challenge current antimicrobial therapies. Quiescent bacteria can re-emerge from their protected intracellular niche and be a source of recurrent urinary-tract infections. Insight into the processes that accompany IBC formation and biofilm dispersal, as well as the factors that drive bacteria into the reservoir, may aid the design of preventive or therapeutic strategies for recurrent infections.
Autotransporter A large family of secreted proteins in Gram-negative bacteria that harbour three functional domains — the amino-terminal signal peptide, the secreted mature protein (passenger domain) and a carboxy-terminal translocator domain — to allow secretion of the passenger protein.
Biofilm A community of cells that are attached to a surface or interface or to each other, and are imbedded in a self-made, protective matrix of extracellular polymeric substances.
here. In this Review, we highlight the diversity of the bacterial virulence mechanisms and consider their consequences in infectious disease. We review recent efforts towards antivirulence-based drug discovery in the framework of marketable drugs, and discuss the challenges that remain and factors that are crucial to developing the antivirulence therapeutic approach.
Microbial attachment and invasion The physical interaction between the pathogen and the host is crucial to the pathogenesis of virtually every bacterial disease. Specific proteins called adhesins mediate the adhesive interactions between the pathogen and a cell-surface ligand on the host cell that is required for host colonization. The ability to impair bacterial adhesion represents an ideal strategy to combat bacterial pathogenesis because of its importance early in the infectious process, and it is also suitable for implementation as a prophylactic to prevent infection.
nature reviews | microbiology
Pathogens are capable of presenting multiple adhesins that can be expressed differentially to permit binding in specific sites and at particular times over the course of a complex infectious cycle. Thus, it may be difficult to develop a universal class of anti-adherence drugs. Nevertheless, several specific pathogenic adhesive strategies have emerged as hallmark requirements for virulence in certain infectious diseases that are immediate targets for drug discovery and development. Some bacteria present non-fimbrial adhesins on their surface. These are expressed as monomeric proteins or protein complexes that assemble at the cell surface — for example, the Dr family of adhesins that are expressed by Escherichia coli and are important for adhesion in the intestine and urinary tract. Adhesive autotransporters represent a class of afimbrial adhesins that are expressed by various unrelated micro organisms, including species of Rickettsia, Bordetella, Neisseria and Helicobacter and many members of the family Enterobacteriaceae. Haemophilus influenzae, a causative agent of sinusitis, bronchitis and otitis media, expresses an adhesive autotransporter called Hap that mediates binding to components of the host-cell extracellular matrix. Most adhesins, however, are incorporated into hetero polymeric extracellular fibres called pili or fimbriae. Many distinct virulence fibres have been described in Gram-negative organisms25. Pili are also produced by Gram-positive organisms (reviewed in REF. 26) and have been linked to virulence in Streptococcus pneumoniae27. Although bacterial fimbriae have diverse functions, many seem to be crucial to the binding and persistence of pathogenic microorganisms in the host (TABLE 1). In the following subsections, we review the importance of pilus-mediated adhesion by E. coli in the pathogenesis of urinary-tract infection (UTI) and discuss strategies to inhibit pilus biogenesis. We also describe the need for new approaches to prevent and treat UTI and examine the market considerations for antivirulence therapeutics. Denying access to uropathogens. Uropathogenic E. coli (UPEC) are the major causative agents of UTI and engage in a coordinated and regulated genetic and mol ecular cascade to assemble type 1 and P pili, which are associated with infections of the bladder and kidney, respectively (reviewed in REF. 28). Virtually all clinical UPEC isolates express type 1 pili29. These are required to bind mannose-containing host receptors, invade host bladder-epithelial (urothelial) cells 30,31 and initiate a pathogenic cascade that involves several distinct phases as examined in the mouse cystitis model32,33 and human UTIs34 (FIG. 1). Within urothelial cells, bacteria first replicate rapidly to form dense biofilm-like communities and are protected from the flow of urine and host defences33,35. UPEC eventually detach and then disperse, or flux, from the intracellular bacterial community (IBC) to initiate new rounds of IBC formation in other cells. Some fluxing bacteria form filaments, evade neutrophil phagocytosis and facilitate bacterial survival33,35. Even after acute infection is resolved, bacteria can remain within the volume 6 | january 2008 | 19
© 2008 Nature Publishing Group
REVIEWS a Uroplakin
E. coli
Type 1 pilus
Superficial facet cell
b
No compound
Plus pilicide
S N N
O
CO2Li
O
Pilicide
Figure 2 | Targeting microbial adhesion. a | Pathogenic Escherichia coli use type 1 pili to bind to hexameric uroplakin protein arrays on the surface of superficial facet cells that line the bladder lumen. Type 1 pili mediate the binding to, and subsequent invasion of, these cells. b | Pyridone-based pilicides inhibit pilus biogenesis by disrupting chaperone–usher Nature Reviews | Microbiology protein interactions and dramatically reduce piliation levels. Image on the left in panel a reproduced, with permission, from REF. 102 (1995) National Academy of Sciences. Image on the right in panel a reproduced, with permission, from REF. 35 (1998) American Association for the Advancement of Science. Images in panel b reproduced, with permission, from REF. 45 (2006) National Academy of Sciences.
Chaperone–usher system A system that facilitates the folding, transport and ordered assembly of pilus subunits at the cell surface.
bladder for many days to weeks, regardless of standard antibiotic treatments, and can be implicated in recurrent infection36. It might be possible to target several factors in the pathogenic cascade to inhibit virulence. Type 1 pili represent an attractive drug target because the bottleneck of invasion and IBC formation selects for fitness in the urothelium and type 1 pili are required for both events. Two general strategies have emerged to inhibit pilus-mediated function. The first is adhesion specific and involves physically precluding pathogen binding to host cells. Carbohydrate derivatives of host ligands, for example, have showed efficacy in blocking the adhesive properties of both type 1 and P pili in biophysical and haemaglutination assays 38–40. This approach can be readily extended to other adherent organisms by tailoring the antiadhesive compounds to their receptor specificities in vivo. The goal of the second strategy is to interrupt pilus assembly, which also blocks pilus-mediated adhesion as well as invasion and intracellular biofilm formation. Type 1 and P pili are both assembled by the chaperone– usher system41. Although the fim and pap operons, which are associated with type 1 and P pili, respectively, are the best studied of these systems, 17 additional chaperone–usher operons have been identified in sequenced E. coli genomes42. The chaperone–usher systems are
20 | january 2008 | volume 6
also necessary for the assembly of extracellular adhesive organelles in a wide range of other pathogens, including species of Salmonella, Pseudomonas, Haemophilus, Klebsiella and Yersinia. Therefore, inhibitors of the chaperone–usher system may serve as broader-range therapeutics, an attractive feature that would enhance the marketability of an effective drug. Pilicides are a class of pilus inhibitors that target chaperone function and inhibit pilus biogenesis43,44. A new class of pilicides, based on a bi-cyclic 2‑pyridone scaffold, inhibit both type 1 and P pili assembly in E. coli by targeting conserved regions on chaperones that are ubiquitous in the chaperone–usher pathways45 (fig. 2). These pilicides also inhibit biofilm formation in E. coli. The synthetic pilicides disrupt an essential protein–protein interaction between the chaperone and usher at a site that has been identified by x‑ray crystallography. Thus, pilicides have the opportunity to work either at the level of attachment and invasion or the level of bacterial aggregation once they are inside the superficial facet cells that line the bladder lumen. Therapeutic outlook for UTIs. The urinary tract is a common site of infection in humans, and UTIs result in more than 8 million outpatient visits per year in the United States and, in the year 2000, expended costs of US$3.5 billion for evaluation and treatment 46,47. Women are the most frequent sufferers — a female has a 50% chance www.nature.com/reviews/micro
© 2008 Nature Publishing Group
REVIEWS of developing an acute UTI during her lifetime — and many experience recurrent infections. With advancing age and co-morbidities, UTI becomes progressively more common among both women and men, particularly in the context of urinary catheterization. Limited treatment options are available for patients with chronic and recurrent UTIs. Although these patients can be treated using prolonged courses of antibiotics, this radically disrupts the symbiotic host–microorganism balance and may be accompanied by the evolution of drug-resistant organisms in the urinary tract48. In addition, UTIs have a strong causal correlation with systemic infection and sepsis if antibiotic therapy is ineffective49. Because infection of the bladder is not limited to extracellular colonization, and can involve both the invasion of host cells and the subversion of host defences and other specific mechanisms, such as filamentation, to promote persistence and re-invasion, it may be essential to target key bottlenecks, such as type 1 pili expression and/or assembly. Current research efforts are underway to identify other candidate nodes in the network of these host–pathogen interactions. The treatment of recurrent or chronic UTIs, in which intracellular bacterial reservoirs already exist, might require a synergistic therapy, such as combining a pilicide, for example, with a stimulator of the immune response or epithelial-cell renewal (to drive bacteria out of intracellular niches). Scientific research and new drug development for UTIs accompany a clear clinical demand, and are associated with large patient populations for clinical-stage development and the potential long-term profits that effective therapies will produce. We anticipate that drug-discovery and development efforts for UTIs and other chronic infections will increase markedly over the next 5–10 years.
a Toxin transcription
Bacterial toxins Toxin-powered pathogens can exert a devastating effect from a distance. For many infectious diseases, the clinical symptoms and cause of tissue damage can be attributed to the action of secreted bacterial toxins. Botulinum, anthrax, diphtheria, tetanus, cholera and Shiga toxins are produced by distinct bacterial pathogens, and each is a major cause of cellular malfunction and morbidity in afflicted individuals50. In addition, their extreme toxicity makes these toxins potential weapons for use in biological warfare or a terrorist attack. Multiple opportunities exist to prevent toxin damage to the host (FIG. 3).
Intracellular toxin target
Targeting a toxin transcription factor. Vibrio cholerae infection is characterized by severe diarrhoea and dehydration, which becomes life threatening if effective treatment of the symptoms is delayed. Cholera toxin provides V. cholerae with its hallmark virulence and triggers the debilitating symptoms of the disease by interacting with G proteins and cyclic AMP in the intestinal lining to interrupt proper ion transport, which results in massive fluid loss51. The potent toxin has been recognized by Hung and colleagues52 as an ideal candidate for drug development, which, as resistance has emerged to the antibiotics of choice, ciprofloxacin and azithromycin, nature reviews | microbiology
V. cholerae
Transcription initiator Bacterial toxin
b Toxin trafficking and function
Receptor mimic Host membrane
Host toxin receptor
Intracellular inhibitor
Figure 3 | Targeting toxin-powered pathogens. a | The inhibition of toxin transcription, as described Vibrio Nature Reviews for | Microbiology cholerae, is one way to inhibit the consequences of toxinmediated virulence. b | Neutralizing toxins, or preventing their trafficking and/or enzymatic activity, at cellular targets is an alternative strategy to inhibit toxin damage to the host.
is particularly needed. The small molecule virstatin was discovered by screening for inhibitors of toxT gene expression. ToxT is a transcription factor that activates the gene transcription of both cholera toxin and another V. cholerae virulence factor, the toxin-co-regulated pilus52. Thus, virstatin inhibits cholerae-gene expression, the earliest step in toxin production. Selective inhibition of gene expression is a general strategy that can be implemented for many virulence factors, as expression is controlled by the environment that is sensed by the organism. In the clinic, however, such a therapy may have a short window of opportunity volume 6 | january 2008 | 21
© 2008 Nature Publishing Group
REVIEWS to work. On the one hand, after an extended time, bacteria may reach high numbers and produce sufficient toxin to overwhelm the host, such that inhibition of toxin production alone may be too late. On the other hand, effective and selective therapeutics could be useful prophylactically in limiting epidemics.
Botulism A rare, but serious illness that is caused by a nerve toxin, botulinum, that is produced by the bacterium Clostridium botulinum.
Chemical genetics The strategy of using small molecules to alter and interrogate biological processes. The small-molecule tools of dissection in this approach harbour the precious chemical scaffolds that may lead directly to new therapeutics.
Project BioShield The Project BioShield Act was incorporated into law by the United States government in July 2004. Through Project BioShield, $5.6 billion will be invested by 2013 in the development of new technological and therapeutic countermeasures against potential bioterrorism agents and to purchase and stockpile effective therapeutics to prevent and treat the illnesses that are related to these threats.
Neutralizing toxins using antibodies. An antivirulence strategy has been the treatment of choice for the postexposure treatment of tetanus, diphtheria and botulinum toxins. Antibodies against the toxins are administered to patients to attempt to neutralize toxins while infection is cleared53. In addition to simply neutralizing toxins, benefit to the host may be ascribed, in part, to a more sophisticated host response that includes the antibodymediated recognition and potential clearance of pathogens and their toxic cargo, as reviewed elsewhere22–24. The neutralizing antibodies are typically obtained from the sera of immunized horses or humans, but the production of monoclonal antibodies is an attractive alternative. Many monoclonal-antibody therapies are already in use for preventing transplant rejection and oncological treatment; an antiviral monoclonal antibody, palivizumab, which prevents respiratory syncytial virus in children, was licensed in 1998. Currently, adult cases of botulism, a rare infectious disease that is caused by Clostridium botulinum, are treated with an antitoxin that contains horse antibodies raised against type A, type B and/or type E strains of botulinum neurotoxins. This drug is available from the Centers for Disease Control and Prevention (CDC), United States, if specially requested by a physician. It is listed in the Official Monographs of the United States Pharmacopeia, even though it has never been examined in controlled human clinical trials. Owing to the equine antitoxin’s potentially serious side effects, it has not typically been used to treat infants suffering from botulism. However, a landmark in antivirulence therapy occurred on 23 October 2003. The Food and Drug Administration in the United States licensed an antivirulence drug that contained antibotulinum-toxin antibodies produced from humans to the California Department of Health Services, based on their seminal clinical research that spanned approximately 15 years. A placebo-controlled clinical trial consisting of 122 infants with botulism demonstrated a profound clinical benefit for patients treated with human botulism immune globulin (BIG), subsequently named BabyBIG54. Clinical results showed a reduction in the length of hospital stay, duration of intensive care, duration of mechanical ventilation, duration of tube or intravenous feeding and hospital charges per patient. This remarkable clinical evidence strongly supports a treatment strategy of neutralizing bacterial toxins post-infection. Targeting toxin trafficking and function. Antibodyneutralization strategies will fall short for some patients and infectious diseases. Toxins that are delivered through the type III secretion system (T3SS; discussed below), for example, are delivered directly into host cells, do not circulate globally and are not ideal candidates
22 | january 2008 | volume 6
for antibody neutralization. The best defences against toxin-empowered pathogens that are encountered in the community, or as potential bioweapons, might require a better understanding of pathogenesis and toxin biochemistry. For many toxins, new research is needed to provide higher-resolution models of toxin trafficking than that presented in FIG. 3 (Ref. 55). In this regard, chemical genetic approaches are driving the discovery of toxin-trafficking inhibitors and, in turn, these small molecules are being used as tools, or molecular scalpels, to dissect the mechanistic details of toxin trafficking. Selective luciferase-based screening assays, for example, have identified lead compounds that inhibit the actions of E. coli’s Shiga toxin once it has gained access to the host cell, and therefore permit protein synthesis56,57. These compounds stop the toxin in real time along its path from the cell surface to the endoplasmic reticulum, and provide a unique opportunity to examine toxin-transport processes in more detail. The ability to target toxins after they have entered host cells, but before they exercise their function, provides an additional opportunity in time and space to control the sequelae of toxin action. Improved models of the assembly and function of anthrax toxin, produced by the Gram-positive Bacillus anthracis, provide several distinct checkpoints that can be targeted for interception55. Peptides and small-molecule inhibitors have been identified that bind to lethal factor (one of the three anthrax toxin components) and inhibit the in vitro enzymatic activity that is linked to pathogenesis in vivo58–60. Several molecules have been reported to block endosomal acidification, and so inhibit the conformational changes that are required for pore formation at the host-cell surface, and anthrax-toxin entry. Having multiple ways to target toxin-powered pathogens encourages the development of new and diverse therapeutics that promise to treat a large number of infectious diseases. Focus on bioterrorism. The dangers that are associated with an anthrax bioterrorism threat were made clear in 2001 when B. anthracis was disseminated throughout the United States postal system leaving five dead and many ill. Anthrax toxins, along with botulinum toxins, are classified in the highest risk category of biological warfare agents by the CDC, and nearly $6 billion have been allocated by Project BioShield to develop and stockpile antibiotics, antitoxins and vaccines for these and other high-threat agents61. Vaccination efforts are not generally promoted, however, because the chance of exposure to any particular bioterrorism agent is remote. PharmAthene was founded in 2001 to develop technologies and products to address these biosecurity needs. Their new antivirulence drug, Valortim, is a human monoclonal antibody that has demonstrated prophylactic and, separately, post-infection protection against B. anthracis toxins in rabbit and non-human primate models 62. It has successfully completed Phase I safety and pharmacokinetic evaluation. In 2006, Cangene Pharmaceuticals received a 5-year, $362-million contract from Project BioShield to develop and produce 200,000 doses of a potentially www.nature.com/reviews/micro
© 2008 Nature Publishing Group
REVIEWS improved botulinum antitoxin, a heptavalent antitoxin that is derived from horses (containing antibodies to 7 toxin subtypes), because of its potential use as a bioterror agent. Its initial order was formally received into the United States Strategic National Stockpile in September 2007.
Haemolytic uraemic syndrome A disease that primarily affects infants and children and is characterized by the loss and destruction of red blood cells. Occurs most commonly in children after a gastrointestinal infection or upper respiratorytract infection and can lead to kidney failure.
Scientific leaps and commercialization hurdles. Antibody-based toxin neutralization (discussed above), could be considered to be a simple approach to target toxin-based virulence that does not require a highresolution roadmap detailing toxin trafficking and subcellular targets. However, toxin-associated diseases are often not that simple. They can be extremely complex, with cascading symptoms that may or may not lead to life-threatening complications. Consequently, the AntiInfective Drugs Advisory Committee meeting held on 12 April 2007, when considering the operational details of executing clinical trials to demonstrate the effectiveness of a new antivirulence therapy, emphasized the pitfalls and hurdles that might impede the commercialization of any antivirulence therapy that targets small patient populations63. At this meeting, research results were presented for two separate monoclonal antibodies that are under development to treat Shiga-toxin-producing bacterial infections. Shiga toxin is produced by Shigella dysenteriae and Shiga-toxin-producing E. coli (STEC), which includes the prototype strain E. coli O157:H7 that has caused several food-borne outbreaks around the world. The toxin is secreted into the host cell and inhibits host protein synthesis, which can lead to multiple clinical manifestations, such as gastrointestinal disease, bloody diarrhoea, the destruction of red blood cells and platelets, and haemolytic uraemic syndrome64. Treatment using antibiotics is controversial owing to the bolus of toxin that could be released as bacteria die, which, potentially, could overwhelm the patient’s defences. Even though evidence of efficacy from animal models was presented at this meeting, representatives from companies and the physicians on the advisory committee were unable to successfully construct an appropriate clinical-trial design to evaluate these potential drugs. The two main challenges for the successful completion of clinical trials seem to be, first, that primary endpoints for this specific disease in humans are unclear and, second, that achieving statistical significance may not be attainable because the low incidence of this disease prevents adequate patient enrolment. The findings of this meeting strongly suggest that researchers should target the virulence mechanisms that are involved in infectious diseases with clear and measurable clinical end-points in sufficiently large patient populations. These realities of commercialization seem to make developing drugs against STEC and other complicated diseases that affect small patient populations less attractive compared with diseases that affect large patient populations, such as UTIs and chronic ear and sinus infections. These hurdles will accompany the development of future antivirulence products, and imminent decisions surrounding the potential Shiga drugs will influence the course of future endeavours.
nature reviews | microbiology
Type III secretion The Shiga and cholera toxins discussed above are exported by type II secretion, a general secretion system through which extracellular enzymes are also secreted65. However, other pathogens deliver their virulent cargo using the T3SS. This system orchestrates the export and delivery of virulence factors, or effector proteins, from the bacterial cytoplasm across the inner membrane, the peptidoglycan and outer membrane and, finally, through the host-cell plasma membrane, directly into the hostcell cytosol in the manner of a molecular syringe66. The T3SS machinery is used by many Gram-negative pathogens, such as E. coli, Salmonella enterica serovar Typhimurium, Shigella flexneri, Pseudomonas aeruginosa and species of Yersinia and Chlamydia. Yersinia species, such as Yersinia pestis, the causative agent of plague, inject effector proteins called YOPs (Yersinia outer proteins) through the T3SS into host cells to inhibit the host immune response and help them to evade a potent host response and subsequent clearance67,68. The conserved elements in the T3SS machinery that allow the interchange of T3SS components among some bacteria69,70 suggest that it may be possible to design broad-range T3SS inhibitors that are effective in treating disparate pathogens, regardless of the unique effector proteins that they deliver. Small-molecule inhibitors of the T3SS have been described in Yersinia pseudotuberculosis71,72. More recently, the discovery that smallmolecule T3SS inhibitors in Yersinia spp. inhibit the T3SS in Chlamydia trachomatis supports the notion of developing broader-range T3SS inhibitors. Interestingly, the role of the T3SS in the pathogenesis of Chlamydia spp. infection is not well understood, owing, in part, to the inability to manipulate Chlamydia spp. genetically. However, in the previously mentioned study, the smallmolecule inhibitors were recruited using chemical genetics to assess the potential importance of the T3SS in C. trachomatis infection. Treatment with the T3SS inhibitor did inhibit the in vivo pathogenic cascade and resulted in a decrease in secreted effector proteins, suggesting that the T3SS was inhibited and is essential to the C. trachomatis infectious cycle73–75. These T3SS inhibitors will be a useful tool for further study of the T3SS in Chlamydia spp., and might represent target leads for future therapeutics. Biofilms and chronic infections Biofilms are complex, organized bacterial assemblies that are highly resistant to antibiotics and host defences. Biofilms can form on abiotic surfaces, such as surgical implants and catheters, and result in persistent infections that are difficult to treat, thereby leading to further health complications and longer hospital stays. Bacteria in chronic wounds, which are particularly prevalent in the elderly and diabetic populations, form biofilms that prevent proper healing76. Indeed, the ability of bacteria to persist robustly in biofilm communities in the human host and the environment, even against antibiotic pressure, is a fundamental observation that extends across the field of microbiology and poses serious challenges to the control and treatment of chronic infectious disease. volume 6 | january 2008 | 23
© 2008 Nature Publishing Group
REVIEWS Table 2 | Selected two-component response systems (TCRSs) involved in virulence Histidine-kinase sensor and response regulator
Organisms
Associated virulence property
AgrC; AgrA
Staphylococcus aureus and Staphylococcus epidermidis
Control of most aggressive virulence factors
AlgD; AlgR
Pseudomonas aeruginosa
Alginate
RcsC; RcsB
Escherichia coli
Capsule
BvgS; BvgA
Bordetella pertussis
Toxins, adhesins and colonization factors
PhoQ; PhoP
P. aeruginosa, Salmonella Divalent cation sensing, modification of enterica serovar Typhimurium, lipopolysaccharide and resistance to antimicrobial Yersinia pestis and Vibrio peptides cholerae
TCRSs involved in the regulation of resistance to antibiotics VanS; VanR
Enterococcus faecalis
Vancomycin resistance
VncS; VncR
Streptococcus pneumoniae
Vancomycin resistance
RprX; RprY
Bacteroides fragilis
Tetracycline resistance
The intracellular biofilms that are formed by UPEC in the urinary tract after their invasion of the bladder epithelium and smaller numbers of quiescent bacteria that are harboured in underlying reservoirs are implicated in the aetiology of recurrent UTIs36. Clinical studies of recurrent infection of the middleear tissue in children indicate that chronic otitis media stems from a persistent biofilm following the first infection77. Helicobacter pylori biofilms have been documented in the biopsied gastric mucosa of patients suffering from gastric ulcers, and it has been suggested that the ulcers are manifestations of these biofilms 78. In addition, patients with cystic fibrosis are threatened by P. aeruginosa biofilm-mediated chronic lung infections, which are responsible for the high morbidity and mortality of patients with cystic fibrosis79,80. Several strategies have emerged to inhibit biofilm formation and eradicate established biofilms. These include, but are not limited to: preventing the initial adherence of bacteria to either a surface or another bacterium (discussed above); interrupting QS mechanisms that are required for the gene expression of biofilm components (discussed below); inhibiting the biosynthesis of the integral polysaccharide and proteinaceous extracellular components of the biofilm matrix81; and identifying or tailoring enzymes that can degrade the biofilm matrix, thereby rendering bacteria accessible to traditional antibiotics and/or immune clearance. If we examine the anti-infective marketplace, the health and economic impact of multidrug-resistant organisms, particularly in the hospital setting, is driving the primary market for drug development. However, markets that encompass recurrent and chronic bacterial infections consist of significantly larger patient populations that have considerable unmet clinical needs and provide attractive opportunities for drug-development efforts. Effective biofilm inhibitors could dramatically change treatment regimens for many infectious diseases and benefit large patient populations. 24 | january 2008 | volume 6
Quorum sensing QS is the illustrative term that is used to describe the chemical signalling that takes place among bacteria to keep track of their cellular density. QS is mediated by the production and subsequent recognition of small molecules called autoinducers, and is used to coordinate gene expression and regulate the numerous processes that are involved in community behaviour and virulence, for example, motility and biofilm formation (reviewed in Refs 82,83). In nature, the red seaweed Delisea pulchra produces chemical compounds called furanones that intercept QS signals and prevent microbial colonization84,85. This natural phenomenon has inspired the search for compounds that selectively inhibit QS, biofilm formation and the virulence of human pathogens during infection86. P. aeruginosa is frequently encountered in nosocomial infections and lung infections in patients with cystic fibrosis, and produces more than 30 QS‑regulated virulence factors. Numerous studies, predominantly carried out over the past 5 years, strongly support the notion that QS inhibitors, including isolated and synthetic furanones, can be effective in treating bacterial infections in vivo87,88. Notably, P. aeruginosa biofilms exhibit increased susceptibility to sodium dodecyl sulphate and the antibiotic tobramycin if treated with a synthetic furanone, compound C‑30 (Ref. 87). In a mouse pulmonary infection model, treatment with compound C‑30 resulted in increased bacterial clearance87. Thus, inhibition of QS can increase the susceptibility of biofilm bacteria to host defences and antibacterial agents. Givskov and colleagues87 have suggested that direct targeting of QS might be effective as an early prophylactic treatment of individuals who have P. aeruginosa infections in the lungs, on implants or in wounds. Other examples underscore the challenge of interrupting the right signals at the right times to reduce bacterial virulence in the host. A non-native N‑acylated‑l-homoserine lactone, for example, was recently discovered that can either inhibit or strongly induce QS in the marine symbiont Vibrio fischeri, depending on the molecule concentration89. Results from a panel of compounds in the www.nature.com/reviews/micro
© 2008 Nature Publishing Group
REVIEWS AIP
AgrC
1
AgrB Cytoplasm Histidinekinase sensor
P
2 3
AgrA
4
Response regulator
Nucleus P
AgrA Gene regulation
RNAIII
agrB
agrD
agrC
agrA
Figure 4 | Pairing quorum sensing and two-component signalling in the Nature Reviews | Microbiology staphylococcal agr system. Staphylococcus aureus uses a two-component response system (TCRS) to mediate quorum sensing (QS). The regulation of QS involves the production of an autoinducer and an increase in its concentration, expression of RNAIII and the subsequent regulation of QS genes. S. aureus produces an autoinducing peptide (AIP) that accumulates extracellularly and activates the TCRS. The TCRS involves signal recognition by a histidine kinase (AgrC) (1), followed by histidine phosphorylation (2) and phosphotransfer to a response regulator (AgrA) (3), which then binds to the RNAIII transcript that encodes a small RNA that functions to modulate gene expression of S. aureus genes (4).
same family elicited varied QS responses, which implies that the rational design of modified QS antagonists or agonists may not be straightforward; new insights are needed to fully understand these responses. In addition, redundancies among QS systems can render inhibitors of a single system ineffective. Indeed, two of the three parallel QS systems in V. cholerae are dispensable for host colonization and the production of two hallmark virulence factors, cholera toxin and the toxin co-regulated pilus90. QS circuits are being examined in numerous models to determine if similar redundancies exist and to what extent individual chemical signals influence multiple QS pathways. The generation of QS circuit diagrams may reveal viable drug targets and small molecules that may prove useful as scalpels to probe the roles of QS in different systems.
Two-component response systems A central requirement of bacterial virulence is the ability to express subsets of genes in response to signals that are specific for a particular environment. Two-component response systems (TCRSs) seem to be the dominant mechanism by which bacteria and fungi respond to nature reviews | microbiology
their environment91. TCRSs can control host invasion, drug resistance, motility, phosphate uptake, osmoregulation, nitrogen fixation and other functions. More than 4,000 TCRSs have been identified in approximately 400 sequenced bacterial genomes92,93. Selected TCRSs that regulate virulence are provided in TABLE 2. The essence of signal transduction lies in the recognition and interpretation of environmental signals that are related to host infection, and conversion of those signals into specific protein–protein interactions and transcriptional activation94. Bacterial transduction systems generally consist of receiver domains that are covalently linked to effector domains (FIG. 4). Specifically, each TCRS is composed of a histidine kinase that is activated by extracellular signals in the host environment and a response regulator that, in turn, transmits the signal to the intracellular target to modulate the gene expression of virulence factors. The interfaces between response regulators and their protein modulators can be targeted to prevent phosphorylation and/or activation, as well as the interfaces of homodimer formation of activated response regulators that are required for the control of gene expression. Broad-spectrum inhibitors could nonselectively target the TCRSs that are common to many bacteria. More selective inhibitors could target the interaction of specific phosphorylated response regulators to prevent expression of a specific virulence gene95. A TCRS is an integral component of the QS system in Gram-positive bacteria that responds to bacterial density, and is named agr for accessory gene regulator. The global regulator agr controls the expression of most virulence genes in staphylococci and is activated by secreted autoinducing peptides (AIPs) called thiolactones that comprise 7–10 amino acids96. AgrC and AgrA are the histidine-kinase sensor and response regulator, respectively, of the TCRS. AIP thiolactones have been suggested as leads for the development of S. aureus therapeutics that are urgently needed, owing to the fact that nosocomial and community-acquired MRSA infections are on the rise. The agr system also functions to downregulate factors that are important in biofilm formation, and agr dysfunction is associated with increased biofilm production97. As discussed above, a more detailed understanding of the organism’s control networks is needed to identify the ideal genetic or molecular checkpoints that need to be targeted to reduce virulence in the host. The prevention of virulence mechanisms that promote antibiotic tolerance could also improve the efficacy of current antibiotics, and, particularly, slow down the processes that lead to drug resistance. Several TCRSs are responsible for inducing the gene expression that confers bacterial tolerance to antibiotics. Vancomycin resistance, for example, is triggered by the VanR–VanS TCRS. VanS detects the glycopeptide antibiotic and VanR activates the expression of the enzymes VanA, VanH and VanX, all of which are required for resistance98,99 (reviewed in REF. 11). The TCRS proteins and the three enzymes cooperate to synthesize an altered peptidoglycan framework that is tolerant to vancomycin exposure, which is a hallmark of the vancomycinresistant enterococci and vancomycin-resistant S. aureus that have emerged in the clinic. volume 6 | january 2008 | 25
© 2008 Nature Publishing Group
REVIEWS Conclusions and perspectives This era may come to be remembered as one in which infectious diseases made a dramatic worldwide resurgence, owing to the rise of antibiotic resistance and emergence of new diseases. We must increase the number of available therapeutics, protect the effectiveness of current antibiotics and, importantly, decrease further pressure for the evolution of new drug-resistance mechanisms. Neutralizing bacterial toxins by using antibodies has emerged as the most-pursued antivirulence therapeutic strategy in industry, with at least six candidates undergoing clinical trials. The initial successes of these antitoxins seem to provide empirical evidence that supports increased research into other antivirulence approaches. Therefore, it is imperative that we determine the mechanisms of virulence and the consequences of host–pathogen interactions from both the pathogen and host perspective. Insights regarding bacterial virulence and pathogenesis have emerged, yet many crucial questions remain unanswered. Major breakthroughs will require multi-disciplinary tools from bacterial and host genetics, structural biology and in vivo imaging, animal models, cell biology, immunology, biochemistry, chemical genetics, functional genomics and systems biology. Atomic-level details of the structural interactions at host–pathogen interfaces are vital for the discovery of effective antivirulence drugs by structure-based drug design for proposed targets.
1. 2. 3. 4. 5. 6. 7. 8. 9.
10.
11. 12.
13. 14.
Alekshun, M. N. & Levy, S. B. Molecular mechanisms of antibacterial multidrug resistance. Cell 128, 1037–1050 (2007). Livermore, D. M. Minimising antibiotic resistance. Lancet Infect. Dis. 5, 450–459 (2005). Palumbi, S. R. Humans as the world’s greatest evolutionary force. Science 293, 1786–1790 (2001). Vicente, M. et al. The fallacies of hope: will we discover new antibiotics to combat pathogenic bacteria in time? FEMS Microbiol. Rev. 30, 841–852 (2006). Levy, S. B. & Marshall, B. Antibacterial resistance worldwide: causes, challenges and responses. Nature Med. 10, S122–S129 (2004). Crowley, P. J. & Martini, L. G. Formulation design: new drugs from old. Drug Discov. Today: Therapeutic Strategies 1, 537–542 (2004). Projan, S. J. Why is big Pharma getting out of antibacterial drug discovery? Curr. Opin. Microbiol. 6, 427–430 (2003). Fernandes, P. Antibacterial discovery and development — the failure of success? Nature Biotechnol. 24, 1497–1503 (2006). Talbot, G. H. et al. Bad bugs need drugs: an update on the development pipeline from the antimicrobial availability task force of the infectious diseases society of America. Clin. Infect. Dis. 42, 657–668 (2006). Blount, K. F., Wang, J. X., Lim, J., Sudarsan, N. & Breaker, R. R. Antibacterial lysine analogs that target lysine riboswitches. Nature Chem. Biol. 3, 44–49 (2007). Identified lysine analogues that bind to lysine riboswitches in vitro and inhibit the growth of Bacillus subtilis. Also discusses the general importance of riboswitches and their suitability as antibacterial drug targets. Wright, G. D. The antibiotic resistome: the nexus of chemical and genetic diversity. Nature Rev. Microbiol. 5, 175–186 (2007). Mwangi, M. M. et al. Tracking the in vivo evolution of multidrug resistance in Staphylococcus aureus by whole-genome sequencing. Proc. Natl Acad. Sci. USA 104, 9451–9456 (2007). Finlay, B. B. & Falkow, S. Common themes in microbial pathogenicity revisited. Microbiol. Mol. Biol. Rev. 61, 136–169 (1997). Lee, Y. M., Almqvist, F. & Hultgren, S. J. Targeting virulence for antimicrobial chemotherapy. Curr. Opin. Pharmacol. 3, 513–519 (2003).
In battling infection, disabling microbial virulence in vivo could shift the advantage to the host and render bacteria impaired in defeating host defences. Alternatively, antivirulence drugs could be used in combination therapy, in which bacterial clearance is mediated by standard antibiotics and the symptoms of virulence are suppressed. The effective deployment of antivirulence drugs will require rapid diagnosis in the clinic of the organism (or organisms) that is responsible for infection and may include profiling of its virulence gene or genes100. Such routine and rapid diagnosis would also improve the use of standard antibiotics, and represents a necessary investment as medicine improves and becomes more personalized. We must succeed in this continuous war against infectious disease. A recent review of the antibacterial drug-development efforts by GlaxoSmithKline reveals, in essence, that we have underestimated our opponent and that arrogance has dominated the search for new antibiotics101. Regaining our competitive advantage will require us to see beyond what we currently accept as dogma. It will also depend on the removal of perceived obstacles to infuse the industry with new opportunities. We need to accept that a one-drug-fits-all strategy will probably fail. We also need to consider each specific disease, pathogen and virulence mechanism, and combine the strengths of synergistic therapies to minimize the evolution by pathogens of resistance.
15. Marra, A. Can virulence factors be viable antibacterial targets? Expert Rev. Anti. Infect. Ther. 2, 61–72 (2004). 16. Savage, D. C. Microbial ecology of the gastrointestinal tract. Annu. Rev. Microbiol. 31, 107–133 (1977). 17. Backhed, F. et al. The gut microbiota as an environmental factor that regulates fat storage. Proc. Natl Acad. Sci. USA 101, 15718–15723 (2004). 18. Miller, J. F., Mekalanos, J. J. & Falkow, S. Coordinate regulation and sensory transduction in the control of bacterial virulence. Science 243, 916–922 (1989). A perspective that emphasizes the need to understand coordinate regulation and sensory transduction to understand the events that occur during the pathogenesis of infectious disease. 19. Virgin, H. W. In vivo veritas: pathogenesis of infection as it actually happens. Nature Immunol. 8, 1143–1147 (2007). An overview and perspective of the benefits that are ascribed to bridging the disciplines of microbiology and immunology, including lessons from Heisenberg. 20. Moxon, E. R., Rainey, P. B., Nowak, M. A. & Lenski, R. E. Adaptive evolution of highly mutable loci in pathogenic bacteria. Curr. Biol. 4, 24–33 (1994). 21. Monack, D. M., Mueller, A. & Falkow, S. Persistent bacterial infections: the interface of the pathogen and the host immune system. Nature Rev. Microbiol. 2, 747–765 (2004). Review and consideration of the dynamic cross-talk at the host–pathogen interface and the delicate balance between protective immunity and immunopathology. 22. Kaufmann, S. H. E. The contribution of immunology to the rational design of novel antibacterial vaccines. Nature Rev. Microbiol. 5, 491–504 (2007). 23. Kokai-Kun, J. F. & Mond, J. J. Antibody therapy for treatment or prevention of infectious diseases. Drug Discov. Today: Therapeutic Strategies 1, 475–481 (2004). 24. Pirofski, L. A. & Casadevall, A. Immunomodulators as an antimicrobial tool. Curr. Opin. Microbiol. 9, 489–495 (2006). 25. Hung, D. L. & Hultgren, S. J. Pilus biogenesis via the chaperone/usher pathway: an integration of structure and function. J. Struct. Biol. 124, 201–220 (1998).
26 | january 2008 | volume 6
26. Telford, J. L., Barocchi, M. le A., Margarit, I., Rappuoli, R. & Grandi, G. Pili in Gram-positive pathogens. Nature Rev. Microbiol. 4, 509–519 (2006). 27. Barocchi, M. A. et al. A pneumococcal pilus influences virulence and host inflammatory responses. Proc. Natl Acad. Sci. USA 103, 2857–2862 (2006). 28. Mulvey, M. A. Adhesion and entry of uropathogenic Escherichia coli. Cell. Microbiol. 4, 257–271 (2002). 29. Garofalo, C. K. et al. Escherichia coli from urine of female patients with urinary tract infections is competent for intracellular bacterial community formation. Infect. Immun. 75, 52–60 (2007). 30. Martinez, J. J., Mulvey, M. A., Schilling, J. D., Pinkner, J. S. & Hultgren, S. J. Type 1 pilus-mediated bacterial invasion of bladder epithelial cells. EMBO J. 19, 2803–2812 (2000). 31. Sauer, F. G., Mulvey, M. A., Schilling, J. D., Martinez, J. J. & Hultgren, S. J. Bacterial pili: molecular mechanisms of pathogenesis. Curr. Opin. Microbiol. 3, 65–72 (2000). 32. Anderson, G. G. et al. Intracellular bacterial biofilmlike pods in urinary tract infections. Science 301, 105–107 (2003). 33. Justice, S. S. et al. Differentiation and developmental pathways of uropathogenic Escherichia coli in urinary tract pathogenesis. Proc. Natl Acad. Sci. USA 101, 1333–1338 (2004). Revealed the multi-step E. coli pathogenic cascade using time-lapse fluorescence videomicroscopy to observe infected mouse-bladder explants. 34. Rosen Da, H. T., Stamm W. E., Humphrey, P. A. & Hultgren, S. J. Detection of intracellular bacterial communities in human urinary tract infection. PLoS Med. (in the press). 35. Mulvey, M. A. et al. Induction and evasion of host defenses by type 1-piliated uropathogenic Escherichia coli. Science 282, 1494–1497 (1998). 36. Mysorekar, I. U. & Hultgren, S. J. Mechanisms of uropathogenic Escherichia coli persistence and eradication from the urinary tract. Proc. Natl Acad. Sci. USA 103, 14170–14175 (2006). 37. Wright, K. J., Seed, P. C. & Hultgren, S. J. Development of intracellular bacterial communities of uropathogenic Escherichia coli depends on type 1 pili. Cell. Microbiol. 9, 2230–2241 (2007). 38. Kihlberg, J. & Magnusson, G. Use of carbohydrates and peptides in studies of adhesion of pathogenic
www.nature.com/reviews/micro © 2008 Nature Publishing Group
REVIEWS bacteria and in efforts to generate carbohydrate-specific T cells. Pure Appl. Chem. 68, 2119–2128 (1996). 39. Firon, N., Ashkenazi, S., Mirelman, D., Ofek, I. & Sharon, N. Aromatic alpha-glycosides of mannose are powerful inhibitors of the adherence of type 1 fimbriated Escherichia coli to yeast and intestinal epithelial cells. Infect. Immun. 55, 472–476 (1987). 40. Bouckaert, J. et al. Receptor binding studies disclose a novel class of high-affinity inhibitors of the Escherichia coli FimH adhesin. Mol. Microbiol. 55, 441–455 (2005). 41. Sauer, F. G., Remaut, H., Hultgren, S. J. & Waksman, G. Fiber assembly by the chaperone–usher pathway. Biochim. Biophys. Acta 1694, 259–267 (2004). 42. Chen, S. L. et al. Identification of genes subject to positive selection in uropathogenic strains of Escherichia coli: a comparative genomics approach. Proc. Natl Acad. Sci. USA 103, 5977–5982 (2006). 43. Svensson, A. et al. Design and evaluation of pilicides: potential novel antibacterial agents directed against uropathogenic Escherichia coli. Chembiochem 2, 915–918 (2001). 44. Larsson, A. et al. Multivariate design, synthesis, and biological evaluation of peptide inhibitors of FimC/FimH protein–protein interactions in uropathogenic Escherichia coli. J. Med. Chem. 48, 935–945 (2005). 45. Pinkner, J. S. et al. Rationally designed small compounds inhibit pilus biogenesis in uropathogenic bacteria. Proc. Natl Acad. Sci. USA 103, 17897–17902 (2006). 46. Report No. 04–5512, 153–183 (US Government Printing Office, Washington DC, 2004). 47. Report No. 04–5512, 187–209 (US Government Printing Office, Washington DC, 2004). 48. Ronald, A. R. et al. Urinary tract infection in adults: research priorities and strategies. Int. J. Antimicrob. Agents 17, 343–348 (2001). 49. Acharya, V. N. Urinary tract infection — a dangerous and unrecognised forerunner of systemic sepsis. J. Postgrad. Med. 38, 52–54 (1992). 50. Schmitt, C. K., Meysick, K. C. & O’Brien, A. D. Bacterial toxins: friends or foes? Emerg. Infect. Dis. 5, 224–234 (1999). 51. Spangler, B. D. Structure and function of cholera toxin and the related Escherichia coli heat-labile enterotoxin. Microbiol. Rev. 56, 622–647 (1992). 52. Hung, D. T., Shakhnovich, E. A., Pierson, E. & Mekalanos, J. J. Small-molecule inhibitor of Vibrio cholerae virulence and intestinal colonization. Science 310, 670–674 (2005). 53. Keller, M. A. & Stiehm, E. R. Passive immunity in prevention and treatment of infectious diseases. Clin. Microbiol. Rev. 13, 602–614 (2000). 54. Arnon, S. S., Schechter, R., Maslanka, S. E., Jewell, N. P. & Hatheway, C. L. Human botulism immune globulin for the treatment of infant botulism. N. Engl. J. Med. 354, 462–471 (2006). 55. Young, J. A. & Collier, R. J. Anthrax toxin: receptor binding, internalization, pore formation, and translocation. Annu. Rev. Biochem. 76, 243–265 (2007). Describes the structural basis and molecular mechanisms of the anthrax toxin and provides insight into toxin function. 56. Saenz, J. B., Doggett, T. A. & Haslam, D. B. Identification and characterization of small molecules that inhibit intracellular toxin transport. Infect. Immun. 75, 4552–4561 (2007). 57. Zhao, L. & Haslam, D. B. A quantitative and highly sensitive luciferase-based assay for bacterial toxins that inhibit protein synthesis. J. Med. Microbiol. 54, 1023–1030 (2005). 58. Tonello, F., Seveso, M., Marin, O., Mock, M. & Montecucco, C. Screening inhibitors of anthrax lethal factor. Nature 418, 386 (2002). 59. Panchal, R. G. et al. Identification of small molecule inhibitors of anthrax lethal factor. Nature Struct. Mol. Biol. 11, 67–72 (2004). 60. Turk, B. E. et al. The structural basis for substrate and inhibitor selectivity of the anthrax lethal factor. Nature Struct. Mol. Biol. 11, 60–66 (2004). 61. Russell, P. K. Project BioShield: what it is, why it is needed, and its accomplishments so far. Clin. Infect. Dis. 45, S68–S72 (2007). 62. US Securities and Exchange Commission (SEC). Pharmathene, Inc. Form 10-Q, Quarter Ended September 30, Commission File Number 001-32587. SEC web site [online], http://www.sec.gov/Archives/ edgar/data/1326190/000110465907083136/a0729212_110q.htm#Item2_103948 (2007). 63. Food and Drug Administration Center for Drug Evaluation and Research. Summary Minutes of the Anti-Infective Drugs Advisory Committee on
April 12, 2007 [online], http://www.fda.gov/ohrms/ dockets/AC/07/minutes/2007-4286m1.pdf (2007). 64. Gyles, C. L. Shiga toxin-producing Escherichia coli: an overview. J. Anim. Sci. 85, E45–E62 (2007). 65. Sandkvist, M. et al. General secretion pathway (eps) genes required for toxin secretion and outer membrane biogenesis in Vibrio cholerae. J. Bacteriol. 179, 6994–7003 (1997). 66. Cornelis, G. R. The type III secretion injectisome. Nature Rev. Microbiol. 4, 811–825 (2006). 67. Cornelis, G. R. & Wolf-Watz, H. The Yersinia Yop virulon: a bacterial system for subverting eukaryotic cells. Mol. Microbiol. 23, 861–867 (1997). 68. Cornelis, G. R. Yersinia type III secretion: send in the effectors. J. Cell. Biol. 158, 401–408 (2002). 69. Cornelis, G. R. & Van Gijsegem, F. Assembly and function of type III secretory systems. Annu. Rev. Microbiol. 54, 735–774 (2000). 70. Rosqvist, R., Hakansson, S., Forsberg, A. & Wolf-Watz, H. Functional conservation of the secretion and translocation machinery for virulence proteins of yersiniae, salmonellae and shigellae. EMBO J. 14, 4187–4195 (1995). 71. Kauppi, A. M., Nordfelth, R., Uvell, H., Wolf-Watz, H. & Elofsson, M. Targeting bacterial virulence: inhibitors of type III secretion in Yersinia. Chem. Biol. 10,241–249 (2003). 72. Nordfelth, R., Kauppi, A. M., Norberg, H. A., Wolf-Watz, H. & Elofsson, M. Small-molecule inhibitors specifically targeting type III secretion. Infect. Immun. 73, 3104–3114 (2005). 73. Wolf, K. et al. Treatment of Chlamydia trachomatis with a small molecule inhibitor of the Yersinia type III secretion system disrupts progression of the chlamydial developmental cycle. Mol. Microbiol. 61, 1543–1555 (2006). 74. Muschiol, S. et al. A small-molecule inhibitor of type III secretion inhibits different stages of the infectious cycle of Chlamydia trachomatis. Proc. Natl Acad. Sci. USA 103, 14566–14571 (2006). An elegant application of chemical genetics to microbial pathogenesis. A small molecule that was identified as a T3SS inhibitor in Yersinia spp. inhibited virulence of C. trachomatis, which supports the notion that the T3SS is important in C. trachomatis pathogenesis. 75. Bailey, L. et al. Small molecule inhibitors of type III secretion in Yersinia block the Chlamydia pneumoniae infection cycle. FEBS Lett. 581, 587–595 (2007). 76. Saye, D. E. Recurring and antimicrobial-resistant infections: considering the potential role of biofilms in clinical practice. Ostomy Wound Manage. 53, 46–62 (2007). 77. Hall-Stoodley, L. et al. Direct detection of bacterial biofilms on the middle-ear mucosa of children with chronic otitis media. JAMA 296, 202–211 (2006). 78. Carron, M. A., Tran, V. R., Sugawa, C. & Coticchia, J. M. Identification of Helicobacter pylori biofilms in human gastric mucosa. J. Gastrointest. Surg. 10, 712–717 (2006). 79. Lam, J., Chan, R., Lam, K. & Costerton, J. W. Production of mucoid microcolonies by Pseudomonas aeruginosa within infected lungs in cystic fibrosis. Infect. Immun. 28, 546–556 (1980). 80. Singh, P. K. et al. Quorum-sensing signals indicate that cystic fibrosis lungs are infected with bacterial biofilms. Nature 407, 762–764 (2000). 81. Benghezal, M. et al. Inhibitors of bacterial virulence identified in a surrogate host model. Cell. Microbiol. 9, 1336–1342 (2007). 82. Keller, L. & Surette, M. G. Communication in bacteria: an ecological and evolutionary perspective. Nature Rev. Microbiol. 4, 249–258 (2006). A holistic perspective of the potential roles of QS signals, from a cooperative to the battleground. 83. Bassler, B. L. & Losick, R. Bacterially speaking. Cell 125, 237–246 (2006). A creative composition and a lively review of the business of bacterial communication. 84. Kjelleberg, S. et al. Do marine natural products interfere with prokaryotic AHL regulatory systems? Aquat. Microb. Ecol. 13, 85–93 (1997). 85. Manefield, M. et al. Evidence that halogenated furanones from Delisea pulchra inhibit acylated homoserine lactone (AHL)-mediated gene expression by displacing the AHL signal from its receptor protein [in process citation]. Microbiology 145, 283–291 (1999). 86. Higgins, D. A. et al. The major Vibrio cholerae autoinducer and its role in virulence factor production. Nature 14 Nov 2007 (doi:10.1038/nature06284). 87. Hentzer, M. et al. Attenuation of Pseudomonas aeruginosa virulence by quorum sensing inhibitors. EMBO J. 22, 3803–3815 (2003).
nature reviews | microbiology
88. Bjarnsholt, T. & Givskov, M. Quorum-sensing blockade as a strategy for enhancing host defences against bacterial pathogens. Phil. Trans. R. Soc. Lond. B 362, 1213–1222 (2007). 89. Geske, G. D., O’Neill, J. C. & Blackwell, H. E. N‑phenylacetanoyl‑l-homoserine lactones can strongly antagonize or superagonize quorum sensing in Vibrio fischeri. ACS Chem. Biol. 2, 315–319 (2007). Identified a non-native N‑acylated‑l-homoserine lactone that can either inhibit or strongly induce QS in the marine symbiont V. fischeri, depending on the molecule concentration. 90. Miller, M. B., Skorupski, K., Lenz, D. H., Taylor, R. K. & Bassler, B. L. Parallel quorum sensing systems converge to regulate virulence in Vibrio cholerae. Cell 110, 303–314 (2002). Revealed the redundant QS capacity in V. cholerae and discusses implications in terms of V. cholerae pathogenicity. 91. Wadhams, G. H. & Armitage, J. P. Making sense of it all: bacterial chemotaxis. Nature Rev. Mol. Cell Biol. 5, 1024–1037 (2004). 92. Galperin, M. Y. Structural classification of bacterial response regulators: diversity of output domains and domain combinations. J. Bacteriol. 188, 4169–4182 (2006). 93. Ulrich, L. E. & Zhulin, I. B. MiST: a microbial signal transduction database. Nucleic Acids Res. 35, D386–D390 (2007). 94. Mascher, T., Helmann, J. D. & Unden, G. Stimulus perception in bacterial signal-transducing histidine kinases. Microbiol. Mol. Biol. Rev. 70, 910–938 (2006). 95. Barrett, J. F. & Hoch, J. A. Two-component signal transduction as a target for microbial anti-infective therapy. Antimicrob. Agents Chemother. 42, 1529–1536 (1998). 96. Mayville, P. et al. Structure-activity analysis of synthetic autoinducing thiolactone peptides from Staphylococcus aureus responsible for virulence. Proc. Natl Acad. Sci. USA 96, 1218–1223 (1999). 97. Otto, M. Quorum-sensing control in Staphylococci — a target for antimicrobial drug therapy? FEMS Microbiol. Lett. 241, 135–141 (2004). 98. Arthur, M., Molinas, C. & Courvalin, P. The VanS–VanR two-component regulatory system controls synthesis of depsipeptide peptidoglycan precursors in Enterococcus faecium BM4147. J. Bacteriol. 174, 2582–2591 (1992). 99. Evers, S. & Courvalin, P. Regulation of VanB-type vancomycin resistance gene expression by the VanS(B)–VanR (B) two-component regulatory system in Enterococcus faecalis V583. J. Bacteriol. 178, 1302–1309 (1996). 100. Sintchenko, V., Iredell, J. R. & Gilbert, G. L. Pathogen profiling for disease management and surveillance. Nature Rev. Microbiol. 5, 464–470 (2007). 101. Payne, D. J., Gwynn, M. N., Holmes, D. J. & Pompliano, D. L. Drugs for bad bugs: confronting the challenges of antibacterial discovery. Nature Rev. Drug Discov. 6, 29–40 (2007). 102. Jones, C. H. et al. FimH adhesin of type 1 pili is assembled into a fibrillar tip structure in the Enterobacteriaceae. Proc. Natl Acad. Sci. USA 92, 2081–2085 (1995).
Acknowledgements
The authors acknowledge funding from the National Institutes of Health to S.J.H. (grant numbers P50-ORWH/DK64540, R01AI029549, R01AI048689 and R01DK51406), G.R.M. (grant number R01GM068460) and L.C. (grant number T32A107172).
Competing interest statement
The authors declare competing financial interests: see web version for details.
DATABASES Entrez Genome Project: http://www.ncbi.nlm.nih.gov/ entrez/query.fcgi?db=genomeprj Bacillus anthracis | Chlamydia trachomatis | Clostridium botulinum | Escherichia coli | Haemophilus influenzae | Helicobacter pylori | Pseudomonas aeruginosa | Shigella dysenteriae | Shigella flexneri | Staphylococcus aureus | Streptococcus pneumoniae | Vibrio cholerae | Vibrio fischeri | Yersinia pestis | Yersinia pseudotuberculosis
FURTHER INFORMATION Scott J. Hultgren’s homepage: http://hultgren.wustl.edu/ public All links are active in the online pdf
volume 6 | january 2008 | 27 © 2008 Nature Publishing Group
REVIEWS
Getting organized — how bacterial cells move proteins and DNA Martin Thanbichler* and Lucy Shapiro‡
Abstract | In recent years, the subcellular organization of prokaryotic cells has become a focal point of interest in microbiology. Bacteria have evolved several different mechanisms to target protein complexes, membrane vesicles and DNA to specific positions within the cell. This versatility allows bacteria to establish the complex temporal and spatial regulatory networks that couple morphological and physiological differentiation with cell-cycle progression. In addition to stationary localization factors, dynamic cytoskeletal structures also have a fundamental role in many of these processes. In this Review, we summarize the current knowledge on localization mechanisms in bacteria, with an emphasis on the role of polymeric protein assemblies in the directed movement and positioning of macromolecular complexes.
*Max Planck Institute for Terrestrial Microbiology, Karl-von-Frisch-Straße, 35043 Marburg, Germany. ‡ Department of Developmental Biology, Stanford University School of Medicine, Beckman Center B300, 279 Campus Drive, Stanford, California 94305, USA. Correspondence to M.T. e-mail: thanbichler@ mpi-marburg.mpg.de doi:10.1038/nrmicro1795 Published online 3 December 2007
Bacteria are among the most successful and widespread organisms on the Earth. In the course of several billion years of evolution, they have adapted to almost every possible biological niche and developed an astonishing range of metabolic pathways, life cycles and cell morphologies. As a result of continuous selection for fast reproduction rates, they have adopted highly streamlined architectures and small genomes that have high coding densities. Their straightforward organization has frequently been regarded as indicative of a primitive cellular state. Consequently, advanced features, such as cytoskeletal structures, intracellular transport processes and spatially regulated transcription and protein localization, were traditionally thought to be restricted to eukaryotic cells. However, this simplistic view was difficult to reconcile with the complexity of many processes that are found in bacteria. Notably, regulation of cell shape, cell polarization and asymmetric cell division are common phenomena in bacterial development that are unlikely to occur without dedicated temporal and spatial regulatory systems or dynamic cytoskeletal elements. Recent work, driven by technological advances that have facilitated the resolution of structural details within bacterial cells, has revealed that bacteria have indeed evolved mechanisms to actively control cell-cycle progression and morphological differentiation. Some of the factors that are involved in these processes are homologues of well characterized eukaryotic proteins, but they are frequently used outside of their established functional contexts. Other factors are found exclusively in bacteria and constitute novel regulatory and structural systems that have evolved to meet the specific
28 | january 2008 | volume 6
physiological requirements of the bacterial cell. In this Review, we highlight key mechanisms of cellular organization in bacteria that reflect the striking momentum that bacterial cell biology has gained in recent years. Specifically, we will discuss pathways that mediate the assembly and positioning of localized protein complexes and set the foundation for temporal and spatial regulatory processes within bacterial cells. In this context, emphasis will be placed on the growing number of dynamic cytoskeletal elements that have been identified in bacteria and their pivotal roles in subcellular organization, DNA segregation and cell division.
Subcellular organization in bacteria The realization that bacteria use specifically localized protein complexes to orchestrate cellular processes led to a surge of research on the mechanisms that structure the bacterial cell (BOX 1). Among the many proteins with uneven subcellular distribution that have been identified in recent years, two major classes can be distinguished. One group forms largely stationary complexes that localize to precisely defined subcellular positions, even though their subunits might be in rapid exchange. Proteins belonging to the second class, by contrast, are part of highly dynamic interaction networks with components that continuously change their subcellular location in a temporally and spatially controlled manner. Assembly of stationary protein complexes. Most stationary complexes that have been characterized so far are based on integral membrane proteins. They usually assemble at the new pole (resulting from the most www.nature.com/reviews/micro
© 2008 Nature Publishing Group
REVIEWS Forespore Precursor of the spore; a resting cell that is highly resistant to a number of environmental stresses, such as heat, ultraviolet irradiation and desiccation.
Single-molecule tracking Microscopic analysis of the movement of individual fluorescently labelled molecules within a cell.
recent division), the old pole, the cell-division site or the forespore septal membrane in the case of Bacillus subtilis sporulation, where they mediate the establishment of cellular organelles, perform localized catalytic activities or serve regulatory functions. But how do these proteins assume a defined intracellular position? Evidence has been presented that at least two localization mechanisms operate in bacteria: diffusion and capture; and targeted membrane insertion. During diffusion and capture, newly synthesized proteins are first inserted randomly throughout the cytoplasmic membrane and then, after diffusion, they are captured by an interaction with a previously localized membrane complex1,2. Targeted membrane insertion, by contrast, describes a process whereby a protein is delivered directly to its destination by translocation to a given cellular site. In all cases, a crucial issue is the identity of the determinant that is responsible for the localization of the membrane protein and how this determinant is positioned in the first place.
Box 1 | Model systems in bacterial cell biology Although an increasing number of species are studied with respect to their cell biology, most information on cellular organization in bacteria is currently based on studies in Escherichia coli, Bacillus subtilis and Caulobacter crescentus. The characteristics of these model systems are listed below.
E. coli • Long history as a bacterial model organism. • Large collection of genetic tools, mutant strains and methods. • Extensive body of knowledge on many aspects of its physiology. B. subtilis • Sporulation as a model for cellular differentiation. • Well-studied physiology. • Large size, which facilitates the resolution of cytoskeletal structures. C. crescentus • Asymmetric cell division (see the figure). • One round of chromosome replication per cell division. • Abundance of dynamically localized regulatory protein complexes. • Synchronizability. In the course of its life cycle, C. crescentus differentiates from a mobile, flagellated swarmer cell into a sessile stalked cell (see the figure). Subsequently, it undergoes an asymmetric cell division that regenerates the stalked cell and creates a new swarmer cell. The stalked cell immediately enters a new round of cell division, whereas the swarmer cell needs to develop into another stalked cell to start the next division cycle. The single polar flagellum is shed during the swarmer-to-stalk-cell transition and re-established at the pole opposite the stalk in the predivisional cell.
Swarmer cell
Stalked cell
Flagellum
Stalk
Nature Reviews | Microbiology nature reviews | microbiology
Diffusion and capture. During B. subtilis sporulation many proteins localize differentially to the septal membrane, which provides an amenable system for investigating the principles that underlie directed protein positioning within the cell (FIG. 1a). SpoIVFB, a proprotein-processing enzyme, is synthesized in the mother cell, but must be localized in the septal membrane to carry out its function. Rudner and colleagues1 reasoned that SpoIVFB is either targeted to the septal membrane directly upon synthesis or is first inserted randomly around the mother-cell cytoplasmic membrane and then, upon diffusion, captured at the septal membrane. To distinguish between these two possibilities, SpoIVFB was expressed from an inducible promoter in vegetatively growing cells, in which it was found to be randomly positioned in the cytoplasmic membrane. During initiation of sporulation in these cells, and in the absence of the inducer, the existing SpoIVFB accumulated at the septum, demonstrating that diffusion and capture is the probable mode of SpoIVFB localization (FIG. 1b). Additional evidence that SpoIVFB diffuses to its site of action was provided by taking advantage of the fact that, when engulfment occurs, the membrane surrounding the forespore is separated topologically from the mother-cell cytoplasmic membrane from which it is derived1. SpoIVFB that was synthesized after engulfment was specifically enriched in the mother-cell cytoplasmic membrane, but was absent from the outer forespore membrane, again suggesting that diffusion and capture, rather than targeted insertion, is the mode of SpoIVFB localization. Evidence for the diffusion-and-capture mode of protein localization also comes from single-molecule tracking of the Caulobacter crescentus membrane histidine kinase PleC, which localizes dynamically to the pole that is opposite the stalk at specific times in the cell cycle2. Following the movement of a PleC–green fluorescent protein (GFP) fusion protein over a timescale of seconds revealed two diffusion coefficients: a high coefficient for molecules that were without directional bias and a lower coefficient for molecules that were targeted to the cell pole. However, the components that mediate the sink at the cell pole have not yet been identified. Diffusion and capture might be a widespread mechanism that operates in several different systems. For example, many components of the bacterial cell-division apparatus are known to assume an even subcellular distribution unless a defined precursor complex has assembled to capture them at mid-cell3. The same principle may also apply to soluble proteins, which can be tethered to a pre-localized targeting factor after three-dimensional diffusion through the cytoplasm or periplasm. Targeted membrane insertion. Whereas SpoIVFB and other proteins that are synthesized in the mother-cell region of sporulating B. subtilis cells seem to find their septal destination by diffusion and capture, the septal localization of the SpoIIQ protein, which is synthesized in the forespore, is achieved by direct insertion and capture4. SpoIIQ was found to act as a localization factor that recruits another mother-cell protein, SpoIIIAH, volume 6 | january 2008 | 29
© 2008 Nature Publishing Group
REVIEWS a B. subtilis sporulation
b Diffusion and capture
c Zipper model A Q
Forespore FB A Q FB A Q
Q A Outer forespore membrane
Q
Figure 1 | Protein localization to the septal membrane during Bacillus subtilis Reviews | Microbiology sporulation. During sporulation (a), B. subtilis undergoesNature an asymmetric cell division that generates two offspring of unequal size and developmental fate — a larger mother cell (green) and a smaller forespore (blue). The mother cell subsequently engulfs the forespore in a phagocytosis-like process, thereby surrounding it with a second membrane, the so-called outer forespore membrane. During this process, several proteins specifically localize to the interface between the two compartments (b). In several cases, their positioning is mediated by diffusion and capture, as exemplified by the recruitment of the freely diffusible mother cell protein SpoIVFB (FB; purple spheres) by a pre-localized complex including the forespore protein SpoIIQ (Q). The interaction of proteins across the mother cell and forespore membranes involves a zipper-like mechanism (c). Using this principle, SpoIIQ engages the mother cell factor SpoIIIA (A), which, in turn, is responsible for the septal localization of other proteins, among them SpoIVFB.
Protofilament The basic polymeric unit of a filamentous structure; consists of a linear row of monomers.
Fluorescence recovery after photobleaching (FRAP). A method used to study the dynamics of polymeric structures. A region within a filament that has been assembled from fluorescently labelled monomers is bleached by illumination with a highintensity laser. Subsequently, fluorescence microscopy is used to monitor the kinetics of fluorescence recovery that results from the substitution of bleached by unbleached subunits in the course of filament turnover.
to the forespore septal membrane5. Based on the direct interaction between the extracytoplasmatic domains of SpoIIQ and SpoIIIAH, a zipper-like mechanism has been proposed as a method that is used to capture SpoIIIAH4–6 (FIG. 1c). This interaction subsequently recruits additional factors that are synthesized by the mother cell (including SpoIVFB)6,7, which yields a protein network that anchors localized proteins at the septum. However, the zipper mechanism alone is not sufficient to direct the forespore protein SpoIIQ to its septal position. Interestingly, it has been shown that all of the membrane proteins that are synthesized in the forespore are initially targeted to the septal membrane, from which they slowly diffuse to the rest of the forespore membrane, unless they are retained by a specific localization factor4. In agreement with this observation, a component of the protein translocation complex, FtsY, is specifically localized to the forespore septal membrane in sporulating B. subtilis cells8. Thus, certain forespore proteins, such as SpoIIQ, appear to be directly targeted to their site of action by localized translocation. Another example of directed protein targeting is provided by the intracellular pathogen Shigella flexneri. To facilitate movement within the host cell, this bacterium positions the protein IcsA at its old cell pole, where it directs polymerization of host actin into a tail-like structure, thereby generating a pushing force. IcsA is an outer membrane protein that has its amino-terminal
30 | january 2008 | volume 6
domain exposed to the host cytoplasm. During synthesis, it is directly targeted to the region around the cell pole. Concomitantly, proteolysis of IcsA occurs over the whole-cell perimeter, in effect clearing away any protein that has diffused away from the pole9,10. IcsA localizes independently of its export from the cytoplasm to the outer membrane10. It has been proposed that a region within the amino‑terminal domain interacts with an as-yet-unidentified factor in the polar region of the cell, thereby targeting IcsA to its defined site of secretion before translocation occurs11. Stationary complexes usually have locally confined activities or serve as regulatory landmarks. By contrast, global processes, such as morphogenesis and macromolecular transport, require the assembly of scaffolds that extend throughout the cell. These structures are formed by dynamic cytoskeletal filaments that adopt specific overall arrangements but are highly variable on a shortterm scale. They provide force as well as directionality, and serve as tracks for the localization of other proteins. In the following section, we will focus specifically on protein filaments that are involved in the determination of bacterial-cell shape; other dynamic structures are discussed in BOX 2 and in the later sections of this article.
Dynamic protein scaffolds and cell shape A major role in subcellular organization has recently been attributed to actin-like proteins that belong to the MreB family. They are found in many bacteria, in which they assemble consistently into spiral-like structures that line the inner face of the cytoplasmic membrane. MreB homologues play an important part in the regulation of cell shape and have been implicated in cell polarity and chromosome segregation12–18. Their function in morphogenesis is dependent on the presence of another scaffold that is formed by the extracytoplasmic protein MreC and is modulated by accessory factors, such as the intermediate filament protein crescentin. The bacterial actin-like cytoskeleton. In eukaryotes, the actin cytoskeleton consists of a highly dynamic network of filaments, the assembly of which is regulated by the binding and hydrolysis of ATP and is further modulated by an abundance of accessory factors. Actin protofilaments possess an intrinsic polarity, whereby monomers are preferentially added to one end and released from the other in a phenomenon that is called treadmilling19. Although bacterial and eukaryotic actin homologues share minimal sequence identity (~15%), studies on MreB from Thermotoga maritima revealed that their three‑dimensional structures and protofilament organization are strikingly similar20,21 (FIG. 2a). Nevertheless, the overall assembly properties of T. maritima MreB deviate from the scheme that has been established for its eukaryotic relatives with respect to several biochemical parameters22. MreB cables are highly dynamic in vivo and assume several different architectures, comprising spiral-like assemblies of varying curvature and length as well as short coils, arcs and rings12,14,15,23. The mechanisms and cellular cues that are responsible for these diverse www.nature.com/reviews/micro
© 2008 Nature Publishing Group
REVIEWS Box 2 | Linear arrangement of subcellular compartments — magnetosomes There are many examples of a intracellular membrane compartments in bacteria, including the chlorosomes and chromatophores of nonoxygenic photosynthetic bacteria and the magnetosomes of magnetotactic bacteria. It is b largely unknown how these structures are established and maintained at their proper subcellular position. In the case of magnetosomes, however, a clearer picture of the underlying mechanisms is starting to emerge. Nature Reviews | Microbiology Magnetosomes are vesicles — formed by invagination of the cytoplasmic membrane123 — that are filled with biomineralized magnetite or greigite and allow bacteria to orient themselves in the Earth’s magnetic field124. The best-studied magnetotactic organisms, Magnetospirillum magneticum and Magnetospirillum gryphiswaldense, contain 15–20 magnetosomes, which are arrayed in linear order parallel to the longitudinal axis of the cell and approximately centred at mid-cell123,125. This arrangement generates a compass needle-like structure that has an overall magnetic dipole moment that equals the sum of the magnetic dipole moments of the individual vesicles126. As magnetosomes have an inherent tendency to agglomerate to reduce their magnetostatic energy, the cell must possess specific mechanisms to stabilize this linear assembly and anchor it within the cell. Electron cryotomography analyses unveiled filamentous structures that run alongside magnetosome chains123,125 (see the figure). The protein that is responsible for their formation has been identified as MamK (purple in the figure), a bacterial actin homologue that belongs to an evolutionarily distinct group of proteins that is clearly separated from the MreB family123. It is conserved in all magnetotactic bacteria and assembles into linear polymers that extend from one cell pole to another123,127. In agreement with a role for MamK in magnetosome alignment, mutants that lack the protein are still able to produce mature magnetosomes, but fail to arrange them into a linear array. Recent work in M. gryphiswaldense has identified a protein, MamJ, that might be involved in attaching magnetosomes to the MamK filament (green in the figure)125. It is an acidic protein, with a repetitive primary sequence which localizes as a linear structure that extends from pole to pole. In a mutant strain that lacks a cluster of genes that are involved in magnetosome biogenesis (among them mamK) this linear arrangement is abolished, and MamJ is dispersed throughout the cytoplasm. Its cooperation with MamK in vesicle positioning is supported by the fact that deletion of mamJ largely reproduces the phenotype of a mamK mutant, leading to the dispersal of magnetosomes without affecting their maturation. As a result, MamK appears to function as a track that recruits vesicles with the help of MamJ, thereby defining the overall orientation of the magnetosome chain. It is thought that empty and immature vesicles are initially randomly distributed along the MamK filament (see the figure, part a). The accumulation of magnetite (beige spheres in the figure) results in magnetostatic interactions between individual vesicles, which leads to their aggregation into densely packed chains (see the figure, part b) that are subsequently stabilized by additional factors (blue in the figure )125. Figure adapted, with permission, from Nature REF. 125 (2006) Macmillan Publishers Ltd.
patterns are unclear. In C. crescentus15,16, and its relative Rhodobacter sphaeroides24,25, the transitions between the different MreB structures were found to be synchronized with cell-cycle progression (FIG. 2b). By contrast, in bacteria from other phylogenetic groups, the actinlike cytoskeleton seems to be largely unaffected by the developmental state of the cell. The first insights into the turnover kinetics of actinlike filaments were provided by studies on B. subtilis. Whereas most Gram-negative bacteria possess one nature reviews | microbiology
mreB gene, this Gram-positive bacterium synthesizes three actin homologues, MreB, Mbl and MreBH, which interact with each other and assemble into a single helical filament12,13,26,27. Fluorescence recovery after photo bleaching (FRAP) studies revealed a rapid exchange of subunits within Mbl cables that occurred uniformly and without any evidence of polarity with respect to filament growth23. Furthermore, MreB and Mbl spirals were observed to rotate in the cell, with one full turn every 50–60 seconds13,26. Additional support for the rapid turnover of MreB cables comes from a study that analysed the movement of single MreB–GFP molecules in live cells of C. crescentus28. The MreB subunits were found to migrate directionally along arc-like trajectories, which is suggestive of treadmilling of individual subunits through stationary MreB filaments. However, the average length of these trajectories was significantly shorter than the MreB cable as a whole, and no common directionality was observed relative to the overall polarity of the cell. In C. crescentus, and possibly other bacteria, the actin-like cytoskeleton is therefore probably composed of numerous short, polarized and highly dynamic protofilaments that associate laterally but in random orientation to each other. The overall dynamics of this structure might be based on the rapid exchange of protofilament bundles rather than individual protein monomers (FIG. 2c). Regulation of cell-wall biosynthesis by actin-like proteins. The arrangement of actin-like proteins into large filaments and their dynamic positioning within the cell are crucial for the spatial regulation of peptidoglycan biosynthesis in most rod-shaped bacteria (BOX 3). This is due, in part, to a direct role in the positioning of the enzymes that are involved in cell-wall formation. In B. subtilis, MreBH was shown to interact with the peptidoglycan hydrolase LytE27. A LytE–GFP fusion is normally found in the form of distinct bands at the division septa and in helical patterns along the lateral cell wall. Upon deletion of the mreBH gene, however, these lateral helical structures are lost. As strains that lack LytE have the same defects in cell shape as mreBH mutant cells, localization of the hydrolase along the helical tracks of MreBH appears to be essential for its proper function. So far, this is the only proven example of an immediate influence of actin-like proteins on peptidoglycan biosynthesis. In many cases, their effect seems to be more indirect and due to a role in the localization of MreC, a protein that is frequently encoded in the same operon as mreB. The role of MreC in bacterial morphogenesis. MreC is a membrane protein that is composed of a single, amino‑terminal transmembrane helix and a large extracytoplasmic domain in Escherichia coli and B. subtilis29,30, but is a soluble periplasmic protein in C. crescentus31. Crystallographic data indicate that it might have the potential to form polymeric structures30. MreC is part of a complex that includes the poorly characterized proteins MreD and RodA, and is essential for survival in most bacteria that have been investigated; its inactivation results in a loss of cell shape and subsequent lysis29,31,32. In support of a role in cell-wall biosynthesis, MreC volume 6 | january 2008 | 31
© 2008 Nature Publishing Group
REVIEWS a Model for MreB treadmilling
(+)
**
(–)
MreB–ATP MreB–ADP
**
**
**
b MreB dynamics in C. crescentus
c Model for the architecture of MreB cables
of a multi-enzyme peptidoglycan biosynthetic complex, thereby organizing the formation of new cell-wall material. The spatial arrangement of MreC is dependent on the positioning of actin-like cables within the cell, although the underlying mechanisms seem to vary among different bacteria. In E. coli, MreC interacts directly with MreB and seems to be required for the proper assembly of MreB into helical filaments29. Similarly, MreC forms helical structures33 that colocalize and physically associate with Mbl cables in B. subtilis26. The functional relevance of this interaction is underscored by the observation that both Mbl and MreC depletion results in the same type of morphological defects33,34; these are caused by the disruption of peptidoglycan biosynthesis in the longitudinal parts of the cell33. In C. crescentus, an extremely different situation is observed. Here, MreC assembles into helical structures that do not associate with MreB cables but, on the contrary, seem to avoid them35. The mechanisms that underlie this effect are still unknown. In accordance with biochemical-interaction studies31, PBP2 was shown to colocalize with MreC in wild-type C. crescentus cells31,35. However, the depletion of MreC does not disrupt the already existing spiral-like structures of PBP2, although it does specifically affect the positioning of newly synthesized molecules35. Thus, MreC seems to guide peptidoglycan-synthesizing enzymes to the sites of active cell-wall growth, where they might subsequently be retained by interaction with other proteins or nascent peptidoglycan strands.
Nature Reviews | Microbiology Figure 2 | Dynamics of actin-like filaments in bacteria. MreB and its orthologues polymerize into bundles of two or more protofilaments that are arranged in parallel to each other (a). Their assembly follows the same principles that have been observed for actin. Polymerization occurs exclusively in the ATP-bound state (blue spheres). However, soon after incorporation into the filament, MreB hydrolyses the nucleotide, thereby changing into its ADP-bound conformation (purple spheres). Subunit exchange is restricted to the ends of the filament, and the addition of new monomers occurs preferentially at the positive-end, that is, the cap which consists of MreB subunits that are still associated with ATP. Subunit release, by contrast, is largely restricted to the negativeend of the filament, which is composed of MreB–ADP. In the steady state, assembly and disassembly proceed at equal rates, resulting in the apparent sliding of individual subunits (marked by asterisks) through the filament while the overall polymer length remains constant — a phenomenon that is called treadmilling. In Caulobacter crescentus, MreB cables exhibit a highly dynamic localization pattern (b). Newborn cells contain spiral-like cables that extend between the two poles. Upon initiation of cell division, these structures are lost and MreB condenses into a ring at the future division site. Later, as cell constriction progresses, the zone of MreB localization gradually expands again, leading finally to the establishment of new spirals in the two incipient daughter cells. Biochemical evidence suggests that actin-like cables consist of numerous treadmilling filaments that are arranged side by side in a random orientation (c).
Crescentin. Recent work in C. crescentus identified a protein, named crescentin, that shares the structural characteristics of intermediate filament proteins and acts as a modulator of cell shape36. Although poorly conserved on the primary-sequence level, these proteins share a common architecture that comprises a long central coiled-coil region that is flanked by variable head and tail domains. Their polymerization occurs spontaneously and does not require nucleotides or exogenous nucleation factors37. Similar to its eukaryotic homologues, crescentin assembles into long filaments in vitro, which illustrates its tendency to associate laterally into small bundles36. Localization studies revealed that the protein is positioned specifically at the inner curvature of the crescent-shaped C. crescentus cell. Upon inactivation of crescentin, the cells lose their typical crescentoid morphology and grow in the form of straight rods, which indicates that crescentin is responsible for establishing cell curvature36. The mechanism that allows this cytoskeletal structure to redirect peptidoglycan biosynthesis remains to be established.
interacts directly with the peptidoglycan synthase penicillinbinding protein (PBP) 2, and several other high- and lowmolecular-weight PBPs (BOX 3), in C. crescentus31. Similar results have been obtained in B. subtilis, although MreC appears to associate preferentially with high-molecularweight PBPs in this organism30. These findings suggest that the protein serves as a scaffold for the formation
Morphogenetic function of the cell-division apparatus. In addition to the systems discussed so far, the cell-division apparatus has an important role in the regulation of cellular morphology. Not only does it comprise peptidoglycan synthases and hydrolases that mediate the formation of the division septum3, but, in C. crescentus, it also directs a division-independent mode of peptidoglycan biosynthesis that leads to zonal growth around the cell centre38. As MurG, the enzyme that catalyses the last cytoplasmic step
32 | january 2008 | volume 6
www.nature.com/reviews/micro © 2008 Nature Publishing Group
REVIEWS Box 3 | Cell-wall biogenesis
Transglycosylase
4
GlcNAc
1
4
MurNAc
1
4
GlcNAc
1
4
MurNAc
1
4
GlcNAc
1
4
MurNAc
1
Transpeptidase
4
GlcNAc
1
4
MurNAc
1
4
GlcNAc
1
4
MurNAc
1
4
GlcNAc
1
4
MurNAc
1
| Microbiology The shape of a bacterium is determined by the architecture of its cell wall (peptidoglycan) —Nature a rigidReviews meshwork that is formed by linear glucan strands that are crosslinked by short peptide bridges128 (see figure). Peptidoglycan biosynthesis initiates in the cytoplasm with the assembly of a lipid-bound disaccharide–pentapeptide precursor. This basic building block comprises the two aminosugars N‑acetylglucosamine (GlcNAc) and N‑acetylmuramic acid (MurNAc), which are connected by a β‑1,4-glycosidic bond and a conserved five-amino-acid peptide that is attached to the lactyl group of the MurNAc unit. Once completed, the precursor is transported across the cytoplasmic membrane and, after release from its lipid carrier, incorporated into the peptidoglycan superstructure. This task is accomplished by a group of enzymes that are called penicillin-binding proteins (PBPs), which include transglycosylases and transpeptidases. Transglycosylases catalyse the formation of another β‑1,4-glycosidic bond between the disaccharide moiety of the precursor and the end of an existing glycan strand. As a result, linear polymers of alternating GlcNAc and MurNAc units are generated, which are all arrayed in parallel to each other. The pentapeptides, which protrude perpendicularly from the sugar backbone, subsequently function to interconnect neighbouring glycan strands. To this end, they are linked in a pair-wise manner by peptide bonds that are formed with the help of transpeptidases. High-molecular-weight PBPs exhibit both transglycosylase and transpeptidase activity, whereas low-molecular-weight PBPs (such as PBP2) function as transpeptidases only. Growth of the bacterium requires continuous remodelling of the peptidoglycan envelope. Bacteria, therefore, also possess a series of autolytic enzymes, such as glycosidases, peptidases and amidases, that selectively cleave the molecular meshwork and thereby facilitate the insertion of additional cell-wall material. Peptidoglycan Meshwork of highly crosslinked glycan strands that constitutes the bacterial cell wall.
Penicillin-binding protein A protein involved in peptidoglycan biosynthesis that is targeted and inactivated by the antibiotic penicillin and its derivatives.
Centromere A region of a DNA molecule that is attached to the DNA-segregation apparatus.
Walker ATPase This ATPase is characterized by the presence of two conserved sequence motifs (Walker A and Walker B motif), which form parts of the nucleotide-binding pocket.
of peptidoglycan formation, was found to be associated with the cell-division apparatus throughout the period of zonal growth38,39, medial cell-wall extension might be promoted by restricting the delivery of peptidoglycan precursors to the mid-cell region. However, future studies are necessary to determine whether this process is additionally supported by the recruitment of a specific set of peptidoglycan synthases from the periphery to the cell centre. MurG is also enriched at the cell-division site of E. coli40, which suggests that mid-cell growth zones could be a common phenomenon in bacteria.
Bacterial DNA segregation Evidence has accumulated that, in addition to protein complexes, plasmids and chromosomes are also actively positioned within bacterial cells. As for morphogenesis, dynamic cytoskeletal filaments have emerged as essential factors in the underlying localization mechanisms.
nature reviews | microbiology
Plasmid segregation. The first indication that replicated DNA might be inherited in a non-random way came from work on low-copy-number replicons. Localization studies revealed that the copies of these plasmids are actively separated from each other and positioned in the incipient daughter-cell compartments before cell division occurs41–46. In all cases that have been investigated in detail, partitioning relies on the activities of three different plasmid-encoded factors — a centromeric sequence, a centromere-binding protein and an ATPase that interacts with the centromeric nucleoprotein complex. Despite this common theme, the nature of the individual factors and the segregation apparatus varies considerably among different plasmids. The most important difference concerns the ATPase component, which can be either a Walker ATPase (type I partitioning system), a member of the actin superfamily (type II plasmid-partitioning system) or a tubulin homologue. volume 6 | january 2008 | 33
© 2008 Nature Publishing Group
REVIEWS a
ParM–ATP Assembly
b
ParM–ADP ParR
GTP hydrolysis
Catastrophe
Stabilization
Figure 3 | Plasmid segregation by the actin homologue ParM. Unlike actin polymers, which eventually reach a steady Nature Reviews | Microbiology state that is characterized by an equilibrium between subunit addition and dissociation (treadmilling), ParM filaments continuously cycle between phases of rapid growth and complete disassembly (a). Biochemical analyses showed that only the ATP-bound form of ParM (blue spheres) is able to polymerize efficiently. However, owing to the cooperative nature of its ATPase activity, ParM rapidly hydrolyses its bound nucleotide once it has become integrated into a filament (purple spheres). Nevertheless, the overall structure remains stable and continues to grow as long as its ends remain capped by ATP-bound subunits. Once these are lost, for example, owing to a shortage in the supply of ParM monomers, the filament starts to depolymerize rapidly (catastrophe). Stabilization of the caps, by contrast, prevents disassembly and allows the polymer to continue elongation. ParR nucleoprotein complexes recruit filaments that have formed spontaneously in the cytoplasm and promote their extension into long polymeric structures. As a consequence, plasmids are pushed apart and moved to opposite ends of the cell (b).
Walker A cytoskeletal ATPases (WACA). A group of Walker ATPases that possess a distinct version of the Walker A motif that deviates from the universal consensus. These proteins share structural similarity with P‑loop GTPases and are recognized as members of the GTPase superfamily.
Plasmid segregation by actin-like proteins. The beststudied partitioning system that is based on an actin-like ATPase is that of the conjugative resistance plasmid R1. Its centromeric region (parC) is bound by the DNAbinding protein ParR47. This complex is then recognized by ParM, an actin-like protein, the ATPase activity of which is essential for plasmid stabilization48,49. In vitro, ParM polymerizes in an ATP-dependent manner, forming filaments that are reminiscent of F‑actin50,51. Despite this similarity, its assembly kinetics have several distinct features. For example, the nucleation of ParM is a rapid and spontaneous process51,52, with filaments elongating in a symmetrical bidirectional fashion that is contrary to the polarized growth of actin52. In addition, ParM filaments never reach an equilibrium between association and dissociation, as is typical for treadmilling actin polymers, but instead continuously cycle between phases of rapid growth and complete disassembly52,53 (FIG. 3a). This behaviour, which is known as dynamic instability, has previously only been observed for eukaryotic tubulin54. Localization studies have revealed that ParM can form axial filaments in E. coli51. Each filament is flanked by two copies of plasmid R1 that appear to be pushed apart towards the cell poles by the polymerization of ParM between them48. The biochemical mechanism of this segregation process has recently been unravelled by reconstituting the R1 partitioning system in vitro53. When a DNA fragment that contained the R1 parS region was attached to beads and mixed with ParR and ParM, the addition of ATP induced the formation of numerous short filaments that extended from the bead
34 | january 2008 | volume 6
surface and, owing to their dynamic instability, went through cycles of rapid growth and shrinkage. However, as soon as two of these beads came into proximity, the filaments between them suddenly started to grow and push them apart. Elongation occurred exclusively at the interface between ParM and the centromeric nucleoprotein complex and was based on the stabilizing effect of the ParR–parS complex on the filament ends, which prevented the spontaneous disassembly of the polymer. These data suggest that the three-component partitioning system of R1 is sufficient to place plasmid copies at opposite cell poles without the help of additional host factors (FIG. 3b). Interestingly, the B. subtilis plasmid pBET131 was found to encode a novel actin-like partitioning ATPase named AlfA that lacks dynamic instability and has polymerization kinetics that are similar to those of MreB55. Thus, actin homologues have adopted different molecular mechanisms to mediate the movement of sister plasmids into the two daughter-cell compartments. Plasmid segregation by Walker-type ATPases. Most non-actin-based partitioning ATPases are members of the so-called Walker A cytoskeletal ATPase (WACA) family (type I partitioning systems)56. Analogous to the R1 partitioning system, each of these factors interacts with a DNA-binding protein that associates cooperatively with a cluster of conserved sites on its cognate plasmid. Despite differences in their primary sequences, overall, WACA ATPases show similar behaviour in vitro. They possess weak cooperative ATPase activity and are capable of polymerizing in an ATP-dependent manner, forming www.nature.com/reviews/micro
© 2008 Nature Publishing Group
REVIEWS massive protofilament bundles that measure up to several micrometres in length. However, detailed analyses of their polymerization dynamics and interactions with the centromeric nucleoprotein complex have not yet been performed. The biology of these systems has recently been described in several excellent reviews57–59. Although type I and type II partitioning systems share some common principles, their modes of action are fundamentally different. Whereas ParM assembles into stable axial filaments that push plasmids to opposite cell poles48,53, WACA ATPases exhibit a highly dynamic localization pattern in vivo. SopA and ParA, which are encoded by the E. coli plasmids F and pB171, respectively, were shown to oscillate back and forth across the nucleoid at ~20 minute intervals60,61. However, the molecular mechanisms that couple the dynamic assembly of partitioning proteins to plasmid stabilization remain to be elucidated. Studies on the localization of plasmids F and pB171 revealed that at the beginning of their division cycle most cells contain a single plasmid cluster that is positioned at the cell centre. Later on, this cluster is duplicated, and the newly generated copies are moved to the quarter positions of the cell. Cytokinesis then occurs halfway between the two clusters, resulting in daughter cells that each carry a single cluster at the mid-cell42,43,45. If the copy number of pB171 is increased artificially, cells accumulate up to seven plasmid clusters that are evenly distributed along the long axis of the cell62. This finding indicates that type I partitioning systems do not move plasmids to defined subcellular positions, as is the case for the R1 partitioning system, but rather distribute them to clusters that are arranged in regular arrays within the nucleoid. Interestingly, the cytoplasmic chemotaxis sensory clusters of R. sphaeroides have a localization pattern that is strikingly similar to that of plasmids that are segregated by type I partitioning systems. The proper positioning of these clusters was found to be dependent on a protein, called PpfA, that is homologous to WACA plasmid partitioning ATPases63, which suggests that bacteria can use similar mechanisms to arrange proteins and DNA within the cell.
Nucleoid A distinct region within the cytoplasm that harbours the chromosomal DNA.
Origin of replication A chromosomal site that serves as the starting point of the bidirectional DNA-replication process.
Terminus A chromosomal region in which the two replication forks meet towards the end of DNA replication.
Plasmid segregation by a tubulin homologue. A third class of plasmid-segregation systems is based on the activity of tubulin homologues. Recent work showed that plasmid pBtoxis from Bacillus thuringiensis serovar israelensis encodes a protein — designated TubZ — that shows significant similarity to tubulin64,65 and is essential for its stable partitioning. TubZ assembles into highly dynamic filaments that translocate rapidly through the cell65. FRAP analyses revealed that filament migration is achieved by an actin-like treadmilling mechanism. However, it is unclear how the translocation of TubZ filaments is related to plasmid stabilization. The unusual characteristics of TubZ, a tubulin homologue with actin-like dynamics, and ParM, an actin homologue with tubulin-like dynamics, demonstrate that bacterial and eukaryotic cytoskeletal proteins have diverged considerably with respect to their function and polymerization mechanisms, even though they share the same evolutionary origin and overall structures.
nature reviews | microbiology
a
ori
ter
b ori
ter
ori
ter C. crescentus
E. coli
Figure 4 | The arrangement of chromosomal DNA in bacteria. Consecutive segments of chromosomal DNA are Nature Reviews | Microbiology folded into supercoiled domains that are stacked on top of each other and arranged into a circular superstructure (a). Consequently, the subcellular position of a locus correlates directly with its location on the circular chromosomal map. In Caulobacter crescentus, the origin of replication (ori) is localized to the flagellated pole of the swarmer cell, whereas the terminus (ter) is found at the opposite pole (b). Therefore, the left and right arms of the chromosome are oriented in parallel to the long axis of the cell. In Escherichia coli, by contrast, ori and ter are both located at mid-cell, such that the two arms of the chromosome flank the transverse cellular axis. Panel a adapted, with permission, from REF. 77 (2006) Elsevier Science.
Arrangement of chromosomal DNA. Bacterial genomes are usually organized into circular chromosomes (one or several) that are up to several megabase pairs in size. It has long been unclear how the cell packages these enormous molecules into the confined space of its cytoplasm and concomitantly ensures their faithful replication and partitioning during each cell-division cycle. With the advent of site-specific labelling methods, it has become possible to probe the arrangement of chromatin in a systematic fashion66,67. An early study in B. subtilis determined the subcellular location of four chromosomal loci and found that their arrangement followed a reproducible pattern68. Conclusive evidence that chromosomal DNA is not distributed randomly within the cell but in fact has a highly conserved organization was provided by a large-scale analysis that investigated the subcellular positioning of 112 chromosomal loci in C. crescentus69 (FIG. 4). This work showed that the origin of replication is invariantly found at the flagellated pole of the bacterium and the terminus is located at the opposite end of the cell. Other loci are arranged between these two fixed points in exactly the same order as on the circular chromosomal map. A similar correlation between the subcellular and chromosomal positions of loci was revealed in E. coli. However, unlike C. crescentus, E. coli places the origin and terminus regions at the cell centre, and the two arms of the chromosome (replichores) form separate domains that are located on opposite sides of its transverse axis70,71. The chromosomal volume 6 | january 2008 | 35
© 2008 Nature Publishing Group
REVIEWS architecture of other organisms has not been analysed in detail yet, but it is likely that the linear ordering of loci is a common principle among bacteria. So, how is this conserved arrangement established? Studies on the movement of newly duplicated chromosomal regions showed that the specific localization of each site is established during the DNA-segregation process. Immediately after the initiation of replication, the newly synthesized origin regions are separated and rapidly moved to their conserved positions in the incipient daughter cells69,72–75. Subsequently, the remaining bulk of the chromosome is duplicated progressively from the origin to the terminus region. Similar to the origin regions, the two new copies of each locus are quickly partitioned. They are condensed immediately after emergence from the replisome, segregated into the daughter-cell compartments and added successively to the newly replicated DNA that has already been deposited there69,70,76. Unlike eukaryotes, therefore, bacteria segregate their DNA while replication is in progress, with the two sister nucleoids growing in layers that reflect the succession of loci on the chromosome. The mechanisms that mediate bacterial-chromosome dynamics are unclear, and several different models have been proposed to account for the rapid and ordered movement of loci during the chromosomesegregation process77. The current understanding is that the two copies of the origin are actively separated from each other and positioned in the incipient daughter cells, so that they serve as landmarks for the establishment of the new sister nucleoids. The bulk of the chromosome then follows as a consequence of DNA condensation, probably supported by the activities of other factors. Finally, the two sister chromosomes are decatenated and actively cleared from the cell-division site to allow cytokinesis to occur78. Chromosome segregation. The existence of an active mechanism that mediates origin segregation is supported by the fact that, in all bacteria investigated, movement of the two origin copies occurs in a highly reproducible and directed manner and is significantly faster than cell elongation69,79,80. One of the factors that has been implicated in this partitioning process is the cytoskeletal protein MreB. Its overproduction, depletion16,34,81 or inactivation by the antibiotic A22 (Refs 17,18) results in misplacement of the origin regions, and chromosome-segregation defects in various bacteria. Moreover, MreB was shown to interact, directly or indirectly, with a chromosomal region that flanks the origin of replication in C. crescentus17. However, it has remained controversial whether these results actually reflect a direct involvement in DNA segregation. Other studies have provided evidence of a role for type I partitioning systems in chromosome segregation. In most bacteria, an operon encoding the WACA ATPase ParA and its cognate centromere-binding protein ParB is located in the vicinity of the replication origin, whereas conserved ParB-binding sites (parS) are scattered throughout the origin-proximal region of the chromosome82. Recent studies in B. subtilis demonstrated that, upon binding to individual sites83, ParB spreads into the flanking chromosomal regions, forming nucleoprotein complexes that cover up to 20 kilobases (kb) of DNA84,85. 36 | january 2008 | volume 6
These individual complexes then aggregate into a single centromere-like superstructure in a ParA-dependent manner86. Chromosomally encoded partitioning systems are able to functionally replace plasmid-borne type I partitioning systems and confer stability to intrinsically unstable plasmids, indicating that they indeed constitute active segregation machineries87,88. In agreement with this finding, the ParAB system was shown to be required for proper transfer of the chromosomal origin region into the forespore compartment at the onset of sporulation in B. subtilis89. In doing so, it cooperates with a sporulation-specific protein, named RacA, that serves to tether the origin-proximal region of the chromosome to the cell pole89–91. Moreover, vegetative cells of B. subtilis that are deficient in ParA or ParB fail to separate their origin regions after duplication, although the bulk of the chromosome is left unaffected92. Evidence for a direct role of ParAB in origin segregation has come from the study of Vibrio cholerae, a bacterium that possesses two circular chromosomes, of which each encodes its own type I partitioning system. The smaller chromosome (ChrII) has the typical segregation pattern that is shared by low-copy-number plasmids, such as F and pB171, and the larger chromosome (ChrI) exhibits dynamics that are similar to those of the C. crescentus chromosome80,93. Interestingly, ParBI, which binds to 3 parS sites that are approximately 65 kb to the right of the replication origin of the large chromosome, is consistently localized closer to the cell poles than to the origin itself and moves ahead of the origin during the segregation process94. These findings suggest that the ParBI–parS complex defines the chromosomal centromere and mediates segregation and polar positioning of the origin regions. ParBI and the origin regions lose their polar localization in mutants that lack ParAI and are instead found near the cell centre before the start of S phase and around the quarter positions in cells that have initiated DNA replication. Therefore, ParAI seems to be necessary for correct positioning of the ParBI–parS complex. Time-lapse analyses indeed showed that the protein assembles into a dynamic polymeric structure that seems to pull the moving ParB I–parS complex from the old pole towards the new pole94. In doing so, the ParAI structure first stretches throughout the cell and then progressively retracts in the direction of the new pole, with its lagging edge colocalizing with the moving ParBI–parS complex. Thus, a mitotic-like mechanism that is based on type I partitioning systems seems to mediate origin segregation in V. cholerae and, presumably, other bacteria. V. cholerae cells that lack ParAI only have minor growth and overall chromosome-partitioning defects94, and similar observations have been made for most other bacteria in which the type I chromosome-partitioning systems had been inactivated82. These findings indicate that active segregation of the newly synthesized origin regions is not necessarily required for the establishment of two fully separated sister nucleoids. As a result, chromosome segregation might rely, in part, on other mechanisms, the precise nature of which remains to be determined. www.nature.com/reviews/micro
© 2008 Nature Publishing Group
REVIEWS a E. coli MinD
MinE FtsZ
b C. crescentus MipZ Origin
Nucleoid
FtsZ ParB MinC
Release of MinCD
Origin segregation
Assembly of a new patch
Displacement of FtsZ
Start of the next cycle
Z-ring formation
Figure 5 | Positioning of the cell-division plane. In Escherichia assembly of the FtsZ Naturecoli, Reviews | Microbiology ring is restricted to the mid-cell by nucleoid occlusion and the MinE-driven pole-to-pole oscillation of the cell-division inhibitor MinCD (a). By contrast, division-plane localization in Caulobacter crescentus is mediated by MipZ, an inhibitor of FtsZ polymerization that forms a complex with the DNA-binding protein ParB at the chromosomal origin of replication, thus exploiting chromosome dynamics for its subcellular positioning (b). Panel b adapted, with permission, from REF. 113 (2006) Elsevier Science.
To generate offspring that contain equivalent DNA, the cell must not only provide a mechanism for proper chromosome segregation, but also ensure that cytokinesis occurs precisely between the two newly formed sister nucleoids. The core machinery that mediates cell division is conserved in most bacteria. Nevertheless, several independent mechanisms have evolved to determine the subcellular site of its assembly.
Nucleoid occlusion The inhibitory effect of the nucleoid on the formation of the septal FtsZ ring.
Division-site placement Formation of the bacterial-cell-division apparatus, called the divisome, occurs in a multistep process that is initiated by assembly of the tubulin homologue FtsZ into a ring-shaped structure at the future division site. This cytoskeletal element mediates the recruitment of all other divisome components, both directly and indirectly, and plays a crucial part in the subsequent constriction process3. Although the three-dimensional structures of FtsZ and tubulin are remarkably similar95, the two proteins show distinct polymerization behaviour. The assembly dynamics of FtsZ have been discussed extensively in several recent reviews56,96,97. Most importantly, FtsZ has a high propensity to polymerize in vivo, forming spiraland arc-like structures that are in a constant process of polymerization and disassembly. Their consolidation at the cell centre, which results from the destabilizing effect of negative regulators in the polar regions of the cell, is thought to provide the basis of septal-ring formation.
nature reviews | microbiology
The Min system. The first mechanism that was shown to control placement of the bacterial division site was discovered in E. coli. It is based on the activity of three proteins, encoded by the minCDE operon, that cooperate to establish a fascinating, self-contained oscillatory system98 (FIG. 5a). MinD is an ATPase of the WACA family and interacts with MinC to form a membrane-associated, topologically unspecific inhibitor of FtsZ-ring formation99,100. Owing to the activity of MinE, however, the MinCD complex is normally restricted to the polar regions of the cell, thus only allowing cell division to occur close to the cell centre101,102. MinCD continuously oscillates between the two cell halves in 40–50 second intervals103–105. In doing so, it first assembles into a large polymeric patch at one end of the cell. Subsequently, this cap-like structure starts to shrink, gradually losing subunits from the pole-distal end, until it has completely disappeared. Concomitantly, a new MinCD assembly forms at the opposite pole, and the cycle starts again. MinE forms a circular structure that follows the retracting edge of the MinCD patches101,106. In its absence, the oscillatory behaviour of the system is lost, and the MinCD complex is evenly dispersed throughout the membrane103. This mechanism is widely used among bacteria, albeit frequently in modified forms. B. subtilis, for example, lacks MinE and positions the MinCD complex statically at its cell poles107,108. Details on the molecular function of the Min proteins are provided in two recent comprehensive reviews96,109. Nucleoid occlusion. In mutants that lack a functional Min system, cell division occurs either close to the poles or between nucleoids, but never on top of regions that contain chromosomal DNA110. Recent studies identified two unrelated proteins, Noc from B. subtilis111 and SlmA from E. coli112, that mediate this so-called nucleoid occlusion effect (FIG. 5a). In their absence, Min-deficient cells accumulate numerous FtsZ clusters that are randomly distributed throughout the cytoplasm. These clusters frequently overlap with the nucleoids and, under certain conditions, promote cell-division events that lead to dissection of the chromosome. However, the inactivation of NocA or SlmA has little effect on FtsZ localization in cells that have a functional Min system, which suggests that nucleoid occlusion might serve as a fail-safe mechanism that ensures proper cell division under conditions of unbalanced growth. However, how do Noc and SlmA function? Both proteins possess a helix–turn–helix DNAbinding motif that allows them to interact nonspecifically with chromosomal DNA and therefore to colocalize with the nucleoid. SlmA was shown to mediate the recruitment of FtsZ to the nucleoid if expressed at high levels and has a strong bundling effect on FtsZ filaments in vitro112. Nucleoid occlusion factors might, therefore, function by out-competing other cell-division proteins that bind to and stabilize the FtsZ ring, thereby efficiently preventing the assembly of a functional divisome. Regulation of cell division by MipZ. There are several organisms that lack MinCDE homologues and nucleoid occlusion but yet still manage to divide properly in the mid-cell region, which suggests the existence of volume 6 | january 2008 | 37
© 2008 Nature Publishing Group
REVIEWS alternative regulatory mechanisms. Recent work in C. crescentus has indeed identified a novel system that couples the assembly and positioning of the FtsZ ring to the initiation of chromosome replication and the bipolar positioning of the duplicated origin regions113 (FIG. 5b). Its central component is MipZ, an essential ATPase that has weak similarity to ParA-like DNA-partitioning proteins and is widely conserved among alphaproteobacteria. MipZ directly interacts with the chromosomepartitioning protein ParB113, which, in turn, binds to a cluster of sites (parS) that are approximately 15 kb away from the chromosomal origin of replication114,115. Together with the origin region, the resulting complex is positioned at the old pole in newborn cells113. Initiation of DNA replication then generates two copies of the parS-containing segment, both of which are immediately decorated with ParB and MipZ. During the subsequent segregation process, one of these segments stays at the original position, while the second copy moves rapidly across the cell to the opposite pole. MipZ interacts with ParB in a highly dynamic manner that is regulated by ATP binding and hydrolysis113. Consequently, it is not stably tethered to the origin regions but rather is distributed in a gradient, with its concentration progressively increasing towards the polar ParB–parS complexes. In vitro studies demonstrated that MipZ acts as an inhibitor of FtsZ polymerization113. Consequently, FtsZ is consistently found in the subcellular region that exhibits the lowest concentration of MipZ. In newborn cells, it initially forms a focus at the new pole, opposite the single MipZ–ParB complex that is located at the old pole. After duplication and segregation of the origin regions, the polar FtsZ aggregate is disassembled and a new one is formed at the cell centre. Synthesis of the other celldivision proteins, which occurs later in the C. crescentus cell cycle116, might then stabilize FtsZ into a septal ring that can establish a functional divisome and initiate cytokinesis. This mechanism ensures that the division site is positioned between the two nascent sister nucleoids. Based on its homology to ParA-type WACA ATPases, and its interaction with ParB, MipZ is probably derived from a plasmid-partitioning protein that later adopted a role in the spatial regulation of cell division.
Concluding remarks Flexibility in the use of dynamic protein structures is a common theme in bacteria. Tubulin filaments are elementary parts of the cell-division apparatus as well as mediators of plasmid segregation. Similarly, actin cables function in DNA partitioning, cell-shape determination and protein and vesicle localization. Comparable diversity is also observed for WACA ATPases, which form filaments that are involved in DNA segregation as well as dynamic
1.
2.
Rudner, D. Z., Pan, Q. & Losick, R. M. Evidence that subcellular localization of a bacterial membrane protein is achieved by diffusion and capture. Proc. Natl Acad. Sci. USA 99, 8701–8706 (2002). Deich, J., Judd, E. M., McAdams, H. H. & Moerner, W. E. Visualization of the movement of single histidine kinase molecules in live Caulobacter cells. Proc. Natl Acad. Sci. USA 101, 15921–15926 (2004).
3. 4.
5.
regulatory complexes that define the cell-division plane. This astonishing range of mechanisms has been identified almost exclusively in a handful of established model organisms that represent only a small fraction of the bacterial world. Therefore, novel systems that are involved in the dynamic temporal and spatial organization of bacteria might await discovery in the future. Although several different systems that mediate the spatial organization of bacteria have been identified, their function is often poorly understood. Notably, the factors that mediate the specific positioning of protein complexes at the bacterial cell poles remain obscure. Recent work in C. crescentus identified a membrane protein, TipN, that interacts with the cell-division apparatus and remains attached to the new poles after cytokinesis, so serving as a landmark that regulates the assembly and positioning of the flagellum in the two daughter cells117,118. However, even though this system provides an elegant way to pass on positional information to the next generation and to differentiate between the old and new cell pole, the primary determinants that retain TipN at its polar position are still unknown. One of the structures that might help to define the cell poles is peptidoglycan. Its architecture at the curved polar regions is probably different from that in the straight longitudinal sections of the cell, thus providing a specific target for peptidoglycanbinding proteins. In addition, there are data that suggest a role for cardiolipin in polar protein localization. This minor lipid species was reported to accumulate, possibly owing to a self-organizing process119, in the polar regions of the cytoplasmic membrane in E. coli120 and B. subtilis121, and it has been suggested that it mediates polar positioning of the osmosensory transporter ProP in E. coli122. However, although both the peptidoglycan- and cardiolipin-mediated localization mechanisms are appealing, future work is required to test their validity. Aside from polar targeting, little information is available on the biochemical mechanisms that are responsible for the assembly of bacterial actin homologues, tubulin homologues and WACA ATPases into dynamic cytoskeletal filaments. The same applies to the temporal and spatial regulation of these structures as well as their interplay with other cellular factors. Finally, further studies are required to fully understand the molecular details that underlie the regulation of division-site placement by nucleoid occlusion and MipZ. As demonstrated by FtsZ positioning in C. crescentus, different spatial regulatory systems can be tightly interconnected. It will, therefore, be a major task to elucidate the interdependence of dynamic processes in the cell and merge the knowledge of individual processes into a global concept of cellular function.
Goehring, N. W. & Beckwith, J. Diverse paths to midcell: assembly of the bacterial cell division machinery. Curr. Biol. 15, R514–R526 (2005). Rubio, A. & Pogliano, K. Septal localization of forespore membrane proteins during engulfment in Bacillus subtilis. EMBO J. 23, 1636–1646 (2004). Blaylock, B., Jiang, X., Rubio, A., Moran, C. P. Jr. & Pogliano, K. Zipper-like interaction between proteins
38 | january 2008 | volume 6
6.
in adjacent daughter cells mediates protein localization. Genes Dev. 18, 2916–2928 (2004). Demonstration of a zipper-like interaction between proteins from the forespore and mother cell that allows the assembly of protein complexes at the septal membrane. Doan, T., Marquis, K. A. & Rudner, D. Z. Subcellular localization of a sporulation membrane protein is achieved through a network of interactions along and
www.nature.com/reviews/micro © 2008 Nature Publishing Group
REVIEWS
7.
8.
9.
10.
11.
12.
13. 14.
15.
16. 17.
18.
19. 20.
21.
22. 23. 24.
25.
26.
27.
across the septum. Mol. Microbiol. 55, 1767–1781 (2005). Jiang, X., Rubio, A., Chiba, S. & Pogliano, K. Engulfment-regulated proteolysis of SpoIIQ: evidence that dual checkpoints control σk activity. Mol. Microbiol. 58, 102–115 (2005). Rubio, A., Jiang, X. & Pogliano, K. Localization of translocation complex components in Bacillus subtilis: enrichment of the signal recognition particle receptor at early sporulation septa. J. Bacteriol. 187, 5000–5002 (2005). Steinhauer, J., Agha, R., Pham, T., Varga, A. W. & Goldberg, M. B. The unipolar Shigella surface protein IcsA is targeted directly to the bacterial old pole: IcsP cleavage of IcsA occurs over the entire bacterial surface. Mol. Microbiol. 32, 367–377 (1999). Charles, M., Perez, M., Kobil, J. H. & Goldberg, M. B. Polar targeting of Shigella virulence factor IcsA in Enterobacteriacae and Vibrio. Proc. Natl Acad. Sci. USA 98, 9871–9876 (2001). Brandon, L. D. et al. IcsA, a polarly localized autotransporter with an atypical signal peptide, uses the Sec apparatus for secretion, although the Sec apparatus is circumferentially distributed. Mol. Microbiol. 50, 45–60 (2003). Jones, L. J., Carballido-Lopez, R. & Errington, J. Control of cell shape in bacteria: helical, actin-like filaments in Bacillus subtilis. Cell 104, 913–922 (2001). Landmark paper showing the helical arrangement of bacterial actin-like filaments in the cell and their role as cytoskeletal elements that are involved in the determination of cell shape. Defeu Soufo, H. J. & Graumann, P. L. Dynamic movement of actin-like proteins within bacterial cells. EMBO Rep. 5, 789–794 (2004). Shih, Y. L., Le, T. & Rothfield, L. Division site selection in Escherichia coli involves dynamic redistribution of Min proteins within coiled structures that extend between the two cell poles. Proc. Natl Acad. Sci. USA 100, 7865–7870 (2003). Figge, R. M., Divakaruni, A. V. & Gober, J. W. MreB, the cell shape-determining bacterial actin homologue, co-ordinates cell wall morphogenesis in Caulobacter crescentus. Mol. Microbiol. 51, 1321–1332 (2004). Gitai, Z., Dye, N. & Shapiro, L. An actin-like gene can determine cell polarity in bacteria. Proc. Natl Acad. Sci. USA 101, 8643–8648 (2004). Gitai, Z., Dye, N. A., Reisenauer, A., Wachi, M. & Shapiro, L. MreB actin-mediated segregation of a specific region of a bacterial chromosome. Cell 120, 329–341 (2005). Kruse, T. et al. Actin homolog MreB and RNA polymerase interact and are both required for chromosome segregation in Escherichia coli. Genes Dev. 20, 113–124 (2006). Pollard, T. D. Regulation of actin filament assembly by arp2/3 complex and formins. Annu. Rev. Biophys. Biomol. Struct. 36, 451–477 (2007). Esue, O., Wirtz, D. & Tseng, Y. GTPase activity, structure, and mechanical properties of filaments assembled from bacterial cytoskeleton protein MreB. J. Bacteriol. 188, 968–976 (2006). van den Ent, F., Amos, L. A. & Löwe, J. Prokaryotic origin of the actin cytoskeleton. Nature 413, 39–44 (2001). A crystallographic analysis that demonstrates high structural similarity between MreB and actin, therefore providing evidence for a common evolutionary origin of the two proteins. Esue, O., Cordero, M., Wirtz, D. & Tseng, Y. The assembly of MreB, a prokaryotic homolog of actin. J. Biol. Chem. 280, 2628–2635 (2005). Carballido-Lopez, R. & Errington, J. The bacterial cytoskeleton: in vivo dynamics of the actin-like protein Mbl of Bacillus subtilis. Dev. Cell 4, 19–28 (2003). Slovak, P. M., Wadhams, G. H. & Armitage, J. P. Localization of MreB in Rhodobacter sphaeroides under conditions causing changes in cell shape and membrane structure. J. Bacteriol. 187, 54–64 (2005). Slovak, P. M., Porter, S. L. & Armitage, J. P. Differential localization of Mre proteins with PBP2 in Rhodobacter sphaeroides. J. Bacteriol. 188, 1691–1700 (2006). Defeu Soufo, H. J. & Graumann, P. L. Dynamic localization and interaction with other Bacillus subtilis actin-like proteins are important for the function of MreB. Mol. Microbiol. 62, 1340–1356 (2006). Carballido-Lopez, R. et al. Actin homolog MreBH governs cell morphogenesis by localization of the cell wall hydrolase LytE. Dev. Cell 11, 399–409 (2006).
28.
29.
30. 31.
32. 33.
34.
35.
36.
37.
38.
39.
40.
41.
42. 43.
44.
45.
46. 47.
48.
This paper demonstrates a direct interaction between the actin-like protein MreBH and the peptidoglycan hydrolase LytE. Kim, S. Y., Gitai, Z., Kinkhabwala, A., Shapiro, L. & Moerner, W. E. Single molecules of the bacterial actin MreB undergo directed treadmilling motion in Caulobacter crescentus. Proc. Natl Acad. Sci. USA 103, 10929–10934 (2006). Analysis of the dynamics of MreB cables in C. crescentus by tracking the movement of single fluorescently labelled MreB subunits within the cell. Kruse, T., Bork-Jensen, J. & Gerdes, K. The morphogenetic MreBCD proteins of Escherichia coli form an essential membrane-bound complex. Mol. Microbiol. 55, 78–89 (2005). van den Ent, F. et al. Dimeric structure of the cell shape protein MreC and its functional implications. Mol. Microbiol. 62, 1631–1642 (2006). Divakaruni, A. V., Loo, R. R., Xie, Y., Loo, J. A. & Gober, J. W. The cell-shape protein MreC interacts with extracytoplasmic proteins including cell wall assembly complexes in Caulobacter crescentus. Proc. Natl Acad. Sci. USA 102, 18602–18607 (2005). Lee, J. C. & Stewart, G. C. Essential nature of the mreC determinant of Bacillus subtilis. J. Bacteriol. 185, 4490–4498 (2003). Leaver, M. & Errington, J. Roles for MreC and MreD proteins in helical growth of the cylindrical cell wall in Bacillus subtilis. Mol. Microbiol. 57, 1196–1209 (2005). Soufo, H. J. & Graumann, P. L. Actin-like proteins MreB and Mbl from Bacillus subtilis are required for bipolar positioning of replication origins. Curr. Biol. 13, 1916–1920 (2003). Dye, N. A., Pincus, Z., Theriot, J. A., Shapiro, L. & Gitai, Z. Two independent spiral structures control cell shape in Caulobacter. Proc. Natl Acad. Sci. USA 102, 18608–18613 (2005). Ausmees, N., Kuhn, J. R. & Jacobs-Wagner, C. The bacterial cytoskeleton: an intermediate filament-like function in cell shape. Cell 115, 705–713 (2003). Discovery of the first bacterial intermediate filament protein. Parry, D. A., Strelkov, S. V., Burkhard, P., Aebi, U. & Herrmann, H. Towards a molecular description of intermediate filament structure and assembly. Exp. Cell Res. 313, 2204–2216 (2007). Aaron, M. et al. The tubulin homologue FtsZ contributes to cell elongation by guiding cell wall precursor synthesis in Caulobacter crescentus. Mol. Microbiol. 64, 938–952 (2007). Divakaruni, A. V., Baida, C., White, C. L. & Gober, J. W. The cell shape proteins MreB and MreC control cell morphogenesis by positioning cell wall synthetic complexes. Mol. Microbiol. 66, 174–188 (2007). Mohammadi, T. et al. The essential peptidoglycan glycosyltransferase MurG forms a complex with proteins involved in lateral envelope growth as well as with proteins involved in cell division in Escherichia coli. Mol. Microbiol. 65, 1106–1121 (2007). Li, Y. & Austin, S. The P1 plasmid is segregated to daughter cells by a ‘capture and ejection’ mechanism coordinated with Escherichia coli cell division. Mol. Microbiol. 46, 63–74 (2002). Gordon, S., Rech, J., Lane, D. & Wright, A. Kinetics of plasmid segregation in Escherichia coli. Mol. Microbiol. 51, 461–469 (2004). Ebersbach, G. & Gerdes, K. Bacterial mitosis: partitioning protein ParA oscillates in spiral-shaped structures and positions plasmids at mid-cell. Mol. Microbiol. 52, 385–398 (2004). Ho, T. Q., Zhong, Z., Aung, S. & Pogliano, J. Compatible bacterial plasmids are targeted to independent cellular locations in Escherichia coli. EMBO J. 21, 1864–1872 (2002). Niki, H. & Hiraga, S. Subcellular distribution of actively partitioning F plasmid during the cell division cycle in E. coli. Cell 90, 951–957 (1997). First analysis of the dynamics of active plasmid segregation in bacteria. Gordon, G. S. et al. Chromosome and low copy plasmid segregation in E. coli: visual evidence for distinct mechanisms. Cell 90, 1113–1121 (1997). Dam, M. & Gerdes, K. Partitioning of plasmid R1. Ten direct repeats flanking the parA promoter constitute a centromere-like partition site parC, that expresses incompatibility. J. Mol. Biol. 236, 1289–1298 (1994). Møller-Jensen, J. et al. Bacterial mitosis: ParM of plasmid R1 moves plasmid DNA by an actin-like insertional polymerization mechanism. Mol. Cell 12, 1477–1487 (2003).
nature reviews | microbiology
49. Jensen, R. B. & Gerdes, K. Partitioning of plasmid R1. The ParM protein exhibits ATPase activity and interacts with the centromere-like ParR–parC complex. J. Mol. Biol. 269, 505–513 (1997). 50. van den Ent, F., Møller-Jensen, J., Amos, L. A., Gerdes, K. & Löwe, J. F‑actin‑like filaments formed by plasmid segregation protein ParM. EMBO J. 21, 6935–6943 (2002). 51. Møller-Jensen, J., Jensen, R. B., Löwe, J. & Gerdes, K. Prokaryotic DNA segregation by an actin-like filament. EMBO J. 21, 3119–3127 (2002). Demonstration of the partitioning of plasmid R1 by a dynamic axial ParM filament that grows between the segregating plasmid copies. 52. Garner, E. C., Campbell, C. S. & Mullins, R. D. Dynamic instability in a DNA-segregating prokaryotic actin homolog. Science 306, 1021–1025 (2004). Elaborate biochemical analysis of the mechanism that underlies the polymerization of the actin-like protein ParM. 53. Garner, E. C., Campbell, C. S., Weibel, D. B. & Mullins, R. D. Reconstitution of DNA segregation driven by assembly of a prokaryotic actin homolog. Science 315, 1270–1274 (2007). This paper reports the first successful reconstitution of a plasmid-partitioning system in vitro and provides detailed information on the R1 partitioning mechanism. 54. Mitchison, T. & Kirschner, M. Dynamic instability of microtubule growth. Nature 312, 237–242 (1984). 55. Becker, E. et al. DNA segregation by the bacterial actin AlfA during Bacillus subtilis growth and development. EMBO J. 25, 5919–5931 (2006). 56. Michie, K. A. & Löwe, J. Dynamic filaments of the bacterial cytoskeleton. Annu. Rev. Biochem. 75, 467–492 (2006). 57. Ebersbach, G. & Gerdes, K. Plasmid segregation mechanisms. Annu. Rev. Genet. 39, 453–479 (2005). 58. Hayes, F. & Barilla, D. The bacterial segrosome: a dynamic nucleoprotein machine for DNA trafficking and segregation. Nature Rev. Microbiol. 4, 133–143 (2006). 59. Leonard, T. A., Møller-Jensen, J. & Löwe, J. Towards understanding the molecular basis of bacterial DNA segregation. Philos. Trans. R. Soc. Lond. B 360, 523–535 (2005). 60. Ebersbach, G. & Gerdes, K. The double par locus of virulence factor pB171: DNA segregation is correlated with oscillation of ParA. Proc. Natl Acad. Sci. USA 98, 15078–15083 (2001). 61. Lim, G. E., Derman, A. I. & Pogliano, J. Bacterial DNA segregation by dynamic SopA polymers. Proc. Natl Acad. Sci. USA 102, 17658–17663 (2005). 62. Ebersbach, G. et al. Regular cellular distribution of plasmids by oscillating and filament-forming ParA ATPase of plasmid pB171. Mol. Microbiol. 61, 1428–1442 (2006). 63. Thompson, S. R., Wadhams, G. H. & Armitage, J. P. The positioning of cytoplasmic protein clusters in bacteria. Proc. Natl Acad. Sci. USA 103, 8209–8214 (2006). 64. Tang, M., Bideshi, D. K., Park, H. W. & Federici, B. A. Minireplicon from pBtoxis of Bacillus thuringiensis subsp. israelensis. Appl. Environmen. Microbiol. 72, 6948–6954 (2006). 65. Larsen, R. A. et al. Treadmilling of a prokaryotic tubulin-like protein, TubZ, required for plasmid stability in Bacillus thuringiensis. Genes Dev. 21, 1340–1352 (2007). First report of a bacterial tubulin homologue that performs a function in plasmid segregation. 66. Gitai, Z., Thanbichler, M. & Shapiro, L. The choreographed dynamics of bacterial chromosomes. Trends Microbiol. 13, 221–228 (2005). 67. Thanbichler, M., Wang, S. C. & Shapiro, L. The bacterial nucleoid: a highly organized and dynamic structure. J. Cell. Biochem. 96, 506–521 (2005). 68. Teleman, A. A., Graumann, P. L., Lin, D. C., Grossman, A. D. & Losick, R. Chromosome arrangement within a bacterium. Curr. Biol. 8, 1102–1109 (1998). 69. Viollier, P. H. et al. Rapid and sequential movement of individual chromosomal loci to specific subcellular locations during bacterial DNA replication. Proc. Natl Acad. Sci. USA 101, 9257–9262 (2004). The first global analysis of chromosome organization and segregation in live bacterial cells. 70. Wang, X., Liu, X., Possoz, C. & Sherratt, D. J. The two Escherichia coli chromosome arms locate to separate cell halves. Genes Dev. 20, 1727–1731 (2006). 71. Nielsen, H. J., Ottesen, J. R., Youngren, B., Austin, S. J. & Hansen, F. G. The Escherichia coli chromosome is organized with the left and right chromosome arms
volume 6 | january 2008 | 39 © 2008 Nature Publishing Group
REVIEWS
72.
73. 74.
75. 76.
77. 78. 79.
80.
81.
82. 83. 84.
85.
86.
87. 88.
89.
90.
91.
92.
93.
in separate cell halves. Mol. Microbiol. 62, 331–338 (2006). Jensen, R. B. & Shapiro, L. The Caulobacter crescentus smc gene is required for cell cycle progression and chromosome segregation. Proc. Natl Acad. Sci. USA 96, 10661–10666 (1999). Li, Y., Sergueev, K. & Austin, S. The segregation of the Escherichia coli origin and terminus of replication. Mol. Microbiol. 46, 985–996 (2002). Roos, M. et al. The replicated ftsQAZ and minB chromosomal regions of Escherichia coli segregate on average in line with nucleoid movement. Mol. Microbiol. 39, 633–640 (2001). Niki, H., Yamaichi, Y. & Hiraga, S. Dynamic organization of chromosomal DNA in Escherichia coli. Genes Dev. 14, 212–223 (2000). Nielsen, H. J., Li, Y., Youngren, B., Hansen, F. G. & Austin, S. Progressive segregation of the Escherichia coli chromosome. Mol. Microbiol. 61, 383–393 (2006). Thanbichler, M. & Shapiro, L. Chromosome organization and segregation in bacteria. J. Struct. Biol. 156, 292–303 (2006). Bigot, S., Sivanathan, V., Possoz, C., Barre, F. X. & Cornet, F. FtsK, a literate chromosome segregation machine. Mol. Microbiol. 64, 1434–1441 (2007). Webb, C. D. et al. Use of time-lapse microscopy to visualize rapid movement of the replication origin region of the chromosome during the cell cycle in Bacillus subtilis. Mol. Microbiol. 28, 883–892 (1998). Fiebig, A., Keren, K. & Theriot, J. A. Fine-scale timelapse analysis of the biphasic, dynamic behaviour of the two Vibrio cholerae chromosomes. Mol. Microbiol. 60, 1164–1178 (2006). Kruse, T., Møller-Jensen, J., Løbner-Olesen, A. & Gerdes, K. Dysfunctional MreB inhibits chromosome segregation in Escherichia coli. EMBO J. 22, 5283–5292 (2003). Bartosik, A. A. & Jagura-Burdzy, G. Bacterial chromosome segregation. Acta Biochim. Pol. 52, 1–34 (2005). Lin, D. C. & Grossman, A. D. Identification and characterization of a bacterial chromosome partitioning site. Cell 92, 675–685 (1998). Breier, A. M. & Grossman, A. D. Whole-genome analysis of the chromosome partitioning and sporulation protein Spo0J (ParB) reveals spreading and origin-distal sites on the Bacillus subtilis chromosome. Mol. Microbiol. 64, 703–718 (2007). Murray, H., Ferreira, H. & Errington, J. The bacterial chromosome segregation protein Spo0J spreads along DNA from parS nucleation sites. Mol. Microbiol. 61, 1352–1361 (2006). Marston, A. L. & Errington, J. Dynamic movement of the ParA-like Soj protein of B. subtilis and its dual role in nucleoid organization and developmental regulation. Mol. Cell 4, 673–682 (1999). Yamaichi, Y. & Niki, H. Active segregation by the Bacillus subtilis partitioning system in Escherichia coli. Proc. Natl Acad. Sci. USA 97, 14656–14661 (2000). Godfrin-Estevenon, A. M., Pasta, F. & Lane, D. The parAB gene products of Pseudomonas putida exhibit partition activity in both P. putida and Escherichia coli. Mol. Microbiol. 43, 39–49 (2002). Wu, L. J. & Errington, J. RacA and the Soj-Spo0J system combine to effect polar chromosome segregation in sporulating Bacillus subtilis. Mol. Microbiol. 49, 1463–1475 (2003). Ben-Yehuda, S. et al. Defining a centromere-like element in Bacillus subtilis by identifying the binding sites for the chromosome-anchoring protein RacA. Mol. Cell 17, 773–782 (2005). Ben-Yehuda, S., Rudner, D. Z. & Losick, R. RacA, a bacterial protein that anchors chromosomes to the cell poles. Science 299, 532–536 (2003). Identification of a B. subtilis sporulation-specific protein that serves to attach the chromosomal origin region to the cell pole in the incipient forespore compartment. Lee, P. S. & Grossman, A. D. The chromosome partitioning proteins Soj (ParA) and Spo0J (ParB) contribute to accurate chromosome partitioning, separation of replicated sister origins, and regulation of replication initiation in Bacillus subtilis. Mol. Microbiol. 60, 853–869 (2006). Fogel, M. A. & Waldor, M. K. Distinct segregation dynamics of the two Vibrio cholerae chromosomes. Mol. Microbiol. 55, 125–136 (2005).
94. Fogel, M. A. & Waldor, M. K. A dynamic, mitotic-like mechanism for bacterial chromosome segregation. Genes Dev. 20, 3269–3282 (2006). This paper proves the involvement of ParAB in the segregation and polar attachment of the chromosomal origin regions in V. cholerae and provides evidence for a partitioning mechanism that is based on a pulling force that is generated by dynamic ParA filaments. 95. Löwe, J. & Amos, L. A. Crystal structure of the bacterial cell-division protein FtsZ. Nature 391, 203–206 (1998). A crystallographic analysis that shows structural similarity between FtsZ and tubulin. 96. Lutkenhaus, J. Assembly dynamics of the bacterial MinCDE system and spatial regulation of the Z ring. Annu. Rev. Biochem. 76, 539–562 (2007). 97. Weiss, D. S. Bacterial cell division and the septal ring. Mol. Microbiol. 54, 588–597 (2004). 98. de Boer, P. A., Crossley, R. E. & Rothfield, L. I. A division inhibitor and a topological specificity factor coded for by the minicell locus determine proper placement of the division septum in E. coli. Cell 56, 641–649 (1989). This work defines the function of the Min system in division-site placement. 99. de Boer, P. A., Crossley, R. E., Hand, A. R. & Rothfield, L. I. The MinD protein is a membrane ATPase required for the correct placement of the Escherichia coli division site. EMBO J. 10, 4371–4380 (1991). 100. de Boer, P. A., Crossley, R. E. & Rothfield, L. I. Roles of MinC and MinD in the site-specific septation block mediated by the MinCDE system of Escherichia coli. J. Bacteriol. 174, 63–70 (1992). 101. Raskin, D. M. & de Boer, P. A. The MinE ring: an FtsZindependent cell structure required for selection of the correct division site in E. coli. Cell 91, 685–694 (1997). 102. Kruse, K. A dynamic model for determining the middle of Escherichia coli. Biophys. J. 82, 618–627 (2002). 103. Raskin, D. M. & de Boer, P. A. Rapid pole‑to‑pole oscillation of a protein required for directing division to the middle of Escherichia coli. Proc. Natl Acad. Sci. USA 96, 4971–4976 (1999). Discovery of the oscillatory behaviour of the Min system. 104. Raskin, D. M. & de Boer, P. A. MinDE-dependent pole‑to‑pole oscillation of division inhibitor MinC in Escherichia coli. J. Bacteriol. 181, 6419–6424 (1999). 105. Hu, Z. & Lutkenhaus, J. Topological regulation of cell division in Escherichia coli involves rapid pole to pole oscillation of the division inhibitor MinC under the control of MinD and MinE. Mol. Microbiol. 34, 82–90 (1999). 106. Hale, C. A., Meinhardt, H. & de Boer, P. A. Dynamic localization cycle of the cell division regulator MinE in Escherichia coli. EMBO J. 20, 1563–1572 (2001). 107. Marston, A. L. & Errington, J. Selection of the midcell division site in Bacillus subtilis through MinDdependent polar localization and activation of MinC. Mol. Microbiol. 33, 84–96 (1999). 108. Marston, A. L., Thomaides, H. B., Edwards, D. H., Sharpe, M. E. & Errington, J. Polar localization of the MinD protein of Bacillus subtilis and its role in selection of the mid-cell division site. Genes Dev. 12, 3419–3430 (1998). 109. Rothfield, L., Taghbalout, A. & Shih, Y. L. Spatial control of bacterial division-site placement. Nature Rev. Microbiol. 3, 959–968 (2005). 110. Yu, X. C. & Margolin, W. FtsZ ring clusters in min and partition mutants: role of both the Min system and the nucleoid in regulating FtsZ ring localization. Mol. Microbiol. 32, 315–326 (1999). 111. Wu, L. J. & Errington, J. Coordination of cell division and chromosome segregation by a nucleoid occlusion protein in Bacillus subtilis. Cell 117, 915–925 (2004). Identification of the nucleoid occlusion protein Noc in B. subtilis. 112. Bernhardt, T. G. & de Boer, P. A. SlmA, a nucleoidassociated, FtsZ binding protein required for blocking septal ring assembly over chromosomes in E. coli. Mol. Cell 18, 555–564 (2005). Discovery of the E. coli nucleoid occlusion protein SlmA. 113. Thanbichler, M. & Shapiro, L. MipZ, a spatial regulator coordinating chromosome segregation with cell division in Caulobacter. Cell 126, 147–162 (2006). Identification and functional analysis of a novel
40 | january 2008 | volume 6
spatial regulator that couples bipolar positioning of the newly synthesized origin regions to formation of the FtsZ ring at mid-cell in C. crescentus. 114. Figge, R. M., Easter, J. & Gober, J. W. Productive interaction between the chromosome partitioning proteins, ParA and ParB, is required for the progression of the cell cycle in Caulobacter crescentus. Mol. Microbiol. 47, 1225–1237 (2003). 115. Mohl, D. A. & Gober, J. W. Cell cycle-dependent polar localization of chromosome partitioning proteins in Caulobacter crescentus. Cell 88, 675–684 (1997). 116. Laub, M. T., McAdams, H. H., Feldblyum, T., Fraser, C. M. & Shapiro, L. Global analysis of the genetic network controlling a bacterial cell cycle. Science 290, 2144–2148 (2000). 117. Huitema, E., Pritchard, S., Matteson, D., Radhakrishnan, S. K. & Viollier, P. H. Bacterial birth scar proteins mark future flagellum assembly site. Cell 124, 1025–1037 (2006). 118. Lam, H., Schofield, W. B. & Jacobs-Wagner, C. A landmark protein essential for establishing and perpetuating the polarity of a bacterial cell. Cell 124, 1011–1023 (2006). Together with reference 117, this paper identifies a polar localization factor that passes on positional information from the mother to the daughter cell. 119. Huang, K. C., Mukhopadhyay, R. & Wingreen, N. S. A curvature-mediated mechanism for localization of lipids to bacterial poles. PLoS Comput. Biol. 2, e151 (2006). 120. Mileykovskaya, E. & Dowhan, W. Visualization of phospholipid domains in Escherichia coli by using the cardiolipin-specific fluorescent dye 10‑N‑nonyl acridine orange. J. Bacteriol. 182, 1172–1175 (2000). 121. Kawai, F. et al. Cardiolipin domains in Bacillus subtilis Marburg membranes. J. Bacteriol. 186, 1475–1483 (2004). 122. Romantsov, T. et al. Cardiolipin promotes polar localization of osmosensory transporter ProP in Escherichia coli. Mol. Microbiol. 64, 1455–1465 (2007). 123. Komeili, A., Li, Z., Newman, D. K. & Jensen, G. J. Magnetosomes are cell membrane invaginations organized by the actin-like protein MamK. Science 311, 242–245 (2006). Analysis of the role of the actin-like protein MamK in the alignment of magnetosome vesicles. 124. Bazylinski, D. A. & Frankel, R. B. Magnetosome formation in prokaryotes. Nature Rev. Microbiol. 2, 217–230 (2004). 125. Scheffel, A. et al. An acidic protein aligns magnetosomes along a filamentous structure in magnetotactic bacteria. Nature 440, 110–114 (2006). Identification of a role for the membrane protein MamJ in the attachment of magnetosomes to the MamK filament. 126. Dunin-Borkowski, R. E. et al. Magnetic microstructure of magnetotactic bacteria by electron holography. Science 282, 1868–1870 (1998). 127. Pradel, N., Santini, C. L., Bernadac, A., Fukumori, Y. & Wu, L. F. Biogenesis of actin-like bacterial cytoskeletal filaments destined for positioning prokaryotic magnetic organelles. Proc. Natl Acad. Sci. USA 103, 17485–17489 (2006). 128. Cabeen, M. T. & Jacobs-Wagner, C. Bacterial cell shape. Nature Rev. Microbiol. 3, 601–610 (2005).
Acknowledgements
This work was supported by funding from the Max Planck Society to M.T. and grants from the National Institutes of Health to L.S.
DATABASES Entrez Genome Project: http://www.ncbi.nlm.nih.gov/ entrez/query.fcgi?db=genomeprj Bacillus subtilis | Caulobacter crescentus | Escherichia coli Magnetospirillum gryphiswaldense | Magnetospirillum magneticum | Rhodobacter sphaeroides | Shigella flexneri | Thermotoga maritima | Vibrio cholerae Entrez Protein: http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=protein FtsY | IcsA | MreC | SpoIVFB | SpoIIQ | SpoIIIAH | TipN
FURTHER INFORMATION Martin Thanbichler’s homepage: http://www.mpi-marburg. mpg.de/thanbichler/ All links are active in the online pdf
www.nature.com/reviews/micro © 2008 Nature Publishing Group
REVIEWS
Drugs versus bugs: in pursuit of the persistent predator Mycobacterium tuberculosis James C. Sacchettini*, Eric J. Rubin‡ and Joel S. Freundlich*
Abstract | Tuberculosis (TB) claims a life every 10 seconds and global mortality rates are increasing despite the use of chemotherapy. But why have we not progressed towards the eradication of the disease? There is no simple answer, although apathy, politics, poverty and our inability to fight the chronic infection have all contributed. Drug resistance and HIV-1 are also greatly influencing the current TB battle plans, as our understanding of their complicity grows. In this Review, recent efforts to fight TB will be described, specifically focusing on how drug discovery could combat the resistance and persistence that make TB worthy of the moniker ‘The Great White Plague’.
*Department of Biochemistry and Biophysics, Texas A&M University, College Station, Texas 77843, USA. ‡ Department of Immunology and Infectious Diseases, Harvard School of Public Health, Boston, Massachusetts 02115, USA. Correspondence to J.C.S. e‑mail:
[email protected] doi:10.1038/nrmicro1816
Many think of tuberculosis (TB) as the scourge that devastated Europe in the seventeenth century and rapidly became a leading cause of death worldwide before being virtually eliminated (BOX 1). TB has been brought back to our thoughts, however, by the recent reports of outbreaks of drug-resistant disease. In fact, TB never ‘went away’, but has remained the ‘Captain of Death’ throughout much of Asia and Africa. TB mortality rates are once again on the rise — this has been attributed to the HIV-1 epidemic, which has produced a new and highly susceptible population, and the inconsistent use of antibiotics, which has led to a new epidemic of drug-resistant disease in many parts of the world. Given that investment in antibiotics in general, including antitubercular drugs, has waned over recent years, we have found ourselves unable to respond to the resurgence of TB. Two factors, persistence (BOX 2) and resistance, have made the treatment of the causative organism, Mycobacterium tuberculosis, difficult. The term persistence describes the survival of M. tuberculosis despite the use of antibiotics (rather than latency, which refers to the ability of apparently dormant bacteria in asymptomatic infected individuals to activate as much as decades after the initial infection). Little concrete information is available on the cellular or metabolic status of persistent mycobacteria. As a consequence of persistence, drug treatment is extended, and current antibiotics (BOX 3) require long courses of treatment to cure patients and prevent relapse. Currently, even the most effective regimens require a combination of at least 3 drugs and last
nature reviews | microbiology
for 6 months. As patients feel better within 1–2 weeks, they have little motivation to complete therapy. Thus, current World Health Organization guidelines call for treatment to be directly observed. The required infrastructure for drug-delivery and treatment supervision can be difficult to provide in much of the world, particularly in areas that are afflicted with poverty and unstable governments. For TB, drug resistance is due to genetic mutations that result in a heritable loss of susceptibility to antibiotics. These mutations generally occur either in the target or the activator of the drug. TABLEs 1,2 summarize the most common mutations for current firstand second-line drugs, and refer to their respective mechanism (or mechanisms) of action and half-lives. Slow-acting drugs, combined with poor health-care systems, have led to incomplete treatment, relapse and the emergence of resistant mycobacteria. Strains of M. tuberculosis that are resistant to at least one of the first-line antibiotics (TABLE 1) have become common, and strains that are resistant to two or more are not uncommon. Although resistance to single agents generally still permits cure (albeit by using even more extended courses of therapy), strains that are resistant to multiple agents cause disease that is far more difficult and costly to treat. This is particularly true for multidrug-resistant (MDR) strains that are resistant to the first-line drugs isoniazid (INH) and rifampicin. Some strains carry far greater levels of drug resistance, including extensively drug-resistant (XDR) bacteria, which are MDR and also resistant to fluoroquinolones volume 6 | january 2008 | 41
© 2008 Nature Publishing Group
REVIEWS Box 1 | Waksman’s vision In his speech at the Nobel Banquet in December 1952, Selman Waksman said “The Great White Plague, which only 10 years ago was thought to be immune to drug therapy, is gradually being eliminated…streptomycin pointed a way. Later supplemented with PAS and more recently with isoniazid, it has brought the control of this disease within sight.” In 1943, Selman Waksman, a microbiologist at Rutgers University, purified a compound from the soil bacterium Streptomyces griseus that could kill a wide spectrum of bacteria, including Mycobacterium tuberculosis, in culture and in animals125,126. Unlike the discovery of penicillin 15 years earlier, Waksman’s discovery of streptomycin was not accidental. His student, Albert Schatz, conducted a directed screen of 10,000 cultures of soil bacteria to identify those that inhibited the growth of co-cultured Gram-negative bacteria. He found only ten cultures that could significantly block the growth of the test bacteria. The Nobel-Prize-winning discovery was heralded by most as the ‘beginning of the end for tuberculosis (TB)’, although the prize itself has been controversial as some have argued that it should have been shared with his student, or with Jorgen Lehmen, who discovered the TB drug para-aminosalicylic acid (PAS) in the same year as the discovery of streptomycin127. The drug was quickly approved by the United States Food and Drug Administration, and within 2 years Merck was producing 25,000 kg per day of streptomycin. Although the discovery of streptomycin proved that a bacterium was the cause of the disease and led to almost miraculous responses in patients infected with TB, it required repeated injection and was associated with significant toxicity. Worse, strains of M. tuberculosis that were resistant to the drug were discovered within a few months of use.
and at least one injectable antibiotic. These infections are extraordinarily difficult to treat using the current agents. Clearly, our current drug armamentarium (TABLES 1,2) has not been sufficient to control the TB epidemic (BOX 4) . New antibiotics, particularly those that are derived from new chemical classes, are more likely to have activity against many drug-resistant strains. The path to creating antibiotics that act against persistent organisms and produce more rapid clearance of infections is less clear. Certainly, a vital part of drug development is the understanding of the physiology of growing and persistent organisms, an area in which little information is available, as reviewed elsewhere1–3. A greater knowledge of persistence could lead to a directed strategy for the development of more rapidly effective antibiotics. However, until a clear picture of persistence and its role in infection is achieved, researchers will have to rely on more empirical approaches, as demonstrated by the discovery of TMC207 (discussed below). Here, we review recent and ongoing efforts to produce new antitubercular drugs, and the properties of current investigational agents. This Review seeks to complement other recent discussions of TB research, such as those by Janin4, Ginsberg and Spigelman5, and Williams and Duncan6. We will follow a ‘drug timeline’, beginning with a discussion of discovery technologies that were designed to provide new clinical antitubercular candidates, before considering compounds that are currently in trials and, finally, already approved drugs that are sought to be used as TB therapies.
Drug-discovery methods Genetic approaches to target identification. Ideal TB drug targets should have three characteristics: they should be required for bacterial growth and persistence (that is, they must be expressed and essential during the time that treatment occurs); it should be possible to inhibit their activity using small molecules (that is, the target should be ‘druggable’)7,8; and they should be accessible to these modulatory compounds. Theoretically, the simplest way to find targets that have these ideal characteristics is to discover an active compound and then 42 | january 2008 | volume 6
define its target. In practice, however, this has not been so simple, and we still do not know the targets of many existing antitubercular drugs that are in clinical use. Recently, two systematic approaches have emerged to define the targets of compounds that have activity against M. tuberculosis. Expression analysis provides a method to profile cellular responses to perturbations, such as small molecules or environmental stress. Several groups have collected a large number of datasets from DNA microarray experiments that were performed under various conditions. In particular, Boshoff and colleagues9 have evaluated bacterial gene expression in response to a number of drugs, toxins and environmental conditions. Although these results do not define a single molecular target, they can be used to identify susceptible pathways that may contain drug targets. The recent availability of affordable whole-genome sequencing also potentially provides a rapid way of finding targets. For example, Andries and colleagues10 selected for mutants that were resistant to the drug TMC207 (discussed below) and sequenced their entire chromosomes. All resilient strains contained mutations in a single gene that encoded a subunit of ATP synthase. However, this success story is not always easily replicated. Box 2 | Bacterial persistence and treatment failure Why does antibiotic treatment fail even when bacteria are not genetically drug resistant? This phenomenon, often termed bacterial ‘persistence’, might be explained in a number of ways, given that Mycobacterium tuberculosis induces a chronic inflammatory response in which bacteria are sequestered from drugs in tissue. The local concentration of antibiotics in lesions, such as granulomas, might not be adequate to cause bacterial death, or some bacteria might adopt a physiological state that renders them less susceptible to antibiotics. This could be a stochastic process, in which a subpopulation of cells adopt a physiological state that renders them drug insensitive. Alternatively, an environmental condition, such as low oxygen or carbon starvation, might induce the persistent state. All of these hypotheses could be true. However, as yet, we do not know if any of them have an important role in treatment failure.
www.nature.com/reviews/micro © 2008 Nature Publishing Group
REVIEWS Box 3 | Lessons from current antitubercular drugs There is an excellent chance that patients who have tuberculosis (TB) can be cured using currently available drugs if they are able to complete the required course of therapy. But what characteristics should new drugs have to improve on current treatment?
Oral bioavailability • Streptomycin is a highly potent drug but is rarely used, partly because it must be injected. Good tolerance • Para-aminosalicylic acid (PAS) was largely abandoned early on because of gastrointestinal intolerance. Usability in multiple populations • Fluoroquinolones are not recommended for the treatment of pregnant women and young children — two populations that are at risk for disease. • Thiacetazone is associated with life-threatening reactions in patients infected with HIV-1. Compatibility with antiretrovirals • Rifampicin alters the metabolism of multiple drugs, particularly protease inhibitors. Infrequent dosing • Prospective antibiotics, such as linezolid, might not be useful if they require more than a single dose each day. Activity against drug-resistant strains • Drugs such as PAS have been revived owing to the increase in multidrug-resistant and extensively drug-resistant disease. Activity that is not necessarily in vitro • Pyrazinamide is a highly effective drug for patients even though it has poor activity under standard laboratory conditions. Rapid clearance of chronic infection • All available drugs, with the exception, possibly, of rifampicin, have limited efficacy in chronic infection. This is particularly true of agents such as isoniazid that act on the cell wall. Affordability • Drugs that are used for other purposes, such as linezolid, are extraordinarily expensive. At current prices, it would be impossible for them to be used in most areas of the world in which TB is prevalent.
In the case of PA‑824, sequencing the genome of resistant mutants using a DNA microarray method showed that mutations in a conserved hypothetical gene produced resistance11. However, the encoded protein was not the target of PA‑824, but instead was a nitroreductase that was responsible for the activation of PA‑824. Genetic analysis might prove to be an alternative method for defining new TB drug targets. Using a range of methods, investigators have found several gene products that, if inhibited, could decrease bacterial growth or increase host survival after infection12–14. Genes that are required for growth in vitro are difficult to define using traditional genetic methods, as mutations in these genes, by definition, result in clones that are unable to grow. However, negative screens that identify genes that cannot be mutated have yielded a set of several hundred candidate genes that are required for in vitro growth14,15. Of course, these are simply screens and do not represent proof that individual gene products are useful as drug targets. The extension of these screens to in vivo animal models of TB would be a giant step forwards in target identification. However, to validate putative targets in the absence of a known inhibitor it is useful to construct conditional mutants. Fortunately, recent work has provided conditional promoters that can be used to construct these informative strains16–18. Is it truly important to define single targets for potential antibiotics? After all, it might be difficult to evolve resistance to drugs that hit multiple targets (so-called nature reviews | microbiology
dirty drugs) and, in fact, there is evidence that some current drugs, such as INH19,20 and para-aminosalicyclic acid21, are capable of inhibiting multiple proteins. However, even in these situations, it is unclear if this provides a strong advantage as resistance can still arise from mutations that seem to affect only single putative targets22,23, perhaps because these represent the points of greatest vulnerability. In any case, for drugs that hit either single or multiple targets, it might be difficult to identify targets using systematic approaches. High-throughput screening. High-throughput screening against target proteins has an important role in modern drug discovery24,25. For M. tuberculosis, biochemical high-throughput screening has contributed to the finding of two clinical drug candidates, TMC207 (Ref. 10) and SQ109 (Ref. 26), and has been used to investigate the viability of small molecules as modulators of a number of mycobacterial targets. Two different approaches — the targeting of wholecell growth and enzyme inhibition — have been used individually and together to apply high-throughput screening to TB drug-discovery efforts. Whole-cell Box 4 | Gates’ vision Bill Gates commented at the 2005 World Health Assembly: “Today, we have tuberculosis drugs you have to take for 9 months. Why can’t we find one that works in 3 days?”
volume 6 | january 2008 | 43 © 2008 Nature Publishing Group
REVIEWS Table 1 | Mechanism of action, resistance and half-life of current first-line antituberculosis agents Antibiotic
Chemical structure
Isoniazid
O
H N
NH2
N
Rifampicin
HO
O O
Mechanism and target
Mutations associated with resistance
Half-life in humans (hours)
Refs
Inhibits mycolic acid synthesis; primary target is InhA and secondary targets are KasA and DfrA
katG (required for drug activation); inhA (promoter mutations); and others
1–3
74
Inhibits transcription; RNA polymerase β-subunit
rpoB
2–3
74
Inhibits arabinogalactan synthesis; possibly EmbB
embB
3–4
74
Unknown (possibly inhibits FAS-I or alters membrane energetics)
pncA (required for drug activation)
10
74
Inhibits protein synthesis; 30S ribosomal subunit
rpsL and rrs
2–3
74
OH O OH OH NH
O H3C
N
O O
Ethambutol
N
OH
O
N
OH H N
N H
CH3
HO
Pyrazinamide
O N
NH2
N
Streptomycin
NH2
H2N N
O O
HO H2N
O
N NH2
Pharmacokinetic profile A quantitative description of the fate of a drug from the moment the treated subject is dosed with the compound to the moment when it (and/or its derivatives) is expelled from the subject.
CHO
OH OH
OH NHCH3 OH O
OH OH
screening against either Mycobacterium smegmatis or M. tuberculosis allows searching for the ultimate goal — potent growth inhibition or killing. As this type of screen is not target based, there is a considerable risk of finding compounds that have generalized toxicity, and the lack of information on the target will certainly complicate the optimization. This risk might be decreased by pre-filtering compound libraries27,28 or designing more targeted screens. High-throughput screening is being pursued at a number of facilities, both in industry and academia. One large National Institutes of Health-funded effort, the Tuberculosis Antimicrobial Acquisition and Coordinating Facility (see Further information), offers a free service to investigators for testing candidate compounds. The facility has evaluated over 79,000 compounds from more than 9,600 researchers for the inhibition of M. tuberculosis growth29. Of these compounds, 130 have demonstrated in vitro efficacy against both drug-sensitive and drug-resistant strains. Such broad screening efforts hold promise for uncovering new leads for antituberculars that, from the outset, display mycobacterial growth inhibition. High-throughput screening has also been used to identify inhibitors of a targeted enzyme in a cell-free environment, through either a binding or functional assay. High-throughput screening has the potential to identify hits that could be potent inhibitors of mycobacterial growth if the compounds have acceptable
44 | january 2008 | volume 6
pharmacokinetic profiles and the target is truly essential.
There are many examples of small molecules that have excellent enzyme inhibition but poor whole-cell potency, possibly owing to their failure to permeate the mycobacterial cell wall. A variant of this strategy relies on a functional assay in which the inhibition of an entire pathway can be screened for. For example, a luciferase reporter can be used to measure the transcription of the iniBAC operon, which is induced by a diverse set of mycobacterial cell-wall biosynthesis inhibitors, including ethambutol and INH30. Barry and colleagues26 used this technology to discover SQ109 (FIG. 1), an ethambutol analogue for which the discovery and clinical progress will be discussed below. GlaxoSmithKline and the Global Alliance for TB Drug Development (TB Alliance; see Further information) have recently conducted a million-compound screen for new inhibitors of an M. tuberculosis enoyl-acyl carrier protein reductase, InhA, which is the target of INH. They found a high hit rate, probably because the crystallographically characterized active site can accommodate hydrophobic groups of varying dimensions, which is consistent with its acceptance of C16–C56 fatty-acid thioester substrates31,32. Importantly, these inhibitors should be active against most of the INH-resistant strains, as they do not require activation by the catalase–peroxidase enzyme KatG33, which is required for the activation of INH. Similar screens carried out by GlaxoSmithKline and the Novartis Institute for Tropical Diseases for inhibitors of isocitrate lyase, an www.nature.com/reviews/micro
© 2008 Nature Publishing Group
REVIEWS Table 2 | Mechanism of action, resistance and half-life of selected second-line antituberculosis agents Antibiotic
Chemical structure
Fluoroquinolones
O
O
F
Mutations associated with resistance
Half-life in humans (hours)
Inhibits DNA gyrase
gyrB
Moxifloxacin: 12; Gatifloxacin: 8
129
Inhibits mycolic acid synthesis; InhA
ethA (required for drug activation) and inhA (promoter mutations)
2
130
Inhibits peptidoglycan synthesis by blocking the synthesis and use of d‑alanine (Ala); Ala racemase and d-Ala-d-Ala ligase
alr (overproduction) and ddl (overproduction)
10
Inhibits folate metabolism; possibly dihydropteroate synthase
thyA
0.75–1
Inhibits protein synthesis; methylated nucleotides in both ribosomal subunits
tlyA and rrs
4–6
Inhibits protein synthesis
rrs
2
132
Inhibits protein synthesis
rrs
3
133
OH
R
N
NH2
S
N
Cycloserine
Refs or sources
O
H3C
Ethionamide
Mechanism and target
O
H2N
NH O
Para‑aminosalicylic acid
NH2
DrugBank (see Further information)
131
OH HO
O
Capreomycin
NH2 R
O NH HN
R = H, OH H N
H2N NH2
HN
O H N
OH HO H2N
HO
Amikacin
HO OH O NH2
O
O
OH NH2
O
OH O
HO H2N HO H2N
Pharmacophore The chemical functional group (or groups) that is present on a molecule and that enables its biological activity.
Chemotype A chemical functional group or classification of a specific array of functional groups.
HO O H N
NH2 N
O
OH
O H2N
H N
HN
O HN
O
H 2N
Kanamycin
O
The Merck Manuals Medical Library (see Further information)
OH HO OH O NH2
O
OH NH2
O
enzyme that is crucial for bacterial persistence during infection34,35, were considerably less successful. Again, this might be explained structurally, as the active site of this enzyme is shallow and highly charged. Small, hydrophilic molecules, such as those that are predicted to bind to isocitrate lyase, are under-represented in most screening libraries, as they are less likely to have desirable pharmacological properties. Finally, a high-throughput screen for inhibitors of pantothenate synthetase, an enzyme that is crucial for the biosynthesis of the essential cofactors acyl carrier protein and coenzyme A, was recently reported by White and colleagues36. Of approximately 4,000 compounds assayed, one lead compound was identified and a preliminary structure characterized in complex with the target enzyme. This successful outcome suggests a path forward for the structure-based design of more potent analogues that inhibit the enzyme and M. tuberculosis. Combining whole-cell and target-based screens might avoid the drawbacks of each. For example, the
nature reviews | microbiology
library that was used to discover SQ109 was screened using both growth-inhibition and cell-wall-biosynthesis assays37. Of nearly 5,000 compounds screened, 25 small molecules were active in both screens, but only one, SQ775, showed significant activity in a mouse model of infection. Structural biology and virtual screening. Describing the proteome of M. tuberculosis has been the focus of much research in the past few years. Primarily owing to the efforts of the TB structural-genomics consortium, more than 260 X‑ray crystal structures of interesting proteins, a large percentage of which were selected on the basis of their being potential drug targets, have been completed, and many are currently being used to facilitate medicinal‑chemistry efforts to rationally design new antibiotics38 (FIG. 1). The availability of these structures provides the opportunity to carry out virtual screening. This volume 6 | january 2008 | 45
© 2008 Nature Publishing Group
REVIEWS O
CH3 OH
Br
H N
NH H
TMC207
O
N
H3C
OH
N
O
HN
N
H3C
Moxifloxacin
O
O
OH O OH O
NO2
O
N
OH O OH OH
O O
O
O
H3C
N
O
HO
N
O O
N
O
N
Rifalazil
N
N
N OH
SQ775
Metronidazole
NH
O
SQ109
Gatifloxacin
HO
NH H3C
N
OH
N
N
HN N N O
Rifapentine
O
N
O
N
O2N
N
N
Sudoterb
O
O
CF3
O
O
H3C
PA-824
N
H3C
O
O
N CH3
N
Cl
O
N
O
S
N
Chlorpromazine
N
CH3
S
Thioridazine
S
S CH3
Trifluoperazine
CF3
O O
N
H N
N S H O O
OH
O
O
Sulbactam OH
Clavulanic acid
O
Linezolid
OH
O
N H
O
N F
O
OPC-67683 N
N
CF3
N
F3C O2N
H N
N H
HO O
O
F
OH
N
O CH3
O
O
F
N CH 3
HO
N
OH
NH HN
S
H H
Imipenem
Figure 1 | Chemical structures of non-approved antituberculars. Nature Reviews | Microbiology
Lipinski’s rules A set of delimited physiochemical properties described by C. A. Lipinski that best fit a studied subset of drugs. In general, compounds that adhere to these guidelines are said to be drug-like.
Shikimate pathway A series of biochemical reactions in plants and microorganisms that are involved in the biosynthesis of aromatic amino acids.
powerful technique has had a beneficial impact on numerous drug‑discovery efforts39–41. Screening using computational methods can be complementary to a biochemical high-throughput screen because of the potential for screening a larger chemical space quickly and inexpensively. Virtual screening can be used in two ways: to identify compounds that are consistent with a pharmacophore model irrespectively of the identity and structure of the pertinent protein target or targets; or to develop inhibitors of a protein based on its known three-dimensional structure. Recent examples of virtual screening that were based on known chemotypes include work from Manetti and colleagues42 and García-García and colleagues43. Manetti et al. used a training set of 471 small molecules that had a range of in vitro activities against M. tuberculosis to construct a model that could be searched using a virtual library of compounds that was filtered to follow Lipinski’s rules44. The virtual hits were assayed against M. tuberculosis, and the two most potent compounds
46 | january 2008 | volume 6
had a minimum inhibitory concentration (MIC) of 25 µg per ml. Using a different mathematical model, García-García et al. tested 5,000 commercially available compounds and found 18 virtual hits43; 5 of these had an MIC90 of less than 50 µM. Other groups have instead focused on specific targets that have known structures. Agrawal and coworkers45 performed a virtual screen for inhibitors of M. tuberculosis chorismate mutase, which is part of the essential shikimate pathway in M. tuberculosis46. They started with known inhibitors of the homologous Saccharomyces cerevisiae enzyme47 and conducted a three-dimensional pharmacophore search of a database that has more than 15,000 members. Of the 15 molecules that scored highest, 4 demonstrated inhibition in an enzyme assay. Lin and colleagues48 focused on the elucidation of new small-molecule inhibitors of AccD5, an acyl-CoA carboxylase essential enzyme that catalyses the transformation of acetyl-CoA and propionyl CoA to the corresponding malonyl-thioester and www.nature.com/reviews/micro
© 2008 Nature Publishing Group
REVIEWS yet to yield a clinical candidate, X‑ray crystallography and structure-based designs have had crucial roles in the discovery of new inhibitors of this validated TB drug target32,50–53. Sullivan and co-workers54 have translated these insights into the discovery of potent triclosanbased antituberculars. FIG. 2 depicts a select subset of X‑ray structures that have had a prominent role in this developing story.
Figure 2 | Overlay of small-molecule inhibitors of InhA. A cross-section through Nature Reviews | Microbiology the surface of the active site of Mycobacterium tuberculosis fatty-acid enoyl-acyl carrier protein reductase (InhA), coloured by atom type (carbon, grey; nitrogen, blue; oxygen, red; and sulphur, yellow). Superimposed are the bound conformations of NADH50 (carbon, yellow; nitrogen, blue; oxygen, red; and phosphorous, aqua), C16 fatty acyl substrate analogue trans‑2-hexadecenoyl-(N-acetylcysteamine) thioester32, shown in its bent conformation (carbon, yellow; sulphur, orange; nitrogen, blue; and oxygen, red) and 12 inhibitors (carbon, grey; nitrogen, blue; oxygen, red; sulphur, yellow; and fluorine, purple) that have half-maximal inhibitory concentration values ranging from 50 nM to 5 µM. Unlike isoniazid and ethionamide, these are non-covalent reversible inhibitors that form ternary complexes with enzymes and NAD+. This figure was created using the molecular modelling system Chimera128.
Lead optimization The process by which a promising small-molecule entity is structurally modified to obtain drug-like pharmacokinetic, pharmacodynamic and safety profiles.
Efflux pump An active transport system for the removal of toxic molecules, such as antibiotics, from cells.
methylmalonyl-thioester, respectively. They screened over 4 million compounds for binding to either the predicted biotin or propionyl-CoA binding pockets and found that 1 of the 9 compounds that scored highest had an enzyme half-maximal inhibitory concentration (IC50) of 10 µM. Approaches such as these underscore the usefulness and potential of virtual screening to enable new hits for drug-discovery efforts. Each of the discovery technologies discussed in this section has potential as a starting point for the discovery of a new antitubercular that, on successful passage through the pre-clinical drug-development pathway, can produce a clinical candidate. Undoubtedly, the most efficient and expeditious way to seed drug discovery lies in the cooperative union of these methodologies in moving from the validated drug target to a clinical candidate. This has been highlighted by examining retrospectively a GlaxoSmithKline antibacterial high-throughput screening campaign, which, over a wide range of targets, provided few compounds for follow up49. Additionally, lead optimization is often a labour- and time-intensive process that many programmes do not endure. A detailed understanding of the drug target (or targets) and how the binding of a drug inhibits target function helps immensely in developing a lead. Whereas efforts to develop new InhA inhibitors, which, in the future, could complement and/or supplant INH, have
nature reviews | microbiology
New TB drugs in clinical trials A chemical entity that has been discovered as an antitubercular by the application of one or more of the methods discussed above and has cleared the considerable hurdles in pre-clinical development can enter clinical trials if it has the approval of the pertinent governmental body (in the United States, the Food and Drug Administration (FDA)). The scientific community is hopeful that the drug candidates described below (the structures of which are shown in FIG. 1) will eventually be used in new treatment regimens that meet the goals discussed earlier. It is clearly a testament to the financial commitment of the involved funding organizations that these candidates are being supported through a process that is incredibly costly. Fluoroquinolones. Fluoroquinolones, which have been used for the treatment of TB since the 1980s55, currently constitute the second line of defence against M. tuberculosis and play a key part in the treatment of MDR disease. Resistance to fluoroquinolones has been attributed to mutations in the genes that encode gyrase (gyrA and gyrB)56,57. As quinolone efflux pumps may also have a role in resistance, it is noteworthy that the annotation of putative pumps in the M. tuberculosis genome58 and the studies of Jacobs and co-workers59 on a probable efflux pump, lfrA. Two approved fluoroquinolones, moxifloxacin and gatifloxacin, have shown promising results both for treating resistant disease and, possibly, shortening the course of therapy. Phase II and III clinical trials that are currently underway are testing the efficacy of moxifloxacin and gatifloxacin as replacements for either INH or ethambutol in first-line therapy60,61. These compounds are particularly attractive, as they have already been approved for use in various other infections and are known to be safe. As fluoroquinolones are prescribed for numerous respiratory infections, many cases of TB could be treated with this monotherapy before they are diagnosed. Theoretically, this could lead to drug resistance, although, as yet, widespread fluoroquinolone-resistance mutations have not been observed. Rifampicin analogues. Rifapentine was approved by the FDA for the treatment of M. tuberculosis infection in June 1998. The piperazinyl hydrazone functionalized rifampicin analogue, initially named DL 473, was first reported in 1975 (Ref. 62) and subsequently has been shown to exhibit in vitro antimycobacterial efficacy that is, in most cases, superior to that of its parent63–65. A noteworthy advantage that rifapentine has compared with rifampicin is its longer serum halflife66, which has encouraged its examination in clinical volume 6 | january 2008 | 47
© 2008 Nature Publishing Group
REVIEWS settings to determine its potential for positively altering the frequency of the antitubercular treatment regimen. An extension of a Phase III trial that is currently evaluating rifapentine and INH in the treatment of latent TB involves recruiting children (C. Dukes Hamilton, personal communication) to assess the efficacy, pharmacokinetic and safety profiles of the drug combination.
Structure–activity relationship (SAR). The relationship between the chemical structure of a compound and its biological or pharmacological activity. This type of relationship can be assessed by considering a series of molecules, each with a slightly different structure, and then noting the effect on the biological activity that is associated with each structural variation.
Fast-track status The FDA status that is reserved for products that demonstrate the potential to treat a serious or life-threatening condition.
F0 subunit of atp synthase
The transmembrane portion of the enzyme complex that is involved in the biosynthesis of ATP, which has a role in the passage of protons through the membrane.
Ames mutagenicity test A sensitive biological method for measuring the mutagenic potency of chemical substances.
Rifalazil. Rifalazil, formerly known as KRM‑1648, has a benzoxazinorifamycin structure67. This heterocyclic modification of ansamycin68 is much more potent against M. tuberculosis clinical isolates than rifampicin69 and is more efficacious in a mouse model70. Rifalazil presumably has a similar mechanism of action to rifampicin68, as mutants are cross-resistant69. Rifalazil has a longer halflife than rifampicin in healthy human volunteers71 and, unlike rifampicin and rifabutin, is neither metabolized by, nor an inducer of, rat and dog hepatic cytochrome P450 enzymes72. These metabolic observations suggest that rifalazil can be co-administered with drugs that are sensitive to oxidative metabolism, thereby reducing the potential for adverse drug–drug interactions. This is a particularly important consideration, as co-infection with HIV-1 and TB is common and rifampicin considerably lowers the serum levels of antiretroviral protease inhibitors73. Rifalazil, however, is not currently registered with the FDA for a clinical trial of TB, probably owing to the toxicity that was observed during a Phase II trial in Brazil74. ActivBiotics is currently investigating significantly lower once-weekly doses of rifalazil in Phase II/ III studies to ascertain the effect on patients that have carotid atherosclerotic disease and have been infected with Chlamydia pneumoniae (A. Sternlicht, personal communication). It remains to be seen whether these lower doses, if safe, would be efficacious for TB. SQ109. SQ109 was discovered using a high-throughput screen of ethambutol analogues that had the dual goals of inhibiting mycobacterial cell-wall synthesis and cell growth in general26. Barry and colleagues26 used the structure–activity relationship (SAR) study results reported by Lederle Laboratories75,76 to prepare more than 63,000 compounds. The compounds were assayed for the inhibition of M. tuberculosis growth and cell-wall biosynthesis. They found 119 hits in the cell-wall-biosynthesis assay and 60 hits in the growth assay that were at least as potent as ethambutol. SQ109, the optimal hit, shares some structural similarity with ethambutol, but has superior in vitro and in vivo activity77. In fact, DNA microarray analysis suggests that this compound has a different mode of action from ethambutol9. SQ109 has achieved fast-track status and is in Phase I clinical trials. TMC207. Originally named R207910 by Johnson and Johnson Pharmaceutical Research and Development10, TMC207 is a new antitubercular agent from the diarylquinoline (DARQ) class. The drug was found by a high-throughput screen of a library of approximately 10,000 compounds (K. Andries, personal communication), during a search for those that inhibited the growth of the rapidly growing environmental bacterium
48 | january 2008 | volume 6
M. smegmatis. It is intriguing to note that the first member of this DARQ class was isolated as a side product in chemical experimentation that was designed to prepare compounds for other early discovery programmes, thus highlighting the importance of screening (targeted compounds and side products) across projects. TMC207 achieves impressive in vitro and in vivo efficacy against drug-sensitive and drug-resistant strains of M. tuberculosis. Resistance studies identified the biological target as the F0 subunit of ATP synthase that is encoded by atpE10,78,79. This finding, together with the lack of crossresistance of TMC207 with existing antitubercular drugs, highlights the potential that undirected screens have for the discovery of compounds that have new mechanisms of action. A Phase II trial is currently underway to investigate the efficacy of TMC207-containing regimens versus standard antitubercular therapy in patients who have MDR TB. Given the previously reported animal studies80, TMC207 could be added to a second-line treatment regimen or could replace a first-line drug to shorten the length of treatment. Sudoterb. The tetra-substituted pyrrole Sudoterb was discovered by building on the previously described antimycobacterial SAR of pyrroles81. The N‑substituent is similar to INH, as both contain an isonicotinoyl hydrazide moiety. Because the only publicly available information about this compound comes from the patent application82, we know little about its mechanism of action. However, the absence of cross-resistance with existing therapies and the unique chemical structure suggests that it could be novel. Sudoterb has been reported to be more potent than INH as a monotherapy in a mouse model for TB and displays an acceptable pharmacological profile in mice and dogs83. The compound is reportedly in Phase I clinical trials83. Nitroimidazoles. Metronidazole has frequently been used to treat infections of microaerophilic or anaerobic bacteria84. Owing to the proposed relevance of limiting oxygen conditions to the latent phase of TB85, metronidazole has also been investigated as an antitubercular. Metronidazole kills only dormant M. tuberculosis and not actively growing cultures86. The drug had no effect on the growth of M. tuberculosis-infected mouse macrophages, but a measurable, although small, efficacy in a mouse model of chronic-stage M. tuberculosis87. Despite these mixed results, a Phase II clinical trial in South Korea is currently recruiting patients to examine the effect of adding metronidazole to a standard second-line therapeutic regimen (see ClinicalTrials.gov in Further information for details of an ongoing clinical trial). Attempts to discover other nitroimidazole-based antituberculars led researchers at Ciba-Geigy to discover CG‑17341 — a nitroimidazooxazole that displays potent activity both in vitro and in vivo against drug-sensitive and drug-resistant strains88,89. A toxicity issue noted in an Ames mutagenicity test hindered this compound from being developed further. Stover and colleagues90 removed this concern, however, with the discovery of the potent antitubercular PA‑824 of the nitroimidazopyran www.nature.com/reviews/micro
© 2008 Nature Publishing Group
REVIEWS class. In vivo studies demonstrated comparable activity between PA‑824 and INH91. Most importantly, PA‑824 displays no cross-resistance with existing TB drugs and is efficacious against non-growing M. tuberculosis under hypoxic conditions. Intriguingly, this activity against persistent mycobacteria is not shared with the original lead compound CG‑17341. However, in vivo tests using a mouse model have failed to demonstrate an advantage in including PA‑824 with first-line antituberculars in terms of shortening the duration of treatment 92. Mechanistically, the mycobacterial target (or targets) of PA‑824 is unknown. Biochemical evidence suggests that both protein and lipid synthesis are inhibited by a compound (or compounds) that is generated from the bioreduction of the imidazole nitro group84,90,93. Fattyacid analysis of treated mycobacteria suggested the inhibition of hydroxymycolate oxidation to ketomycolate90. Further mechanistic work is required to understand the mechanism of action of PA‑824. Overall, the likelihood of a novel mechanism of action and activity against persistent mycobacteria bodes well for the positive impact of PA‑824 on antitubercular chemotherapy. This compound is in Phase I clinical trials that are supported by the TB Alliance. OPC‑67683. Building on the work of Ciba-Geigy in nitroimidazoles and the discovery of PA‑824, researchers at the Otsuka Pharmaceutical Company in Japan examined the SAR around PA‑824 and proposed that further variation of the furan portion of the heterocycle could yield potent antituberculars that are free of mutagenicity93,94. Whereas PA‑824 features a pendant 4‑trifluoromethoxybenzyloxy group, the Otsuka researchers introduced a 4‑piperidine moiety in place of the trifluoromethoxy group to improve oral bioavailability. OPC‑67683 was eventually prepared in an effort to decorate the piperidine 4‑position with hydrophobic groups such as 4‑trifluoromethoxyphenyl. OPC‑67683 is free of the mutagenicity of its nitroimidazole progenitor, is orally bioavailable in mice and provides efficacy that is equivalent to rifampicin against M. tuberculosis Kurono at less than one-fifteenth of the dose93. OPC‑67683 was equally efficacious against cultures of drug-sensitive and drug-resistant strains of M. tuberculosis and superior to INH, ethambutol, rifampicin, streptomycin, CGI‑17341 and PA‑824, and also produced rapid eradication of infection in mice. OPC‑67683 was not metabolized by a panel of human microsomes and did not positively or negatively interfere with their catalytic activities, thereby lending hope to the idea that this drug could be used in combination with HIV‑1 therapies. Mechanistically, OPC‑67683 seems to inhibit mycolic-acid biosynthesis and, more specifically, methoxy- and keto-mycolic-acid biosynthesis93, although it probably requires biotransformation for its activity95. Given the limited amount of target information that is currently available, it is unclear if OPC‑67683 and PA‑824 have different mechanisms of action. OPC‑67683 successfully completed Phase I clinical trials, demonstrating satisfactory safety and pharmacokinetic profiles in healthy individuals. The small molecule is now being evaluated nature reviews | microbiology
for its early bactericidal efficacy in Phase II studies in combination with the standard front-line regimen. The clinical candidates outlined in this section hold promise as next-generation antituberculars, but each must now pass through the phases of clinical trials to receive approval. This process is not without considerable uncertainty, as demonstrated by the high attrition rates for clinical compounds in the United States.
Approved non-TB drugs as antituberculars A strategy to reduce the risk that is associated with failure in clinical trials owing to an inadequate humansafety profile lies in the use of already-approved drugs as antituberculars. In this section, we will discuss the potential for three classes of non-TB therapeutics in the fight against TB. The main hurdle has become the demonstration of sufficient efficacy at a dosage level that was previously deemed to be safe. Linezolid. Linezolid received FDA approval for MDR Gram-positive bacterial infections in 2000. This 3‑aryl‑2-oxazolidinone antibiotic, and its analogues, has displayed promising in vitro and in vivo efficacy against drug-sensitive and drug-resistant M. tuberculosis96,97. In Staphylococcus aureus, and presumably other bacteria such as M. tuberculosis, linezolid inhibits protein synthesis by binding to the 23S ribosome to prevent translation98. Clinical examination of linezolid as a potential antitubercular, although demonstrating promising efficacy against MDR and XDR TB, has detected significant toxicity99–102. The rate of incidence and severity of these adverse events might be reduced by decreasing the dosage amount of linezolid. A clinical trial that is expected to be completed in late 2007 is currently being conducted in Brazil to determine the efficacy of lower drug doses (J. Johnson, personal communication). In addition, further optimization of the oxazolidinone series for M. tuberculosis activity is underway103. The potential for using already approved drugs as antitubercular agents holds considerable promise. Given that marketed drugs have well-documented and acceptable safety profiles, a major hurdle has been removed in terms of finding new therapies for TB. However, as most antibiotics have little activity against M. tuberculosis, the crucial step in using these drugs is to demonstrate good efficacy.
β-lactam. Although the β‑lactam class of antibiotics has been used in clinics for over 60 years, none of its representatives has been used for the treatment of TB. β‑lactams act by inhibiting bacterial-cell-wall biosynthesis and would be a welcome addition to the antitubercular arsenal. Unfortunately, M. tuberculosis produces only a single β‑lactamase that has broad specificity104–106. Other genes may also have a role in resistance by affecting cellwall permeability and/or binding affinity for the pertinent penicillin-binding proteins107. Although β‑lactams can penetrate the cell and inhibit their targets108, their potency is primarily limited by β‑lactamase-mediated degradation. Two strategies could avoid this problem. First, in some cases, β‑lactamase inhibitors, such as volume 6 | january 2008 | 49
© 2008 Nature Publishing Group
REVIEWS clavulanic acid and sulbactam, allowed growth inhibition below the µg per ml level if used in combination with β‑lactams. Second, the carbapenem imipenem is resistant to cleavage. Unfortunately, both imipenem108–110 and amoxicillin in combination with clavulanic acid111 have produced mixed results in early bactericidalefficacy assessments in humans. However, this remains a promising area for investigation and recent structural work using M. tuberculosis and Mycobacterium fortuitum β‑lactamases106,112 might prompt the further design of β‑lactamase inhibitors and/or β‑lactams that have reduced β‑lactamase susceptibility. Phenothiazines. Phenothiazines have a rich and diverse history of medicinal uses113 and antitubercular activity against drug-sensitive and drug-resistant strains of M. tuberculosis114,115. The major examples are chlorpromazine115, thioridazine116 and trifluoperazine117,118. These drugs have substantial, and sometimes even disabling, side effects that limit their use at effective plasma concentrations. However, they might be more effective in vivo than in vitro as they are concentrated in macrophages119,120. Phenothiazines appear to target the type‑2 nicotinamide adenine dinucleotide (NADH):menaquinone dehydrogenase (NDH‑2) 121. NDH‑2 is crucial to the mycobacterial electron-transport chain and is the only such enzyme in M. tuberculosis that is absent in humans. Notably, this target seems to be important in starved cultures122, suggesting that NDH‑2 could be Boshoff, H. I. M. & Barry 3rd, C. E. Tuberculosismetabolism and respiration in the absence of growth. Nature Rev. Microbiol. 3, 70–80 (2005). 2. Russell, D. G. Who puts the tubercle in tuberculosis? Nature Rev. Microbiol. 5, 39–47 (2007). 3. Stewart, G. R., Roberston, B. D. & Young, D. B. Tuberculosis: a problem with persistence. Nature Rev. Microbiol. 1, 97–105 (2003). Focuses on the biology of persistence in mycobacteria. 4. Janin, Y. L. Antituberculosis drugs: ten years of research. Bioorg. Med. Chem. 15, 2479–2513 (2007). Discusses antitubercular small molecules, from early discovery compounds to approved drugs. 5. Ginsberg, A. M. & Spigelman, M. Challenges in tuberculosis drug research and development. Nature Med. 13, 290–294 (2007). An enlightening discussion of the current hurdles in TB drug discovery. 6. Williams, K. J. & Duncan, K. Current strategies for identifying and validating targets for new treatmentshortening drugs for TB. Curr. Mol. Med. 7, 297–307 (2007). An informative exposition that focuses on drugdiscovery strategies. 7. Hopkins, A. L. & Groom, C. R. The druggable genome. Nature Rev. Drug Discov. 1, 727–730 (2002). 8. Cheng, A. C. et al. Structure-based maximal affinity model predicts small-molecule druggability. Nature Biotechnol. 25, 71–75 (2007). 9. Boshoff, H. I. M. et al. The transcriptional responses of Mycobacterium tuberculosis to inhibitors of metabolism. J. Biol. Chem. 279, 40174–40184 (2004). 10. Andries, K. et al. A diarylquinoline drug active on the ATP synthase of Mycobacterium tuberculosis. Science 307, 223–227 (2005). 11. Manjunatha, U. H. et al. Identification of a nitroimidazo‑oxazine‑specific protein involved in PA‑824 resistance in Mycobacterium tuberculosis. Proc. Natl Acad. Sci. USA 103, 431–436 (2006). 12. Camacho, L. R., Ensergueix, D., Perez, E., Gicquel, B. & Guilhot, C. Identification of a virulence gene cluster 1.
13.
14.
15.
16. 17.
18. 19.
20.
21.
22.
an attractive target for other compounds. One way to achieve this would be to start with the known active phenothiazines and construct analogues that have more selectivity, a goal that is currently being pursued123,124.
Conclusions M. tuberculosis would seem to be a vulnerable organism, given that it has no noteworthy animal or environmental reservoir and limited genetic diversity. However, despite the availability of several effective antibiotics, TB continues to be a widespread and devastating disease. The need for new fast-acting drugs is clear. Fortunately, several chemical entities are currently in clinical trials, and numerous promising compounds are in the earlier stages of drug development. It is likely that new drugs will become available in the near future that can cope with the resistance problem as it currently exists. However, resistance continually evolves. Persistence is currently an even more formidable enemy, requiring a much more thorough understanding of its basic biological underpinnings and the small molecules that can modulate its role in the disease state. Therefore, little hope exists in the short term for a drastic reduction in the time for treatment, given the targets and drugs that are under clinical investigation. Despite the considerable challenges that are posed by these ‘bugs’, the recent improvement in the scale of efforts to find new drugs lends hope that science will deliver on the promise of eradicating ‘The Great White Plague’ that was celebrated prematurely by some owing to Selman Waksman’s Nobel Prize.
of Mycobacterium tuberculosis by signature-tagged transposon mutagenesis. Mol. Microbiol. 34, 257–267 (1999). Cox, J. S., Chen, B., McNeil, M. & Jacobs, W. R. Complex lipid determines tissue-specific replication of Mycobacterium tuberculosis in mice. Nature 402, 79–83 (1999). Sassetti, C. M., Boyd, D. H. & Rubin, E. J. Comprehensive identification of conditionally essential genes in mycobacteria. Proc. Natl Acad. Sci. USA 2001, 12712–12717 (2001). Lamichane, G. et al. A postgenomic method for predicting essential genes at subsaturation levels of mutagenesis: application to Mycobacterium tuberculosis. Proc. Natl Acad. Sci. USA 12, 7213–7218 (2003). Ehrt, S. et al. Controlling gene expression in mycobacteria with anhydrotetracycline and Tet repressor. Nucleic Acids Res. 33, e21 (2005). Carroll, P., Muttucumaru, D. G. & Parish, T. Use of a tetracycline-inducible system for conditional expression in Mycobacterium tuberculosis and Mycobacterium smegmatis. Appl. Environ. Microbiol. 71, 3077–3084 (2005). Blokpoel, M. C. et al. Tetracycline-inducible gene regulation in mycobacteria. Nucleic Acids Res. 33, e22 (2005). Argyrou, A., Jin, L., Siconilfi-Baez, L., Angeletti, R. H. & Blanchard, J. S. Proteome-wide profiling of isoniazid targets in Mycobacterium tuberculosis. Biochemistry 45, 13947–13953 (2006). Argyrou, A., Vetting, M. W., Aladegbami, B. & Blanchard, J. S. Mycobacterium tuberculosis dihydrofolate reductase is a target for isoniazid. Nature Struct. Mol. Biol. 13, 408–413 (2006). Nopponpunth, V., Sirawaraporn, W., Greene, P. J. & Santi, D. V. Cloning and expresion of Mycobacterium tuberculosis and Mycobacterium leprae dihydropteroate synthase in Escherichia coli. J. Bacteriol. 181, 6814–6821 (1999). Rengarajan, J. et al. The folate pathway is a target for resistance to the drug para-aminosalicyclic acid (PAS) in mycobacteria. Mol. Microbiol. 53, 275–282 (2004).
50 | january 2008 | volume 6
23. Vilcheze, C. et al. Transfer of a point mutation in Mycobacterium tuberculosis inhA resolves the target of isoniazid. Nature Med. 12, 1027–1029 (2006). 24. Lipinski, C. & Hopkins, A. Navigating chemical space for biology and medicine. Nature 432, 855–861 (2004). 25. Nwaka, S. & Hudson, A. Innovative lead discovery strategies for tropical diseases. Nature Rev. Drug Discov. 5, 941–955 (2006). 26. Lee, R. E. et al. Combinatorial lead optimization of [1,2]-diamines based on ethambutol as potential antituberculosis preclinical candidates. J. Comb. Chem. 5, 172–187 (2003). Demonstrates a combinatorial chemistry expansion around ethambutol that was paired with a functional-assay approach to produce novel chemical entities that modulate mycobacterial cell-wall biosynthesis, including SQ109. 27. Johnson, D. E. & Rodgers, A. D. Computational toxicology: heading toward more relevance in drug discovery and development. Curr. Opin. Drug Discov. Devel. 9, 29–37 (2006). 28. Pearl, G. M., Livingston-Carr, S. & Durham, S. K. Integration of computational analysis as a sentinel tool in toxicological assessments. Curr. Top. Med. Chem. 1, 247–255 (2001). 29. Goldman, R. et al. in Annual Conference on Antimicrobial Resistance (National Foundation for Infectious Diseases, Bethesda, 2006). 30. Alland, D., Steyn, A. J., Weisbrod, T., Aldrich, K. & Jacobs, W. R. Jr. Characterization of the Mycobacterium tuberculosis iniBAC promoter, a promoter that responds to cell wall biosynthesis inhibition. J. Bacteriol. 182, 1802–1811 (2000). 31. Qureshi, N., Sathyamoorthy, N. & Takayama, K. Biosynthesis of C30 to C56 fatty acids by an extract of Mycobacterium tuberculosis H37Ra. J. Bacteriol. 157, 46–52 (1984). 32. Rozwarski, D. A., Vilchéze, C., Sugantino, M., Bittman, R. & Sacchettini, J. C. Crystal structure of the Mycobacterium tuberculosis enoyl-ACP reductase, InhA, in complex with NAD+ and a C16 fatty acyl substrate. J. Biol. Chem. 274, 15582–15589 (1999).
www.nature.com/reviews/micro © 2008 Nature Publishing Group
REVIEWS 33. Zhang, Y., Heym, B., Allen, B., Young, D. & Cole, S. The catalase-peroxidase gene and isoniazid resistance of Mycobacterium tuberculosis. Nature 358, 591–593 (1992). 34. Muñoz-Elías, E. J. & McKinney, J. D. Mycobacterium tuberculosis isocitrate lyases 1 and 2 are jointly required for in vivo growth and virulence. Nature Med. 11, 638–644 (2005). 35. Sharma, V. et al. Structure of isocitrate lyase, a persistence factor of Mycobacterium tuberculosis. Nature Struct. Biol. 7, 663–668 (2000). 36. White, E. L. et al. A novel inhibitor of Mycobacterium tuberculosis pantothenate synthetase. J. Biomol. Screen. 12, 100–105 (2007). 37. Bogatcheva, E. et al. Identification of new diamine scaffolds with activity against Mycobacterium tuberculosis. J. Med. Chem. 49, 3045–3048 (2006). 38. Murillo, A. C. et al. High throughput crystallography of TB drug targets. Infect. Disord. Drug Targets 7, 127–139 (2007). Discusses the use of X‑ray crystallography to catalyze TB drug-discovery efforts by the elucidation of the structures of essential M. tuberculosis gene products. 39. Bajorath, J. Integration of virtual and high-throughput screening. Nature Rev. Drug Discov. 1, 882–894 (2002). 40. Shoichet, B. K. Virtual screening of chemical libraries. Nature 432, 862–865 (2004). Provides a good introduction to virtual screening using examples from drug discovery. 41. Shoichet, B. K., McGovern, S. L., Wei, B. & Irwin, J. J. Lead discovery using molecular docking. Curr. Opin. Chem. Biol. 6, 439–446 (2002). 42. Manetti, F. et al. Ligand-based virtual screening, parallel solution-phase and microwave-assisted synthesis as tools to identify and synthesize new inhibitors of Mycobacterium tuberculosis. ChemMedChem 1, 973–989 (2006). 43. García-García, A. et al. Search of chemical scaffolds for novel antituberculosis agents. J. Biomol. Screen. 10, 206–214 (2005). 44. Lipinski, C. A., Lombardo, F., Dominy, B. W. & Feeney, P. J. Experimental and computational approaches to estimate solubility and permeability in drug discovery and development settings. Adv. Drug Deliv. Rev. 23, 3–25 (1997). 45. Agrawal, H., Kumar, A., Bal, N. C., Siddiqi, M. I. & Arora, A. Ligand based virtual screening and biological evaluation of inhibitors of chorismate mutase (Rv1885c) from Mycobacterium tuberculosis H37Rv. Bioorg. Med. Chem. Lett. 17, 3053–3058 (2007). 46. Parish, T. & Stoker, N. G. The common aromatic amino acid biosynthesis pathway is essential in Mycobacterium tuberculosis. Microbiology 148, 3069–3077 (2002). 47. Hediger, M. E. Design, synthesis, and evaluation of aza inhibitors of chorismate mutase. Bioorg. Med. Chem. 15, 4995–5010 (2004). 48. Lin, T. W. et al. Structure-based inhibitor design of ACCD5, an essential acyl-CoA carboxylase carboxyltransferase domain of Mycobacterium tuberculosis. Proc. Natl Acad. Sci. USA 103, 3072–3077 (2006). 49. Payne, D. J., Gwynn, M. N., Holmes, D. J. & Pompliano, D. L. Drugs for bad bugs: confronting the challenges of antibacterial discovery. Nature Rev. Drug Discov. 6, 29–40 (2007). 50. Dessen, A., Quemard, A., Blanchard, J. S., Jacobs, W. R. Jr & Sacchettini, J. C. Crystal structure and function of the isoniazid target of M. tuberculosis. Science 267, 1638–1641 (1995). 51. Kuo, M. R. et al. Targeting tuberculosis and malaria through inhibition of enoyl reductase. J. Biol. Chem. 278, 20851–20859 (2003). 52. Rozwarski, D. A., Grant, G. A., Barton, D. H. R., Jacobs, W. R. Jr & Sacchettini, J. C. Modification of the NADH of the isoniazid target (InhA) from Mycobacterium tuberculosis. Science 279, 98–102 (1998). 53. Wang, F. et al. Mechanism of thionamide drug action against tuberculosis and leprosy. J. Exp. Med. 204, 73–78 (2007). 54. Sullivan, T. J. et al. High affinity InhA inhibitors with activity against drug-resistant strains of Mycobacterium tuberculosis. ACS Chem. Biol. 1, 43–53 (2006). 55. Tsukamura, M., Nakamura, E., Yoshii, S. & Amano, H. Therapeutic effect of a new antibacterial substance ofloxacin (DL8280) on pulmonary tuberculosis. Am. Rev. Respir. Dis. 131, 352–356 (1985).
56. Aubry, A. et al. Novel gyrase mutations in quinoloneresistant and -hypersusceptible clinical isolates of Mycobacterium tuberculosis: functional analysis of mutant enzymes. Antimicrob. Agents Chemother. 50, 104–112 (2006). 57. Veziris, N. et al. Treatment failure in a case of extensively drug-resistant tuberculosis associated with selection of a GyrB mutant causing fluoroquinolone resistance. Eur. J. Clin. Microbiol. Infect. Dis. 26, 423–425 (2007). 58. Cole, S. T. et al. Deciphering the biology of Mycobacterium tuberculosis from the complete genome sequence. Nature 393, 537–544 (1998). Determined the genome of M. tuberculosis H37Rv. 59. Takiff, H. E. et al. Efflux pump of the proton antiporter family confers low-level fluoroquinolone resistance in Mycobacterium smegmatis. Proc. Natl Acad. Sci. USA 9, 362–366 (1996). 60. Cynamon, M., Sklaney, M. R. & Shoen, C. Gatifloxacin in combination with rifampicin in a murine tuberculosis model. J. Antimicrob. Chemother. 60, 429–432 (2007). 61. Spigelman, M. K. New tuberculosis therapeutics: a growing pipeline. J. Infect. Dis. 196, S28–S34 (2007). 62. Cricchio, R., Arioli, V. & Lancini, G. C. Hydrazones of 3‑formylrifamycin SV. I — hydrazones with N‑amino‑N′substituted piperazines: synthesis, antibacterial activity and other biological properties. Farmaco [Sci] 30, 605–619 (1975). 63. Arioli, V. et al. Antibacterial activity of DL 473, a new semisynthetic rifamycin derivative. J. Antibiot. (Tokyo) 34, 1026–1032 (1981). 64. Bemer-Melchior, P., Bryskier, A. & Drugeon, H. B. Comparison of the in vitro activities of rifapentine and rifampicin against Mycobacterium tuberculosis. J. Antimicrob. Chemother. 46, 571–575 (2000). 65. Dickinson, J. M. & Mitchison, D. A. In vitro properties of rifapentine (MDL473) relevant to its use in intermittent chemotherapy of tuberculosis. Tubercle 68, 113–118 (1987). 66. Burman, W. J., Gallicano, K. & Peloquin, C. Comparative pharmacokinetics and pharmacodynamics of the rifamycin antibacterials. Clin. Pharmacokinet. 40, 327–341 (2001). 67. Yamane, T. et al. Synthesis and biological activity of 3′‑hydroxy‑5′-aminobenzoxazinorifamycin derivatives. Chem. Pharm. Bull. (Tokyo) 41, 148–155 (1993). 68. Fuji, K., Saito, H., Tomioka, H., Mae, T. & Hosoe, K. Mechanism of action of antimycobacterial activity of the new benzoxazinorifamycin KRM‑1648. Antimicrob. Agents Chemother. 39, 1489–1492 (1995). 69. Hirata, T. et al. In vitro and in vivo activities of the benzoxazinorifamycin KRM‑1648 against Mycobacterium tuberculosis. Antimicrob. Agents Chemother. 39, 2295–2303 (1995). 70. Klemens, S. P. & Cynamon, M. Activity of KRM‑1648 in combination with isoniazid against Mycobacterium tuberculosis in a murine model. Antimicrob. Agents Chemother. 40, 298–301 (1996). 71. Dietze, R. et al. Safety and bactericidal activity of rifalazil in patients with pulmonary tuberculosis. Antimicrob. Agents Chemother. 45, 1972–1976 (2001). 72. Mae, T. et al. Effect of a new rifamycin derivative, rifalazil, on liver microsomal enzyme induction in rat and dog. Xenobiotica 28, 759–766 (1998). 73. Burman, W. J. & Jones, B. E. Treatment of HIV-related tuberculosis in the era of effective antiretroviral therapy. Am. J. Respir. Crit. Care Med. 164, 7–12 (2001). 74. Report No. TDR/PRD/TB/03.1W (World Health Organization, Geneva, 2003). 75. Shepherd, R. G. & Wilkinson, R. G. Antituberculous agents. II. N,N′-diisopropylethylenediamine and analogs. J. Med. Pharm. Chem. 5, 823–835 (1962). 76. Wilkinson, R. G., Cantrall, M. B. & Shepherd, R. G. Antituberculous agents. III. (+)‑2,2-(ethylene diimino)‑di‑1-butanol and some analogs. J. Med. Pharm. Chem. 5, 835–845 (1962). 77. Jia, L. et al. Pharmacodynamics and pharmacokinetics of SQ109, a new diamine-based antitubercular drug. Br. J. Pharmacol. 144, 80–87 (2005). 78. Koul, A. et al. Diarylquinolines target subunit c of mycobacterial ATP synthase. Nature Chem. Biol. 3, 323–324 (2007). 79. Petrella, S. et al. Genetic basis for natural and acquired resistance to the diarylquinoline R207910 in mycobacteria. Antimicrob. Agents Chemother. 50, 2853–2856 (2006). 80. Lounis, N. et al. Combinations of R207910 with drugs used to treat multidrug-resistant tuberculosis have the
nature reviews | microbiology
potential to shorten treatment duration. Antimicrob. Agents Chemother. 50, 3543–3547 (2006). 81. Deidda, D. et al. Bactericidal activities of the pyrrole derivative BM212 against multidrug-resistant and intramacrophagic Mycobacterium tuberculosis strains. Antimicrob. Agents Chemother. 42, 3035–3037 (1998). 82. Arora, S. K., Sinha, N., Sinha, R. & Upadhyaya, R. S. U.S. Patent Application. US 2005/0256128 A1 (2005). 83. Casenghi, M. Development of new drugs for TB chemotherapy. Campaign for access to essential medicines [online], http://www.aerzte-ohne-grenzen.at/ img/db/msfmedia-3701.pdf (2006). 84. Edwards, D. I. Mechanism of antimicrobial action of metronidazole. J. Antimicrob. Chemother. 5, 499–502 (1979). 85. Wayne, L. G. & Sohaskey, C. D. Nonreplicating persistence of Mycobacterium tuberculosis. Annu. Rev. Microbiol. 55, 139–163 (2001). 86. Wayne, L. G. & Sramek, H. A. Metronidazole is bactericidal to dormant cells of Mycobacterium tuberculosis. Antimicrob. Agents Chemother. 38, 2054–2058 (1994). 87. Brooks, J. V., Furney, S. K. & Orme, I. M. Metronidazole therapy in mice infected with tuberculosis. Antimicrob. Agents Chemother. 43, 1285–1288 (1999). 88. Ashtekar, D. R. et al. In vitro and in vivo activities of the nitroimidazole CGI 17341 against Mycobacterium tuberculosis. Antimicrob. Agents Chemother. 37, 183–186 (1993). 89. Nagrajan, K., Shankar, R. G., Rajappa, S., Shenoy, S. J. & Costa-Pereira, R. Nitroimidazoles XXI 2,3‑dihydro‑6nitroimidazo [2,1‑b] oxazoles with antitubercular activity. Eur. J. Med. Chem. 24, 631–633 (1989). 90. Stover, C. K. et al. A small-molecule nitroimidazopyran drug candidate for the treatment of tuberculosis. Nature 405, 962–966 (2000). 91. Tyagi, S. et al. Bactericidal activity of the nitroimidazopyran PA‑824 in a murine model of tuberculosis. Antimicrob. Agents Chemother. 49, 2289–2293 (2005). 92. Nuermberger, E. et al. Combination chemotherapy with the nitroimidazopyran PA‑824 and first-line drugs in a murine model of tuberculosis. Antimicrob. Agents Chemother. 50, 2621–2625 (2006). 93. Matsumoto, M. et al. OPC‑67683, a nitro‑dihydro‑imidazooxazole derivative with promising action against tuberculosis in vitro and in mice. PLoS Med. 3, 2131–2144 (2006). 94. Sasaki, H. et al. Synthesis and antituberculosis activity of a novel series of optically active 6-nitro‑2,3dihydroimidazo[2,1‑b]oxazoles. J. Med. Chem. 49, 7854–7860 (2006). 95. Raether, W. & Hänel, H. Nitroheterocyclic drugs with broad spectrum activity. Parasitol. Res. 90, S19–S39 (2003). 96. Brickner, S. J. et al. Synthesis and antibacterial activity of U‑100592 and U‑100766, two oxazolidinone antibacterial agents for the potential treatment of multidrug-resistant gram-positive bacterial infections. J. Med. Chem. 39, 673–679 (1996). 97. Zurenko, G. E. et al. In vitro activities of U‑100592 and U‑100766, novel oxazolidinone antibacterial agents. Antimicrob. Agents Chemother. 40, 839–845 (1996). 98. Colca, J. R. et al. Cross-linking in the living cell locates the site of action of oxazolidinone antibiotics. J. Biol. Chem. 278, 21972–21979 (2003). 99. Fortún, J. et al. Linezolid for the treatment of multidrug-resistant tuberculosis. J. Antimicrob. Chemother. 56, 180–185 (2005). 100. Ntziora, F. & Falagas, M. E. Linezolid for the treatment of patients with atypical mycobacterial infection: a systematic review. Int. J. Tuberc. Lung Dis. 11, 606–611 (2007). 101. Park, I. N. et al. Efficacy and tolerability of daily-half dose linezolid in patients with intractable multidrugresistant tuberculosis. J. Antimicrob. Chemother. 58, 701–704 (2006). 102. von der Lippe, B., Sandven, P. & Brubakk, O. Efficacy and safety of linezolid in multidrug resistant tuberculosis (MDR-TB) — a report of ten cases. J. Infect. 52, 92–96 (2006). 103. Sood, R., Rao, M., Singhal, S. & Rattan, A. Activity of RBx 7644 and RBx 8700, new investigational oxazolidinones, against Mycobacterium tuberculosis infected murine macrophages. Int. J. Antimicrob. Agents 25, 464–468 (2005).
volume 6 | january 2008 | 51 © 2008 Nature Publishing Group
REVIEWS 104. Fisher, J. F., Meroueh, S. O. & Mobashery, S. Bacterial resistance to β‑lactam antibiotics: compelling opportunism, compelling opportunity. Chem. Rev. 105, 395–424 (2005). 105. Flores, A. R., Parsons, L. M. & Pavelka, M. S. Jr. Genetic analysis of the β-lactamases of Mycobacterium tuberculosis and Mycobacterium smegmatis and susceptibility to β-lactam antibiotics. Microbiology 151, 521–532 (2005). 106. Wang, F., Cassidy, C. & Sacchettini, J. C. Crystal structure and activity studies of the Mycobacterium tuberculosis β‑lactamase reveal its critical role in resistance to β‑lactam antibiotics. Antimicrob. Agents Chemother. 50, 2762–2771 (2006). 107. Flores, A. R., Parsons, L. M. & Pavelka, M. S. Jr. Characterization of novel Mycobacterium tuberculosis and Mycobacterium smegmatis mutants hypersusceptible to β-lactam antibiotics. J. Bacteriol. 187, 1892–1900 (2005). 108. Chambers, H. F. et al. Can penicillins and other betalactam antibiotics be used to treat tuberculosis? Antimicrob. Agents Chemother. 39, 2620–2624 (1995). 109. Chambers, H. F., Turner, J., Schecter, G. F., Kawamura, M. & Hopewell, P. C. Imipenem for treatment of tuberculosis in mice and humans. Antimicrob. Agents Chemother. 49, 2816–2821 (2005). 110. Rodloff, A. C., Goldstein, E. J. C. & Torres, A. Two decades of imipenem therapy. J. Antimicrob. Chemother. 58, 916–929 (2006). 111. Chambers, H. F., Kocagoz, S., Sipit, T., Turner, J. & Hopewell, P. C. Activity of amoxicillin/clavulanate in patients with tuberculosis. Clin. Infect. Dis. 26, 874–877 (1998). 112. Sauvage, E. et al. Crystal structure of the Mycobacterium fortuitum class A β-lactamase: structural basis for broad substrate specificity. Antimicrob. Agents Chemother. 50, 2516–2521 (2006). 113. Lopez-Munoz, F. et al. History of the discovery and clinical introduction of chlorpromazine. Ann. Clin. Psychiatry 17, 113–135 (2006). 114. Amaral, L., Kristiansen, J. E., Viveiros, M. & Atouguia, J. Activity of phenothiazines against antibiotic-resistant Mycobacterium tuberculosis: a review supporting further studies that may elucidate the potential use of thioridazine as anti-tuberculosis therapy. J. Antimicrob. Chemother. 47, 505–511 (2001). 115. Hollister, L. E., Eikenberry, D. T. & Raffel, S. Chlorpromazine in nonpsychotic patients with
pulmonary tuberculosis. Am. Rev. Respir. Dis. 81, 562–566 (1960). 116. Amaral, L., Kristiansen, J. E., Abebe, L. S. & Millet, W. Inhibition of the respiration of multi-drug resistant clinical isolates of Mycobacterium tuberculosis by thioridazine: potential use for initial therapy of freshly diagnosed tuberculosis. J. Antimicrob. Chemother. 38, 1049–1053 (1996). 117 Ratnakar, P. & Murthy, P. S. Antitubercular activity of trifluoperazine, a calmodulin antagonist. FEMS Microbiol. Lett. 1, 73–76 (1992). 118. Reddy, M. V., Nadadhur, G. & Gangadharam, P. R. In-vitro and intracellular antimycobacterial activity of trifluoperazine. J. Antimicrob. Chemother. 37, 196–197 (1996). 119. Crowle, A. J., Douvas, G. S. & May, M. H. Chlorpromazine: a drug potentially useful for treating mycobacterial infections. Chemotherapy 38, 410–419 (1992). 120. Ordway, D. et al. Clinical concentrations of thioridazine kill intracellular multidrug-resistant Mycobacterium tuberculosis. Antimicrob. Agents Chemother. 47, 917–922 (2003). 121. Weinstein, E. A. et al. Inhibitors of type II NADH: menaquinone oxidoreductase represent a class of antitubercular drugs. Proc. Natl Acad. Sci. USA 102, 4548–4553 (2005). 122. Xie, Z., Siddiqi, N. & Rubin, E. J. Differential antibiotic susceptibilities of starved Mycobacterium tuberculosis isolates. Antimicrob. Agents Chemother. 49, 4778–4780 (2005). 123. Bate, A. B. et al. Synthesis and antitubercular activity of quaternized promazine and promethazine derivatives. Bioorg. Med. Chem. Lett. 17, 1346–1348 (2007). 124. Madrid, P. B., Polgar, W. E., Toll, L. & Tanga, M. J. Synthesis and antitubercular activity of phenothiazines with reduced binding to dopamine and serotonin receptors. Bioorg. Med. Chem. Lett. 17, 3014–3017 (2007). 125. Schatz, A., Bugie, E. & Waksman, S. A. Streptomycin, a substance exhibiting antibiotic activity against Gram-positive and Gram-negative bacteria. Proc. Soc. Exp. Biol. Med. 55, 66–69 (1944). 126. Schatz, A., Bugie, E. & Waksman, S. A. Effect of streptomycin and other antibiotic substances upon Mycobacterium tuberculosis and related organisms. Proc. Soc. Exp. Biol. Med. 57, 244– 248 (1944). 127. Kingston, W. Streptomycin, Schatz v. Waksman, and the balance of credit for discovery. J. Hist. Med. Allied Sci. 60, 218–220 (2005).
52 | january 2008 | volume 6
128. Pettersen, E. F. et al. UCSF chimera — a visualization system for exploratory research and analysis. J. Comput. Chem. 25, 1605–1612 (2004). 129. Wright, D. H., Brown, G. H., Peterson, M. L. & Rotschafer, J. C. Application of fluoroquinolone pharmacodynamics. J. Antimicrob. Chemother. 46, 669–683 (2000). 130. McIlleron, H. et al. Determinants of rifampin, isoniazid, pyrazinamide, and ethambutol pharmacokinetics in a cohort of tuberculosis patients. Antimicrob. Agents. Chemother. 50, 1170–1177 (2006). 131. Berning, S. E. & Peloquin, C. A. in Antimicrobial Chemotherapy (eds Yu, V. L., Merigan, T. C., Barriere, S. & White, N. J.) 663–668 (Williams and Wilkins, Maryland, 1998). 132. Doluisio, J. T., Dittert, L. W. & LaPiana, J. C. Pharmacokinetics of kanamycin following intramuscular administration. J. Pharmacokinet. Biopharm. 1, 253–265 (1973). 133. Adamis, G. et al. Pharmacokinetic interactions of ceftazidime, imipenem and aztreonam with amikacin in healthy volunteers. Int. J. Antimicrob. Agents 23, 144–149 (2004).
Acknowledgements
The authors thank M. Spigelman, R. Goldman, J. Garcia, K. Andries, C. Dukes Hamilton, J. Guilemont, J. Johnson and A. Sternlicht for insightful conversations and, in some cases, providing unpublished results. The authors are supported by a grant from the National Institutes of Health (PO1A1068135) and the Robert A. Welch Foundation.
DATABASES Entrez Genome Project: http://www.ncbi.nlm.nih.gov/ entrez/query.fcgi?db=genomeprj Chlamydia pneumoniae | Mycobacterium smegmatis | Mycobacterium tuberculosis | Saccharomyces cerevisiae | Staphylococcus aureus
FURTHER INFORMATION James C. Sacchettini’s homepage: http://puffer.tamu.edu/ ClinicalTrials.gov: http://clinicaltrials.gov/ct2/show/NCT0 0425113?term=Metronidazole&rank=3 DrugBank: http://redpoll.pharmacy.ualberta.ca/drugbank/ TB Alliance: http://www.tballiance.org The Merck Manuals Medical Library: http://www.merck. com/mmpe/index.html Tuberculosis Antimicrobial Acquisition and Coordinating Facility: http://www.taacf.org All links are active in the online pdf
www.nature.com/reviews/micro © 2008 Nature Publishing Group
REVIEWS
Salmonellae interplay with host cells Andrea Haraga*, Maikke B. Ohlson‡ and Samuel I. Miller*‡§
Abstract | Salmonellae are important causes of enteric diseases in all vertebrates. Characterization of the molecular mechanisms that underpin the interactions of salmonellae with their animal hosts has advanced greatly over the past decade, mainly through the study of Salmonella enterica serovar Typhimurium in tissue culture and animal models of infection. Knowledge of these bacterial processes and host responses has painted a dynamic and complex picture of the interaction between salmonellae and animal cells. This Review focuses on the molecular mechanisms of these host–pathogen interactions, in terms of their context, significance and future perspectives. Pinocytosis A nonspecific process by which small volumes of extracellular fluid are taken up by certain eukaryotic cells owing to the engulfment of fluid in small membrane vesicles.
Tight junction The connection between two adjacent cells in a monolayer that is formed by extracellularmatrix and protein complexes; impermeable to water and other molecules.
Macropinocytosis Used to refer to the endocytosis of large volumes of extracellular fluid and particles by membrane ruffles.
*Department of Genome Sciences, ‡Department of Microbiology, §Department of Medicine, University of Washington, Seattle, Washington 98195, USA. Correspondence to S.I.M. e‑mail:
[email protected]. edu doi:10.1038/nrmicro1788 Published online 19 November 2007
Salmonellae are Gram-negative bacterial pathogens that are capable of infecting a wide range of animals, which can result in several manifestations of disease1. The host specificities and the disease symptoms that are caused by some of the thousands of Salmonella serovars are listed in TABLE 1. Salmonellae are typically acquired by the oral ingestion of contaminated food or water; however, exposure to pet reptiles and amphibians, which are often carriers of the bacteria, can also pose a risk for salmonellosis in humans (FIG. 1). Although conditions that increase the gastric pH reduce the infectious dose of the bacterium, salmonellae have an adaptive acid-tolerance response that might promote their survival in the low pH milieu of the stomach2. After entering the small intestine, salmonellae traverse the intestinal mucous layer and evade being killed by digestive enzymes, bile salts, secretory IgA, antimicrobial peptides and other innate immune defences in order to gain access to the underlying epithelium3–5. Salmonellae have the ability to invade the non-phagocytic enterocytes of the intestinal epithelium by bacterial-mediated endocytosis6. After adherence to the apical surface of the cell, using various fimbrial adhesins, the bacteria disrupt the epithelial brush border and induce membrane ruffles that engulf the organisms7,8. However, salmonellae preferentially enter microfold (M) cells, which are specialized epithelial cells that sample intestinal antigens through pinocytosis and transport them to lymphoid cells in the underlying Peyer’s patches9,10. Alternatively, they can translocate through the intestinal epithelia after uptake by CD18expressing phagocytes11. Moreover, in vitro, salmonellae are able to disrupt tight junctions, which seal the epithelial cell layer and restrict the paracellular passage of ions, water and immune cells12. This, in addition to intestinal inflammatory responses, probably contributes to the induction of diarrhoea. The fact that there are
nature reviews | microbiology
multiple mechanisms for crossing the intestinal barrier indicates the importance of this strategy to the lifestyle of salmonellae. M-cell-mediated uptake also allows salmonellae to interact with, and possibly enter, intestinal epithelial cells through their basolateral surface. Epithelial turnover and shedding then, in turn, could return the bacteria to the lumen of the intestine, simply requiring that they survive and, perhaps, replicate intracellularly. Such mechanisms are possibly important for intestinal colonization and even for acute disease by salmonellae, as a great deal of data, from both animal and cell culture models, indicate that the specialized ability to invade and survive in epithelial cells is essential for their pathogenesis 13–17. By contrast, the importance of phagocytosis and trafficking through CD18-positive cells is unknown and might represent another mechanism, redundant to M-cell-mediated entrance, for salmonellae to cross the intestinal barrier without actively interacting with the epithelial surface. It is interesting to speculate that this method may be important for typhoid fever, in which so few organisms can cause disease but induce only minimal intestinal disturbance in the majority of cases. This could allow Salmonella enterica serovar Typhi (S. typhi) to disseminate systemically in a relatively silent fashion, as opposed to mechanisms that involve bacterial-mediated endocytosis, which are associated with intestinal inflammation and diarrhoea13,14. Once the epithelial barrier has been breached, Salmonella serotypes that are associated with systemic illness can enter intestinal macrophages by inducing macropinocytosis , sensing the phagosomal environment and activating various virulence mechanisms in order to survive in the microbicidal environment of the macrophage 18–25 . This promotes bacterial volume 6 | january 2008 | 53
© 2008 Nature Publishing Group
REVIEWS Table 1 | Examples of Salmonella enterica serovars; their hosts and diseases Salmonella enterica serovar
Host specificity
Disease and symptoms
Humanrestricted
Enteric fever: fever; abdominal pain; transient diarrhoea or constipation; and a salmoncoloured maculopapular rash on the trunk
Broad-range
Gastroenteritis: abdominal pain; vomiting; and inflammatory diarrhoea
Typhoid Typhi; Paratyphi
Non-typhoid Typhimurium; Enteriditis
replication and subsequent dissemination throughout the reticuloendothelial system (RES). Infection with non‑typhoidal strains in healthy human adults, however, is usually limited to the intestine, where the bacteria induce an early inflammatory response that results in the infiltration of polymorphonuclear leukocytes (PMNs) into the intestinal lumen1,26. Production of the potent PMN chemokine interleukin (IL)‑8, and, perhaps, other similar molecules, by the infected epithelia is thought to be required for this process27. The release of cytotoxic granules by PMNs, as well as the effects of various bacterial molecules in stimulating innate immune inflammatory responses and manipulating hostcell processes, may then result in the destruction or turnover of the intestinal mucosa, which contributes to inflammatory diarrhoea. Recent progress has been made in our understanding of the molecular mechanisms by which salmonellae interact with host cells and manipulate various host responses that result in the induction of intestinal inflammatory responses and systemic illness. Specifically, the identification of bacterial and host proteins that are involved in such processes as the invasion of epithelial cells, stimulation and repression of signalling cascades, sensing of the intracellular environment and establishment of a niche for intracellular replication have contributed to insights into the complex pathogenesis of this bacterium, and will be discussed here in detail. Reticuloendothelial system (RES). The meshwork of connective tissue that contains immune cells, such as macrophages, and surrounds tissues that are associated with the immune system, such as the spleen and lymph nodes. Immune cells in the RES provide surveillance of antigens that the body encounters and can be quickly recruited to sites of infection.
Pathogenicity island A large region of genomic DNA that encodes genes that are associated with virulence. A pathogenicity island is typically transferred horizontally between bacterial strains and is often inserted into tRNA genes within the genome.
The two virulence-associated T3SSs A specialized apparatus, named the type III secretion system (T3SS), is essential to salmonellae pathogenesis and the colonization of host tissues. The T3SS mediates the transfer of bacterial virulence proteins, known as effectors, from the bacterial cell into the host-cell cytoplasm28. Once inside the eukaryotic cell, these effectors can alter host cellular functions, such as cytoskeletal architecture, membrane trafficking, signal transduction and cytokine gene expression, in order to promote bacterial survival and colonization. The T3SS is evolutionarily related to the flagellar export system and is present in multiple Gram-negative animal and plant pathogens. It comprises more than 20 proteins, some of which form a supramolecular structure that is known as the needle complex (NC) (FIG. 2). Electron micrographs of purified T3SSs revealed that the NC consists of a multiring base that spans the inner and outer membranes of the bacterial envelope, an inner rod that joins the rings
54 | january 2008 | volume 6
and a hollow needle29,30. Another set of proteins comprise the translocon, which is thought to be involved in the translocation of effectors by forming a translocation pore in the host cellular membrane31. The effector proteins have been shown to contain specific targeting signals located in their amino termini that route them to the T3SS32–35. In addition to the short secretion signal, many effectors have a binding site for a specific chaperone, the function of which is thought to be to stabilize and target its cognate substrate to the translocation apparatus34,36,37. Most chaperones are specific for a single effector, but some can facilitate the secretion of more than one effector protein38–41. An ATPase that is located in the base of the NC not only drives the export process but also facilitates the release of effectors from chaperones before transport42. Although much of the T3SS is similar among Gram-negative bacteria, the set of effector proteins is unique to each species. For a complete list of Salmonella spp. effectors, their host-cell functions and binding partners, see Table 2. Salmonellae encode two distinct virulence‑associated T3SSs within Salmonella pathogenicity islands 1 and 2 (SPI1 and SPI2) that function at different times during infection28. Whereas the SPI1-encoded T3SS is active on contact with the host cell and translocates bacterial proteins across the plasma membrane, the SPI2 T3SS is expressed within the phagosome and translocates effectors across the vacuolar membrane. The SPI1 system has been shown to be required for invasion of non-phagocytic cells, induction of intestinal inflammatory responses and diarrhoea, as well as colonization of the intestine. The SPI2 T3SS, by contrast, has an important role in bacterial survival in macrophages and establishment of systemic disease. Interestingly, Brown and colleagues43 have recently shown by RIVET (recombination-based in vivo expression technology) that expression of the SPI2 T3SS in mice might begin in the early stages of Salmonella enterica serovar Typhimurium (S. typhimurium) infection, before intestinal penetration. As there is currently no evidence that this T3SS is also involved in intestinal colonization, the authors speculated that its early induction in the intestinal lumen might prepare the bacterium for the inhospitable intracellular environment of the macrophage and ease the transition to the later systemic phase of the disease. However, it is also possible that early expression of this T3SS is required for the optimal function of the invasion-associated T3SS, as mutations in the SPI2 apparatus can cause a significant reduction in the expression of several SPI1 T3SS genes and impair the ability of the bacterium to invade epithelial cells44,45. Conversely, there is mounting evidence that some of the effectors of the SPI1 T3SS are expressed, or persist within, host cells long after bacterial internalization and contribute to events that were previously attributed exclusively to effectors of the SPI2 T3SS46–51. Although it is unknown whether mutations in the SPI1 T3SS could cause pleiotropic effects, these findings indicate that the two Salmonella T3SSs do not operate in an independent manner, as previously thought, and possibly cooperate to facilitate the intracellular lifestyle of the bacteria. This is particularly intriguing, because the www.nature.com/reviews/micro
© 2008 Nature Publishing Group
REVIEWS
Salmonella spp.
Epithelial cell
M cell
Macrophage T cell
B cell
Gastroenteritis: PMN influx
Enteric fever: dissemination to lymph nodes, liver and spleen
Figure 1 | Biology of salmonellae infection. Orally ingested salmonellae at the Nature Reviews |survive Microbiology low pH of the stomach and evade the multiple defences of the small intestine in order to gain access to the epithelium. Salmonellae preferentially enter M cells, which transport them to the lymphoid cells (T and B) in the underlying Peyer’s patches. Once across the epithelium, Salmonella serotypes that are associated with systemic illness enter intestinal macrophages and disseminate throughout the reticuloendothelial system. By contrast, non-typhoidal Salmonella strains induce an early local inflammatory response, which results in the infiltration of PMNs (polymorphonuclear leukocytes) into the intestinal lumen and diarrhoea.
sequence information and genetic organization of the two Salmonella T3SSs, as well as their phylogenetic distribution among Salmonella species, indicate that they were acquired independently, at different periods and from different sources44,52–57. Therefore, it is likely that selection pressure from animal environments led to the cooperation between the two T3SSs, to optimize colonization, invasion and intracellular survival of salmonellae in animals. The most extensively used animal model for investigating the contributions of the two T3SSs to salmonellae virulence is the highly sensitive natural-resistanceassociated macrophage protein 1 (Nramp1)‑null mouse, which is susceptible to mortal infection if inoculated with as few as ten bacteria58. Nramp1 is a macrophage-specific ion exporter that removes ions from the Salmonellacontaining vacuole (SCV) and, thus, restricts bacterial replication59. The Nramp1-null mouse model system has been used successfully to identify many important genes that are required for T3SS-associated virulence60,61. However, it is too sensitive to detect phenotypes for T3SS effectors that contribute to long-term persistence, because these mice do not survive long enough to study the proteins that are necessary for systemic colonization. The resurgence of interest in the resistant Nramp1-positive mouse model has enabled studies to be performed that are more physiologically relevant to long-term systemic infection49,62. These reports confirmed that SPI2 T3SS effectors are important for persistent infection and colonization, but also validated nature reviews | microbiology
older data that had suggested a role for the SPI1 T3SS during systemic disease17,63. This, combined with the previously mentioned putative function of the SPI2 T3SS prior to intestinal penetration, further demonstrates that the dogmatic compartmentalization of the roles of the two T3SSs is simplistic. In addition, the recent resurrection of the streptomycin-treated mouse model of acute intestinal infection with inflammation and diarrhoea might enable dissection of the roles of the host and bacteria in acute intestinal disease64. It is likely that, in the near future, models of salmonellae-induced chronic intestinal disease at the intestinal mucosal surface will be further developed and studied. In fact, our group has demonstrated that the A/J mouse strain develops this type of disease (L. Peiser, K. Smith and S.I.M., unpublished observations). These animal models should be useful in understanding the contributions of mucosal immune defences, as well as bacterial factors such as T3SS effectors, to chronic intestinal disease. However, these experimental systems also have limitations, as specific disease outcomes may be different in different hosts for broad host-range salmonellae. A future challenge in understanding these animal models will be to determine the effects of specific effectors in both the model and natural hosts. In addition, the diversity and evolution of the effector content of different Salmonella strains isolated from natural sources might also reveal different functions or roles that cannot be explored by these model systems. In this regard, many S. typhimurium effectors are present in only a percentage of strains, and in some serotypes, such as Typhi, are only present as pseudogenes. Selection for the presence or loss of effectors in specific animal populations or human disease settings might, ultimately, be more informative of the actual natural disease or colonization than the phenotypes of deletion mutants in model systems. Therefore, the acceptance of the limitations of these animal models is important to prevent dogmatic viewpoints about the roles of such mechanisms or specific T3SS effector proteins. This is particularly relevant to the study of effector proteins with redundant functions, as the necessity for their similar tasks during pathogenesis might not be apparent in model systems. A complete picture will probably require the integration of animal model information with real world information about effector content and disease course in nature.
Bacterial-mediated endocytosis The concerted function of at least five SPI1 T3SS effectors is required for efficient invasion of cultured epithelial cells, although optimal invasion in animal tissues might be more complex and diverse (FIG. 3). SopE, SopE2 and SopB (also known as SigD) activate the host Rho GTPases Cdc42, Rac1 and RhoG, which leads to actin cytoskeletal reorganization, membrane ruffling and bacterial internalization by macropinocytosis65–70. However, recent evidence suggests that only Rac1 and RhoG are indispensable for the actin remodelling events that are generated by Salmonella spp. during host-cell entry70. SopE, SopE2 and SopB are all essential for the invasion of epithelial cells, as an S. typhimurium mutant that is defective in all three effectors cannot induce actin volume 6 | january 2008 | 55
© 2008 Nature Publishing Group
REVIEWS a
Needle Base
b Base
Needle
Figure 2 | Structure of the Salmonella type III Nature Reviews | Microbiology secretion system (T3SS). a | Electron micrographs of a purified Salmonella enterica serovar Typhimurium (S. typhimurium) needle complex (NC). b | Electron cryomicroscopy and surface images of the structures of the base and NC of the S. typhimurium T3SS. Part a reproduced, with permission, from Ref. 31 (2002) Elsevier Science. Part b reproduced, with permission, from Ref. 185 (2004) American Association for the Advancement of Science.
rearrangements and, therefore, become intracellular69. Whereas SopE and SopE2 are potent guanine nucleotide exchange factors (GEFs) for all three GTPases, SopB only stimulates Cdc42 and RhoG indirectly through its phosphoinositide phosphatase activity65,67–70. Although the mechanism of action of SopB is not yet understood, Patel and colleagues70 have recently shown that activation of RhoG by SopB occurs by a cellular exchange factor called SGEF. As SGEF has a phosphoinositidebinding pleckstrin homology domain, it is plausible that it is activated by the phosphoinositide fluxes that are generated by the enzymatic action of SopB71. Activation of Rho GTPases then results in the activation of the Wiskott–Aldrich Syndrome protein (WASP) family members N‑WASP and WAVE2, which leads to recruitment of the actin-related protein‑2/3 (Arp2/3) complex to sites of membrane ruffles and stimulation of actin polymerization67,72–74. There are two SPI1 T3SS effectors that promote bacterial internalization by binding to actin and 56 | january 2008 | volume 6
modulating actin dynamics directly. SipA (also called SspA) helps to initiate actin polymerization at the site of S. typhimurium entry by decreasing the critical concentration and increasing the stability of actin filaments, whereas SipC (also called SspC), an effector that also functions as a translocon and is inserted in the host cell’s plasma membrane with a cytoplasmic domain that may bind to intermediate filaments, can nucleate and bundle actin 75–77. In addition, SipA can enhance the activity of SipC independently of any host cellular protein, which indicates that there might be a unique collaboration between the two effectors78. Although SipA and SipC might act in concert with SopE, SopE2 and SopB to mediate the cytoskeletal changes that are required for the formation of membrane ruffles, they cannot induce membrane ruffling and invasion by themselves 69,76,77,79. Instead, they seem to facilitate efficient bacterial uptake by directing the spatial localization of actin foci beneath the invading bacteria and, perhaps, preventing disassembly of the S. typhimurium-induced actin structures.
Intestinal inflammatory responses The stimulation of Cdc42 by SopE, SopE2 and SopB also triggers several mitogen-activated protein kinase (MAPK) pathways, including the Erk, Jnk and p38 pathways, which results in the activation of the transcription factors activator protein‑1 (AP‑1) and nuclear factor-κB (NF-κB)16,70,80 (FIG. 3). These transcription factors then direct the production of pro-inflammatory cytokines, such as IL‑8, which stimulate PMN transmigration and the inflammatory response leading to diarrhoea. Boyle and colleagues81 have recently reported that disruption of tight junction structure and function — one aspect of S. typhimurium-induced enteritis that is likely to be important — by the effectors SopB, SopE, SopE2 and SipA also occurs by activation of Rho family GTPases. In addition, SopB also promotes intestinal disease by increasing the intracellular concentration of d‑myo-inositol 1,4,5,6-tetrakisphosphate, a compound that stimulates cellular chloride secretion and fluid flux69,82. SipB (also called SspB), an effector protein that is also part of the SPI1 T3SS translocation machinery, adds to the inflammatory response by increasing production of the pro-inflammatory cytokines IL‑1β and IL‑18 through binding and activating caspase‑1 (Ref. 83). Caspase‑1 activation can also occur by the recognition of flagellin by the Ipaf cytoplasmic signalling pathway84,85. It is plausible that this is the result of the accidental translocation of flagellin into the host cell’s cytoplasm by the SPI1 T3SS owing to the conserved evolution of the flagellar and the SPI1-encoded secretion systems. The observation that caspase‑1-deficient mice are more susceptible to orally administered S. typhimurium than wild-type mice indicates that SPI1 T3SS-mediated caspase‑1 activation could be essential to diarrhoeal disease and intestinal survival86,87. SopA and SopD also contribute to enteritis in calves13,88,89. Recent evidence indicates that SopA has a HECT (homologous to E6-AP carboxyl terminus)-like E3 ubiquitin ligase activity, which is www.nature.com/reviews/micro
© 2008 Nature Publishing Group
REVIEWS Table 2 | Effectors of the Salmonella SPI1- and SPI2-encoded type III secretion systems (T3SSs) Effector
Cellular function
Host-cell target
References
AvrA
Inhibits nuclear factor (NF)‑κB signalling and interleukin (IL)‑8 production; also prevents ubiquitination of β‑catenin
Unknown
SipA or SspA
Decreases the critical concentration of G‑actin and increases the stability of F‑actin; also induces PMN transepithelial migration and disrupts tight junctions
F-actin; T‑plastin
SipB or SspB*
Binds and activates caspase‑1 and induces autophagy in macrophages
Caspase‑1; cholesterol
SipC or SspC*
Nucleates and bundles actin
F-actin; cytokeratin‑8 and cytokeratin-18
SopA
Stimulates PMN transmigration by HECT-like E3 ubiquitin ligase activity
Unknown
89,90
SopB or SigD
Activates Cdc42, RhoG, AktA and chloride secretion through its inositol phosphatase activity and disrupts tight junctions
Unknown
69,70,81,82,192
SopD
Stimulates fluid accumulation in bovine ligated ileal loops and contributes to diarrhoea in calves and systemic disease in mice
Unknown
13,82,91
SopE
Activates Cdc42, Rac1 and RhoG by its GEF activity and disrupts tight junctions
Cdc42, Rac1 and Rab5
SopE2
Activates Cdc42, Rac1 and RhoG by its GEF activity and disrupts tight junctions
Cdc42 and Rac1
67,68,81
SptP
Inhibits Cdc42 and Rac1 by its GAP activity and MAPK signalling and IL‑8 secretion through its tyrosine phosphatase activity
Rac1
92,94,95
GogB
Unknown
Unknown
PipB
Unknown
Unknown
157
PipB2
Contributes to Sif formation
Kinesin‑1
157,165,170
SifA
Induces Sif formation, maintains integrity of the SCV and downregulates kinesin recruitment to the SCV
SKIP and Rab7
SifB
Unknown
Unknown
SopD2
Contributes to Sif formation
Unknown
SpiC*
Interferes with endosomal trafficking
Hook3
SpvB‡
Actin-specific ADP-ribosyltransferase and downregulates Sif formation
Actin
SseF
Contributes to Sif formation and microtubule bundling
Unknown
162,169
SseG
Contributes to Sif formation and microtubule bundling
Unknown
162,169
SseI or SrfH
Contributes to host-cell dissemination
Filamin and TRIP6
160,197
SseJ
Maintains integrity of the SCV and has deacylase activity
Unknown
155,180
SseK1
Unknown
Unknown
198
SseK2
Unknown
Unknown
198
SseL
Deubiquitinase
Ubiquitin
199
SspH2
Inhibits the rate of actin polymerization and contributes to virulence in calves
Filamin and profilin
SteA
Unknown
Unknown
200
SteB
Unknown
Unknown
200
SteC
Unknown
Unknown
200
SlrP
Contributes to virulence in calves
Unknown
SspH1
Inhibits NF‑κB signalling and IL‑8 secretion, contributes to virulence in calves and has E3 ubiquitin ligase activity
PKN1
SPI1 T3SS 99,186 76, 81,187,188
83,189,190 75,77,191
65,81,193
SPI2 T3SS 194
154,164,166, 181,195 156 91 150,151–153 160, 175,176,196
96,160
SPI1 and SPI2 T3SS 201 95–98
*Also a component of the secretion apparatus. ‡Has not been definitively shown to be an SPI2 T3SS effector. GAP, GTPase-activating protein; GEF, guanine nucleotide exchange factor; HECT, homologous to E6-AP carboxyl terminus; MAPK, mitogen-activated protein kinase; PMN, polymorphonuclear leukocyte; SCV, Salmonella-containing vacuole; Sif, Salmonella-induced filament; SPI, Salmonella pathogenicity island.
nature reviews | microbiology
volume 6 | january 2008 | 57 © 2008 Nature Publishing Group
REVIEWS Membrane ruffle
Tight junction Salmonella spp. Actin Epithelial cell
SopE SopE2 SopB
SopA
SipA SipC
SptP SopB
Rho GTPases
MAPK
Rho GTPases
CI–
NF-κB AP-1
IL-8
NF-κB
IL-1β IL-18
SipB
PMN
SspH1 AvrA
MAPK
CI–
Macrophage
Figure 3 | SPI1 T3SS-induced changes in host cells. On contact with the epithelial Nature Reviews | Microbiology cell, salmonellae assemble the Salmonella pathogenicity island 1(SPI1)-encoded type III secretion system (T3SS) and translocate effectors (yellow spheres) into the eukaryotic cytoplasm. Effectors, such as SopE, SopE2 and SopB, then activate host Rho GTPases, which results in the rearrangement of the actin cytoskeleton into membrane ruffles, induction of mitogen-activated protein kinase (MAPK) pathways and destabilization of tight junctions. Changes in the actin cytoskeleton, which are further modulated by the actin-binding proteins SipA and SipC, lead to bacterial uptake. MAPK signalling activates the transcription factors activator protein‑1 (AP‑1) and nuclear factor-κB (NF‑κB), which turn on production of the pro-inflammatory polymorphonuclear leukocyte (PMN) chemokine interleukin (IL)‑8. SipB induces caspase‑1 activation in macrophages, with the release of IL‑1β and IL‑18, so augmenting the inflammatory response. In addition, SopB stimulates Cl– secretion by its inositol phosphatase activity. The destabilization of tight junctions allows the transmigration of PMNs from the basolateral to the apical surface, paracellular fluid leakage and access of bacteria to the basolateral surface. However, the transmigration of PMNs also occurs in the absence of tight-junction disruption and is further promoted by SopA. The actin cytoskeleton is restored and MAPK signalling is turned off by the enzymatic activities of SptP. This also results in the down-modulation of inflammatory responses, to which SspH1 and AvrA also contribute by inhibiting activation of NF‑κB. Figure reproduced from an original kindly provided by A. Haraga, University of Washington, USA.
involved in the S. typhimurium-induced transepithelial migration of PMNs90. However, how SopD, an effector that is also expressed under SPI2 T3SS-inducing conditions and continues to contribute to later stages of infection, exerts its effects remains unknown33,91. In summary, all of these mechanisms strongly implicate the SPI1 T3SS in the induction of gastroenteritis and intestinal disease.
Transepithelial migration The movement of cells, such as neutrophils and invading bacteria, from the basolateral (bottom) to the apical (top) surface, or the reverse, of an epithelial cell layer. Migration can also occur between two adjacent cells through tight junctions.
Reversal of cytoskeletal and signalling responses Interestingly, shortly after bacterial invasion the actin cytoskeleton regains its normal architecture7. SptP has been shown to participate in this reversal process by its GTPase-activating protein activity on Cdc42 and Rac1 (Ref. 92) (FIG. 3). The opposing activities of SopE or SopE2 and SptP are mediated by the temporal regulation of these proteins93. SopE has a shorter half-life in eukaryotic cells than SptP owing to its ubiquitination and rapid proteasome-dependent degradation. SptP is
58 | january 2008 | volume 6
a multifunctional protein that is also involved in the reversal of the pro-inflammatory signalling cascade that is associated with invasion94. It has tyrosine phosphatase activity, which has a role in downregulation of the S. typhimurium-induced activation of the MAPK Erk. In addition, S. typhimurium can downregulate IL‑8 production after invasion of intestinal epithelial cells, and SptP and SspH1 participate in this process95. SspH1 is a leucine-rich-repeat protein that is localized to the mammalian nucleus and inhibits NF‑κB-dependent gene expression95,96. It has also been shown to have ubiquitin ligase activity97. As it interacts with a host serine/threonine protein kinase called PKN1, which if activated also inhibits NF‑κB-dependent gene expression, SspH1 could interfere with inflammatory signalling by binding to and activating PKN1 (Ref. 98). In fact, PKN1 is activated in epithelial cells during infection with S. typhimurium. However, as this also occurs in the absence of SspH1, it is possible that SspH1 is not required for this process or that there are additional effectors or bacterial mechanisms that are capable of activating PKN1. Another SPI1 T3SS effector, AvrA, has been reported to inhibit NF‑κB activity and proinflammatory cytokine secretion99. However, unlike its Yersinia spp. homologue, YopJ, which binds and acetylates MAPK kinases and the inhibitor of NF-κB kinase (IKK), thereby preventing their phosphorylation and activation, AvrA has not been shown to interact with or affect the activity of any host proteins that are involved in the NF‑κB pathway100,101. The fact that salmonellae use multiple effectors to dampen the host’s inflammatory response indicates that this function is important for their pathogenesis, perhaps by contributing to colonization of the intestinal tract over prolonged periods of time. Furthermore, it is interesting to speculate that the evolutionary pressure on the host–pathogen interaction for non-typhoidal salmonellae is for asymptomatic intestinal colonization, whereas for typhoidal strains, which have poor intestinal colonization, the selective pressure is for replication within professional phagocytes, a condition that is also less inflammatory than extracellular growth. The bacteria might induce a mild state of inflammatory response and their goal might be to evolve towards parasitism or commensalism. In fact, the long period of intestinal colonization greatly exceeds that of the acute disease phase for non-typhoidal strains102. Also, typhoidal disease in the absence of antibiotic therapy is often characterized by periods of few or no symptoms, followed by a relapse that is more of a chronic, prolonged febrile illness, in which intracellular replication is the predominant bacterial lifestyle1. Thus, it is likely that a dynamic induction of inflammatory responses, followed by their dampening, is important to the diseases caused by salmonellae.
The SCV as a niche for replication Salmonellae are capable of infecting and replicating in many cell types, but are thought to primarily replicate in macrophages, as they are found in the lymphatic tissues and organs of the RES during systemic infection1,103. Salmonella strains that are defective for macrophage www.nature.com/reviews/micro
© 2008 Nature Publishing Group
REVIEWS Salmonella spp.
SspH2 SpvB Ssel
Spacious phagosome
SifA PipB2
s tor mo e l u tub icro
M
SCV Actin
Sif
SseJ
Epithelial cell Microtubules Nucleus SseF SseG Golgi
Secretory vesicles
Nature Microbiology Figure 4 | Formation of the SCV and induction of the SPI2 T3SSReviews within| the host cell. Shortly after internalization by macropinocytosis, salmonellae are enclosed in a spacious phagosome that is formed by membrane ruffles. Later, the phagosome fuses with lysosomes, acidifies and shrinks to become adherent around the bacterium. This is called the Salmonella-containing vacuole (SCV), which contains the endocytic marker lysosomal associated membrane protein 1 (LAMP‑1; purple). The Salmonella pathogenicity island 2 (SPI2) T3SS (type III secretion system) is induced within the SCV and translocates effector proteins (yellow spheres) across the phagosomal membrane several hours after phagocytosis. The SPI2 T3SS effectors SifA and PipB2 contribute to Salmonella-induced filament (Sif) formation along microtubules (green) and regulate microtubule-motor (yellow star shape) accumulation on the Sif and the SCV. SseJ is a deacylase that is active on the phagosome membrane. SseF and SseG cause microtubule bundling adjacent to the SCV and direct Golgi-derived vesicle traffic toward the SCV. Actin accumulates around the SCV in a SPI2 T3SS-dependent manner, in which SspH2, SpvB and SseI are thought to have a role.
Auxotrophic An organism that cannot synthesize certain organic compounds, such as amino acids, that are necessary for its metabolism. For growth, auxotrophic organisms must be able to take up the lacking compound from the surrounding environment.
replication are avirulent in mouse models of infection, which underscores the importance of bacterial survival and replication in macrophages to disease104. In addition, the observation that many attenuated strains of salmonellae are auxotrophic, particularly for purines, pyrimidines, amino acids and other metabolites that are required for their replication but that are not readily available within mammalian cells, supports the concept that intracellular replication is essential to salmonellae virulence105. Unlike non-phagocytic cells, salmonellae can enter macrophages by several endocytic processes, including SPI1 and non-SPI1 T3SS-induced macropinocytosis20. Following SPI1 T3SS-induced macropinocytosis, a few SPI1 T3SS effectors, such as SipA, SopB, SopD and SopE2, persist within the cell and have recently been implicated in contributing to the intracellular stages of the infection process47,48,50,51,91,106. Once intracellular, salmonellae remain inside a vacuolar compartment, which has been named the spacious phagosome (SP)20 (FIG. 4). The SP shrinks over minutes to hours to form an adherent membrane around one or more bacteria, which is then referred to as the SCV. The SCV can persist
nature reviews | microbiology
intracellularly from hours to days, making it a unique phagosome with respect to the normal progression of phagolysosomal maturation and recycling. Although there has been controversy within the field, several reports have shown that salmonellae can survive within macrophages in which the lysosomal compartments have fused with the SCV107–110. Consequently, the avoidance of phagolysosomal fusion is unlikely to be a major pathogenic strategy of salmonellae. Studies in various cell types have also demonstrated that the vacuole acidifies; however, depending on the mechanism of host‑cell entry, vacuolar acidification may be delayed in both macrophages and epithelial cells19,110–112. The SCV has been shown to interact transiently with the early endocytic pathway and quickly gain and lose early endocytic markers, such as EEA1 (early endosomal antigen 1), TfR (transferrin receptor) and the early endocytic trafficking GTPases Rab5 and Rab11 (Refs 113,114). Several late endosomal markers are commonly associated with the SCV at later time points, including the GTPase Rab7, LAMP1 (lysosomal associated membrane protein 1), LAMP2, LAMP3 and the vacuolar ATPase110,113,115,116. There is conflicting data concerning the occurrence of M6PR (mannose‑6-phosphate receptor), LBPA (lysobisphosphatidic acid) and the lysosomal hydrolase cathep sin D on the SCV115,117–120. Furthermore, cholesterol has been reported to accumulate on the SCV118,121. The presence or absence of different markers on the persistent SCV may simply indicate that they are variably detected rather than reflect whether or not the SCV has matured through a normal endocytic pathway. As it has been shown that the SCV can fuse with lysosomes and acidify, the ability of salmonellae to survive exposure to lysosomal contents reveals that resistance to antimicrobial peptides, nitric oxide and oxidative killing are important to its survival within macrophages and to virulence122–129. This is supported by the observations that Salmonella spp. mutants that are sensitive to these chemicals are attenuated for mouse virulence, whereas mice that are deficient for the production of these compounds have increased susceptibility to Salmonella spp.23,104,129–131
Sensing and response to the vacuole Salmonellae sense the acidic environment of the SCV, resulting in the induction of various regulatory systems that promote intracellular survival, for example, by surface remodelling of the protein, carbohydrate and membrane components of the bacterial envelope19. Such regulatory systems include OmpR/EnvZ, PhoP/PhoQ, RpoS/RpoE, PmrA/PmrB, Cya/Cyp and cyclic diGMP, all of which confer resistance to antimicrobial peptides and oxidative stress18,124,127,129,132–135. The phagosomal environment is acidic, with a pH range of <5 to 5.5, has a concentration of magnesium and calcium in the 1 mM range and contains antimicrobial peptides and oxygen and nitrogen radicals that can damage the bacterial cell19,111,112. Various studies indicate that both pH and antimicrobial peptides are important signatures of the phagosomal environment and such conditions activate many of the regulators that are implicated in salmonellae virulence24,25,125,136. It is likely that several sensory systems volume 6 | january 2008 | 59
© 2008 Nature Publishing Group
REVIEWS respond to the phagosome environment and cooperate to orchestrate the complex cascade of events that are necessary to alter the bacterial surface and promote intracellular survival137. The best characterized of these is the PhoQ sensor, which promotes resistance to antimicrobial peptides18,25. PhoQ contains a novel acidic domain that is bridged to the inner membrane by interactions with metal ions and binds to and initiates responses to antimicrobial peptides138,139. It also responds to pH by structural changes that are determined by regions of the protein separate from the metal ion bridges involved in antimicrobial peptide sensing136. After phagocytosis, salmonellae undergo extensive bacterial surface remodelling, as has been shown for the lipid A component of lipopolysaccharide (LPS) during growth within macrophages137,140. Bacterial molecules that the host can recognize as indicators of infection, such as the SPI1 T3SS and flagellin, are repressed and the LPS structure is altered85,141. Some of the specific surface modifications include: decreasing the length of the O antigen, which is the repeating carbohydrate polymer of LPS; alterations to the number of acyl chains in the structure of the lipid A component of LPS; and changes in the protein content of the outer membrane, the inner membrane and the peptidoglycan layer142–145. Synthesis of enzymes that allow the bacteria to cope with oxidative and nitrogenous radicals also occurs146. Microarray studies have shown that up to 919 S. typhimurium genes are differentially regulated in response to the phagosomal environment, demonstrating that dramatic transcriptional and post-translational changes probably occur when salmonellae make the transition from a nutrient-rich extracellular environment to the intracellular environment147.
Protein delivery across the vacuolar membrane As a result of sensing the phagosomal environment, the expression and assembly of the SPI2-encoded T3SS is induced134,148. Although the function of this T3SS in pathogenesis remains poorly defined, it has been shown to be essential for virulence in the mouse model of infection22,149. As mutants of the SPI2 T3SS cannot replicate in tissue-culture cells and animal models, it is likely that the role of this T3SS during disease is to promote intracellular replication within the intestine during the acute phase of the infection and in other organs during the chronic state. A reasonable hypothesis for the main function of the SPI2 T3SS is that it promotes intracellular replication by altering host vesicular trafficking, so that useful metabolic molecules, such as amino acids and lipids, are routed to the SCV and the vesicular compartment membrane is expanded. To date, at least 20 Salmonella spp. effector proteins are known to be translocated by the SPI2 T3SS across the phagosomal membrane into the eukaryotic-cell cytoplasm; however, their specific roles in promoting intracellular replication or modifying vesicular movement are not yet understood (for a list of SPI2 T3SS effectors, their functions and binding partners, see TABLE 2). In addition, no individual translocated effector has definitively been shown to alter vesicular trafficking. Although it has been reported that SpiC alters 60 | january 2008 | volume 6
endosome–endosome fusion in vitro and binds to proteins that are implicated in vesicular trafficking, its role as a T3SS effector protein remains controversial150,151. As it has not been universally observed to be translocated into eukaryotic cells and it is required for the translocation of several, if not all, SPI2 T3SS effectors, as well as the surface expression of translocon proteins, it is more likely that SpiC is part of the SPI2 secretion apparatus and is not an effector152,153. The most important translocated effectors, by virtue of causing virulence defects in mice when mutated, are SifA, SseJ, SseF, SseG, SopD2 and PipB2 (Refs 91,154–158) . The observation that S. typhimurium that lacks any single SPI2 T3SS effector protein cannot cause the same virulence attenuation in mice as a mutant strain that lacks the entire SPI2 T3SS suggests that many effectors function cooperatively to exert their effects on the host cell. Furthermore, the deletion or mutation of many effector genes has no virulence phenotype, which implies that their functions might be redundant. Early studies of SPI2 T3SS effectors, which primarily focused on determining their subcellular localization in mammalian cells, revealed that they might have specific targeting sequences that direct localization to endosomal compartments, the Golgi apparatus, the actin cytoskeleton and the microtubule network33,155–157,159–162. This indicates that components of these host-cell structures might be the intracellular targets of the SPI2 T3SS.
Salmonella-induced tubular endosomes In epithelial cells and interferon‑γ-primed macrophages, the SPI2 T3SS induces the formation of long filamentous membrane structures — Salmonella-induced filaments (Sifs) — that originate from the SCV and extend throughout the cell157,163 (FIG. 4). These structures are LAMP1-positive and have a similar membrane composition to the SCV118. Sif formation requires microtubules, but not the actin cytoskeleton, although these filaments can be decorated with actin159. Possible mechanisms that contribute to Sif formation include repetitive initiation of vesicular budding from the SCV, in which fission events are incomplete, or continuous fusion of endocytic vesicles with the SCV, which would result in endosomal tubulation or elongation of the SCV. Regardless of the mechanism of Sif formation, the filaments may function to increase the size of the SCV to accommodate bacterial replication during systemic infection. Sif formation is dependent on the SPI2 T3SS effector SifA, and to a lesser extent on SseF, SseG, SopD2 and PipB2 (Refs 91,164,165). However, it is a dynamic phenotype that may also be modulated by other effectors. SifA was recently shown to bind to the host cell protein SKIP (SifA and kinesin interacting protein)166. The same authors also found that, if SKIP was depleted by RNA interference in S. typhimurium-infected cells, the cells failed to form Sifs, suggesting that Sif formation requires SKIP. In addition, Alto and colleagues167 identified SifA as a member of the WxxxE (tryptophan (W)-variable (x)-xx-glutamate) family of bacterial effectors that function as mimics of GTPases. As SifA contains a carboxy-terminal Caax (cysteine (C)-aliphatic residue (a)-a-x) motif that www.nature.com/reviews/micro
© 2008 Nature Publishing Group
REVIEWS
Prenylated The post-translational addition of lipid chains, such as farnesyl or geranylgeranyl, to cysteine residues in proteins that contain a prenylation motif called a CaaX box. This process facilitates membrane localization and/or protein– protein interactions.
is prenylated and S‑acylated by host-cell enzymes, similar to GTPases, it is possible that SifA uses a GTPase-type mechanism for membrane localization and, perhaps, Sif formation168. Furthermore, the region of SKIP that interacts with SifA is a pleckstrin homology domain, which is commonly found in proteins that bind to signalling lipids and other regulatory molecules, including those that modulate Rho GTPases166. The roles of PipB2, SopD2, SseF and SseG in Sif formation are not entirely clear. S. typhimurium strains that lack any of these effectors do not induce Sifs as efficiently as wild-type S. typhimurium91,165,169. Instead, infection of cultured cells with these mutants tends to lead to the formation of ‘ pseudo-Sifs’, which extend from the SCV and co-localize with effectors, but do not contain LAMP1. This suggests that in the absence of these proteins the ability to form Sifs — which is defined as being entirely LAMP1-positive — is impaired and that induction of Sifs involves multiple steps that are orchestrated by different effectors. Transient expression of PipB2 in mammalian cells induces the movement of LAMP1-positive compartments to the cell periphery, which is probably the result of its interaction with the plus-end-directed microtubule motor kinesin165,170. This activity might contribute to the outward extension of the SCV, which would promote Sif formation. SopD2, which has homo logy to the SPI1 T3SS translocated effector SopD, has also been shown to cooperate with SifA to induce Sifs, but by an as yet unidentified mechanism33,91. SseF and SseG promote the aggregation of endosomal vesicles and recruit Golgi-derived exocytic vesicles to the SCV, which suggests that salmonellae are able to usurp both endocytic and exocytic cellular transport processes171–174. However, it is not known how these activities directly influence Sif formation. The deletion of SseJ and SpvB can cause an increase in Sif formation at later time points during the infection of cultured cells, which suggests that these proteins have Sif downregulatory functions175. Furthermore, these effectors have virulence defects in the Nramp1-null animal model, which indicates that their activities contribute to systemic infection156,176. Although SpvB has not been formally demonstrated to be an SPI2 T3SS effector, some studies suggest that the SPI2 T3SS is required for its activity in infected cells177,178. SpvB is an actin-specific ADP-ribosyltransferase that promotes actin depolymerization176. SseJ has homology to glycerophospholipid cholesterol acyl transferase enzymes, which can remove acyl chains from phospholipids and transfer them to cholesterol in a two-step deacylase-acyltranferase reaction32,179. In fact, purified SseJ has deacylase activity in vitro, which has been shown to be required for its virulence in mice180. It has been proposed that SseJ and SifA have complementary roles in maintaining the integrity of the SCV membrane, because sifA-mutant S. typhimurium tends to lose the SCV membrane but does not do so if SseJ is also lacking. This suggests that, in the absence of SifA, SseJ could cause damage to the SCV by its enzymatic activity. In addition, our laboratory has obtained evidence that SseJ has phospholipase activity in vivo that might be localized to the endosomal
nature reviews | microbiology
membrane, which has led us to propose that its most likely function during S. typhimurium infection is to cleave phospholipids from the SCV membrane (Megha, M.B.O. and S.I.M., unpublished observations). The cleavage of acyl chains from phospholipids could then promote curvature of the membrane or produce discrete lipid environments that serve as platforms for promoting vesicle fusion and binding scaffolding proteins and such activity could modulate Sif formation. However, how these functions, and Sif formation in general, are important for Salmonella-induced disease is currently not understood. It is plausible that Sifs increase the size of the SCV in a specific and directional fashion that promotes bacterial replication and/or redirect nutrient-rich organelles to the SCV. In this regard, the SPI2 T3SS could help salmonellae to replicate inside the phagosome and gain important nutrients for growth, but yet avoid some of the host-defence mechanisms and inflammatory responses that would result from their release into the cytoplasm.
Microtubule motors and movement of the SCV As Sif formation requires microtubules — which have been observed to accumulate and bundle around the SCV — there has been an increased interest in understanding how microtubules and microtubule motors contribute to the intracellular life cycle of salmonellae and whether targeting vesicular trafficking along microtubule networks in infected cells is indeed one of their pathogenic strategies162 (FIG. 4). SifA has been reported to downregulate the recruitment of the plus-end directed microtubule motor kinesin to the SCV, which is mediated by PipB2 (Refs 166,170). It does so by interacting with SKIP, which, if in a complex with SifA, displaces kinesin from the SCV166. This interference is thought to be crucial for maintaining the integrity of the SCV. The inhibitory activity of SifA on kinesin seems to be dominant over the kinesin-recruiting activity of PipB2, which is suggestive of a potential coordinate or temporal regulation between the two effectors170. In addition, SseF and SseG have been found to co-localize with microtubules and cause microtubule bundling in S. typhimuriuminfected cells. In particular, SseF has been implicated in recruiting the minus-end directed motor dynein to the SCV162,173. Therefore, it would appear that the SCV accumulates dynein, but not kinesin. However, as SifA inhibits the interaction between Rab7 and its effector RILP, a protein that is associated with the dynein motor complex, it is plausible that SifA also prevents dynein accumulation on the SCV181,182. Although these results paint a paradoxical picture of the need to promote or inhibit microtubule-motor accumulation on the SCV, salmonellae might employ a carefully choreographed combination of both the recruitment and inhibition of both types of microtubule motors. For example, at specific times during infection, the recruitment of dynein could promote minus-end-directed movement of the SCV to stabilize it near the nucleus or the Golgi apparatus, where the SCV would have increased access to trafficking vesicles that contain nutrients and membrane, whereas at other times the recruitment of kinesin volume 6 | january 2008 | 61
© 2008 Nature Publishing Group
REVIEWS to the SCV could induce plus-end-directed movement towards the cell periphery to promote the enlargement of the SCV and/or elongation of Sifs. Recent studies have also shown that the SCV in infected cultured epithelial cells is localized in a perinuclear position and that effectors from both the SPI1- and the SPI2-encoded T3SSs are involved in this process50,161,166,170,173,181. In particular, SifA, SseF, SseG and SipA seem to control this phenomenon, because the SCV containing S. typhimurium strains lacking any of these effectors does not localize near the nucleus. As the disruption of perinuclear localization of the SCV is correlated with decreased bacterial replication, it has been proposed that salmonellae require juxtanuclear positioning for optimal growth. In fact, Salcedo and Holden161 have shown that SCVs in close proximity to the nucleus contain more bacteria on average than SCVs further away from the nucleus. However, it has not been directly demonstrated that non-nuclear localization of the SCV, per se, is inhibitory to bacterial replication. Rather, it seems more likely that the lack of these effectors or disruption of vesicle trafficking along the microtubule network correlates with defective replication and that non-nuclear SCV positioning is a by-product of this more direct effect. Furthermore, the positioning of the SCV might be different in polarized epithelial cells, in which movement of the vacuole in a directional fashion, such as during transcytosis, could be more physiologically relevant. As discussed above, it is possible that positioning the SCV in close proximity to the Golgi, or other organelles, has some advantages for salmonellae in terms of acquiring nutrients, inhibiting immune responses, prolonging the maturation of intestinal epithelial cells and allowing more intracellular time before epithelial shedding into the intestinal lumen. All of these effects could then contribute to intestinal colonization or the intracellular growth of salmonellae.
Actin polymerization around the SCV SPI2 T3SS-dependent actin condensation around the SCV has been observed, which suggests that this T3SS is also involved in cytoskeletal modifications160,183 (FIG. 4). In addition, the treatment of S. typhimurium-infected cells with inhibitors of actin polymerization results in decreased bacterial replication, which indicates that the manipulation of the actin cytoskeleton is important for bacterial replication183. Although the effectors that are responsible for the recruitment of actin to the SCV have not been identified, several have been shown to interact with actin-binding proteins or to directly manipulate actin polymerization. However, it is also possible that SCV-associated actin polymerization is not mediated by effectors, but is due to the insertion of the T3SS translocon into the SCV membrane, which has also been observed with the similar Yersinia spp. T3SSs184. Indeed, most SPI2 T3SS effectors that have been shown to interact with components of the actin cytoskeleton have actin-inhibitory activities, which suggests that the most likely goal of salmonellae is to reduce or reverse vacuoleassociated actin polymerization (VAP). For example, SpvB has been shown to be involved in the disassembly 62 | january 2008 | volume 6
of VAP, presumably by ADP-ribosylating actin and promoting actin depolymerization160,176. SspH2 has also been reported to inhibit the rate of actin polymerization in vitro160. In addition, SspH2 can interact with newly polymerized actin, probably by binding to the actinbinding proteins filamin and profilin, through its amino and carboxy termini, respectively. Another effector, SseI (also called SrfH), shares a high degree of identity with the amino terminus of SspH2 and has been shown to bind to polymerized actin, probably through filamin. Although transiently expressed SspH2 and SseI are associated with VAP during S. typhimurium infection, they are not required for the formation of this structure. Thus, unlike the clearly defined and temporally controlled roles of the SPI1 T3SS effectors in manipulating the actin cytoskeleton during invasion of epithelial cells, the biological significance of the SPI2 T3SS-mediated actin recruitment to the SCV, and the interaction of the effectors of this T3SS with the actin network, remains to be understood. It might be important for translocon stability and translocation, thereby allowing the full effector complement to be delivered consistently, and it could explain the importance of this process to intracellular replication183. Subsequently, as T3SS effectors seem to decrease or remodel actin polymerization, there might be a period in which actin removal is important, for example, to decrease inflammatory responses or allow fusion with, or remodelling of, the endosomal membrane. Alternatively, actin removal could be essential for the effectors to interact with the membrane surface, as certain protein and phospholipid domains that are covered in polymerized actin may inhibit their binding to important regions within the phagosome membrane.
Future perspectives The study of the molecular basis of Salmonella spp. pathogenesis in mammals has advanced greatly over the past 15 years owing to the identification of many of the key molecules and mechanisms in mammalian model systems and the discovery of important innate immune receptors and responses to the bacteria. For example, important findings have been made in our understanding of the mechanisms of entrance into nonphagocytic cells, the way bacteria sense the intracellular environment and remodel their surfaces, and the inflammatory responses to salmonellae. Concurrently, the identification of innate immune receptors for bacterial products and characterization of ligand binding by these receptors has progressed greatly. We anticipate a further explosion of information in the next decade about the details of the intracellular lifestyle of salmonellae, the function of effectors that are translocated across the phagosome membrane and the response of the host to the activity of these bacterial proteins. Studying how salmonellae manipulate endosomal trafficking through the endocytic pathway should lead to important observations about unknown mechanisms of endocytic trafficking. In addition, how effectors function to bring together mammalian protein components that may not normally be together could offer important findings with relevance for non-infectious diseases. The www.nature.com/reviews/micro
© 2008 Nature Publishing Group
REVIEWS study of salmonellae in more-resistant mouse models, coupled with sophisticated imaging, should also lead to a greater understanding of the pathophysiological role of identified virulence mechanisms and the development of chronic disease. This may have relevance for other diseases at mucosal surfaces, such as inflammatory bowel disease. The future decade may also take the field of salmonellae pathogenesis and host responses beyond model systems to the diversity of bacteria and hosts in nature. Emerging technologies, such as whole-genome sequencing, will allow us to investigate the diversity of effectors that are used by the many pathogenic Salmonella species, contributing to our view of host specificity. Specific hypotheses will be generated by statistical association, which will then need to be tested in
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11. 12.
13.
14.
15.
Pegues, D. A., Ohl, M. E. & Miller, S. I. in Principles and Practice of Infectious Diseases (eds Mandell, G. L., Bennet, J. E. & Dolin, R.) 2636–2654 (Churchill Livingstone, New York, 2005). Garcia-del Portillo, F., Foster, J. W. & Finlay, B. B. Role of acid tolerance response genes in Salmonella typhimurium virulence. Infect. Immun. 61, 4489–4492 (1993). Michetti, P., Mahan, M. J., Slauch, J. M., Mekalanos, J. J. & Neutra, M. R. Monoclonal secretory immunoglobulin A protects mice against oral challenge with the invasive pathogen Salmonella typhimurium. Infect. Immun. 60, 1786–1792 (1992). Selsted, M. E., Miller, S. I., Henschen, A. H. & Ouellette, A. J. Enteric defensins: antibiotic peptide components of intestinal host defense. J. Cell Biol. 118, 929–936 (1992). Prouty, A. M., Brodsky, I. E., Falkow, S. & Gunn, J. S. Bile‑salt‑mediated induction of antimicrobial and bile resistance in Salmonella typhimurium. Microbiology 150, 775–783 (2004). Francis, C. L., Starnbach, M. N. & Falkow, S. Morphological and cytoskeletal changes in epithelial cells occur immediately upon interaction with Salmonella typhimurium grown under low-oxygen conditions. Mol. Microbiol. 6, 3077–3087 (1992). Takeuchi, A. Electron microscope studies of experimental Salmonella infection. I. Penetration into the intestinal epithelium by Salmonella typhimurium. Am. J. Pathol. 50, 109–136 (1967). This reference showed membrane ruffling that is induced by Salmonella spp. in intestinal cells by microscopy for the first time. Baumler, A. J., Tsolis, R. M. & Heffron, F. Contribution of fimbrial operons to attachment to and invasion of epithelial cell lines by Salmonella typhimurium. Infect. Immun. 64, 1862–1865 (1996). Kohbata, S., Yokoyama, H. & Yabuuchi, E. Cytopathogenic effect of Salmonella typhi GIFU 10007 on M cells of murine ileal Peyer’s patches in ligated ileal loops: an ultrastructural study. Microbiol. Immunol. 30, 1225–1237 (1986). Jones, B. D., Ghori, N. & Falkow, S. Salmonella typhimurium initiates murine infection by penetrating and destroying the specialized epithelial M cells of the Peyer’s patches. J. Exp. Med. 180, 15–23 (1994). Vazquez-Torres, A. et al. Extraintestinal dissemination of Salmonella by CD18-expressing phagocytes. Nature 401, 804–808 (1999). Jepson, M. A., Collares-Buzato, C. B., Clark, M. A., Hirst, B. H. & Simmons, N. L. Rapid disruption of epithelial barrier function by Salmonella typhimurium is associated with structural modification of intercellular junctions. Infect. Immun. 63, 356–359 (1995). Zhang, S. et al. The Salmonella enterica serotype Typhimurium effector proteins SipA, SopA, SopB, SopD, and SopE2 act in concert to induce diarrhea in calves. Infect. Immun. 70, 3843–3855 (2002). Raffatellu, M. et al. SipA, SopA, SopB, SopD, and SopE2 contribute to Salmonella enterica serotype Typhimurium invasion of epithelial cells. Infect. Immun. 73, 146–154 (2005). Stecher, B. et al. Comparison of Salmonella enterica serovar Typhimurium colitis in germfree mice and mice
model systems. In addition, the availability of human and animal genotyping will lead to an increase in data about the diversity of human and animal responses to infection with various salmonellae. Model systems in various animals that have diverse innate immune modifications will also need to be developed to further understand the natural world in the context of our available tools and to, ultimately, reach the stage in which the evolution of microorganisms and the emergence of infectious diseases can be observed in real time. Therefore, the study of salmonellae pathogenesis should yield a rich treasure trove of information for those who are interested in microbial pathogenesis, innate immunity, cell biology and genomics, and should remain an important model system of host–pathogen interactions for many years to come.
pretreated with streptomycin. Infect. Immun. 73, 3228–3241 (2005). 16. Hobbie, S., Chen, L. M., Davis, R. J. & Galan, J. E. Involvement of mitogen-activated protein kinase pathways in the nuclear responses and cytokine production induced by Salmonella typhimurium in cultured intestinal epithelial cells. J. Immunol. 159, 5550–5559 (1997). 17. Galan, J. E. & Curtiss, R. 3rd Cloning and molecular characterization of genes whose products allow Salmonella typhimurium to penetrate tissue culture cells. Proc. Natl Acad. Sci. USA 86, 6383–6387 (1989). 18. Miller, S. I., Kukral, A. M. & Mekalanos, J. J. A twocomponent regulatory system (phoP phoQ) controls Salmonella typhimurium virulence. Proc. Natl Acad. Sci. USA 86, 5054–5058 (1989). This reference identified the PhoP/PhoQ twocomponent system as being important for the regulation of virulence. 19. Alpuche-Aranda, C. M., Swanson, J. A., Loomis, W. P. & Miller, S. I. Salmonella typhimurium activates virulence gene transcription within acidified macrophage phagosomes. Proc. Natl Acad. Sci. USA 89, 10079–10083 (1992). 20. Alpuche-Aranda, C. M., Racoosin, E. L., Swanson, J. A. & Miller, S. I. Salmonella stimulate macrophage macropinocytosis and persist within spacious phagosomes. J. Exp. Med. 179, 601–608 (1994). 21. Ochman, H., Soncini, F. C., Solomon, F. & Groisman, E. A. Identification of a pathogenicity island required for Salmonella survival in host cells. Proc. Natl Acad. Sci. USA 93, 7800–7804 (1996). References 21 and 22 were the first to show that the T3SS that is encoded on SPI2 is important for virulence. 22. Shea, J. E., Hensel, M., Gleeson, C. & Holden, D. W. Identification of a virulence locus encoding a second type III secretion system in Salmonella typhimurium. Proc. Natl Acad. Sci. USA 93, 2593–2597 (1996). 23. Vazquez-Torres, A. et al. Salmonella pathogenicity island 2‑dependent evasion of the phagocyte NADPH oxidase. Science 287, 1655–1658 (2000). 24. Miao, E. A., Freeman, J. A. & Miller, S. I. Transcription of the SsrAB regulon is repressed by alkaline pH and is independent of PhoPQ and magnesium concentration. J. Bacteriol. 184, 1493–1497 (2002). 25. Bader, M. W. et al. Regulation of Salmonella typhimurium virulence gene expression by cationic antimicrobial peptides. Mol. Microbiol. 50, 219–230 (2003). 26. McCormick, B. A., Miller, S. I., Carnes, D. & Madara, J. L. Transepithelial signaling to neutrophils by salmonellae: a novel virulence mechanism for gastroenteritis. Infect. Immun. 63, 2302–2309 (1995). 27. McCormick, B. A. et al. Surface attachment of Salmonella typhimurium to intestinal epithelia imprints the subepithelial matrix with gradients chemotactic for neutrophils. J. Cell Biol. 131, 1599–1608 (1995). 28. Hansen-Wester, I. & Hensel, M. Salmonella pathogenicity islands encoding type III secretion systems. Microbes Infect. 3, 549–559 (2001).
nature reviews | microbiology
29. Kubori, T. et al. Supramolecular structure of the Salmonella typhimurium type III protein secretion system. Science 280, 602–605 (1998). This study provided the first electron microscopy images of the T3SS needle complex. 30. Kimbrough, T. G. & Miller, S. I. Contribution of Salmonella typhimurium type III secretion components to needle complex formation. Proc. Natl Acad. Sci. USA 97, 11008–11013 (2000). 31. Kimbrough, T. G. & Miller, S. I. Assembly of the type III secretion needle complex of Salmonella typhimurium. Microbes Infect. 4, 75–82 (2002). 32. Miao, E. A. & Miller, S. I. A conserved amino acid sequence directing intracellular type III secretion by Salmonella typhimurium. Proc. Natl Acad. Sci. USA 97, 7539–7544 (2000). 33. Brumell, J. H. et al. SopD2 is a novel type III secreted effector of Salmonella typhimurium that targets late endocytic compartments upon delivery into host cells. Traffic 4, 36–48 (2003). 34. Lee, S. H. & Galan, J. E. Salmonella type III secretionassociated chaperones confer secretion-pathway specificity. Mol. Microbiol. 51, 483–495 (2004). 35. Karavolos, M. H. et al. Type III secretion of the Salmonella effector protein SopE is mediated via an N‑terminal amino acid signal and not an mRNA sequence. J. Bacteriol. 187, 1559–1567 (2005). 36. Fu, Y. & Galan, J. E. Identification of a specific chaperone for SptP, a substrate of the centisome 63 type III secretion system of Salmonella typhimurium. J. Bacteriol. 180, 3393–3399 (1998). 37. Hong, K. H. & Miller, V. L. Identification of a novel Salmonella invasion locus homologous to Shigella ipgDE. J. Bacteriol. 180, 1793–1802 (1998). 38. Bronstein, P. A., Miao, E. A. & Miller, S. I. InvB is a type III secretion chaperone specific for SspA. J. Bacteriol. 182, 6638–6644 (2000). 39. Tucker, S. C. & Galan, J. E. Complex function for SicA, a Salmonella enterica serovar Typhimurium type III secretion-associated chaperone. J. Bacteriol. 182, 2262–2268 (2000). 40. Ehrbar, K., Friebel, A., Miller, S. I. & Hardt, W. D. Role of the Salmonella pathogenicity island 1 (SPI‑1) protein InvB in type III secretion of SopE and SopE2, two Salmonella effector proteins encoded outside of SPI‑1. J. Bacteriol. 185, 6950–6967 (2003). 41. Higashide, W. & Zhou, D. The first 45 amino acids of SopA are necessary for InvB binding and SPI‑1 secretion. J. Bacteriol. 188, 2411–2420 (2006). 42. Akeda, Y. & Galan, J. E. Chaperone release and unfolding of substrates in type III secretion. Nature 437, 911–915 (2005). 43. Brown, N. F. et al. Salmonella pathogenicity island 2 is expressed prior to penetrating the intestine. PLoS Pathog. 1, e32 (2005). 44. Hensel, M. et al. Functional analysis of ssaJ and the ssaK/U operon, 13 genes encoding components of the type III secretion apparatus of Salmonella pathogenicity island 2. Mol. Microbiol. 24, 155–167 (1997). 45. Deiwick, J. et al. Mutations in Salmonella pathogenicity island 2 (SPI2) genes affecting transcription of SPI1 genes and resistance to antimicrobial agents. J. Bacteriol. 180, 4775–4780 (1998).
volume 6 | january 2008 | 63 © 2008 Nature Publishing Group
REVIEWS 46. Steele-Mortimer, O. et al. The invasion-associated type III secretion system of Salmonella enterica serovar Typhimurium is necessary for intracellular proliferation and vacuole biogenesis in epithelial cells. Cell. Microbiol. 4, 43–54 (2002). 47. Hernandez, L. D., Hueffer, K., Wenk, M. R. & Galan, J. E. Salmonella modulates vesicular traffic by altering phosphoinositide metabolism. Science 304, 1805–1807 (2004). 48. Drecktrah, D., Knodler, L. A., Galbraith, K. & SteeleMortimer, O. The Salmonella SPI1 effector SopB stimulates nitric oxide production long after invasion. Cell. Microbiol. 7, 105–113 (2005). 49. Lawley, T. D. et al. Genome-wide screen for Salmonella genes required for long-term systemic infection of the mouse. PLoS Pathog. 2, e11 (2006). 50. Brawn, L. C., Hayward, R. D. & Koronakis, V. Salmonella SPI1 effector SipA persists after entry and cooperates with a SPI2 effector to regulate phagosome maturation and intracellular replication. Cell Host Microbe 1, 63–75 (2007). 51. Giacomodonato, M. N. et al. SipA, SopA, SopB, SopD and SopE2 effector proteins of Salmonella enterica serovar Typhimurium are synthesized at late stages of infection in mice. Microbiology 153, 1221–1228 (2007). 52. Li, J. et al. Relationship between evolutionary rate and cellular location among the Inv/Spa invasion proteins of Salmonella enterica. Proc. Natl Acad. Sci. USA 92, 7252–7256 (1995). 53. Ochman, H. & Groisman, E. A. Distribution of pathogenicity islands in Salmonella spp. Infect. Immun. 64, 5410–5412 (1996). 54. Boyd, E. F., Li, J., Ochman, H. & Selander, R. K. Comparative genetics of the inv-spa invasion gene complex of Salmonella enterica. J. Bacteriol. 179, 1985–1991 (1997). 55. Hensel, M. et al. Analysis of the boundaries of Salmonella pathogenicity island 2 and the corresponding chromosomal region of Escherichia coli K‑12. J. Bacteriol. 179, 1105–1111 (1997). 56. Hensel, M., Nikolaus, T. & Egelseer, C. Molecular and functional analysis indicates a mosaic structure of Salmonella pathogenicity island 2. Mol. Microbiol. 31, 489–498 (1999). 57. Porwollik, S., Wong, R. M. & McClelland, M. Evolutionary genomics of Salmonella: gene acquisitions revealed by microarray analysis. Proc. Natl Acad. Sci. USA 99, 8956–8961 (2002). 58. Hormaeche, C. E. Natural resistance to Salmonella typhimurium in different inbred mouse strains. Immunology 37, 311–318 (1979). 59. Vidal, S. M., Malo, D., Vogan, K., Skamene, E. & Gros, P. Natural resistance to infection with intracellular parasites: isolation of a candidate for Bcg. Cell 73, 469–485 (1993). 60. Hensel, M. et al. Genes encoding putative effector proteins of the type III secretion system of Salmonella pathogenicity island 2 are required for bacterial virulence and proliferation in macrophages. Mol. Microbiol. 30, 163–174 (1998). 61. Beuzon, C. R. & Holden, D. W. Use of mixed infections with Salmonella strains to study virulence genes and their interactions in vivo. Microbes Infect. 3, 1345–1352 (2001). 62. Monack, D. M., Bouley, D. M. & Falkow, S. Salmonella typhimurium persists within macrophages in the mesenteric lymph nodes of chronically infected Nramp1+/+ mice and can be reactivated by IFNgamma neutralization. J. Exp. Med. 199, 231–241 (2004). 63. Behlau, I. & Miller, S. I. A PhoP-repressed gene promotes Salmonella typhimurium invasion of epithelial cells. J. Bacteriol. 175, 4475–4484 (1993). 64. Barthel, M. et al. Pretreatment of mice with streptomycin provides a Salmonella enterica serovar Typhimurium colitis model that allows analysis of both pathogen and host. Infect. Immun. 71, 2839–2858 (2003). 65. Hardt, W. D., Chen, L. M., Schuebel, K. E., Bustelo, X. R. & Galan, J. E. S. typhimurium encodes an activator of Rho GTPases that induces membrane ruffling and nuclear responses in host cells. Cell 93, 815–826 (1998). This work elegantly demonstrated the ability of SopE to activate GTPases and subsequent actin rearrangements and transcriptional responses. 66. Bakshi, C. S. et al. Identification of SopE2, a Salmonella secreted protein which is highly homologous to SopE and involved in bacterial invasion of epithelial cells. J. Bacteriol. 182, 2341–2344 (2000).
67. Stender, S. et al. Identification of SopE2 from Salmonella typhimurium, a conserved guanine nucleotide exchange factor for Cdc42 of the host cell. Mol. Microbiol. 36, 1206–1221 (2000). 68. Friebel, A. et al. SopE and SopE2 from Salmonella typhimurium activate different sets of RhoGTPases of the host cell. J. Biol. Chem. 276, 34035–34040 (2001). 69. Zhou, D., Chen, L. M., Hernandez, L., Shears, S. B. & Galan, J. E. A Salmonella inositol polyphosphatase acts in conjunction with other bacterial effectors to promote host cell actin cytoskeleton rearrangements and bacterial internalization. Mol. Microbiol. 39, 248–259 (2001). 70. Patel, J. C. & Galan, J. E. Differential activation and function of Rho GTPases during Salmonella-host cell interactions. J. Cell Biol. 175, 453–463 (2006). 71. Ellerbroek, S. M. et al. SGEF, a RhoG guanine nucleotide exchange factor that stimulates macropinocytosis. Mol. Biol. Cell 15, 3309–3319 (2004). 72. Criss, A. K. & Casanova, J. E. Coordinate regulation of Salmonella enterica serovar Typhimurium invasion of epithelial cells by the Arp2/3 complex and Rho GTPases. Infect. Immun. 71, 2885–2891 (2003). 73. Unsworth, K. E., Way, M., McNiven, M., Machesky, L. & Holden, D. W. Analysis of the mechanisms of Salmonella-induced actin assembly during invasion of host cells and intracellular replication. Cell. Microbiol. 6, 1041–1055 (2004). 74. Shi, J., Scita, G. & Casanova, J. E. WAVE2 signaling mediates invasion of polarized epithelial cells by Salmonella typhimurium. J. Biol. Chem. 280, 29849–29855 (2005). 75. Hayward, R. D. & Koronakis, V. Direct nucleation and bundling of actin by the SipC protein of invasive Salmonella. EMBO J. 18, 4926–4934 (1999). References 75 and 76 detected direct manipulation of actin by the SPI1 T3SS effectors SipA and SipC. 76. Zhou, D., Mooseker, M. S. & Galan, J. E. Role of the S. typhimurium actin-binding protein SipA in bacterial internalization. Science 283, 2092–2095 (1999). 77. Scherer, C. A., Cooper, E. & Miller, S. I. The Salmonella type III secretion translocon protein SspC is inserted into the epithelial cell plasma membrane upon infection. Mol. Microbiol. 37, 1133–1145 (2000). 78. McGhie, E. J., Hayward, R. D. & Koronakis, V. Cooperation between actin-binding proteins of invasive Salmonella: SipA potentiates SipC nucleation and bundling of actin. EMBO J. 20, 2131–2139 (2001). 79. Higashide, W., Dai, S., Hombs, V. P. & Zhou, D. Involvement of SipA in modulating actin dynamics during Salmonella invasion into cultured epithelial cells. Cell. Microbiol. 4, 357–365 (2002). 80. Chen, L. M., Hobbie, S. & Galan, J. E. Requirement of CDC42 for Salmonella-induced cytoskeletal and nuclear responses. Science 274, 2115–2118 (1996). 81. Boyle, E. C., Brown, N. F. & Finlay, B. B. Salmonella enterica serovar Typhimurium effectors SopB, SopE, SopE2 and SipA disrupt tight junction structure and function. Cell. Microbiol. 8, 1946–1957 (2006). 82. Norris, F. A., Wilson, M. P., Wallis, T. S., Galyov, E. E. & Majerus, P. W. SopB, a protein required for virulence of Salmonella dublin, is an inositol phosphate phosphatase. Proc. Natl Acad. Sci. USA 95, 14057–14059 (1998). 83. Hersh, D. et al. The Salmonella invasin SipB induces macrophage apoptosis by binding to caspase‑1. Proc. Natl. Acad. Sci. USA 96, 2396–2401 (1999). 84. Franchi, L. et al. Cytosolic flagellin requires Ipaf for activation of caspase‑1 and interleukin 1b in Salmonella-infected macrophages. Nature Immunol. 7, 576–582 (2006). 85. Miao, E. A. et al. Cytoplasmic flagellin activates caspase‑1 and secretion of interleukin 1b via Ipaf. Nature Immunol. 7, 569–575 (2006). 86. Lara-Tejero, M. et al. Role of the caspase‑1 inflammasome in Salmonella typhimurium pathogenesis. J. Exp. Med. 203, 1407–1412 (2006). 87. Raupach, B., Peuschel, S. K., Monack, D. M. & Zychlinsky, A. Caspase‑1‑mediated activation of interleukin-1b (IL-1b) and IL‑18 contributes to innate immune defenses against Salmonella enterica serovar Typhimurium infection. Infect. Immun. 74, 4922–4926 (2006). 88. Jones, M. A. et al. Secreted effector proteins of Salmonella dublin act in concert to induce enteritis. Infect. Immun. 66, 5799–5804 (1998).
64 | january 2008 | volume 6
89. Wood, M. W. et al. The secreted effector protein of Salmonella dublin, SopA, is translocated into eukaryotic cells and influences the induction of enteritis. Cell Microbiol. 2, 293–303 (2000). 90. Zhang, Y., Higashide, W. M., McCormick, B. A., Chen, J. & Zhou, D. The inflammation-associated Salmonella SopA is a HECT-like E3 ubiquitin ligase. Mol. Microbiol. 62, 786–793 (2006). 91. Jiang, X. et al. The related effector proteins SopD and SopD2 from Salmonella enterica serovar Typhimurium contribute to virulence during systemic infection of mice. Mol. Microbiol. 54, 1186–1198 (2004). 92. Fu, Y. & Galan, J. E. A Salmonella protein antagonizes Rac‑1 and Cdc42 to mediate host-cell recovery after bacterial invasion. Nature 401, 293–297 (1999). References 92 and 93 identified the GTPase-antagonizing activity of SptP and showed how the activities of SopE and SptP are temporally regulated by proteasome-dependent degradation. 93. Kubori, T. & Galan, J. E. Temporal regulation of Salmonella virulence effector function by proteasomedependent protein degradation. Cell 115, 333–342 (2003). 94. Murli, S., Watson, R. O. & Galan, J. E. Role of tyrosine kinases and the tyrosine phosphatase SptP in the interaction of Salmonella with host cells. Cell. Microbiol. 3, 795–810 (2001). 95. Haraga, A. & Miller, S. I. A Salmonella enterica serovar Typhimurium translocated leucine-rich repeat effector protein inhibits NF‑kappa B‑dependent gene expression. Infect. Immun. 71, 4052–4058 (2003). 96. Miao, E. A. et al. Salmonella typhimurium leucinerich repeat proteins are targeted to the SPI1 and SPI2 type III secretion systems. Mol. Microbiol. 34, 850–864 (1999). 97. Rohde, J. R., Breitkreutz, A., Chenal, A., Sansonetti, P. J. & Parsot, C. Type III secretion effectors of the IpaH family are E3 ubiquitin ligases. Cell Host Microbe 1, 77–83 (2007). 98. Haraga, A. & Miller, S. I. A Salmonella type III secretion effector interacts with the mammalian serine/threonine protein kinase PKN1. Cell. Microbiol. 8, 837–846 (2006). 99. Collier-Hyams, L. S. et al. Cutting edge: Salmonella AvrA effector inhibits the key proinflammatory, antiapoptotic NF‑kappaB pathway. J. Immunol. 169, 2846–2850 (2002). 100. Mittal, R., Peak-Chew, S. Y. & McMahon, H. T. Acetylation of MEK2 and I kappa B kinase (IKK) activation loop residues by YopJ inhibits signaling. Proc. Natl Acad. Sci. USA 103, 18574–18579 (2006). 101. Mukherjee, S. et al. Yersinia YopJ acetylates and inhibits kinase activation by blocking phosphorylation. Science 312, 1211–1214 (2006). 102. Buchwald, D. S. & Blaser, M. J. A review of human salmonellosis: II. Duration of excretion following infection with nontyphi Salmonella. Rev. Infect. Dis. 6, 345–356 (1984). 103. Richter-Dahlfors, A., Buchan, A. M. & Finlay, B. B. Murine salmonellosis studied by confocal microscopy: Salmonella typhimurium resides intracellularly inside macrophages and exerts a cytotoxic effect on phagocytes in vivo. J. Exp. Med. 186, 569–580 (1997). 104. Fields, P. I., Swanson, R. V., Haidaris, C. G. & Heffron, F. Mutants of Salmonella typhimurium that cannot survive within the macrophage are avirulent. Proc. Natl Acad. Sci. USA 83, 5189–5193 (1986). This important study demonstrated that survival within macrophages is important for S. typhimurium virulence. 105. O’Callaghan, D., Maskell, D., Liew, F. Y., Easmon, C. S. & Dougan, G. Characterization of aromatic- and purine-dependent Salmonella typhimurium: attention, persistence, and ability to induce protective immunity in BALB/c mice. Infect. Immun. 56, 419–423 (1988). 106. Dukes, J. D. et al. The secreted Salmonella dublin phosphoinositide phosphatase, SopB, localizes to PtdIns(3)P-containing endosomes and perturbs normal endosome to lysosome trafficking. Biochem. J. 395, 239–247 (2006). 107. Carrol, M. E., Jackett, P. S., Aber, V. R. & Lowrie, D. B. Phagolysosome formation, cyclic adenosine 3′:5′-monophosphate and the fate of Salmonella typhimurium within mouse peritoneal macrophages. J. Gen. Microbiol. 110, 421–429 (1979). 108. Buchmeier, N. A. & Heffron, F. Inhibition of macrophage phagosome–lysosome fusion by Salmonella typhimurium. Infect. Immun. 59, 2232–2238 (1991).
www.nature.com/reviews/micro © 2008 Nature Publishing Group
REVIEWS 109. Oh, Y. K. et al. Rapid and complete fusion of macrophage lysosomes with phagosomes containing Salmonella typhimurium. Infect. Immun. 64, 3877–3883 (1996). 110. Drecktrah, D., Knodler, L. A., Howe, D. & SteeleMortimer, O. Salmonella trafficking is defined by continuous dynamic interactions with the endolysosomal system. Traffic 8, 212–225 (2007). 111. Rathman, M., Sjaastad, M. D. & Falkow, S. Acidification of phagosomes containing Salmonella typhimurium in murine macrophages. Infect. Immun. 64, 2765–2773 (1996). 112. Martin-Orozco, N. et al. Visualization of vacuolar acidification-induced transcription of genes of pathogens inside macrophages. Mol. Biol. Cell 17, 498–510 (2006). 113. Steele-Mortimer, O., Meresse, S., Gorvel, J. P., Toh, B. H. & Finlay, B. B. Biogenesis of Salmonella typhimurium-containing vacuoles in epithelial cells involves interactions with the early endocytic pathway. Cell. Microbiol. 1, 33–49 (1999). 114. Smith, A. C., Cirulis, J. T., Casanova, J. E., Scidmore, M. A. & Brumell, J. H. Interaction of the Salmonellacontaining vacuole with the endocytic recycling system. J. Biol. Chem. 280, 24634–24641 (2005). 115. Garcia-del Portillo, F. & Finlay, B. B. Targeting of Salmonella typhimurium to vesicles containing lysosomal membrane glycoproteins bypasses compartments with mannose 6‑phosphate receptors. J. Cell Biol. 129, 81–97 (1995). 116. Meresse, S., Steele-Mortimer, O., Finlay, B. B. & Gorvel, J. P. The rab7 GTPase controls the maturation of Salmonella typhimurium-containing vacuoles in HeLa cells. EMBO J. 18, 4394–4403 (1999). 117. Hashim, S., Mukherjee, K., Raje, M., Basu, S. K. & Mukhopadhyay, A. Live Salmonella modulate expression of Rab proteins to persist in a specialized compartment and escape transport to lysosomes. J. Biol. Chem. 275, 16281–16288 (2000). 118. Brumell, J. H., Tang, P., Mills, S. D. & Finlay, B. B. Characterization of Salmonella-induced filaments (Sifs) reveals a delayed interaction between Salmonellacontaining vacuoles and late endocytic compartments. Traffic 2, 643–653 (2001). 119. Garvis, S. G., Beuzon, C. R. & Holden, D. W. A role for the PhoP/Q regulon in inhibition of fusion between lysosomes and Salmonella-containing vacuoles in macrophages. Cell. Microbiol. 3, 731–744 (2001). 120. Cuellar-Mata, P. et al. Nramp1 modifies the fusion of Salmonella typhimurium-containing vacuoles with cellular endomembranes in macrophages. J. Biol. Chem. 277, 2258–2265 (2002). 121. Catron, D. M. et al. The Salmonella-containing vacuole is a major site of intracellular cholesterol accumulation and recruits the GPI-anchored protein CD55. Cell. Microbiol. 4, 315–328 (2002). 122. Fields, P. I., Groisman, E. A. & Heffron, F. A Salmonella locus that controls resistance to microbicidal proteins from phagocytic cells. Science 243, 1059–1062 (1989). 123. Miller, S. I. & Mekalanos, J. J. Constitutive expression of the PhoP regulon attenuates Salmonella virulence and survival within macrophages. J. Bacteriol. 172, 2485–2490 (1990). 124. Nickerson, C. A. & Curtiss, R. Role of sigma factor RpoS in initial stages of Salmonella typhimurium infection. Infect. Immun. 65, 1814–1823 (1997). 125. Bearson, B. L., Wilson, L. & Foster, J. W. A low pHinducible, PhoPQ-dependent acid tolerance response protects Salmonella typhimurium against inorganic acid stress. J. Bacteriol. 180, 2409–2417 (1998). 126. Shiloh, M. U. et al. Phenotype of mice and macrophages deficient in both phagocyte oxidase and inducible nitric oxide synthase. Immunity 10, 29–38 (1999). 127. Gunn, J. S., Ryan, S. S., Van Velkinburgh, J. C., Ernst, R. K. & Miller, S. I. Genetic and functional analysis of a PmrA–PmrB‑regulated locus necessary for lipopolysaccharide modification, antimicrobial peptide resistance, and oral virulence of Salmonella enterica serovar Typhimurium. Infect. Immun. 68, 6139–6146 (2000). 128. Vazquez-Torres, A., Jones-Carson, J., Mastroeni, P., Ischiropoulos, H. & Fang, F. C. Antimicrobial actions of the NADPH phagocyte oxidase and inducible nitric oxide synthase in experimental salmonellosis. I. Effects on microbial killing by activated peritoneal macrophages in vitro. J. Exp. Med. 192, 227–236 (2000). 129. Hisert, K. B. et al. A glutamate‑alanine‑leucine (EAL) domain protein of Salmonella controls bacterial
survival in mice, antioxidant defence and killing of macrophages: role of cyclic diGMP. Mol. Microbiol. 56, 1234–1245 (2005). 130. Humphreys, S., Stevenson, A., Bacon, A., Weinhardt, A. B. & Roberts, M. The alternative sigma factor, sigmaE, is critically important for the virulence of Salmonella typhimurium. Infect. Immun. 67, 1560–1568 (1999). 131. Rosenberger, C. M., Gallo, R. L. & Finlay, B. B. Interplay between antibacterial effectors: a macrophage antimicrobial peptide impairs intracellular Salmonella replication. Proc. Natl Acad. Sci. USA 101, 2422–2427 (2004). 132. Curtiss, R. & Kelly, S. M. Salmonella typhimurium deletion mutants lacking adenylate cyclase and cyclic AMP receptor protein are avirulent and immunogenic. Infect. Immun. 55, 3035–3043 (1987). 133. Fang, F. C. et al. The alternative sigma factor katF (rpoS) regulates Salmonella virulence. Proc. Natl Acad. Sci. USA 89, 11978–11982 (1992). 134. Lee, A. K., Detweiler, C. S. & Falkow, S. OmpR regulates the two-component system SsrA-SsrB in Salmonella pathogenicity island 2. J. Bacteriol. 182, 771–781 (2000). 135. Testerman, T. L. et al. The alternative sigma factor sE controls antioxidant defences required for Salmonella virulence and stationary-phase survival. Mol. Microbiol. 43, 771–782 (2002). 136. Prost, L. R. et al. Activation of the bacterial sensor kinase PhoQ by acidic pH. Mol. Cell 26, 165–174 (2007). References 136 and 138 show that PhoQ directly senses antimicrobial peptides and low pH. 137. Heithoff, D. M. et al. Coordinate intracellular expression of Salmonella genes induced during infection. J. Bacteriol. 181, 799–807 (1999). 138. Bader, M. W. et al. Recognition of antimicrobial peptides by a bacterial sensor kinase. Cell 122, 461–472 (2005). 139. Cho, U. S. et al. Metal bridges between the PhoQ sensor domain and the membrane regulate transmembrane signaling. J. Mol. Biol. 356, 1193–1206 (2006). 140. Gibbons, H. S., Kalb, S. R., Cotter, R. J. & Raetz, C. R. Role of Mg2+ and pH in the modification of Salmonella lipid A after endocytosis by macrophage tumour cells. Mol. Microbiol. 55, 425–440 (2005). 141. Ernst, R. K., Guina, T. & Miller, S. I. Salmonella typhimurium outer membrane remodeling: role in resistance to host innate immunity. Microbes Infect. 3, 1327–1334 (2001). 142. Guo, L. et al. Regulation of lipid A modifications by Salmonella typhimurium virulence genes phoP-phoQ. Science 276, 250–253 (1997). This study demonstrates that PhoP/PhoQ regulates the remodelling of the LPS structure in a way that makes it less immunostimulatory by TLR4 signalling. 143. Gunn, J. S. et al. PmrA‑PmrB‑regulated genes necessary for 4‑aminoarabinose lipid A modification and polymyxin resistance. Mol. Microbiol. 27, 1171–1182 (1998). 144. Guo, L. et al. Lipid A acylation and bacterial resistance against vertebrate antimicrobial peptides. Cell 95, 189–198 (1998). 145. Hilbert, F., Garcia-del Portillo, F. & Groisman, E. A. A periplasmic d‑alanyl‑d-alanine dipeptidase in the Gram-negative bacterium Salmonella enterica. J. Bacteriol. 181, 2158–2165 (1999). 146. Fang, F. C. et al. Virulent Salmonella typhimurium has two periplasmic Cu, Zn-superoxide dismutases. Proc. Natl Acad. Sci. USA 96, 7502–7507 (1999). 147. Eriksson, S., Lucchini, S., Thompson, A., Rhen, M. & Hinton, J. C. Unravelling the biology of macrophage infection by gene expression profiling of intracellular Salmonella enterica. Mol. Microbiol. 47, 103–118 (2003). 148. Cirillo, D. M., Valdivia, R. H., Monack, D. M. & Falkow, S. Macrophage-dependent induction of the Salmonella pathogenicity island 2 type III secretion system and its role in intracellular survival. Mol. Microbiol. 30, 175–188 (1998). 149. Hensel, M. et al. Simultaneous identification of bacterial virulence genes by negative selection. Science 269, 400–403 (1995). 150. Uchiya, K. et al. A Salmonella virulence protein that inhibits cellular trafficking. EMBO J. 18, 3924–3933 (1999). 151. Shotland, Y., Kramer, H. & Groisman, E. A. The Salmonella SpiC protein targets the mammalian Hook3 protein function to alter cellular trafficking. Mol. Microbiol. 49, 1565–1576 (2003).
nature reviews | microbiology
152. Freeman, J. A., Rappl, C., Kuhle, V., Hensel, M. & Miller, S. I. SpiC is required for translocation of Salmonella pathogenicity island 2 effectors and secretion of translocon proteins SseB and SseC. J. Bacteriol. 184, 4971–4980 (2002). 153. Yu, X. J. et al. SpiC is required for secretion of Salmonella pathogenicity island 2 type III secretion system proteins. Cell. Microbiol. 4, 531–540 (2002). 154. Beuzon, C. R. et al. Salmonella maintains the integrity of its intracellular vacuole through the action of SifA. EMBO J. 19, 3235–3249 (2000). 155. Ruiz-Albert, J. et al. Complementary activities of SseJ and SifA regulate dynamics of the Salmonella typhimurium vacuolar membrane. Mol. Microbiol. 44, 645–661 (2002). 156. Freeman, J. A., Ohl, M. E. & Miller, S. I. The Salmonella enterica serovar Typhimurium translocated effectors SseJ and SifsB are targeted to the Salmonella-containing vacuole. Infect. Immun. 71, 418–427 (2003). 157. Knodler, L. A. et al. Salmonella type III effectors PipB and PipB2 are targeted to detergent-resistant microdomains on internal host cell membranes. Mol. Microbiol. 49, 685–704 (2003). 158. Deiwick, J. et al. The translocated Salmonella effector proteins SseF and SseG interact and are required to establish an intracellular replication niche. Infect. Immun. 74, 6965–6972 (2006). 159. Brumell, J. H., Goosney, D. L. & Finlay, B. B. SifA, a type III secreted effector of Salmonella typhimurium, directs Salmonella-induced filament (Sif) formation along microtubules. Traffic 3, 407–415 (2002). 160. Miao, E. A. et al. Salmonella effectors translocated across the vacuolar membrane interact with the actin cytoskeleton. Mol. Microbiol. 48, 401–415 (2003). 161. Salcedo, S. P. & Holden, D. W. SseG, a virulence protein that targets Salmonella to the Golgi network. EMBO J. 22, 5003–5014 (2003). 162. Kuhle, V., Jackel, D. & Hensel, M. Effector proteins encoded by Salmonella pathogenicity island 2 interfere with the microtubule cytoskeleton after translocation into host cells. Traffic 5, 356–370 (2004). 163. Garcia-del Portillo, F., Zwick, M. B., Leung, K. Y. & Finlay, B. B. Salmonella induces the formation of filamentous structures containing lysosomal membrane glycoproteins in epithelial cells. Proc. Natl Acad. Sci. USA 90, 10544–10548 (1993). References 163 and 164 were the first to show that S. typhimurium produces Sifs in infected cultured cells and that the effector that is responsible for this activity is SifA. 164. Stein, M. A., Leung, K. Y., Zwick, M., Garcia-del Portillo, F. & Finlay, B. B. Identification of a Salmonella virulence gene required for formation of filamentous structures containing lysosomal membrane glycoproteins within epithelial cells. Mol. Microbiol. 20, 151–164 (1996). 165. Knodler, L. A. & Steele-Mortimer, O. The Salmonella effector PipB2 affects late endosome/lysosome distribution to mediate Sif extension. Mol. Biol. Cell. 16, 4108–4123 (2005). 166. Boucrot, E., Henry, T., Borg, J. P., Gorvel, J. P. & Meresse, S. The intracellular fate of Salmonella depends on the recruitment of kinesin. Science 308, 1174–1178 (2005). 167. Alto, N. M. et al. Identification of a bacterial type III effector family with G protein mimicry functions. Cell 124, 133–145 (2006). 168. Reinicke, A. T. et al. A Salmonella typhimurium effector protein SifA is modified by host cell prenylation and S‑acylation machinery. J. Biol. Chem. 280, 14620–14627 (2005). 169. Guy, R. L., Gonias, L. A. & Stein, M. A. Aggregation of host endosomes by Salmonella requires SPI2 translocation of SseFG and involves SpvR and the fmsaroE intragenic region. Mol. Microbiol. 37, 1417–1435 (2000). 170. Henry, T. et al. The Salmonella effector protein PipB2 is a linker for kinesin‑1. Proc. Natl Acad. Sci. USA 103, 13497–13502 (2006). 171. Hansen-Wester, I., Stecher, B. & Hensel, M. Type III secretion of Salmonella enterica serovar Typhimurium translocated effectors and SseFG. Infect. Immun. 70, 1403–1409 (2002). 172. Kuhle, V. & Hensel, M. SseF and SseG are translocated effectors of the type III secretion system of Salmonella pathogenicity island 2 that modulate aggregation of endosomal compartments. Cell. Microbiol. 4, 813–824 (2002).
volume 6 | january 2008 | 65 © 2008 Nature Publishing Group
REVIEWS 173. Abrahams, G. L., Muller, P. & Hensel, M. Functional dissection of SseF, a type III effector protein involved in positioning the Salmonella-containing vacuole. Traffic 7, 950–965 (2006). 174. Kuhle, V., Abrahams, G. L. & Hensel, M. Intracellular Salmonella enterica redirect exocytic transport processes in a Salmonella pathogenicity island 2‑dependent manner. Traffic 7, 716–730 (2006). 175. Birmingham, C. L., Jiang, X., Ohlson, M. B., Miller, S. I. & Brumell, J. H. Salmonella-induced filament formation is a dynamic phenotype induced by rapidly replicating Salmonella enterica serovar Typhimurium in epithelial cells. Infect. Immun. 73, 1204–1208 (2005). 176. Lesnick, M. L., Reiner, N. E., Fierer, J. & Guiney, D. G. The Salmonella spvB virulence gene encodes an enzyme that ADP-ribosylates actin and destabilizes the cytoskeleton of eukaryotic cells. Mol. Microbiol. 39, 1464–1470 (2001). 177. Browne, S. H., Lesnick, M. L. & Guiney, D. G. Genetic requirements for Salmonella-induced cytopathology in human monocyte-derived macrophages. Infect. Immun. 70, 7126–7135 (2002). 178. Gotoh, H. et al. Extracellular secretion of the virulence plasmid-encoded ADP-ribosyltransferase SpvB in Salmonella. Microb. Pathog. 34, 227–238 (2003). 179. Brumlik, M. J. & Buckley, J. T. Identification of the catalytic triad of the lipase/acyltransferase from Aeromonas hydrophila. J. Bacteriol. 178, 2060–2064 (1996). 180. Ohlson, M. B., Fluhr, K., Birmingham, C. L., Brumell, J. H. & Miller, S. I. SseJ deacylase activity by Salmonella enterica serovar Typhimurium promotes virulence in mice. Infect. Immun. 73, 6249–6259 (2005). 181. Harrison, R. E. et al. Salmonella impairs RILP recruitment to Rab7 during maturation of invasion vacuoles. Mol. Biol. Cell 15, 3146–3154 (2004). 182. Marsman, M., Jordens, I., Kuijl, C., Janssen, L. & Neefjes, J. Dynein-mediated vesicle transport controls intracellular Salmonella replication. Mol. Biol. Cell 15, 2954–2964 (2004). 183. Meresse, S. et al. Remodelling of the actin cytoskeleton is essential for replication of intravacuolar Salmonella. Cell. Microbiol. 3, 567–577 (2001). 184. Viboud, G. I. & Bliska, J. B. A bacterial type III secretion system inhibits actin polymerization to prevent pore formation in host cell membranes. EMBO J. 20, 5373–5382 (2001).
185. Marlovits, T. C. et al. Structural insights into the assembly of the type III secretion needle complex. Science 306, 1040–1042 (2004). 186. Sun, J., Hobert, M. E., Rao, A. S., Neish, A. S. & Madara, J. L. Bacterial activation of b-catenin signaling in human epithelia. Am. J. Physiol. Gastrointest. Liver Physiol. 287, G220–227 (2004). 187. Zhou, D., Mooseker, M. S. & Galan, J. E. An invasionassociated Salmonella protein modulates the actinbundling activity of plastin. Proc. Natl Acad. Sci. USA 96, 10176–10181 (1999). 188. Lee, C. A. et al. A secreted Salmonella protein induces a proinflammatory response in epithelial cells, which promotes neutrophil migration. Proc. Natl Acad. Sci. USA 97, 12283–12288 (2000). 189. Hernandez, L. D., Pypaert, M., Flavell, R. A. & Galan, J. E. A Salmonella protein causes macrophage cell death by inducing autophagy. J. Cell Biol. 163, 1123–1131 (2003). 190. Hayward, R. D. et al. Cholesterol binding by the bacterial type III translocon is essential for virulence effector delivery into mammalian cells. Mol. Microbiol. 56, 590–603 (2005). 191. Carlson, S. A., Omary, M. B. & Jones, B. D. Identification of cytokeratins as accessory mediators of Salmonella entry into eukaryotic cells. Life Sci. 70, 1415–1426 (2002). 192. Knodler, L. A., Finlay, B. B. & Steele-Mortimer, O. The Salmonella effector protein SopB protects epithelial cells from apoptosis by sustained activation of Akt. J. Biol. Chem. 280, 9058–9064 (2005). 193. Mukherjee, K., Parashuraman, S., Raje, M. & Mukhopadhyay, A. SopE acts as an Rab5-specific nucleotide exchange factor and recruits nonprenylated Rab5 on Salmonella-containing phagosomes to promote fusion with early endosomes. J. Biol. Chem. 276, 23607–23615 (2001). 194. Coombes, B. K. et al. Genetic and molecular analysis of GogB, a phage-encoded type III-secreted substrate in Salmonella enterica serovar Typhimurium with autonomous expression from its associated phage. J. Mol. Biol. 348, 817–830 (2005). 195. Brumell, J. H., Rosenberger, C. M., Gotto, G. T., Marcus, S. L. & Finlay, B. B. SifA permits survival and replication of Salmonella typhimurium in murine macrophages. Cell. Microbiol. 3, 75–84 (2001). 196. Tezcan-Merdol, D. et al. Actin is ADP-ribosylated by the Salmonella enterica virulence-associated protein
66 | january 2008 | volume 6
SpvB. Mol. Microbiol. 39, 606–619 (2001). 197. Worley, M. J., Nieman, G. S., Geddes, K. & Heffron, F. Salmonella typhimurium disseminates within its host by manipulating the motility of infected cells. Proc. Natl Acad. Sci. USA 103, 17915–17920 (2006). 198. Kujat Choy, S. L. et al. SseK1 and SseK2 are novel translocated proteins of Salmonella enterica serovar Typhimurium. Infect. Immun. 72, 5115–5125 (2004). 199. Rytkonen, A. et al. SseL, a Salmonella deubiquitinase required for macrophage killing and virulence. Proc. Natl Acad. Sci. USA 104, 3502–3507 (2007). 200. Geddes, K., Worley, M., Niemann, G. & Heffron, F. Identification of new secreted effectors in Salmonella enterica serovar Typhimurium. Infect. Immun. 73, 6260–6271 (2005). 201. Tsolis, R. M., Adams, L. G., Ficht, T. A. & Baumler, A. J. Contribution of Salmonella typhimurium virulence factors to diarrheal disease in calves. Infect. Immun. 67, 4879–4885 (1999).
Acknowledgments
We would like to thank the entire Salmonella pathogenesis community for its work that made this Review possible. In particular, we are grateful to members of the Miller laboratory, past and present, for their contributions to the ideas that are presented here. We would also like to apologize to those authors whose work was not cited owing to space limitations. A.H. is supported by a Career Development Award from the Northwest Regional Center of Excellence for Biodefense and Emerging Infectious Diseases Research (National Institute of Allergy and Infectious Diseases (NIAID) grant U54 AI057141). M.B.O. is supported by the Comprehensive Training in Inter-Disciplinary Oral Health Research T32 grant DE07132. S.I.M. is supported by the National Institutes of Health, NIAID grants R01 AI30479, R01 AI048683 and U54 AI057141 for the Northwest Regional Center of Excellence for Biodefense and Emerging Infectious Diseases Research.
DATABASES Entrez Genome Project: http://www.ncbi.nlm.nih.gov/ entrez/query.fcgi?db=geneomeprj Salmonella typhi | Salmonella typhimurium
All links are active in the online pdfF
www.nature.com/reviews/micro © 2008 Nature Publishing Group
REVIEWS
An integrated model of the recognition of Candida albicans by the innate immune system Mihai G. Netea*, Gordon D. Brown‡, Bart Jan Kullberg* and Neil A. R. Gow §
Abstract | The innate immune response was once considered to be a limited set of responses that aimed to contain an infection by primitive ‘ingest and kill’ mechanisms, giving the host time to mount a specific humoral and cellular immune response. In the mid‑1990s, however, the discovery of Toll-like receptors heralded a revolution in our understanding of how microorganisms are recognized by the innate immune system, and how this system is activated. Several major classes of pathogen-recognition receptors have now been described, each with specific abilities to recognize conserved bacterial structures. The challenge ahead is to understand the level of complexity that underlies the response that is triggered by pathogen recognition. In this Review, we use the fungal pathogen Candida albicans as a model for the complex interaction that exists between the host pattern-recognition systems and invading microbial pathogens. Innate immune system The suite of host responses to microbial invaders that results in rapid defence without requiring prior stimulation.
*Department of Medicine and Nijmegen University Centre for Infectious Diseases, Radboud University Nijmegen Medical Centre, Nijmegen, The Netherlands. ‡ Institute of Infectious Diseases and Molecular Medicine, Division of Immunology, CLS, University of Cape Town, Rondebosch 7925, South Africa. § School of Medical Sciences, Institute of Medical Sciences, University of Aberdeen, Foresterhill, Aberdeen AB25 2ZD, UK. Correspondence to M.G.N. & N.A.R.G. e-mails:
[email protected];
[email protected] doi:10.1038/nrmicro1815
The clinical spectrum of Candida spp. infections ranges from benign colonization of the skin and mucosal surfaces to mucocutaneous forms of candidiasis and systemic infections (candidaemia and deep-seated organ candidiasis). Despite the fact that fungal infections are typically self-limiting, the number of life-threatening systemic fungal infections has risen steadily over the past three decades1, although this increase might now be reaching a plateau2. The increased prevalence of fungi as agents of disseminated infection has been restricted to patients in whom surgical or chemotherapeutic interventions and/or underlying immunological deficiencies have allowed fungi to overwhelm the protective host defence mechanisms2. In healthy and immunologically normal individuals, the innate immune system is an efficient sentinel that provides protection from the thousands of fungal species that humans regularly encounter. Candida albicans is a ubiquitous fungal organism that often colonizes the skin and the mucosal surfaces of normal individuals, without causing disease. However, when the normal host defence mechanisms are impaired (for example, in patients who are undergoing chemotherapy for malignancies, receiving immunosuppressants after an organ transplant, or patients with AIDS), C. albicans can become a pathogen. Candidaemia has an incidence of between 1.1 and 24 cases per 100,000 individuals, and an associated mortality of more than 30%3,4. It seems
nature reviews | microbiology
unlikely that antifungal agents will have an impact on the grim mortality statistics for systemic fungal infections without the aid of new clinical approaches, such as combining chemotherapy with immunotherapy5. However, augmenting the ability of the immune system to eliminate a pathogen requires a sophisticated understanding of the molecular mechanisms that are involved in pathogen recognition and in the host immune response. In this Review we describe the recent progress that has been made in research into the mechanisms that are involved in the recognition of fungal pathogens, and present how this progress has changed our understanding of the pathogenesis of infections in general, and invasive candidiasis in particular.
Innate immunity and host defence Until recently, little was known about the ways in which neutrophils and macrophages, the major players in innate immunity, recognized C. albicans as a pathogenic microorganism, or how the fungal–leukocyte interaction triggers an inflammatory response. The dogma that had been blindly accepted over the past 50 years was that although effective, innate immunity was non-specific and “rather primitive and dumb” (Ref. 6). However, this simplistic model, in which innate immunity performs only simple ‘ingest and destroy’ tasks, could not explain how innate immune cells recognize microbial pathogens volume 6 | january 2008 | 67
© 2008 Nature Publishing Group
REVIEWS Box 1 | The recognition of microorganisms by pattern-recognition receptors The innate recognition of pathogens is achieved through germ-line-encoded receptors, the pattern-recognition receptors (PRRs), which sense conserved chemical signatures called pathogen-associated molecular patterns (PAMPs). Four major classes of PRRs have been identified. • Toll-like receptors (TLRs) are cell-membrane-associated (TLR1, TLR2, TLR4, TLR5 and TLR6) or intracellular (TLR3, TLR7, TLR8 and TLR9) receptors. Several TLRs have been implicated in the recognition of fungal components: TLR2 recognizes phospholipomannan, TLR4 recognizes O‑linked mannans, TLR6 is involved in the recognition of zymosan and TLR9 detects fungal DNA. • C-type lectin receptors (CLRs) are mainly membrane-bound receptors that recognize polysaccharide structures from Candida albicans: dectin 1 recognizes β-glucans, whereas the macrophage mannose receptor (MR) and DC‑SIGN recognize N‑linked mannans. • Nucleotide-binding oligomerization domain (NOD)-like receptors (NLRs) and retinoic-acid-inducible gene I (RIGI) receptors are intracellular receptors for bacterial peptidoglycans and viral nucleic acids. So far, no studies have documented the involvement of NLRs or RIGI receptors in the recognition of fungi. • The blueprint for the structure of PRRs is represented by the extracellular pathogen-recognition domain (the Leu-rich repeat (LRR) domain in TLRs and the C‑type lectin domain (CLD) in CLRs). The intracellular signalling domain (the TLRinterleukin 1 receptor, TIR domain) of TLRs and the immunoreceptor Tyr-based activation-like motif (ITAM) from some CLRs, such as dectin 1, ultimately induce the intracellular signals responsible for the functional activity of the receptor. The four major classes of PRRs discussed here are restricted to families of mammalian molecules with a PRR function at a cellular level that have a proven signalling pathway that leads to activation of the innate host response. There are other classes of molecules with less well established functions as PRRs. (For example, peptidoglycanrecognition proteins (PGRPs), which recognize peptidoglycans in insect and mammalian cells. Although PGRPs induce defensins in Drosophila melanogaster, their function in mammalian cells has not yet been established.) Some circulating proteins that bind bacterial structures (for example, pentraxins and mannose-binding lectin (MBL)), are also considered by some to be PRRs.
as ‘non-self ’, or why different responses are triggered by different classes of microorganisms. Only in the past decade has it become clear that the innate immune system not only specifically recognizes various classes of microorganisms, it also initiates and modulates the subsequent adaptive responses that are delivered by Tcells and B cells through their interactions with antigenpresenting cells (especially dendritic cells (DCs))7. The tasks of recognizing an invading pathogen and activating the host response are accomplished by patternrecognition receptors (PRRs), which recognize conserved microbial chemical signatures called pathogen-associated molecular patterns (PAMPs) (BOX 1).
Dendritic cells ‘Professional’ antigenpresenting cells that are found in the T-cell areas of lymphoid tissues and as minor cellular components in most tissues. They have a branched or dendritic morphology and are important stimulators of T-cell responses.
The fungal cell wall The fungal cell wall is an essential organelle that maintains the viability of fungal cells. The regulation of fungal cell wall biosynthesis and glycosylation has been extensively reviewed elsewhere8–10, but a basic outline of the components of the cell wall is necessary to understand how it is recognized by the host immune system. To be strong, yet plastic, fungal cell walls combine skeletal and matrix components in a way that resembles the engineering principles that are involved in constructing elaborate structures, such as reinforced concrete buildings, that are made of mesh and mortar. The cells of the innate immune system recognize elements of both the skeletal and matrix components of the C. albicans cell wall. The skeletal component of the cell wall of the majority of fungal pathogens, including C. albicans, is based on a core structure of β-(1,3)-glucan covalently linked to β-(1,6)-glucan and chitin (a β-(1,4)-linked polymer of N‑acetylglucosamine (GlcNAc)) (FIG.1). These polymers form hydrogen bonds between adjacent polysaccharide
68 | january 2008 | volume 6
chains to form a tough three-dimensional network of microfibrils. Most models suggest that the skeletal components of the cell wall are found close to the cell membrane in an inner layer, although some chitin and glucan can be present throughout the thickness of the wall. In budding yeast cells, a scar is left on the mother cell after separation of the bud, and at this site the components of the inner layers of the cell wall, such as chitin and β-(1,3)-glucan, can become exposed at the surface11. In addition to the glucan and chitin skeleton, the C. albicans cell wall contains a matrix that mainly comprises glycosylated proteins. In C. albicans, the major class of cell wall proteins are glycosylphosphatidylinositol (GPI)-anchor-dependent cell wall proteins (GPI-CWPs), which are attached through a GPI remnant to β-(1,3)glucan or chitin by a highly branched β-(1,6)-glucan linker. The CWPs are normally highly glycosylated with mannose-containing polysaccharides (sometimes called mannan), and carbohydrates can account for up to 90% of their molecular mass. Many CWPs have a lollipop structure with a globular domain that is presented to the outside of the cell and a Ser/Thr-rich polypeptide stem-like domain that is stabilized in the cell wall by O‑linked mannan side chains 8. These ether-linked O‑mannans are relatively short, linear polysaccharides that, in C. albicans, consist of one to five mannose (Man) sugars that are almost exclusively α-(1,2)-linked (FIG. 1). N‑mannan consists of a core Man8GlcNAc2 triantennary complex to which a highly branched structure is attached, comprising up to 150 mannose sugars arranged as an α-(1,6)-linked backbone with side chains of α(1,2)-, α-(1,3)-mannose and phosphomannan9 (FIG. 1). The mannan structures define many of the serotypes of Candida spp.12 For example, in serotype B strains www.nature.com/reviews/micro
© 2008 Nature Publishing Group
REVIEWS Mannan Cell wall protein β-(1,6)-glucan β-(1,3)-glucan Chitin
N-acetylglucosamine α-(1,3)-mannose α-(1,2)-mannose α-(1,6)-mannose β-(1,2)-mannose β-(1,4)-mannose β-(1,6)-glucose β-(1,3)-glucose
P
n NH
O
Asn X Ser/Thr
Ser/Thr
Pir-CWP
CWP
Figure 1 | The structure of the Candida albicans cell wall. The schematic shows the major components of the cell wall and their distributions. β-(1,3)-glucan and chitin (poly-β-(1,4)-N‑acetylglucosamine) are the main structural components, Nature Reviews | Microbiology and these are located towards the inside of the cell wall. The outer layer is enriched with cell wall proteins (CWP) that are attached to this skeleton mainly via glycosylphosphatidylinositol remnants to β-(1,6)-glucan or, for mannoproteins with internal repeat domains (Pir-CWP), via alkali-sensitive linkages to β-(1,3)-glucan. The insets show the structure of the glucan and mannan components.
β-(1,2)-mannose is found exclusively in the phosphomannan, whereas in serotype A strains β-(1,2)-mannose also occurs in the acid-stable side-chain fraction9. Cytokines Biologically active molecules that are released by cells and that affect the function of other cells.
Fcg receptor A surface molecule on various cells that binds to the Fc regions of immunoglobulins, thereby initiating effector functions.
T helper 1 An immune response that is characterized by a subset of helper T-cells that secrete a particular set of cytokines, including interleukin 2, interferon-γ and TNFα, the main function of which is to stimulate phagocytosismediated defences against intracellular pathogens.
C-type lectin C-type lectins are largely calcium-dependent animal lectins that are carbohydratebinding proteins. The binding activity of C-type lectins is based on the structure of the carbohydrate-recognition domain, which is highly conserved in this family.
Fungal recognition Owing to the localization of mannoproteins and mannans in the outermost part of the cell wall, mannan detection would be expected to be one of the first steps in the recognition of C. albicans by the host innate immune system. However, the presence of β‑glucans and chitin, especially at the level of the bud scar, is also likely to influence the recognition of C. albicans by leukocytes. So how is the recognition of the C. albicans cell surface achieved by the immune system? The main cells of the host innate immune response that recognize invading pathogens are monocytes and neutrophils in the circulation, together with macrophages in infected tissues. Monocytes express high levels of Toll-like receptors (TLRs) on their cell membranes, as well as moderate levels of lectin receptors (LRs). During differentiation into macrophages, they retain expression of TLRs while strongly upregulating their expression of LRs. Important differences in the expression of several receptors have also been reported between resident and inflammatory macrophages, as the expression of PRRs can be modified by cytokines or microbial products. DCs, which are crucial for antigen processing and presentation, also express most of the PRRs that are important for the recognition of fungal pathogens. Neutrophils show moderate expression of TLRs and strongly express phagocytic receptors such as complement receptor 3 (CR3) and Fcγ receptors (FcγRs), but the expression of PRRs on T cells is much more restricted (FIG. 2). The mosaic of PRRs that is expressed by each of these cell
nature reviews | microbiology
types ultimately determines the type of response that is initiated following recognition of C. albicans. Mannans and mannoproteins. Both mannans and mannoproteins from the C. albicans cell wall have important immunostimulatory activities, ranging from stimulation of cytokine production13,14 to induction of DC maturation15 and T-cell immunity16,17. Mannoproteins induce mainly T helper 1 (TH1)-type cytokine profiles, which have protective effects against disseminated C. albicans infection18. The first receptor on the surface of macrophages to be described as a mannan receptor was the C‑type-lectin mannose receptor (MR)19,20. The MR recognizes oligosaccharides that terminate in mannose, fucose and GlcNAc21, and this binding is mediated by carbohydrate-recognition domains (CRDs) 4 to 8 in the extracellular region of the receptor22. In vitro studies have shown that the MR preferentially recognizes α-linked oligomannoses with branched, rather than linear, structures23, and this was supported by data that demonstrated that in monocytes and macrophages, the MR recognizes branched N‑bound mannans from C. albicans24. By contrast, recognition of the shorter linear structures of O‑bound mannan is performed by TLR4 (Ref. 24), and results in cytokine production25. Interestingly, TLR4 stimulation is lost during the germination of yeast into hyphae, which leads to a loss of interferon-γ (IFNγ) production capacity26. On DCs, however, the binding of C. albicans mannans is mediated through the MR and DC‑SIGN (DC‑SIGN is a receptor that is specifically expressed on the cell membrane of DCs)27. The α-linked mannose structures on the surface of C. albicans are recognized by the MR, TLR4 and volume 6 | january 2008 | 69
© 2008 Nature Publishing Group
REVIEWS Monocyte TLR4
Dectin 1
TLR2
MR
TLR6
Neutrophil TLR2
Macrophage
TLR9
TLR6
TLR2
TLR4
Dectin 2
CR3
Dectin 1 TLR9
Galectin 3 Dectin 1
FcγR
TLR4 FcγR
Dendritic cell CD4
TLR2
TLR2 TLR9 TLR4
MR CR3
Dectin 2 Dectin 1
DCSIGN
TLR9
MR
TLR4 FcγR
CR3
Figure 2 | Cell populations and pattern-recognition receptors involved in Nature Reviews | Microbiology Candida albicans recognition. The main populations involved in the recognition of C. albicans during the innate immune response are the monocytes, neutrophils and macrophages. Dendritic cells are crucial for processing of, and antigen presentation to, T cells, and thus to activation of specific immunity. The differential expression of patternrecognition receptors by these cell types is shown. CR3, complement receptor 3; FcγR, Fcγ receptor; MR, mannose receptor; TLR, Toll-like receptors.
Chemokines Small, inducibly secreted proteins that induce the activation and migration of lymphocytes.
Complement A part of the innate immune system that comprises serum proteins that can protect against infection.
DC‑SIGN. By contrast, β-(1,2)-mannosides, which are present in the acid-stable and acid-labile components of mannoproteins and in phospholipomannan (PLM), are recognized by other mechanisms. On the one hand, PLM reportedly stimulates cytokine production through TLR2 (Ref. 28). On the other hand, it has been suggested that recognition of the acid-labile β-(1,2)-mannosides is redundant in human monocytes24, although it might play a role in tissue macrophages, particularly in the gut mucosa. In this respect, a recent study has shown that galectin 3 on the surface of murine macrophages can discriminate between pathogenic C. albicans and non-pathogenic Saccharomyces cerevisiae, and that an association between galectin 3 and TLR2 is involved in this process29. Another lectin family member, dectin 2, has also been described to function as a receptor for C. albicans mannans30, although this interaction is less well characterized. Dectin 2 is mainly expressed on myeloid cells and maturing inflammatory monocytes31. Owing to its short intracytoplasmic tail, dectin 2 must interact with the FcγR to induce intracellular signals and, interestingly, seems to be mainly involved in the recognition of C. albicans hyphae32. Dectin 2 is encoded by a different gene cluster to the β-glucan receptor dectin 1 (see below) and does not recognize β‑glucans32; rather, dectin 2 has binding specificity for mannose33.
β-glucans. The C. albicans cell wall consists of approximately 60% β-glucan34. Although initially thought to be buried beneath the mannoprotein layer, recent evidence
70 | january 2008 | volume 6
suggests that β-glucans are exposed on the cell surface35, although possibly restricted to specific regions, such as bud scars11. β‑glucans can stimulate leukocytes in vitro, which induces cytotoxic and antimicrobial activities as well as the production of pro-inflammatory mediators, cytokines and chemokines36. β‑glucans are released into the circulation during systemic fungal infections37. The recognition of β‑glucans has primarily been ascribed to two receptors, CR3 and, more recently, dectin 1. Although other β‑glucan receptors have been described, including lactosylceramide and scavenger receptors, the physiological role of these receptors is still unclear. CR3 is a widely expressed β2-integrin that recognizes several endogenous and exogenous ligands, pathogens that have been opsonized by iC3b (the inactivated form of complement component C3b) and carbohydrates, including β‑glucans. Carbohydrate recognition is mediated by a lectin domain38,39, which is distinct from the normal ligand-binding site (the I domain) of CR3 (Ref. 39). The lectin domain mediates recognition of both the yeast and hyphal forms of C. albicans40,41, as well as several other fungi. Recognition by CR3 does not trigger protective host responses, such as the respiratory burst42, and can repress pro-inflammatory signals43,44. Dectin 1 is a myeloid-expressed transmembrane receptor and possesses a single extracellular, nonclassical C‑type-lectin-like domain that specifically recognizes β-(1,3)-glucans. Dectin 1 can recognize several fungi, including C. albicans yeast, although it does not appear to recognize C. albicans hyphae11,45. The cytoplasmic tail of dectin 1 contains an immunoreceptor tyrosine-based activation-like motif (ITAM), which can mediate various protective responses through spleen tyrosine kinase and caspase recruitment domain protein 9 (Syk–CARD9)-dependent pathways, such as the stimulation of interleukin 2 (IL-2) and IL‑10 (Ref. 46), IL‑6 (Ref. 47), and IL‑17 production48. Although Sykdependent signalling from dectin 1 is sufficient for these responses, stimulation of the mitogen-activated protein kinase (MAPK) and nuclear factor (NF)-κB pathways, with subsequent production of pro-inflammatory cytokines, such as tumour necrosis factor (TNF), requires collaborative signalling with the TLR2 receptor49,50. The precise nature of the TLR2 ligand from C. albicans has not yet been completely elucidated, although PLM might be recognized by TLR2 and TLR6 (Ref. 28). It has recently been suggested that phagocytosis of C. albicans by neutrophils can be mediated by the minor cell wall component β-(1,6)-glucan51. Beads coated with β-(1,6)-glucan are ingested by neutrophils, and the phagocytosis of yeast cells treated with β-(1,6)-glucanase is reduced. This recognition appears to be mediated by CR3 following opsonization by C3d fragments that bind β-(1,6)-glucan. Other C. albicans components. In addition to manno proteins, mannans and β‑glucans, other structures of C. albicans can also be recognized as fungal PAMPs. Chitin, a less studied polysaccharide component of the C. albicans cell wall, is a β-(1,4)-linked homopolymer of GlcNAc that forms antiparallel hydrogen-bonded chains www.nature.com/reviews/micro
© 2008 Nature Publishing Group
REVIEWS called microfibrils52. Recently, it has been demonstrated that chitin induces recruitment of immune cells that mainly release IL‑4 and IL‑13 (Ref. 53). Little is known about the recognition pathways of chitin and its role during C. albicans infections, and no recognition receptors for chitin have yet been described. Bacterial and fungal DNA are poorly methylated, in contrast to mammalian DNA. This difference has been proposed to be instrumental in the recognition of nonself DNA by TLR9 (Ref. 54). Although the recognition of fungal DNA has not been formally demonstrated, the involvement of TLR9 in the recognition of C. albicans is supported by the observation that cytokine production from CD4+ T-cells from TLR9–/– mice is skewed (higher IL‑4, lower IFNγ) compared to cytokine production from CD4+ T cells from wild-type mice, upon challenge with C. albicans yeast55. As yet, no studies have investigated the possible role of RNA recognition systems (that is, TLR3, and TLR7 and TLR8) in the host response to C. albicans infection.
Activation of host defence by PRRs At first glance, ligation of the various host immune receptors by C. albicans PAMPs appears to lead to a set of standard, and possibly redundant, pathways that stimulate cytokine production, phagocytosis and fungal killing. However, this early model of a ‘standard’ PRR response has been refined over the past few years. Various PRRs enable the innate immune system not only to recognize specific PAMPs, but also to specifically modulate the response that follows. In addition, by inducing specific cytokine profiles, PRRs bring a certain degree of specificity to the innate response. Type I interferons Interferons that are rapidly induced by virus replication, as well as by some bacterial and fungal infections.
TH2
A type of activated T helper cell that participates in phagocytosis-independent responses and that downregulates proinflammatory responses that are induced by TH1 cells. TH2 cells secrete interleukin 4 (IL-4), IL-5, IL-6 and IL-10.
Regulatory T-cell A small population of CD4+ T-cells that expresses the transcription factor forkhead box P3 and that has suppressor activity towards other T-cells either by cell–cell contact or cytokine release.
Zymosan Particulate preparation of S. cerevisiae cell walls that is rich in β-glucans and mannan and that can be used to activate the innate immune system.
C. albicans uptake. Phagocytosis of C. albicans is mediated by the concerted action of several opsonic and nonopsonic receptors. Complement binding and activation is mediated by the alternative pathway, and complement activation is mainly important for the chemotaxis and opsonization of C. albicans, but not for C. albicans lysis, which is prevented by the thick and complex cell wall56. Furthermore, although mannose-binding lectin does bind and recognize C. albicans, the lectin pathway of complement activation probably plays only a minor role in C. albicans uptake57. Several membrane-bound receptors contribute to the phagocytosis of C. albicans, among which dectin 1 (Ref. 58), the MR59,60 and DC‑SIGN27 have been directly shown to mediate uptake of fungal particles in transfected cell systems, although the ability of the MR to mediate phagocytosis has recently been questioned61. TLRs do not mediate fungal uptake, but they might be involved in directing the subsequent maturation of the phagosome62 and the presentation of antigens, although this is still controversial62. C. albicans killing. Following uptake, killing of C. albicans occurs through both oxidative and nonoxidative mechanisms63. Although the receptors that are involved in triggering these events are largely unknown, dectin 1 induces the respiratory burst in response to
nature reviews | microbiology
fungi, an activity that can be enhanced by TLR signalling50,64. The respiratory burst is an essential antifungal effector mechanism that results in the production of toxic oxidants65,66 and in the activation of granule proteases that can kill C. albicans65–67. Killing of C. albicans also occurs extracellularly, through the as-yet-undefined actions of PRRs such as galectin 3 (Ref. 68). Cytokine induction. More is known about the receptors that are involved in the induction of cytokine production by C. albicans. At least four TLRs (TLR2, TLR4, TLR6 and TLR9) are involved in triggering these responses. Upon recognition of microbial structures, TLRs activate either the NF‑κB or MAPK pathway, which lead to the stimulation of pro-inflammatory cytokine production69. The balance between the signals that are induced by TLR2 and TLR4 seems to have a crucial role in the regulation of the immune response. TLR4 can strongly stimulate pro-inflammatory cytokines through two pathways: one involves the myeloid differentiation primary response gene 88 (MyD88)–Mal-mediated induction of the NF‑κB-dependent release of pro-inflammatory cytokines and chemokines; the other involves the interferon regulatory factor 3 (IRF3)-dependent release of type I interferons70, which induce the secondary production of TH1-type cytokines such as IFNγ 26,55,71. By contrast, several studies have demonstrated that although TLR2 ligation can induce pro-inflammatory cytokines, this effect is weaker than that mediated by TLR4 ligation72. Moreover, TLR2 ligands fail to induce the release of IL‑12 and TH1-type IFNγ, thus promoting conditions that are favourable for TH2- or regulatory T cell (TReg)-type responses73. Stabilization of the transcription factor c‑Fos, which is a suppressor of IL‑12 expression, is an important step in this process74. In support of this, a recent in vitro study has reported that zymosan induces the development of a tolerigenic DC population through TLR2 and dectin 1 (Ref. 75). The induction of this tolerigenic cytokine profile depends on extracellular signal related kinase (ERK)/MAPK phosphorylation, a mechanism that is distinct from the p38/Jun N‑terminal kinase (JNK) pathway that is induced by TLR4 stimulation75,76. Also, in vivo studies have shown that knockout mice that lack TLR2 are more resistant to disseminated candidiasis, and this is accompanied by a TH1 bias55,77. C. albicans induces immunosuppression through TLR2-mediated IL‑10 release, and this leads to the generation of CD4+CD25+ TReg cells with immunosuppressive potential77,78. Similar data have been reported in models of Schistosoma mansoni and Borrelia burgdorferi infection79,80. The role of TLR6 and TLR9 in the induction of cytokines is less prominent. Although TLR2–TLR6 heterodimers are involved in the recognition of zymosan81, cytokine production is only moderately reduced in TLR6–/– macrophages that are stimulated with C. albicans, and this does not result in increased susceptibility to disseminated candidiasis82. Similar to TLR2, the absence of TLR9 also results in a slight shift of the cytokine production that is induced by C. albicans towards a more anti-inflammatory profile55. volume 6 | january 2008 | 71
© 2008 Nature Publishing Group
REVIEWS
Tetraspanin A family of transmembrane proteins that have four transmembrane domains and two extracellular domains of different sizes. The function of tetraspanins is unclear, but they seem to interact with many other transmembrane proteins and form large multimeric protein networks.
TH17 response
The TH17 subpopulation are T-cells that release mainly IL-17, which has both neutrophil chemotactic activity and the potential to promote autoimmunity.
Of the non-TLR receptors, dectin 1, dectin 2 and the MR have been implicated in initiating responses to C. albicans. Not much is known about dectin 2, but this receptor does preferentially bind C. albicans hyphae and induces the production of TNFα and IL-1 receptor antagonist (IL-1Ra)32. Dectin 1 induces the production of numerous cytokines and chemokines in response to fungi, including TNFα, macrophage inflammatory protein 2 (MIP2), macrophage inflammatory protein 1α (MIP1α), granulocyte–macrophage colony-stimulating factor (GM-CSF), granulocyte colony-stimulating factor (G-CSF), IL‑10, IL‑2, IL‑1α, IL‑1β, IL‑23 and IL‑6 (Refs 46,50,83–86). Whereas the production of IL‑10, IL‑6 and IL‑2 can be mediated by dectin 1 directly87, the induction of pro-inflammatory cytokines and chemokines requires collaborative signalling with TLR2 (Refs 50,83), although evidence suggests that dectin 1 is sufficient for these responses in certain cell types, such as alveolar macrophages85. By contrast, the interaction of dectin 1 with the tetraspanin CD37 seems to inhibit the function of dectin 1 and stimulation of IL‑6 (Ref. 88). The intracellular pathways that are activated by dectin 1 for the induction of cytokines involve the recruitment of Syk46 and downstream signalling via CARD9 (Ref. 89). The dectin 1–Syk–CARD9 pathway induces DC maturation and directs TH17 responses that are independent of the interaction with TLRs48, and this effect has been particularly linked with hyphal stimulation of DCs90. The role of IL‑17 in the host response to disseminated candidiasis has been suggested to be protective, by inducing neutrophil recruitment91, whereas it exerts a predominantly inflammatory pathological effect in gastric C. albicans infections92. In addition to its mediation of the dectin 1–Syk pathway, CARD9 is also a central adaptor molecule that mediates the pro-inflammatory signals that are induced by other classes of receptors, such as the nucleotide-binding oligomerization domain (NOD)-like receptors93, ITAM-associated receptors and possibly the TLRs94. Recently it has also been suggested that TRIF-dependent pathways modulate the balance between deleterious TH17 responses and protective TReg cells in mice with gastric candidiasis95. The receptors and intracellular pathways that are involved in the recognition of C. albicans by leukocytes are depicted in FIG. 3. Although all of these pathways have been demonstrated to be involved in the recognition of fungal components, it has not been verified that identical mechanisms are responsible for the recognition of intact, live C. albicans cells. For example, zymosan has been used extensively for immune stimulation, but the relationship between the response to zymosan and the response to native fungal cells has not yet been clarified. The specificity of the antifungal response is evident in differences between the kinetics of the various immunological effects and in the magnitude of each of these responses. Epithelial and endothelial surfaces also have PRRs that recognize surface PAMPs of invading and colonizing microorganisms such as C. albicans. Studies of the role of epithelial cell immunity have been significantly enhanced by the availability of model epithelial and endothelial tissues in which a cell line of keratinocytes is grown on the
72 | january 2008 | volume 6
surface of a polycarbonate support membrane. Immune cells such as polymorphonuclear leukocytes (PMNs) can be added to such ex vivo reconstituted human epithelial (RHE) tissue cultures along with an inoculum of fungal cells. The expression of TLRs in RHE models is similar to the TLR expression profile in vivo96. Using such approaches, PMNs have been shown to be active in the innate immune response at these primary barriers to systemic invasion97, and differences have been noted between the local immunity of the oral and vaginal mucosae97,98. For example, PMNs enhance an inflammatory TH1 response in the oral epithelia, which results in the induction of proinflammatory cytokines, such as IFNγ and TNFα, effectively mimicking the situation that is observed in vivo97. Vaginal epithelial cells express TLR2 and TLR4 in vivo, and the presence of heat-inactivated C. albicans cells and zymosan induces these cells to secrete inflammatory cytokines, chemokines and to produce β-defensin 2 — an antimicrobial compound that also induces PMN chemotaxis99. Work to characterize fungal PAMPs and PRRs that induce local innate responses is ongoing. For example, it has recently been shown that human epithelial TLR4 is directly involved in the PMN-dependent protection of oral mucosa96.
In vivo models in TLR- and LR‑deficient mice In vitro stimulation models that use either cell lines or primary cells have suggested specific roles for the various PRRs in the recognition of C. albicans, as detailed above. However, whether these pathways are indeed important for the antifungal host response in vivo can only be demonstrated in experimental models of infections in genetically modified mice that lack particular receptors, or in patients with specific immune defects. Since the first observation that TLRs recognize C. albicans and activate the innate immune response74, additional studies have confirmed the important role that TLRs have in the recognition of C. albicans and the anti-candidal host response. A global role for TLRs in defence against disseminated candidiasis was demonstrated by the increased susceptibility of MyD88–/– mice to C. albicans infection55,100,101. A subsequent study reported increased susceptibility of TLR2–/– mice to disseminated candidiasis, and this effect was attributed to decreased production of TNF and chemokines102. However, the increased susceptibility of TLR2–/– mice to infection with C. albicans was not confirmed in two other studies, which found reduced mortality and a decreased fungal burden in TLR2–/– mice, accompanied by decreased production of IL‑10 and increased production of IL‑12 and IFNγ 55,77. An additional in vitro study also confirmed the increased ability of macrophages from TLR2–/– mice to contain C. albicans103. These latter studies55,77,103 suggest that TLR2, by inducing mainly an anti-inflammatory response, inhibits the innate immune response to C. albicans infections in certain circumstances. It has also been suggested that TLR4 is involved in the recognition of C. albicans24,25,71, and TLR4-defective C3H/HeJ mice have been reported to be more susceptible to disseminated candidiasis71. However, other studies www.nature.com/reviews/micro
© 2008 Nature Publishing Group
REVIEWS Monocytes/macrophages
Dendritic cells
Mannan β-Glucans (O-linked) PLM? TLR4
IL-8 TNF
TGFβ IL-10
Mannan (O-linked)
Chitin R? Dectin 2 FcγR
Syk
IκB NF-κB
TNF
+
CARD9
NF-κB NF-AT
IL-10 IL-2
TReg
TNF IL-1β
IL-4
TNFα
β-Glucans Mannan β-Glucans (N-linked) PLM?
TLR4
+
NF-κB
NF-κB
MR
Mannan (N-linked)
TRIF
Mal
Mal
IκB p38/JNK NF-κB
ERK
IRF3
NF-κB
Type I IFN
TNF
TH2
TH1
TLR2
MyD88
ERK
Dectin 1
Mannan (N-linked) Chitin
MyD88
Mal
β-Glucans
TRAM
IκB NF-κB
TLR6 TLR2 + Galectin 3
MyD88
MyD88
Mal
β-Mannosides
Dectin 1
Mannan (N-linked)
MR
DCSIGN
Syk
CARD9
c-Fos
IL-12 IFNγ
TGFβ IL-10
IL-23 IL-6
TH1
TReg
TH17
TNF IL-1β
IL-10
Figure 3 | Recognition of Candida albicans at the membrane level. Recognition of C. albicans at the level of the cell membrane is mediated by TLRs and LRs. TLR4 induces mainly pro-inflammatory signals in monocytic cell types Nature Reviews | Microbiology (monocytes, macrophages and DCs) through the MyD88–Mal-mediated NF‑κB and MAPK pathways, while stimulating TH1 responses through IRF3-dependent mechanisms mainly in plasmacytoid DCs. TLR2 stimulates the production of moderate amounts of pro-inflammatory cytokines and strong IL‑10 and TGFβ responses. On the one hand, this leads to the induction of a tolerant phenotype in DCs, through an ERK/MAPK-dependent mechanism121. On the other hand, in monocytes and macrophages it induces TGFb and IL-10, and subsequent proliferation of TReg cells and immunosuppression78. The proinflammatory effects of TLR2 can be amplified by dectin 1 and galectin 3 — the latter especially in macrophages. In addition to amplifying the effects of TLR2, the non-classical lectin-like receptor dectin 1 induces IL‑2, IL‑10 and TH17 responses through a Syk–CARD9 cascade, independently of its interaction with TLR2. The classical lectin-like receptor, the MR, induces pro-inflammatory effects in monocytes and macrophages, whereas chitin-dependent stimulation of these cells induces mainly TH2 responses122, although this effect has yet to be demonstrated for C. albicans, and the identity of the chitin receptor is unknown. Other less well characterized pathways include stimulation of TNF and IL‑1Ra by dectin 2, while engagement of DC-SIGN in DCs induces production of the immunosuppressive cytokine IL‑10 . CARD9, caspase recruitment domain protein 9; DC, dendritic cell; ERK, extracellular signal related kinase; FcγR, Fcγ receptor; IL, interleukin; IL‑1Ra, interleukin‑1 receptor antagonist; IFNγ, interferon-γ; IRF3, interferon regulatory factor 3; JNK, Jun N‑terminal kinase; LR, lectin receptor; MAPK, mitogen-activated protein kinase; MR, mannose receptor; MyD88, myeloid differentiation primary response gene 88; NF‑κB, nuclear factor κB; PLM, phospholipomannan; Syk, spleen tyrosine kinase; TGFβ, transforming growth factor-β; TH, T helper; TLR, Toll-like receptor; TNF, tumour necrosis factor; TReg, regulatory T-cell.
have observed variable results, with TLR4–/– mice being more susceptible in models of intragastric infection or intravenous re-infection 55, not different from wildtype animals in models of intravenous infection with C. albicans yeast104, or even surviving longer in a model of intravenous infection with C. albicans hyphae55. The differences between the experimental models and/or the C. albicans strains used are thought to be responsible for these differences. Although no increased susceptibility of TLR9–/– mice to disseminated candidiasis has been observed, the fungal burden in the organs of TLR9-deficient animals tends to be lower than in control mice55. Although pro-inflammatory cytokine production was slightly lower in TLR9–/– macrophages that had been stimulated with C. albicans, this mild difference did not result in nature reviews | microbiology
a deleterious outcome of infection, probably owing to compensatory pathways that are mediated either by other TLRs or by LR‑dependent mechanisms. The study of the role of LRs in the in vivo response to C. albicans infections is still in its infancy. Recently, it has been shown that dectin 1–/– mice are more susceptible to C. albicans infection, and infection results in increased fungal growth in the organs and accelerated death105. This is consistent with the increased susceptibility of CARD9–/– mice to disseminated candidiasis89. Notably, all of the major components of antifungal defence — cytokine release, phagocyte recruitment, phagocytosis and microbial killing — are impaired in the knockout mice, which suggests an important role for dectin-1mediated recognition of β-glucans for anti-candidal defence in vivo105. However, these results are at odds with volume 6 | january 2008 | 73
© 2008 Nature Publishing Group
REVIEWS a
C. albicans yeast TLR2
c
C. albicans hyphae Dectin 1
CR3
FcγR
TLR4
Mal
Syk β-Glucans shielded from dectin 1 recognition
NF-κB
C. albicans yeast
MyD88
ERK
Mal
MyD88
MyD88
Mal
TLR4
b
NF-κB
c-Fos
Anti-inflammatory signals
Pro-inflammatory signals
Pro-inflammatory signals
Figure 4 | Candida albicans mechanisms to escape the innate response using pattern-recognition receptors. Nature Reviews | Microbiology a | The balance between the signals that are induced by Toll-like receptor 2 (TLR2) and TLR4 seems to have a crucial role in the regulation of the immune response. TLR2-mediated signals are mainly anti-inflammatory and can be deleterious when activated prematurely or excessively. TLR4-mediated signals are mainly pro-inflammatory. b | Shielding of the β‑glucans by mannans in hyphae prevents activation of the dectin 1 signalling pathways. c | Inhibitory signals from complement receptor 3 (CR3) and Fcγ receptor II (FcγRII)/FcγRIII have inhibitory effects on the activation of the immune system via TLR4. ERK, extracellular signal related kinase; MyD88, myeloid differentiation primary response gene 88; NF‑κB, nuclear factor κB; Syk, spleen tyrosine kinase, TLR, Toll-like receptor.
another study in dectin 1–/– mice, which failed to discern increased susceptibility to candidiasis106. Differences in the experimental host or between different C. albicans strains might account for these discrepancies, and additional studies are needed to pinpoint the origin of these differences. Studies on the in vivo roles of other LRs in the response to C. albicans defences are also scarce. In one study, intraperitoneal administration of C. albicans resulted in an increased fungal burden in the organs of MR–/– mice, compared to controls, although survival was not affected107. However, intraperitoneal injection of C. albicans is not an established experimental model of invasive candidiasis, as it mainly induces peritonitis rather than a disseminated infection108. A recent study has demonstrated an important role for galectin 3 in the recognition of C. albicans in the gut29, with β-(1,2)mannosides able to block colonization and invasion of C. albicans in the intestines9. Finally, like mice defective in TLR2, mice defective in CR3 or FcγR are more resistant to disseminated candidiasis, and display more vigorous cytokine release and antifungal killing, in line with the immunosuppressive effect of these receptor pathways on the inflammatory response60. It has become clear that PRRs are important for the host response to C. albicans, with various TLRs and LRs having distinctive roles in innate immunity: some receptors have a more pro-inflammatory role (for example, TLR4, dectin 1 and the MR), whereas others exert immunosuppressive effects (for example, TLR2, CR3 and FcγR) (FIGS 3,4). In addition, the choice of experimental model and the C. albicans strain used might have an important impact on the outcome of infection. More work must 74 | january 2008 | volume 6
still be done to fully understand the role of PRRs in the various types of C. albicans infections in vivo. Studies of disseminated candidiasis are more advanced, but little work has been done in other forms of C. albicans infection, such as oropharyngeal or vaginal candidiasis, and little attention has been paid to the mechanisms that lead to colonization rather than infection.
Escape mechanisms based on PRRs As evidence supporting the important role of PRRs in the recognition of C. albicans for the induction of an immune response accumulates, data have also emerged to show that certain fungal pathogens have evolved strategies to curtail recognition by these receptors, or to use it to their own advantage to evade the host response. For example, it was recently shown in Drosophila melanogaster that Gram-negative binding protein 3 (GNBP3) acts as a dectin-1-like lectinrecognition mechanism for fungal glucans, including that of C. albicans109. Certain fungal insect pathogens secrete proteases that degrade GNBP3 to avoid detection, but the D. melanogaster persophone protease is activated by the fungal protease virulence factor, which results in the bypass of GNBP3 and the activation of the Toll pathway downstream of the PRR. It remains to be seen whether there is a resonance of this co-evolution between recognition mechanisms and virulence factors in human–fungal interactions. In the mammalian host, inappropriate or premature activation of immunomodulatory receptors can be used as an escape route by C. albicans, as detailed above for the outcome of systemic candidiasis in TLR2–/– mice (FIG. 4a). Modulation of DC function can also be a target www.nature.com/reviews/micro
© 2008 Nature Publishing Group
REVIEWS for escape mechanisms, with the MR and DC‑SIGN possibly inducing anti-inflammatory cytokine production and immunosuppression110,111. Thus it seems that the MR might have a dual role, either pro-inflammatory (in monocytes and macrophages) or immunosuppressive (in DCs), although this assumption is still controversial. In addition, the recognition of β‑glucans by CR3 does not trigger protective host responses, such as the respiratory burst42, and can repress pro-inflammatory signals44 (FIG. 4). In line with these immunosuppressive effects in vitro, CR3-knockout mice are more resistant to disseminated candidiasis60, and are also more resistant to Blastomyces dermatitidis infection43. Some studies have suggested that the differential activation of stimulatory and inhibitory receptors is caused by the dimorphic nature of C. albicans yeast and hyphae, which induce different DC activation profiles. The recognition of yeasts by DCs is mediated mainly by the MR, dectin 1 (Ref. 60) and DC‑SIGN27,112, and hyphae are recognized mainly through dectin 2, CR3, FcγRII and FcγRIII60. In addition to inducing direct anti-inflammatory effects, C. albicans has developed strategies to either block or avoid recognition by stimulatory PRRs (FIG. 4b) . The morphogenetic change of C. albicans from a yeast to a hyphal form, after adhesion of yeast cells to the intravascular endothelium and invasion of the surrounding tissue, might represent a mechanism whereby the pathogen can avoid recognition of β‑glucans by the immune system113. Hyphal filaments lack surfaceexposed β‑glucan and have a range of other cell-surface modifications that could influence immune detection11. In addition, the induction of the pro-inflammatory response through dectin 1 can be prevented, because the mannoprotein layer covers the β‑glucan layer of the cell wall, thus blocking the β‑glucan–dectin 1 interaction11. Accordingly, exposure of β‑glucan through treatment with the antifungal caspofungin, by heat treatment of cells or by mutation of glycosylation genes that result in depletion of surface mannan all led to an enhanced dectin-1-mediated pro-inflammatory cytokine signal. This suggests that the mannoprotein layer is a mask that downregulates the pro-inflammatory dectin-1-mediated response114. As the exposed β‑glucan of intact C. albicans yeast cells is thought to be presented mainly at the surface of the wall as bud scars, the attenuation of dectin‑1mediated recognition during germ tube evagination could be due to the expansion of the hyphal cell surface that lacks any bud scars. In line with this, Rappleye et al. showed that the β‑glucan–dectin 1 interaction between Histoplasma capsulatum and macrophages can be shielded by a different cell wall component — α-(1,3)glucan115. The presence of α-(1,3)-glucan in yeast cells prevents recognition through dectin 1, and mutation of AGS1 (which encodes α-(1,3)-glucan synthase) results in greatly enhanced TNFα production115. Therefore, both mannans and α‑glycans can act as shields that mask dectin-1-mediated immune responses. It is important to note that mannans are in themselves immunostimulatory 15,24,116–118. How can the immunostimulatory properties of mannans be reconciled with the apparent inhibitory effect of masking nature reviews | microbiology
the β-glucan–dectin 1 interactions? The answer to this paradox might reside in the complex interaction between TLRs and LRs. Although both mannans and glucans can induce pro-inflammatory signals, concomitant stimulation of the mannan–TLR and β-glucan–dectin 1 pathways has a synergistic effect on the amplification of the immune response. Amplification mechanisms between dectin 1 and TLR2 (Refs 49,50), and between dectin 1 and TLR4 (K. M. Dennehy and colleagues, unpublished observations) have recently been documented. If the interaction of β-glucans with dectin 1 is shielded by mannans, this synergism may be lost, with inefficient activation of host defence, despite recognition of mannans. It is therefore mainly the loss of synergistic effects, rather than the inhibition of an individual recognition pathway, that may lead to defective cytokine release.
An integrated model of innate recognition Progress in our understanding of the mechanisms responsible for C. albicans recognition, as described above, allows us to suggest an integrated view of how C. albicans is recognized and of how the host innate immune response is activated during candidiasis. Although we are still in the early stages of understanding the complexity of fungal recognition, and there is still much to learn, there are several principles that characterize recognition of C. albicans. • First, recognition depends on ‘tasting’ several PAMPs in the fungal cell wall. Specific receptor systems have evolved for the recognition of the major polysaccharide cell wall components, such as the MR and DC‑SIGN for the recognition of branched N‑linked mannan, TLR4 for linear O‑linked mannan, galectin 3 for β-mannosides, dectin 1 and TLR2 for β-glucan and PLM, and CR3 for β-(1,6)-glucan. • Second, despite overlapping and sometimes redundant functions, each ligand–receptor system activates specific intracellular signalling pathways, and this has distinct consequences for the activation of the various components of the host immune response. • Third, differential expression of the various PRRs is an important mechanism for the cell-type-specific response to fungal pathogens (FIG. 2). • Fourth, the fully integrated response to a specific pathogen depends on the mosaic of PRRs and receptor complexes that are engaged. Co-stimulation via multiple PAMP–PRR interactions can increase both the sensitivity and the specificity of the immune recognition process. These four principles can be extrapolated to immune recognition of any microorganism. For example, the final response upon stimulation of monocytes by C. albicans depends on integrating the signals that are received from TLR2, dectin 1, TLR4 and the MR. This response differs from the stimulation that is induced by the Gram-negative bacterium Escherichia coli, which can be ascribed to recognition of lipopolysaccharide by TLR4, membrane lipopeptides by TLR2 and flagellin by TLR5. By contrast, Gram-positive bacteria are mainly recognized by TLR2 (which recognizes lipoteichoic acid) and NOD2 (which recognizes peptidoglycans). The different volume 6 | january 2008 | 75
© 2008 Nature Publishing Group
REVIEWS intracellular events that are stimulated by these pathways are ultimately responsible for the tailored innate response to infection by different classes of microorganisms. From this perspective, although described here for C. albicans, these general principles can be considered as a blueprint for pattern recognition of all pathogenic microorganisms by the innate immune response.
Future directions In parallel with the general renaissance in the field of innate immunity, our understanding of C. albicans recognition by the innate immune system has improved considerably over the past decade. Despite this progress, more challenges lie ahead. Although we have gained a much better understanding of the basal (polysaccharide) structures that are recognized by the PRRs, more work must be done to understand the details of the interactions of these PAMPs with their receptors on the host cell membrane. Recent evidence, for example, demonstrates that dectin 1 associates with tetraspanins, which suggests that dectin 1 forms receptor complexes on the cell surface that might be involved in regulating immune responses88. In addition, little is known about the recognition of the other components of the C. albicans cell wall, such as the proteins themselves, or the possible sensing of products that are secreted by C. albicans. The differential recognition of the yeast and hyphal forms also promises to be a fertile area for future research. An additional challenge is to provide evidence that the pattern-recognition pathways identified in experimental models are also applicable during C. albicans infections in humans. This challenge can be partly met by experiments
Edmond, M. B. et al. Nosocomial bloodstream infections in United States hospitals: a three-year analysis. Clin. Infect. Dis. 29, 239–244 (1999). 2. Enoch, D. A., Ludlam, H. A. & Brown, N. M. Invasive fungal infections: a review of epidemiology and management options. J. Med. Microbiol. 55, 809–818 (2006). 3. 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). 4. Gudlaugsson, O. et al. Attributable mortality of nosocomial candidemia, revisited. Clin. Infect. Dis. 37, 1172–1177 (2003). 5. Stevens, D. A. Combination immunotherapy and antifungal chemotherapy. Clin. Infect. Dis. 26, 1266–1269 (1998). 6. Mitchison, A. Will we survive? Sci. Am. 269, 136–144 (1993). 7. Hoebe, K., Janssen, E. & Beutler, B. The interface between innate and adaptive immunity. Nature Immunol. 5, 971–974 (2004). 8. Ernst, J. F. & Prill, S. K. O‑glycosylation. Med. Mycol. 39 1, 67–74 (2001). 9. Cutler, J. E. N‑glycosylation of yeast, with emphasis on Candida albicans. Med. Mycol. 39, 75–86 (2001). 10. Klis, F. M., Boorsma, A. & De Groot, P. W. Cell wall construction in Saccharomyces cerevisiae. Yeast 23, 185–202 (2006). 11. Gantner, B. N., Simmons, R. M. & Underhill, D. M. Dectin‑1 mediates macrophage recognition of Candida albicans yeasts but not filaments. EMBO J. 24, 1277–1286 (2005). Describes the differential recognition of C. albicans yeast and hyphae by dectin 1 as a major escape mechanism. 12. Suzuki, A. in Candida and Candidosis (ed. Calderone, R. A.) 29–36 (ASM Press, Washington, 2002). 1.
in which human primary cells are exposed to C. albicans. However, valuable information can also be gathered from immunogenetic studies. For example, TLR4 polymorphisms have recently been identified as a susceptibility trait for systemic candidiasis, which supports a role for this receptor in the host immune response to C. albicans119. Studies of additional functional polymorphisms in larger cohorts of patients are needed to confirm the role of the various pathways in the pathogenesis of candidiasis The ultimate ambition of research into the biology of host–fungus interactions is to be able to translate findings at the molecular level of the interactions into new treatments for patients who are vulnerable to lethal fungal infections. Although direct blockade or stimulation of PRRs might be problematic as a viable therapeutic approach, vaccine biology has become a domain in which our newly acquired knowledge of PRRs is already demonstrating potential120. The understanding of the modulation of DC function by PRRs has led to the design of novel and specific vaccine adjuvants, and this work is likely to have a major impact on the development of future antifungal vaccines. In conclusion, in this Review we have presented an integrated model of innate pattern recognition of an important human pathogen, the fungus C. albicans, including a description of the PAMPs of C. albicans that are recognized by the host, to the receptor pathways that are stimulated by specific fungal molecular structures. We have also discussed and speculated on how these signals are integrated to bring about efficient activation of the host innate response. This model, which focuses on the interaction of C. albicans with host defences, could be conceptually pertinent to the modelling of various other host–pathogen interactions.
13. Vecchiarelli, A., Puliti, M., Torosantucci, A., Cassone, A. & Bistoni, F. In vitro production of tumor necrosis factor by murine splenic macrophages stimulated with mannoprotein constituents of Candida albicans cell wall. Cell. Immunol. 134, 65–76 (1991). 14. Garner, R. E., Rubanowice, K., Sawyer, R. T. & Hudson, J. A. Secretion of TNF-α by alveolar macrophages in response to Candida albicans mannan. J. Leuk. Biol. 55, 161–168 (1994). 15. Pietrella, D., Bistoni, G., Corbucci, C., Perito, S. & Vecchiarelli, A. Candida albicans mannoprotein influences the biological function of dendritic cells. Cell. Microbiol. 8, 602–612 (2006). 16. Gomez, M. J. et al. Purification and biochemical characterization of a 65-kilodalton mannoprotein (MP65), a main target of anti-Candida cell-mediated immune responses in humans. Infect. Immun. 64, 2577–2584 (1996). 17. Gomez, M. J. et al. Biochemical and immunological characterization of MP65, a major mannoprotein antigen of the opportunistic human pathogen Candida albicans. Infect. Immun. 68, 694–701 (2000). 18. Mencacci, A. et al. A mannoprotein constituent of Candida albicans that elicits different levels of delayed-type hypersensitivity, cytokine production, and anticandidal protection in mice. Infect. Immun. 62, 5353–5360 (1994). 19. Wileman, T. E., Lennartz, M. R. & Stahl, P. D. Identification of the macrophage mannose receptor as a 175-kDa membrane protein. Proc. Natl Acad. Sci. USA 83, 2501–2505 (1986). 20. Stephenson, J. D. & Shepherd, V. L. Purification of the human alveolar macrophage mannose receptor. Biochem. Biophys. Res. Commun. 148, 883–889 (1987). 21. Stahl, P. D., Rodman, J. S., Miller, M. J. & Schlesinger, P. H. Evidence for receptor-mediated binding of
76 | january 2008 | volume 6
22. 23.
24.
25.
26.
27. 28. 29.
glycoproteins, glycoconjugates, and lysosomal glycosidases by alveolar macrophages. Proc. Natl Acad. Sci. USA 75, 1399–1403 (1978). Linehan, S. A., Martinez-Pomares, L. & Gordon, S. Macrophage lectins in host defence. Microbes Infect. 2, 279–288 (2000). Kery, V., Krepinsky, J. J., Warren, C. D., Capek, P. & Stahl, P. D. Ligand recognition by purified human mannose receptor. Arch. Biochem. Biophys. 298, 49–55 (1992). Netea, M. G. et al. Immune sensing of Candida albicans requires cooperative recognition of mannans and glucans by lectin and Toll-like receptors. J. Clin. Invest. 116, 1642–1650 (2006). Proposed, for the first time, an integrated model of fungal recognition. Tada, H. et al. Saccharomyces cerevisiae- and Candida albicans-derived mannan induced production of tumor necrosis factor α by human monocytes in a CD14- and Toll-like receptor 4‑dependent manner. Microbiol. Immunol. 2002, 503–512 (2002). van der Graaf, C. A. A., Netea, M. G., Verschueren, I., van der Meer, J. W. M. & Kullberg, B. J. Differential cytokine production and Toll-like receptor signaling pathways by Candida albicans blastoconidia and hyphae. Infect. Immun. 73, 7458–7464 (2005). Cambi, A. et al. The C‑type lectin DC‑SIGN (CD209) is an antigen-uptake receptor for Candida albicans on dendritic cells. Eur. J. Immunol. 33, 532–538 (2003). Jouault, T. et al. Candida albicans phospholipomannan is sensed through Toll-like receptors. J. Infect. Dis. 188, 165–172 (2003). Jouault, T. et al. Specific recognition of Candida albicans by macrophages requires galectin‑3 to discriminate Saccharomyces cerevisiae and needs association with TLR2 for signaling. J. Immunol. 177, 4679–4687 (2006).
www.nature.com/reviews/micro © 2008 Nature Publishing Group
REVIEWS 30. Ariizumi, K. et al. Cloning of a second dendritic cellassociated C‑type lectin (dectin‑2) and its alternatively spliced isoforms. J. Biol. Chem. 275, 11957–11963 (2000). 31. Taylor, P. R. et al. Dectin‑2 is predominantly myeloid restricted and exhibits unique activation-dependent expression on maturing inflammatory monocytes elicited in vivo. Eur. J. Immunol. 35, 2163–2174 (2005). 32. Sato, K. et al. Dectin‑2 is a pattern recognition receptor for fungi that couples with the Fc receptor γ chain to induce innate immune responses. J. Biol. Chem. 281, 38854–38866 (2006). 33. McGreal, E. P. et al. The carbohydrate-recognition domain of Dectin‑2 is a C‑type lectin with specificity for high mannose. Glycobiology 16, 422–430 (2006). 34. Klis, F. M., de Groot, P. & Hellingwerf, K. Molecular organization of the cell wall of Candida albicans. Med. Mycol. 39, 1–8 (2001). A model of the C. albicans cell wall. 35. Torosantucci, A. et al. A novel glyco-conjugate vaccine against fungal pathogens. J. Exp. Med. 202, 597–606 (2005). 36. Brown, G. D. & Gordon, S. Immune recognition of fungal β-glucans. Cell. Microbiol. 7, 471–479 (2005). 37. Obayashi, T. et al. Plasma (1,3)‑β‑d-glucan measurement in diagnosis of invasive deep mycosis and fungal febrile episodes. Lancet 345, 17–20 (1995). 38. Thornton, B. P., Vetvicka, V., Pitman, M., Goldman, R. C. & Ross, G. D. Analysis of the sugar specificity and molecular location of the β‑glucanbinding lectin site of complement receptor type 3 (CD11b/CD18). J. Immunol. 156, 1235–1246 (1996). 39. Diamond, M. S., Garcia-Aguilar, J., Bickford, J. K., Corbi, A. L. & Springer, T. A. The I domain is a major recognition site on the leukocyte integrin Mac‑1 (CD11b/CD18) for four distinct adhesion ligands. J. Cell Biol. 120, 1031–1043 (1993). 40. Forsyth, C. B. & Mathews, H. L. Lymphocyte adhesion to Candida albicans. Infect. Immun. 70, 517–527 (2002). 41. Forsyth, C. B., Plow, E. F. & Zhang, L. Interaction of the fungal pathogen Candida albicans with integrin CD11b/CD18: recognition by the I domain is modulated by the lectin-like domain and the CD18 subunit. J. Immunol. 161, 6198–6205 (1998). 42. Wright, S. D. & Silverstein, S. C. Receptors for C3b and C3bi promote phagocytosis but not the release of toxic oxygen from human phagocytes. J. Exp. Med. 158, 2016–2023 (1983). 43. Brandhorst, T. T., Wuthrich, M., Finkel-Jimenez, B., Warner, T. & Klein, B. S. Exploiting type 3 complement receptor for TNF-α suppression, immune evasion, and progressive pulmonary fungal infection. J. Immunol. 173, 7444–7453 (2004). 44. Wright, S. D. & Silverstein, S. C. Tumor-promoting phorbol esters stimulate C3b and C3b′ receptormediated phagocytosis in cultured human monocytes. J. Exp. Med. 156, 1149–1164 (1982). 45. Brown, G. D. & Gordon, S. Fungal β-glucans and mammalian immunity. Immunity 19, 311–315 (2003). A review that describes the role of β‑glucan recognition in host defence. 46. Rogers, N. C. et al. SYK-dependent cytokine induction by dectin‑1 reveals a novel pattern recognition pathway for C type lectins. Immunity 22, 507–517 (2005). 47. Gow, N. A. R. et al. Recognition of Candida albicans β‑glucan by dectin‑1 induces cytokines and has nonredundant effects on the activation of innate immunity. J. Infect. Dis. (in the press). 48. Leibundgut-Landmann, S. et al. SYK- and CARD9dependent coupling of innate immunity to the induction of T helper cells that produce interleukin 17. Nature Immunol. 8, 630–638 (2007). This study demonstrates the specific activation of TH17 cells by dectin 1. 49. Brown, G. D. et al. Dectin‑1 mediates the biological effects of β-glucans. J. Exp. Med. 197, 1119–1124 (2003). 50. Gantner, B. N., Simmons, R. M., Canavera, S. J., Akira, S. & Underhill, D. M. Collaborative induction of inflammatory responses by dectin‑1 and Toll-like receptor 2. J. Exp. Med. 197, 1107–1117 (2003). References 49 and 50 show for the first time the collaboration between TLRs (TLR2) and LRs (dectin 1).
51. Rubin-Bejerano, I., Abeijon, C., Magnelli, P., Grisafi, P. & Fink, G. R. Phagocytosis by human neutrophils is stimulated by a unique fungal cell wall component. Cell Host Microbe 2, 55–67 (2007). 52. Minke, R. & Blackwell, J. The structure of α-chitin. J. Mol. Biol. 120, 167–181 (1978). 53. van der Graaf, C. A. et al. Nucleotide oligomerization domain 2 (Nod2) is not involved in the pattern recognition of Candida albicans. Clin. Vaccine Immunol. 13, 423–425 (2006). 54. Wagner, H. Toll meets bacterial CpG-DNA. Immunity 14, 499–502 (2001). 55. Bellocchio, S. et al. The contribution of Toll-like/IL‑1 receptor superfamily to innate and adaptive immunity to fungal pathogens in vivo. J. Immunol. 172, 3059–3069 (2004). 56. Kozel, T. R. Activation of the complement system by pathogenic fungi. Clin. Microbiol. Rev. 9, 34–46 (1996). 57. Kozel, T. R., Weinhold, L. C. & Lupan, D. M. Distinct characteristics of initiation of the classical and alternative complement pathways by Candida albicans. Infect. Immun. 64, 3360–3368 (1996). 58. Herre, J. et al. Dectin‑1 uses novel mechanisms for yeast phagocytosis in macrophages. Blood 104, 4038–4045 (2004). 59. Ezekowitz, R. A. B. et al. Restoration of phagocyte function by interferon‑γ in X‑linked chronic granulomatous disease occurs at the level of a progenitor cell. Blood 76, 2443–2448 (1990). 60. Romani, L. et al. The exploitation of distinct recognition receptors in dendritic cells determines the full range of host immune relationships with Candida albicans. Int. Immunol. 16, 149–161 (2004). 61. Le Cabec, V., Emorine, L. J., Toesca, I., Cougoule, C. & Maridonneau-Parini, I. The human macrophage mannose receptor is not a professional phagocytic receptor. J. Leuk. Biol. 77, 934–943 (2005). 62. Blander, J. M. & Medzhitov, R. Toll-dependent selection of microbial antigens for presentation by dendritic cells. Nature 440, 808–812 (2006). 63. Romani, L. Immunity to fungal infections. Nature Rev. Immunol. 4, 1–13 (2004). 64. Kennedy, A. D. et al. Dectin‑1 promotes fungicidal activity of human neutrophils. Eur. J. Immunol. 37, 467–478 (2007). 65. Aratani, Y. et al. Severe impairment in early host defense against Candida albicans in mice deficient in myeloperoxidase. Infect. Immun. 67, 1828–1836 (1999). 66. Mansour, M. K. & Levitz, S. M. Interactions of fungi with phagocytes. Curr. Opin. Microbiol. 5, 359–365 (2002). 67. Reeves, E. P. et al. Killing activity of neutrophils is mediated through activation of proteases by K+ flux. Nature 416, 291–297 (2002). 68. Kohatsu, L., Hsu, D. K., Jegalian, A. G., Liu, F. T. & Baum, L. G. Galectin‑3 induces death of Candida species expressing specific β-1, 2-linked mannans. J. Immunol. 177, 4718–4726 (2006). 69. Akira, S. & Hemmi, H. Recognition of pathogenassociated molecular patterns by TLR family. Immunol. Letters 85, 85–95 (2003). 70. Toshchakov, V. et al. TLR4, but not TLR2, mediates IFNβ-induced STAT I α/β-dependent gene expression in macrophages. Nature Immunol. 3, 392–398 (2002). 71. Netea, M. G. et al. The role of Toll-like receptors in the defense against disseminated candidiasis. J. Infect. Dis. 185, 1483–1489 (2002). Demonstrates, for the first time, the role of TLRs in C. albicans recognition. 72. Hirschfeld, M. et al. Signaling by Toll-like receptor 2 and 4 agonists results in differential gene expression in murine macrophages. Infect. Immun. 69, 1477–1482 (2001). 73. Re, F. & Strominger, J. L. Toll-like receptor 2 (TLR2) and TLR4 differentially activate human dendritic cells. J. Biol. Chem. 276, 37692–37699 (2001). 74. Agrawal, S. et al. Different Toll-like receptor agonists instruct dendritic cells to induce distinct Th responses via differential modulation of extracellular signalregulated kinase‑mitogen‑activated protein kinase and c‑Fos. J. Immunol. 171, 4984–4989 (2003). 75. Dillon, S. et al. Yeast zymosan, a stimulus for TLR2 and dectin‑1, induces regulatory antigen-presenting cells and immunological tolerance. J. Clin. Invest. 116, 916–928 (2006). 76. Pulendran, B. et al. Lipopolysaccharides from distinct pathogens induce different classes of immune
nature reviews | microbiology
77.
78. 79.
80.
81.
82.
83.
84. 85.
86.
87.
88. 89. 90.
91.
92.
93.
94.
95.
96. 97.
responses in vivo. J. Immunol. 167, 5067–5076 (2001). Netea, M. G. et al. Toll-like receptor 2 inhibits cellular responses against Candida albicans through pathways mediated by IL‑10 and CD4+ CD25+ regulatory T cells. J. Immunol. 172, 3712–3718 (2004). Demonstrates, for the first time, induction of TReg cells through TLR2 signals. Sutmuller, R. P. et al. Toll-like receptor 2 controls expansion and function of regulatory T cells. J. Clin. Invest. 116, 485–494 (2006). Van der Kleij, D. et al. A novel host-parasite lipid cross-talk: schistosomal lyso-phosphatidylserine activates Toll-like receptor 2 and affects immune polarization. J. Biol. Chem. 277, 48122–48129 (2002). Diterich, I., Rauter, C., Kirschning, C. J. & Hartung, T. Borrelia burgdorferi-induced tolerance as a model of persistence via immunosuppression. Infect. Immun. 71, 3979–3985 (2003). Underhill, D. M. et al. The Toll-like receptor 2 is recruited to macrophage phagosomes and discriminate between pathogens. Nature 401, 811–815 (1999). One of the first studies to demonstrate the discrimination of pathogens by TLRs. Netea, M. G., van de Veerdonk, F., Verschueren, I., van der Meer, J. W. & Kullberg, B. J. Role of TLR1 and TLR6 in the host defense against disseminated candidiasis. FEMS Immnol. Med. Microbiol. (in the press). Brown, G. R., Silva, M. D., Thompson, P. A. & Beutler, B. Lymphoid hyperplasia, CD45RBhigh to CD45RBlow T‑cell imbalance, and suppression of type I diabetes mellitus result from TNF blockade in NOD–NOD–scid adoptive T cell transfer. Diabetologia 41, 1502–1510 (1998). Steele, C. et al. The β-glucan receptor dectin‑1 recognizes specific morphologies of Aspergillus fumigatus. PLoS Pathog. 4, 323–334 (2005). Steele, C. et al. Alveolar macrophage-mediated killing of Pneumocystis carinii f. sp. muris involves molecular recognition by the Dectin‑1 β-glucan receptor. J. Exp. Med. 198, 1677–1688 (2003). Viriyakosol, S., Fierer, J., Brown, G. D. & Kirkland, T. N. Innate immunity to the pathogenic fungus Coccidioides posadasii is dependent on Toll-like receptor 2 and dectin‑1. Infect. Immun. 73, 1553–1560 (2005). Slack, E. C. et al. Syk-dependent ERK activation regulates IL‑2 and IL‑10 production by DC stimulated with zymosan. Eur. J. Immunol. 37, 1600–1612 (2007). Meyer-Wentrup, F. et al. Dectin‑1 interaction with tetraspanin CD37 inhibits IL‑6 production. J. Immunol. 178, 154–162 (2007). Gross, O. et al. Card9 controls a non-TLR signalling pathway for innate anti-fungal immunity. Nature 442, 651–656 (2006). Acosta-Rodriguez, E. V. et al. Surface phenotype and antigenic specificity of human interleukin 17producing T helper memory cells. Nature Immunol. 8, 639–646 (2007). Huang, W., Na, L., Fidel, P. L. & Schwarzenberger, P. Requirement for interleukin‑17A for systemic antiCandida albicans host defense in mice. J. Infect. Dis. 190, 524–631 (2004). Zelante, T. et al. IL‑23 and the TH17 pathway promote inflammation and impair antifungal immune resistance. Eur. J. Immunol. 37, 2695–2706 (2007). Hsu, Y. M. et al. The adaptor protein CARD9 is required for innate immune responses to intracellular pathogens. Nature Immunol. 8, 198–205 (2007). Hara, H. et al. The adaptor protein CARD9 is essential for the activation of myeloid cells through ITAMassociated and Toll-like receptors. Nature Immunol. 8, 619–629 (2007). De Luca, A. et al. Functional yet balanced reactivity to Candida albicans requires TRIF, MyD88, and IDOdependent inhibition of Rorc. J. Immunol. 179, 5999–6008 (2007). Weindl, G. et al. Human epithelial cells establish direct antifungal defense through TLR4-mediated signaling. J. Clin. Invest. (in the press). Schaller, M. et al. Polymorphonuclear leukocytes (PMNs) induce protective Th1-type cytokine epithelial responses in an in vitro model of oral candidosis. Microbiology 150, 2807–2813 (2004).
volume 6 | january 2008 | 77 © 2008 Nature Publishing Group
REVIEWS 98. Fidel, P. L. Jr. History and update on host defense against vaginal candidiasis. Am. J. Reprod. Immunol. 57, 2–12 (2007). 99. Pivarcsi, A. et al. Microbial compounds induce the expression of pro-inflammatory cytokines, chemokines and human β‑defensin‑2 in vaginal epithelial cells. Microbes Infect. 7, 1117–1127 (2005). 100. Gozalbo, D., Roig, P., Villamon, E. & Gil, M. L. Candida and candidiasis: the cell wall as a potential molecular target for antifungal therapy. Curr. Drug Targets Infect. Disord. 4, 117–135 (2004). 101. Marr, K. A. et al. Differential role of MyD88 in macrophage-mediated responses to opportunistic fungal pathogens. Infect. Immun. 71, 5280–5286 (2003). 102. Villamon, E. et al. Toll-like receptor‑2 is essential in murine defenses against Candida albicans infections. Microbes Infect. 6, 1–7 (2004). 103. Blasi, E. et al. Biological importance of the two Tolllike receptors, TLR2 and TLR4, in macrophage response to infection with Candida albicans. FEMS Immunol. Med. Microbiol. 44, 69–79 (2005). 104. Murciano, C. et al. Toll-like receptor 4 defective mice carrying point or null mutations do not show increased susceptibility to Candida albicans in a model of hematogenously disseminated infection. Med. Mycol. 44, 149–157 (2006). 105. Taylor, P. R. et al. Dectin‑1 is required for β-glucan recognition and control of fungal infection. Nature Immunol. 8, 31–38 (2007). 106. Saijo, S. et al. Dectin‑1 is required for host defense against Pneumocystis carinii but not against Candida albicans. Nature Immunol. 8, 39–46 (2007). References 105 and 106 are the first reports of in vivo fungal infections in dectin 1 knockout mice. 107. Lee, S. J., Zheng, N. Y., Clavijo, M. & Nussenzweig, M. C. Normal host defense during systemic candidiasis in mannose receptor-deficient mice. Infect. Immun. 71, 437–445 (2003).
108. Vonk, A. G., Netea, M. G., van Krieken, J. H., Van der Meer, J. W. M. & Kullberg, B. J. Delayed clearance of intraabdominal abcesses caused by Candida albicans in tumor necrosis factor-α and lymphotoxin‑α deficient mice. J. Infect. Dis. 186, 1815–1822 (2002). 109. Gottar, M. et al. Dual detection of fungal infections in Drosophila via recognition of glucans and sensing of virulence factors. Cell 127, 1425–1437 (2006). 110. Geijtenbeek, T. B. et al. Mycobacteria target DC‑SIGN to suppress dendritic cell function. J. Exp. Med. 197, 7–17 (2003). 111. van Kooyk, Y., Engering, A., Lekkerkerker, A. N., Ludwig, I. S. & Geijtenbeek, T. B. Pathogens use carbohydrates to escape immunity induced by dendritic cells. Curr. Opin. Immunol. 16, 488–493 (2004). 112. Taylor, P. R., Brown, G. D., Geldhof, A. B., MartinezPomares, L. & Gordon, S. Pattern recognition receptors and differentiation antigens define murine myeloid cell heterogeneity ex vivo. Eur. J. Immunol. 33, 2090–2097 (2003). 113. d’Ostiani, C. F. et al. Dendritic cells discriminate between yeasts and hyphae of the fungus Candida albicans. Implications for initiation of T helper cell immunity in vitro and in vivo. J. Exp. Med. 191, 1661–1674 (2000). 114. Wheeler, R. T. & Fink, G. R. A drug-sensitive genetic network masks fungi from the immune system. PLoS Pathog. 2, 328–339 (2006). 115. Rappleye, C. A., Eissenberg, L. G. & Goldman, W. E. Histoplasma capsulatum α‑(1,3)-glucan blocks innate immune recognition by the β-glucan receptor. Proc. Natl Acad. Sci. USA 104, 1366–1370 (2007). 116. Mangeney, M., Fischer, A., Le Deist, F., Latge, J. P. & Durandy, A. Direct activation of human B lymphocytes by Candida albicans-derived mannan antigen. Cell. Immunol. 122, 329–337 (1989). 117. Lillegard, J. B., Sim, R. B., Thorkildson, P., Gates, M. A. & Kozel, T. R. Recognition of Candida albicans
78 | january 2008 | volume 6
by mannan-binding lectin in vitro and in vivo. J. Infect. Dis. 193, 1589–1597 (2006). 118. Sheng, K. C. et al. Mannan derivatives induce phenotypic and functional maturation of mouse dendritic cells. Immunology 118, 372–383 (2006). 119. Van der Graaf, C. A. et al. Toll-like receptor 4 Asp299Gly/Thr399Ile polymorphisms are a risk factor for Candida bloodstream infection. Eur. Cytokine Netw. 17, 29–34 (2006). 120. Cassone, A. & Torosantucci, A. Opportunistic fungi and fungal infections: the challenge of a single, general antifungal vaccine. Expert Rev. Vaccines 5, 859–867 (2006). 121. Dillon, S. et al. Yeast zymosan, a stimulus for TLR2 and dectin‑1, induces regulatory antigen-presenting cells and immunological tolerance. J. Clin. Invest. 116, 916–928 (2006). 122. Reese, T. A. et al. Chitin induces accumulation in tissue of innate immune cells associated with allergy. Nature 447, 92–96 (2007).
Acknowledgements
This work was supported by a Vidi Grant from Netherlands Organization for Scientific Research to M.G.N., and by the Wellcome Trust to N.A.R.G. and G.D.B.
DATABASES Entrez Genome project: http://www.ncbi.nlm.nih.gov/sites/ entrez?db=genomeprj Borrelia burgdorferi | Candida albicans | Saccharomyces cerevisiae | Schistosoma mansoni UniProtKB: http://ca.expasy.org/sprot CARD9 | dectin 1 | dectin 2 | galectin 3 | IL-2 | IL‑6 | IL‑10 | Syk | TLR2 | TLR4 | TLR6 | TLR9 All links are active in the online pdf
www.nature.com/reviews/micro © 2008 Nature Publishing Group
REVIEWS
Kiss and spit: the dual roles of Toxoplasma rhoptries John C. Boothroyd* and Jean-Francois Dubremetz‡
Abstract | Toxoplasma gondii is a single-celled, eukaryotic parasite that can only reproduce inside a host cell. Upon entry, this Apicomplexan parasite co-opts host functions for its own purposes. An unusual set of apical organelles, named rhoptries, contain some of the machinery that is used by T. gondii both for invasion and to commandeer host functions. Of particular interest are a group of injected protein kinases that are among the most variable of all the T. gondii proteins. At least one of these kinases has a major effect on host-gene expression, including the modulation of key regulators of the immune response. Here, we discuss these recent findings and use them to propose a model in which an expansion of host range is a major force that drives rhoptry-protein evolution.
Apicomplexa A phylum of unicellular eukaryotes that are obligate parasites and defined by a collection of apical organelles that are involved in invasion of a host cell.
Microneme A small, cylindrical organelle that is found at the periphery of the anterior end of Apicomplexan parasites that secretes its contents onto the surface of a gliding or invading parasite.
*Department of Microbiology and Immunology, Stanford University School of Medicine, Stanford, California 94305‑5124, USA. ‡ UMR 5539 CNRS, Université de Montpellier 2, CP 107, Place Eugène Bataillon, Montpellier 34090, France. Correspondence to J.C.B. e‑mail: john.boothroyd@ stanford.edu doi:10.1038/nrmicro1800 Published online 3 December 2007
The phylum Apicomplexa includes a large number of obligate intracellular parasites. Among these are some notorious human and animal pathogens from genera such as Plasmodium (the causative agent of malaria), Toxoplasma (an important cause of congenital disease and infection in immuno-compromised patients1; see BOX 1 for a brief description of the Toxoplasma life cycle and the clinical spectrum of human toxoplasmosis), Cryptosporidium (a cause of serious gastrointestinal disease) and Eimeria (a major problem in the poultry industry and cause of chicken coccidiosis). The phylum is defined by the presence of an apical complex that comprises a microtubule anchoring ring through which dedicated secretory organelles release their contents2. There are two types of apical secretory organelle — the small, rod-shaped micronemes and the much larger, bulb-shaped rhoptries (Greek for club). FIG. 1a provides an electron micrograph of one of the asexual forms of Toxoplasma gondii, which clearly shows an apical cytoskeleton and a substantial complement of micronemes and rhoptries. Given their conservation across much of the phylum, it has long been suspected that rhoptries have a key role in the intracellular lifestyle of these pathogens. Recently, exciting results have shed light on at least two functions of the rhoptries. In this Review, we will discuss the ultrastructure and content of rhoptries, their role in invasion and function in subverting host-cell processes, as well as the evolutionary pressures that might underlie the extreme variability of rhoptry proteins.
nature reviews | microbiology
Rhoptry ultrastructure and content The size, electron density and number of rhoptries varies among Apicomplexan species and between the different developmental stages of a single species. In this Review, we will mainly focus on the rapidly dividing, asexual form of T. gondii that is known as the tachyzoite3. Tachyzoites predominate during the acute stages of infection in an intermediate host, which can be virtually any warm-blooded animal (BOX 1). This breadth of host range also extends to the cellular level, as tachyzoites can invade and replicate in almost any host cell that they encounter, at least in vitro. In vivo, the cellular tropism of tachyzoites has not been well characterized as, until recently4,5, methods to survey parasite growth in the entire animal have been lacking. Most in vitro studies use fibroblasts as the host cell because of their availability, the fact that they are a primary cell line (making studies on host-gene expression more relevant than in a transformed cell line) and the ease with which they can be examined by light microscopy (as they have full contact inhibition and grow as a uniform, flat monolayer). Tachyzoites typically have ~12 rhoptries, each of which is ~2 to 3 micrometres in length6 (FIG. 1). Using light microscopy, the rhoptries can be visualized as an elongated cluster that is present at the anterior end of the parasite and is frequently located on one side. They are visible using differential interference contrast microscopy, but are more easily observed by immunofluorescence using a range of antibodies that are specific for their protein contents. Interestingly, most of these antibodies volume 6 | january 2008 | 79
© 2008 Nature Publishing Group
REVIEWS Box 1 | Toxoplasma gondii biology Asexual Warm-blooded intermediate
Bradyzoite 3
1
Sexual Feline
4
2
Gamete 5
Tachyzoite
7
6
Oocyst
T. gondii is an extremely common protozoan parasite of warm-blooded animals that has a host range that extends from birds to humans. Its life cycle comprises two, potentially Nature Reviewsbetween | Microbiology independent cycles, one asexual and one sexual, although movement the two is almost certainly key to efficient transmission (see the figure). The asexual cycle (see the figure, left-hand panel) can occur in virtually any warmblooded animal and ~100 mammalian and avian species have been documented as being infected with T. gondii in nature. In these ‘intermediate’ hosts, the parasite population initially expands by rapid proliferation of the tachyzoite form (‘tachy’ means fast in Greek and in this context refers to speedy replication). Once an immune response is elicited, the parasite differentiates to the more slowly growing bradyzoite form (‘brady’ means slow in Greek) (arrow 1). The bradyzoite form can persist for the life of the host in cyst-like structures that are present deep in the brain and other tissues. During immunosuppression (for example, in patients with AIDS) the parasite can resume rapid replication in the form of tachyzoites (arrow 2). Transmission between intermediate hosts is by the ingestion of raw or under-cooked meat and other organs that contain the infectious, encysted bradyzoites (for example, a mouse eaten by a hawk or an undercooked lamb chop eaten by a human) (arrow 3). The sexual cycle (see the figure, righthand panel), which begins when bradyzoite-bearing tissue from an intermediate host is ingested by a feline (arrow 4), involves gametogenesis and fertilization (arrow 5) in the gut epithelium of felines. The cat family is the only group of animals that is known to serve as a definitive host for this parasite; that is, it is the only host in which sexual reproduction occurs, although felines can also support asexual reproduction. The sexual cycle culminates in the shedding of up to 100,000,000 highly stable oocysts in the faeces of an infected cat (these oocysts are initially shed in an immature state and require approximately 2 days to mature in the environment and gain full infectivity). Mature oocysts are highly infectious and can either infect another cat (arrow 6) or, more probably, a grazing intermediate host (arrow 7) that is foraging in a farm. T. gondii causes serious disease in humans, but this disease is primarily confined to the developing foetus of a woman who acquires her first infection during pregnancy and individuals who are immunocompromised as a result of HIV-1 infection, lymphoma or other immune-suppressive syndromes. Disease pathologies range from asymptomatic to severe and, sometimes, fatal55. Occasionally, otherwise healthy adults can also experience acute symptoms, especially in the eye54. These different disease outcomes may be related to which of the three most common strains of T. gondii (Type I, II or III) is responsible for the infection46,47,55.
Rhoptry A club-shaped secretory organelle that is found at the anterior end of Apicomplexan parasites that releases its contents during invasion; subdivided into a bulbous base and tapering-neck.
stain either the bulbous base7,8 or the tapering neck of the rhoptry9, but not both. It seems, therefore, that the rhoptry contents are not a random mixture but are sorted into discrete subcompartments. No internal membranes seem to separate these domains and the mechanism by which proteins are sorted to a particular region is not yet known. After translation, at least initially, rhoptry proteins move through a conventional eukaryotic secretory
80 | january 2008 | volume 6
pathway that involves rough endoplasmic reticulum and the Golgi apparatus, but the process of rhoptry formation and the signals that ultimately target proteins to this organelle have yet to be precisely identified10–14. There is some evidence that rhoptries are related to exosomes15, which are membrane-limited bodies that are extruded by some cells. This has interesting implications for the final topology of a rhoptry protein during release, as exosome formation might involve an invagination of the rhoptry membrane and, therefore, place an integral membrane protein in an inverted orientation relative to their position by alternative models16. There are 29 proven rhoptry proteins, of which 24 are present in the rhoptry bulb (most of these are therefore termed ROP proteins) and 5 are present in the rhoptry neck (termed RON proteins) (TABLE 1). Many of these were first discovered by proteomic analyses of purified rhoptries9, which identified an additional 28 proteins that have yet to be verified as being truly rhoptry in origin. Given the high percentage of proteins that were verified as being rhoptry proteins from the first group that was analysed, most of these 28 proteins are probably also from this compartment. Further evidence of a rhoptry origin comes from the fact that many of these 28 proteins are paralogues of known rhoptry proteins, as are the predicted products of several additional genes that have yet to be analysed in detail (TABLE 2). The largest rhoptry gene family shows clear homology to protein kinases17. The canonical member of this family is ROP2 and, in at least one instance, a member of this family (ROP18) has been confirmed to have kinase activity17,18, although the vast majority seem to have lost the key catalytic residues and, hence, the ability to phosphorylate proteins19 (TABLE 1). Apart from their kinase domains, these proteins seem to be unique to the Toxoplasma genus and its close relatives (for example, species of Neospora); no bona fide homologues have been described in the distantly related Plasmodium spp. This could reflect an incomplete knowledge of the rhoptry proteome in species of Plasmodium and/or a difference in the intracellular niches that they occupy. Other ROPs are homologous to phosphatases20 and proteases21,22, but several ROPs, including ROP1, ROP6 and ROP9, seem to be unique to the Toxoplasma genus and are of unknown function23,24. In contrast to the extensive ROP2 family, each of the RON proteins is encoded either by a unique gene that has no similarity to any other gene in the T. gondii genome (for example, RON1 and RON5) or by a gene that has one or two paralogues elsewhere in the T. gondii genome (for example, RON2, RON3 and RON4) (TABLE 2). The initial discovery of the sequence of the RON proteins gave no clue as to their biological function9. Several T. gondii RONs have clear orthologues in related genera, including Plasmodium, which suggested their involvement in processes that are common to the Apicomplexa phylum. Although their exact function has not been determined, the unusual secretion of RON4 and its subsequent migration down the length of the parasite during invasion has led to clear models about the overall role of RON4 and other RONs, as discussed below. www.nature.com/reviews/micro
© 2008 Nature Publishing Group
REVIEWS Finally, in addition to their protein complement, rhoptries also contain lipids25 and this lipid complement has a high ratio of cholesterol to phospholipids. Interestingly, these lipids are sometimes present as membrane whorls inside the rhoptry organelle, which can be visualized using electron microscopy. This might explain the unusual topology of ROP proteins that are released into the host cell during invasion; as discussed below, some ROPs seem to enter the host-cell cytosol in a freely soluble form whereas other ROPs are associated with unusual vesicle-like bodies that seem to fuse with the nascent parasitophorous vacuole.
Paralogue A gene that shares a common evolutionary origin and has evolved in parallel with another gene that is located in the same genome or organism, typically to serve different but related functions.
Orthologue A gene in one species that shares a common evolutionary origin with a related gene in a different species and that serves essentially the same function.
Parasitophorous vacuole The vacuole that harbours the parasite.
Moving junction (MJ). A migrating ring of contact between a host-cell plasma membrane and the surface of an invading parasite.
Role of rhoptries in cell invasion The only circumstance in which rhoptries are known to secrete their contents is during the process of invasion into a host cell (FIGS 2,3; see Supplementary information S1 (movie)). The trigger for release is unknown, but it evidently depends on a direct recognition between the apical surface of the parasite and the receptor molecule (or molecules) on the host cell. The identities of the molecules on either side of this interaction are also unknown and no stimulus has yet been identified that will induce rhoptry secretion in the absence of host-cell contact. The mechanics of the fusion event that allows rhoptry protein release are a mystery. It could be as simple as fusion of the rhoptry to the parasite’s plasma membrane or, perhaps, an intermediate, anterior compartment. Whatever the process, a distinct opening at the anterior-most tip of the tachyzoite is clearly observed during host-cell invasion26 and this is presumed to be the opening through which the rhoptry contents ultimately flow. Once released, rhoptry proteins have various destinations (FIG. 3). The RON proteins, RON2, 4 and 5, form a complex with the micronemal protein AMA1 and this multimeric complex colocalizes with the so-called moving junction (MJ)27,28. The MJ is a ring-like structure that represents the circular point of contact between the parasite surface and host plasma membrane29. During invasion, the MJ migrates down the length of the parasite (FIGS 2,3; see Supplementary information S1 (movie)). AMA1 seems to be necessary for the MJ to form, because parasites in which AMA1 expression has been reduced to ~1% of the wild type (using parasites that harbour a tetracycline-regulated copy of the AMA1 gene) release the RONs, but the MJ fails to form and the parasites do not invade host cells28. The MJ might represent the mechanism by which the parasite makes contact with the host cytoskeleton. The most widely accepted model is one in which the parasite anchors itself on integral proteins of the host plasma membrane and drags them backward (relative to the parasite surface), effectively converting the host plasma membrane into parasitophorous vacuole membrane that surrounds the intracellular portion of the parasite (FIG. 2). This wrapping of the parasite in parasitophorous vacuole membrane is simultaneous with a clear forward motion of the parasite into the host cell, relative to the host-cell’s normal perimeter (Supplementary information S1 (movie)), which strongly argues that the parasite must have a firm
nature reviews | microbiology
a AC RON PVM
M A
ROP G
N
b PV RON
HC
HPM PVM
ROP
Mito
Figure 1 | Toxoplasma gondii Nature ultrastructure. a | An Reviews | Microbiology intracellular T. gondii tachyzoite inside a parasitophorous vacuole membrane (PVM), showing the apical cytoskeleton (AC) and neighbouring micronemes (M), rhoptry bulbs (ROP) and rhoptry necks (RON). Other components of the parasite, such as the nucleus (N), the Golgi apparatus (G) and the plastid that is specific to the Apicomplexa phylum, the ‘Apicoplast’ (A), are shown. The scale bar represents 0.5 µm. b | A lower magnification of a rosette of T. gondii in the parasitophorous vacuole (PV) inside an infected cell (HC). The host plasma membrane (HPM), as well as host mitochondria (Mito) in close apposition to the PVM, can also be seen. The scale bar represents 1 µm. Image in part a reproduced, with permission, from REF. 56 Elsevier Science. Image in part b reproduced, with permission, from REF. 57 Elsevier Science.
grip on something that is connected to fixed anchors within the host cell; that is, the host plasma membrane integral proteins must be connected to the host-cell cytoskeleton. Finally, the model predicts that the volume 6 | january 2008 | 81
© 2008 Nature Publishing Group
REVIEWS Table 1 | Known rhoptry proteins Protein name
Gene identification number*; GenBank accession number
Final Predicted destination‡ coding function
Biological function
Comments
Plasmodium falciparum§
ROP1
583.m00003; M71274
PV
Unknown
Unknown
No
23
ROP2A¶
33.m01398; Z36906
PVM
Protein kinase
No
37
ROP2B¶
63.m00146
PVM
Protein kinase
Mitochondria recruitment Unknown
No
37,58
ROP4¶
83.m02145; Z71787 or AY662677
PVM
Non-catalytic kinase
Unknown
No
19,37
ROP5¶
551.m00238; EF466101 or DQ116423
PVM
Non-catalytic kinase
Unknown
No
9,35,59
ROP7¶
83.m02145; AM056071
PVM
Unknown
No
36
ROP8¶
33.m00005; AF011377
PVM
No
58
ROP9
49.m00048; AJ401616
Unknown
Non-catalytic kinase Non-catalytic kinase Unknown
Knock-out in a type I strain is still virulent More adjacent genes probably exist but are missing from ToxoDB4.2 Should be a tandem, identical copy of ROP2A but appears as a fragment in the ME49 sequence in ToxoDB4.2 Phosphorylated; mis-annotated in ToxoDB4.2 as one gene (fusion with ROP7) Approximately 3 or 4 more tandem copies exist that are not present in the ME49 sequence in ToxoDB4.2 Mis-annotated in ToxoDB4.2 as one gene (fusion with ROP4) None
No
24
ROP10 ROP11¶
Unknown Unknown
Unknown Protein kinase
Unknown Unknown
No No
9 9
ROP12 ROP13
583.m05686; DQ124368 42.m03584; AAZ29607 or DQ077905 20.m08222; DQ096559 583.m09115; DQ096560
Distinct from another protein named ROP9 (Ref. 60), for which the sequence and gene are unidentified None None
Unknown Unknown
Unknown Unknown
Unknown Unknown
No No
9 9
ROP14 ROP15 ROP16¶
583.m00692; DQ096565 Unknown 27.m00091; DQ096561 Unknown 55.m08219; DQ116422 Host nucleus
Yes No No
9 9 9
ROP17¶ ROP18¶
Unknown PVM
No No
17 17
TgSUB2
55.m08191; AM075203 20.m03896; AM075204 or EF092842 583.m00011; AF420596
None Mis-annotated in ToxoDB4.2; encoded on the opposite strand None None Extremely different in T. gondii type I, II and III strains None Extremely different in T. gondii type I, II and III strains None
Yes
22
Toxopain1
50.m00008; AY071839
Unknown
None
No
21
Injected into host cell where it ends up in the nucleus; knock-out has a slight growth defect in vitro Knock-out in type I strain is still virulent Possible interference with host actin Knock-out in type II strain is still virulent
No
20
Yes
61
No No
62 63
None
Yes
9
None
Yes
9
Covalently attached to RON4 None Covalently attached to RON2 Cleaved into three major fragments
Yes Yes Yes Yes
9 9 9 None
Unknown
TgPP2C-hn 74.m00766 + 74m.00767; Host nucleus EF450457 TgNHE2
129.m00252; AY735393
Unknown
Toxofilin BRP1
33.m02185; AJ132777 583.m09133
Unknown Unknown
TgRAB11
80.m00009
Unknown
RON1
583.m11443 + 583. m00597; DQ096562 145.m00331; DQ096563 42.m00026; DQ096564 44.m06355; DQ096566 583.m09191 + 583. m09192 + 583.m00636
Unknown
RON2 RON3 RON4 RON5#
MJ Unknown MJ Possibly MJ
Unknown Unknown
Transmembrane Unknown Unknown Unknown Protein kinase Host subversion and virulence Protein kinase Unknown Protein kinase Virulence Serine protease ROP protein processing Cathepsin B ROP protein protease processing Protein Unknown phosphatase 2C Na+ and H+ Ionic exchanger homeostasis Actin-binding Unknown Unknown Bradyzoite- and merozoitespecific Small GTPase Protein trafficking Possibly GPIUnknown anchored Transmembrane Invasion Transmembrane Unknown Unknown Invasion Unknown Invasion
Refs||
*The ToxoDB4.2 (Toxoplasma gondii Genome resource; see Further information64) gene identifier is provided for the T. gondii ME49 strain, along with the GenBank entry for (typically, but not always) the T. gondii RH strain, if such an entry exists. The GenBank entry should be a verified coding sequence and, therefore, encode the definitive amino-acid sequence of the primary translation product. The gene identifiers in ToxoDB4.2 are tentative predictions that in most cases have yet to be experimentally confirmed. Consequently, although they are approximately correct, the transcription start sites, splice junctions and predicted protein sequences might be incorrect. This is why some genes have multiple identifiers — for example, the large RON1 gene is currently incorrectly annotated as two different genes (indicated by the + symbol, which separates the gene identifiers listed). ‡ Final location of the protein after invasion. § Significant orthologue, with homology that extends beyond the presence of a conserved enzymatic domain (for example, more than just a kinase or protease active site is conserved throughout evolution), exists in P. falciparum. ||The reference that reports the definitive sequence and/or identification of this protein as a rhoptry protein. ¶Homologue of ROP2. No gene has yet been linked to the ROP2 family member that has been dubbed ROP3. ROP3 could simply be a post-translational modification of one of the known ROP2 family members or encoded by one of the ROP2-like genes for which the protein product has yet to be detected. #P. Bradley and M. Lebrun, unpublished observations. GPI, glycophosphatidylinositol; MJ, moving junction; PV, parasitophorous vacuole; PVM, parasitophorous vacuole membrane.
82 | january 2008 | volume 6
www.nature.com/reviews/micro © 2008 Nature Publishing Group
REVIEWS Table 2 | Suspected rhoptry proteins Protein name
Gene identification number*; GenBank accession number
Final Predicted coding destination‡ function
Comments
Plasmodium falciparum§
ROP6
55.m00092; AY792971
Possibly HPM
Possible protease and GPI-anchored
None
No
665
ROP2L3¶
55.m08224
Unknown
Protein kinase
Mass spectrometry data for expression**; detected in rhoptry proteome‡‡
No
9,17
ROP2L4¶
57.m01774
Unknown
Protein kinase
None
No
17
ROP2L5¶
49.m03275
Unknown
Protein kinase
None
No
17
¶
ROP2L6
80.m02343
Unknown
Non-catalytic kinase
Mass spectrometry data for expression**; detected in rhoptry proteome‡‡
No
9,17
ROP2L7¶
49.m03159
Unknown
Protein kinase
None
No
None
¶
ROP2L8
52.m01543
Unknown
Protein kinase
Mass spectrometry data for expression**; detected in rhoptry proteome‡‡
No
9
ROP2L9¶
55.m04788
Unknown
Protein kinase
Mass spectrometry data for expression**
No
None
ROP2L10¶
59.m06126
Unknown
Protein kinase
None
No
None
¶
ROP2L11
86.m00398
Unknown
Unknown
Truncated pseudogene
No
None
ROP2L12¶
86.m00844
Unknown
Unknown
Truncated pseudogene
No
None
ROP2L13
25.m01746
Unknown
Non-catalytic kinase
None
No
None
RON2L1#
83.m01266
Unknown
Unknown
None
Yes
28
RON2L2#
57.m01722
Unknown
Unknown
None
Yes
28
#
RON3L1
20.m03905
Unknown
Unknown
Mass spectrometry data for expression**
Yes
None
RON4L1#
52.m01582
Unknown
Unknown
Mass spectrometry data for expression**
Yes
28
¶
Refs||
*The ToxoDB4.2 (Toxoplasma gondii Genome resource; see Further information64) gene identifier is provided for the T. gondii ME49 strain, along with the GenBank entry for (typically, but not always) the T. gondii RH strain, if such an entry exists. The GenBank entry should be a verified coding sequence and, therefore, encode the definitive amino-acid sequence of the primary translation product. The gene identifiers in ToxoDB4.2 are tentative predictions that in most cases have yet to be experimentally confirmed. Consequently, although they are approximately correct, the transcription start sites, splice junctions and predicted protein sequences might be incorrect in their detail. ‡Final location of the protein after invasion. § Significant orthologue, with homology that extends beyond the presence of a conserved enzymatic domain (for example, more than just a kinase or protease active site is conserved throughout evolution), exists in P. falciparum. ||Reference that originally reported this gene or protein. ¶ROP2-like (ROP2L) protein predicted in ToxoDB4.2 but not yet confirmed to be rhoptry localized and, except as noted, not even confirmed to be expressed. #RON2-, RON3- or RON4-like protein (RON2L, RON3L and RON4L, respectively) predicted in ToxoDB4.2 but not yet confirmed to be rhoptry localized and, except as noted, not even confirmed to be expressed. **Mass spectrometry has revealed that one or more peptides are present in tachyzoites based on data presented in ToxoDB4.2. ‡‡Detected by mass spectrometry in rhoptry-enriched fraction but not yet confirmed to be rhoptrylocalized. The biological function of these proteins is unknown. GPI, glycophosphatidylinositol; HPM, host plasma membrane.
region of contact between the parasite and the host cell needs to be a circular band, as otherwise a gliding parasite would not be able to take a ‘dive’ into the host cell — it would simply keep moving along the surface. AMA1, which was first described in Plasmodium falciparum, might organize the MJ into a ring28. The association of AMA1 with RON4 is also observed in P. falciparum, which indicates that the overall collaboration of rhoptry RONs with micronemal AMA1 is a conserved feature of most, if not all, species of the Apicomplexa phylum30. How the RON proteins (RON2, 4 and 5) that form the MJ complex associate with each other is not known, although RON2 and RON4 seem to be linked by disulphide bonds 28. How these proteins further associate with a micronemal protein such as AMA1 is also unclear; it is presumed, however, that they come into contact with one another at, or just below, the apical surface of the parasite. Likewise, the topology of the various components of the MJ complex within the membrane (or membranes) has not yet been determined. RON2 has at least 2, and possibly 3, predicted transmembrane domains, but it is unclear which membrane (host or parasite) they nature reviews | microbiology
are embedded within. One exciting possibility is that RON2, or a different MJ protein, is inserted into the host plasma membrane during the first steps of invasion and that it then contacts the host cytoskeleton to provide the anchoring that is described above. This is analogous to a phenomenon that has been reported for enteropathogenic Escherichia coli, in which the bacterium inserts a protein (Tir) into the host cell that then plays a part in attachment and subsequent invasion31. An appealing feature of such a model is that it would explain the ability of T. gondii to invade almost any cell from a wide range of warm-blooded animals; if they provide their own anchor, they could make a home in almost any port, as long as they are able to make contact with a highly conserved cytoskeletal protein. Alternatively, of course, the MJ proteins could remain anchored in the parasite’s plasma membrane and associate with host-plasmamembrane structures that are, in turn, anchored to the host-cell cytoskeleton. If so, a portion, or domain, of that host‑plasma‑membrane molecule would need to be structurally conserved among many species and cell types to explain the extraordinary range of hosts and cells that T. gondii can productively infect. volume 6 | january 2008 | 83
© 2008 Nature Publishing Group
REVIEWS
HC
AC
* MJ
PVM MJ
Figure 2 | Toxoplasma gondii invasion. A T. gondii Nature Reviews | Microbiology tachyzoite invading an HeLa cell (HC). An irregularly shaped organelle that is derived from rhoptry exocytosis (asterisk) is found near the apical cytoskeleton (AC). The parasitophorous vacuole membrane (PVM) that is derived from the host cell membrane is found around the portion of the parasite that has invaded the host cell. The invasion is thought to be driven by parasite motors acting at the moving junction (MJ). The scale bar represents 0.5 µm. Image reproduced, with permission, from REF. 56 Elsevier Science.
ROP proteins are injected into host cells ROPs seem to have completely different functions from RONs. Like the neck proteins, ROPs are released during invasion but they do not form organized structures and are not found at the MJ. Instead, following release, they migrate to one of three general locations: the lumen of the nascent parasitophorous vacuole; the parasitophorous vacuole membrane; or the interior of the host cell (FIG. 3). ROP1 is an example of migration to the first location: it is released during invasion and accumulates within the lumen of the nascent parasitophorous vacuole23. Remarkably, based on studies that used a combination of ROP1 knock-out parasites and parasite lines that express epitope-tagged versions of ROP1, it seems that ROP1 can be synthesized in one parasite but end up in the parasitophorous vacuole of another32. This suggests that ROP1 is not simply ‘dumped’ into the nascent parasitophorous vacuole during invasion, but instead is released into the host cell where it then migrates to the nearest parasitophorous vacuole. Because in nature most cells will be infected by only 84 | january 2008 | volume 6
one parasite, the nearest parasitophorous vacuole is usually that of the same parasite that released the ROP1 during invasion. Small vesicle-like structures referred to as evacuoles have been observed that contain ROP1 but are devoid of parasites (hence they have also been named ‘empty’-vacuoles or ‘e’-vacuoles32). Evacuoles are observed in a small, but significant, percentage of invasion events in tissue culture and their frequency can be increased if invasion is blocked by drug treatment. For example, treatment with cytochalasin D prevents the actin and myosin motors of the parasite from exerting their propulsive force, although attachment is not affected (the host cell’s actin and myosin motors appear to be irrelevant to parasite invasion17,33). The result is a ‘frustrated’ parasite that is stuck to the outside of a host cell that contains many evacuoles but lacks a developing parasitophorous vacuole membrane for the ROP1-containing evacuoles to fuse to. The ROP2 family of proteins generally migrates to the second location for rhoptry proteins, the parasitophorous vacuole membrane19,34–37. It seems that several members of this protein family are intimately associated with the parasitophorous vacuole membrane, and are possibly even integral membrane proteins. Early suggestions that a hydrophobic alpha helix functions as a transmembrane domain34 have been called into question now that we know that this helix is a conserved feature of most protein kinases and its hydrophobicity is a necessary feature of its being buried within the interior of the protein. Although no crystal structure of a ROP2 family member has been reported, it seems unlikely that a helix would be used to span a membrane, especially given the clear conservation of sequence (and presumably structure) on either side of the hydrophobic portion. ROP2 has also been implicated in the recruitment of host-cell mitochondria 38,39. This has been postulated to be through recognition of the processed amino terminus of ROP2, which resembles a mitochondrial-import signal (that is, an amphipathic helix). The proposal is that host mitochondria mistakenly attempt to import ROP2, which is somehow firmly tethered to the parasitophorous vacuole membrane. The result is that as the mitochondria attempt to ‘reel in’ the ROP2 protein, they ratchet down onto the parasitophorous vacuole membrane, which is where they are routinely observed by electron microscopy. Recent data on ROP18, a member of the ROP2 family, might indicate an interesting aspect of its association with the parasitophorous vacuole membrane; when expressed inside an infected host cell, by direct transfection of the ROP18 gene (minus the portion encoding a signal peptide), the protein is eventually found concentrated on the parasitophorous vacuole membrane, presumably on the face that is exposed to the host cytosol18. This is also the location of a putative parasite-derived kinase that phosphorylates host iκB40. Whether ROP18 or another member of the ROP2 family is responsible for iκB phosphorylation remains to be directly investigated and the mechanism by which these proteins associate with the parasitophorous vacuole membrane is likewise unknown. www.nature.com/reviews/micro
© 2008 Nature Publishing Group
REVIEWS Nucleus
Rhoptry bulb
Moving junction
Microneme Rhoptry neck
Toxofilin?
Host plasma membrane
PVM
ROP2 family ROP1
ROP16 PP2C-hn Nucleus
Figure 3 | Schematic model for rhoptry contribution to invasion. Rhoptry bulbs Nature Reviews | Microbiology (grey) and rhoptry necks (red) release their contents during the invasion process in concert with simultaneous release by micronemes (green). RON2, RON4 and RON5 collaborate with micronemal AMA1 to create the moving junction, which migrates down the surface of the parasite, forming a ring of contact with the host plasma membrane. This effectively excludes many host plasma membrane integral proteins and results in the generation of a parasitophorous vacuole membrane (PVM), which envelopes the parasite. ROP2 family members are injected during invasion, perhaps in association with small vesicles, and ultimately end up on the host cytosolic side of the PVM. ROP1 is also observed in association with the injected vesicles, but most of this protein ends up inside the parasitophorous vacuole lumen. ROP16 and PP2C-hn (hn is the abbreviation for host nucleus) are not observed in the vesicles, but accumulate inside the host nucleus. Other soluble rhoptry proteins (for example, toxofilin) are also presumed to be injected, but in the absence of a concentrating mechanism they will be present at too low a concentration to be detectable (only a few rhoptries secrete their contents during invasion and any proteins that they release will be diluted by up to a million-fold or more in the host cytosol). Figure adapted from Ref. 28.
The third known destination for ROPs is the interior of the host cell or, more specifically, the host-cell nucleus. So far, two rhoptry proteins have been observed in the nucleus, a protein phosphatase of the 2C class (PP2C-hn; hn is the abbreviation for host nucleus)20 and a putative protein kinase that is a highly divergent member of the ROP2 family (ROP16 (Ref. 41)). As for the other ROPs, how these proteins enter the host cell is a mystery. The only clues come from patch-clamp experiments, which show that there is a break in the continuity of the host plasma membrane at the earliest times of invasion42, perhaps reflecting the moment when injection occurs (FIGS 2,3). The process is clearly an early one, as both ROP16 and PP2C-hn are detectable in the host nucleus within 15 minutes of allowing infection to commence, which is close to the time that such proteins would need to reach the host nucleus if they were introduced at time zero. It should be noted that once ROP16 and PP2C-hn enter the host cytosol, conventional nuclear localization signals (NLSs) are used to traffic them to the nucleus. This has been demonstrated by showing that ablation of a classic NLS signature, which both proteins have, stops these nature reviews | microbiology
proteins from concentrating in the nucleus41,43. There is no evidence for the existence in T. gondii of genes or proteins that are related to those used by bacteria for introducing proteins into a host cell (for example, type III or type IV secretion systems44), although there are many parallels between these processes. The function of the rhoptry PP2C-hn is unknown, as a knock-out strain showed no changes in host-gene expression (based on microarray analysis of infected cells) or virulence (based on infection studies in mice), although the knock-out strain was partially compromised in its ability to grow in fibroblasts in vitro43. Any conclusion concerning the lack of a virulence phenotype must be qualified, however, because the knock-out of PP2C-hn was in an RH strain in which virulence is so high (LD100 of 1 parasite) that anything short of a dramatic reduction in virulence might still yield a parasite that is capable of killing a mouse.
ROP16 and ROP18 — roles in virulence It has long been known that different strains of T. gondii produce radically different pathologies in mice and maybe even in humans45. Using F1 progeny from crosses between two strains that differ in their virulence, two research groups have mapped the parasite loci that are responsible for the different pathogenicities46,47. The results showed that virulence in a given host (in these cases, mice) is influenced by which allele is present at each of at least five loci. Two of the most important loci are ROP16 and ROP18; for example, depending on which allele of ROP18 a strain carries, its LD50 in mice can vary by over 4 logs. The impact of the ROP16 locus is less dramatic but still significant. The identities of the other three virulence loci have yet to be determined. For ROP18, there is still much to be learned about the exact mechanism by which the different alleles effect such dramatic differences in disease outcome. It is known that the allele that is associated with low virulence yields a tiny fraction (~0.1%) of ROP18 mRNA compared with the amount that is produced by the alleles found in more virulent strains46,47. This seems to be due to the presence of a large insertion and a small deletion within the presumptive promoter region of the low virulence allele, which seems to render the promoter inactive. The precise function of the ROP18 protein, however, is not yet known, although it clearly does have potent kinase activity, and a parasite substrate has been observed but not yet identified18. ROP16 subverts host gene expression Microarray experiments using infected host cells in vitro have shown that the infected host cell responds differently depending on which strain of T. gondii is used and that these differences also segregate as distinct phenotypes among the F1 progeny41. Importantly, ROP16 is among the main parasite loci that have been shown to be responsible for these differences. Testing of specifically engineered strains showed that this putative protein kinase somehow intersects the host signal transducer and activator of transcription (STAT) pathways41. These pathways are central to the regulation of many host genes, volume 6 | january 2008 | 85
© 2008 Nature Publishing Group
REVIEWS including several cytokines and other immune-response mediators. This suggested the obvious possibility that differences in how a host cell responds to T. gondii might result in differences in the disease pathology. One of the main immune mediators that ROP16 affects by the perturbation of STAT pathways seems to be interleukin (IL)-12, which is central to the host response to T. gondii infection48,49. Too much, or too little, IL-12 can have serious and negative consequences for the host by causing too much or too little of the necessary immune responses, and so its variable expression, which depends on the allele of ROP16 that is carried by the invading parasite, is an attractive explanation for the differences in the disease that are caused by the various strains. This possibility has been explored in vivo and ROP16 does indeed seem to have a role in determining serum IL-12 levels and, therefore, the overall course of infection41. Using the terminology that is favoured by those studying bacterial pathogens, ROPs are crucial ‘effector proteins’. They might equally be called ‘negotiator proteins’, owing to their role in managing the interaction between host and parasite, as the vast majority of T. gondii infections result in a persistent infection with little, if any, disease.
Rhoptries and T. gondii evolution ROP16 and ROP18 are members of a large gene family (the ROP2 family) and are two of the most variable loci in the entire T. gondii genome. The ratio of synonymous to nonsynonymous substitutions is also extremely high, thus strengthening the argument that these two genes are subject to a strong positive selection for change41,46,47. What is the pressure that drives this process? One possibility is simply that it is the immune pressure of the sort that drives variation in many pathogen proteins. ROP16 and ROP18 may particularly be subject to such selection, relative to other T. gondii antigens, because they seem to be accessible to the host cytosol and are, therefore, freely available to the host cell for class I major histocompatibility complex presentation. As yet, however, there are no data to indicate that these proteins are efficiently presented or are crucial as targets of the immune response in the conventional sense. This is in contrast to the well-characterized family of surface antigens that is known as the SAG1 related sequences or SRS family, which are immunodominant antigens (at least for the humoral response) that show a more modest level of variability between strains50. An alternative hypothesis for the diversification of ROP16 and ROP18 is selection for an expanded host range. This model posits that if a strain found itself in a new ecological niche, and therefore in a new host, as long as it could infect to some degree (and be transmitted), there would be a strong pressure to optimize the ‘fit’ between a negotiator protein and the new host. This could lead to a powerful and rapid selection for new gene variants or, as is the case for ROP18, selection for dramatic events that downregulate or upregulate its expression (as discussed above, the ROP18 locus in type III strains of T. gondii has a large 86 | january 2008 | volume 6
disruption in the promoter region that is thought to be responsible for the massive decrease in its expression relative to type I and II strains). The hypothesis is that, in some hosts, high levels of expression of ROP18 are so problematic to the host or parasite that either the host or parasite is killed prematurely, with the result that there is little, if any, transmission. For example, the highly expressed type II allele of ROP18, if present in a type III background, generally causes that strain to be fatal to mice46. Hence, the promoter disruption of ROP18 might have been necessary for T. gondii to productively infect mice, whereas full-on expression might be needed for the infection of some other intermediate hosts. The differences in the coding region (as opposed to the promoter) might be evidence for more subtle changes that are needed for optimization of the interaction of ROP18 with its respective target in related host species. Ultimately, the situation might be somewhat analogous to that of the influenza virus, in which both immune pressure and host range seem to play important parts in the evolution of several of its genes. For example, the haemagglutinin gene of H3N2 viruses is constantly evolving under selective pressure from the immune system, but for an H5N1 virus to become transmissible between humans other mutations may need to occur that allow binding to the particular receptors that are present in the relevant human tissue51. When did these differences arise? To answer this question, additional sequences of ROP16 and ROP18 genes, from many more strains that infect diverse hosts and, therefore, have diverse ecological niches, are clearly needed. It is possible that the extremely large number of coding-sequence differences are a vestige of a long evolutionary time period, during which time the sequences became more divergent as the genes were optimized to different niches. The disruption of the promoter might be a more recent event that in one stroke expanded the parasite’s host range to accommodate the emergence of a new niche, which, perhaps, is related to human migrations. One scenario might be that an ancestral North American strain that experienced the promoter disruption suddenly became a productive infector of Norwegian roof rats or European house sparrows, both of which were introduced to North America only in the past few centuries. Genotyping of some of the currently most common strains has shown that one or two matings can have a dramatic impact on the population biology of T. gondii 52 and that recent mixing of gene pools from distant geographic locations has probably occurred; South America appears to be the source of some of the greatest diversity53. Such matings may have given rise to recombinant strains that have just the right mix of crucial rhoptry proteins, which has enabled the productive infection of hosts that were previously not susceptible to this parasite.
Clinical implications of rhoptry biology A detailed understanding of rhoptry-protein functions could have profound clinical implications that are well www.nature.com/reviews/micro
© 2008 Nature Publishing Group
REVIEWS beyond the usual opportunities for the development of new vaccines or drugs to block parasite metabolism. A compound that interferes with assembly or migration of the MJ, for example, would clearly be lethal to these obligate intracellular parasites. Much of the MJ machinery is conserved in the malaria parasite9,30, so the benefit of such a drug could also extend to treating this worldwide scourge. ROP16 and ROP18 are kinases that are injected into the host cell, where their different allelic forms substantially alter the host–pathogen interaction18,41,46,47. This finding might be key to appropriate decision making if treating severe ocular toxoplasmosis with steroids that suppress inflammation54. On the one hand, if infection is with a strain in which the ROP16 allele causes an excessive IL-12 response, dampening of inflammation might be clinically appropriate to prevent damage to the eye itself. On the other hand, suppressing the immune response to a strain that expresses a ROP16 allele, which is associated with a weak IL-12 response (and therefore is naturally producing little inflammation), might result in an inadequate immune defence, leading to uncontrolled parasite growth. Knowing which strain is infecting a patient, therefore, might allow future clinicians who are armed with a refrigerator full of immune-modulators (for example, recombinant cytokines), to tweak the immune response in the right direction and to the right degree. Luft, B. J. & Remington, J. S. Toxoplasmic encephalitis in AIDS. Clin. Infect. Dis. 15, 211–222 (1992). 2. Dubremetz, J. F., Garcia-Reguet, N., Conseil, V. & Fourmaux, M. N. Apical organelles and host-cell invasion by Apicomplexa. Int. J. Parasitol. 28, 1007–1013 (1998). 3. Dubey, J. P. in Parasitic Protozoa (ed. Kreier, J. P.) 1–158 (Academic, California, 1993). 4. Pepper, M., Dzierszinski, F., Crawford, A., Hunter, C. A. & Roos, D. Development of a system to study CD4+‑T‑cell responses to transgenic ovalbuminexpressing Toxoplasma gondii during toxoplasmosis. Infect. Immun. 72, 7240–7246 (2004). 5. Gubbels, M. J., Striepen, B., Shastri, N., Turkoz, M. & Robey, E. A. Class I major histocompatibility complex presentation of antigens that escape from the parasitophorous vacuole of Toxoplasma gondii. Infect. Immun. 73, 703–711 (2005). 6. Dubey, J. P., Lindsay, D. S. & Speer, C. A. Structures of Toxoplasma gondii tachyzoites, bradyzoites, and sporozoites and biology and development of tissue cysts. Clin. Microbiol. Rev. 11, 267–299 (1998). 7. Sadak, A., Taghy, Z., Fortier, B. & Dubremetz, J. F. Characterization of a family of rhoptry proteins of Toxoplasma gondii. Mol. Biochem. Parasitol. 29, 203–211 (1988). 8. Saffer, L. D., Mercereau-Puijalon, O., Dubremetz, J. F. & Schwartzman, J. D. Localization of a Toxoplasma gondii rhoptry protein by immunoelectron microscopy during and after host cell penetration. J. Protozool. 39, 526–530 (1992). 9. Bradley, P. J. et al. Proteomic analysis of rhoptry organelles reveals many novel constituents for host– parasite interactions in Toxoplasma gondii. J. Biol. Chem. 280, 34245–34258 (2005). 10. Hoppe, H. C., Ngo, H. M., Yang, M. & Joiner, K. A. Targeting to rhoptry organelles of Toxoplasma gondii involves evolutionarily conserved mechanisms. Nature Cell Biol. 2, 449–456 (2000). 11. Bradley, P. J. & Boothroyd, J. C. The pro region of Toxoplasma ROP1 is a rhoptry-targeting signal. Int. J. Parasitol. 31, 1177–1186 (2001). 12. Striepen, B., Soldati, D., Garcia-Reguet, N., Dubremetz, J. F. & Roos, D. S. Targeting of soluble proteins to the 1.
Conclusions Rhoptries contain many of the key molecules that are used for parasite entry into host cells and subversion of host functions. The MJ, which contains four known proteins, AMA1, RON2, RON4 and RON5, and probably several other proteins that have yet to be identified, plays a central part during invasion. Determining how these various molecules interact to create the MJ is, therefore, of great importance to understanding how the parasitophorous vacuole forms. Such knowledge is also likely to yield some novel findings about interactions between membranes, as migration of a circular ring of contact between two membranes without fusion is an unusual process in biology. The ROP2 family (which includes ROP16 and ROP18) is extensive and we have almost no understanding of how this large family of proteins interacts with itself, let alone other host or parasite proteins. Clearly, based on their high sequence divergence and role in virulence they are crucial to the host–pathogen interaction, but the molecular details remain to be discovered. Even those members of the ROP2 family that seem to have lost kinase activity might be important and could mediate an effect by regulating the activity of those members that retain the ability to phosphorylate other proteins. The prediction that rhoptries contain many of the molecules that are key to an intracellular lifestyle is proving correct and recent results are a tantalizing glimpse of an exciting future.
rhoptries and micronemes in Toxoplasma gondii. Mol. Biochem. Parasitol. 113, 45–53 (2001). 13. Ngo, H. M. et al. AP‑1 in Toxoplasma gondii mediates biogenesis of the rhoptry secretory organelle from a post-Golgi compartment. J. Biol. Chem. 278, 5343–5352 (2003). 14. Bradley, P. J., Li, N. & Boothroyd, J. C. A GFP-based motif-trap reveals a novel mechanism of targeting for the Toxoplasma ROP4 protein. Mol. Biochem. Parasitol. 137, 111–120 (2004). 15. Ngo, H. M., Yang, M. & Joiner, K. A. Are rhoptries in Apicomplexan parasites secretory granules or secretory lysosomal granules? Mol. Microbiol. 52, 1531–1541 (2004). 16. Fevrier, B. & Raposo, G. Exosomes: endosomalderived vesicles shipping extracellular messages. Curr. Opin. Cell Biol. 16, 415–421 (2004). 17. El Hajj, H. et al. The ROP2 family of Toxoplasma gondii rhoptry proteins: proteomic and genomic characterization and molecular modeling. Proteomics 6, 5773–5784 (2006). 18. El Hajj, H. et al. ROP18 is a rhoptry kinase controlling the intracellular proliferation of Toxoplasma gondii. PLoS Pathog. 3, e14 (2007). The first direct demonstration of the kinase activity of the rhoptry protein ROP18 and its role in T. gondii proliferation in vitro. 19. Carey, K. L., Jongco, A. M., Kim, K. & Ward, G. E. The Toxoplasma gondii rhoptry protein ROP4 is secreted into the parasitophorous vacuole and becomes phosphorylated in infected cells. Eukaryot. Cell 3, 1320–1330 (2004). 20. Gilbert, L. A., Ravindran, S., Turetzky, J. M., Boothroyd, J. C. & Bradley, P. J. Toxoplasma gondii targets a protein phosphatase 2C to the nuclei of infected host cells. Eukaryot. Cell 6, 73–83 (2007). The first study to reveal the targeting of a rhoptry protein to the host-cell nucleus after invasion. 21. Que, X. et al. The cathepsin B of Toxoplasma gondii, toxopain‑1, is critical for parasite invasion and rhoptry protein processing. J. Biol. Chem. 277, 25791–25797 (2002). 22. Miller, S. A., Thathy, V., Ajioka, J. W., Blackman, M. J. & Kim, K. TgSUB2 is a Toxoplasma gondii rhoptry organelle processing proteinase. Mol. Microbiol. 49, 883–894 (2003).
nature reviews | microbiology
23. Ossorio, P. N., Schwartzman, J. D. & Boothroyd, J. C. A Toxoplasma gondii rhoptry protein associated with host cell penetration has unusual charge asymmetry. Mol. Biochem. Parasitol. 50, 1–15 (1992). 24. Reichmann, G., Dlugonska, H. & Fischer, H. G. Characterization of TgROP9 (p36), a novel rhoptry protein of Toxoplasma gondii tachyzoites identified by T cell clone. Mol. Biochem. Parasitol. 119, 43–54 (2002). 25. Foussard, F., Leriche, M. A. & Dubremetz, J. F. Characterization of the lipid content of Toxoplasma gondii rhoptries. Parasitology 102 Pt 3, 367–370 (1991). 26. Nichols, B. A., Chiappino, M. L. & O’Connor, G. R. Secretion from the rhoptries of Toxoplasma gondii during host-cell invasion. J. Ultrastruct. Res. 83, 85–98 (1983). First visualization of apparent rhoptry secretion into an infected host cell. 27. Lebrun, M. et al. The rhoptry neck protein RON4 re-localizes at the moving junction during Toxoplasma gondii invasion. Cell. Microbiol. 7, 1823–1833 (2005). 28. Alexander, D. L., Mital, J., Ward, G. E., Bradley, P. & Boothroyd, J. C. Identification of the moving junction complex of Toxoplasma gondii: a collaboration between distinct secretory organelles. PLoS Pathog. 1, e17 (2005). This paper was the first to identify a cooperation between rhoptry and microneme proteins in invasion. Together with reference 27 it was also the first to show the involvement of RON proteins in the MJ. 29. Aikawa, M., Miller, L. H., Johnson, J. & Rabbege, J. Erythrocyte entry by malarial parasites: a moving junction between erythrocyte and parasite. J. Cell Biol. 77, 72–82 (1978). 30. Alexander, D. L., Arastu-Kapur, S., Dubremetz, J. F. & Boothroyd, J. C. Plasmodium falciparum AMA1 binds a rhoptry neck protein homologous to TgRON4, a component of the moving junction in Toxoplasma gondii. Eukaryot. Cell 5, 1169–1173 (2006). 31. Kenny, B. et al. Enteropathogenic E. coli (EPEC) transfers its receptor for intimate adherence into mammalian cells. Cell 91, 511–520 (1997).
volume 6 | january 2008 | 87 © 2008 Nature Publishing Group
REVIEWS 32. Hakansson, S., Charron, A. J. & Sibley, L. D. Toxoplasma evacuoles: a two-step process of secretion and fusion forms the parasitophorous vacuole. EMBO J. 20, 3132–3144 (2001). First direct visualization of rhoptry-protein trafficking within an infected cell. 33. Dobrowolski, J. M. & Sibley, L. D. Toxoplasma invasion of mammalian cells is powered by the actin cytoskeleton of the parasite. Cell 84, 933–939 (1996). 34. Beckers, C. J. M., Dubremetz, J. F., MercereauPuijalon, O. & Joiner, K. A. The Toxoplasma gondii rhoptry protein ROP2 is inserted into the parasitophorous vauole membrane, surrounding the intracellular parasite, and is exposed to the host cell cytoplasm. J. Cell Biol. 127, 947–961 (1994). First report that rhoptry proteins can be directly exposed to the infected host-cell cytosol. 35. El Hajj, H., Lebrun, M., Fourmaux, M. N., Vial, H. & Dubremetz, J. F. Inverted topology of the Toxoplasma gondii ROP5 rhoptry protein provides new insights into the association of the ROP2 protein family with the parasitophorous vacuole membrane. Cell. Microbiol. 9, 54–64 (2007). 36. El Hajj, H., Lebrun, M., Fourmaux, M. N., Vial, H. & Dubremetz, J. F. Characterization, biosynthesis and fate of ROP7, a ROP2 related rhoptry protein of Toxoplasma gondii. Mol. Biochem. Parasitol. 146, 98–100 (2006). 37. Herion, P., Hernandez-Pando, R., Dubremetz, J. F. & Saavedra, R. Subcellular localization of the 54-kDa antigen of Toxoplasma gondii. J. Parasitol. 79, 216–222 (1993). 38. Sinai, A., Webster, P. & Joiner, K. Association of host cell endoplasmic reticulum and mitochondria with the Toxoplasma gondii parasitophorous vacuole membrane: a high affinity interaction. J. Cell. Sci. 110, 2117–2128 (1997). 39. Sinai, A. P. & Joiner, K. A. The Toxoplasma gondii protein ROP2 mediates host organelle association with the parasitophorous vacuole membrane. J. Cell. Biol. 154, 95–108 (2001). 40. Molestina, R. E., Payne, T. M., Coppens, I. & Sinai, A. P. Activation of NF‑κB by Toxoplasma gondii correlates with increased expression of antiapoptotic genes and localization of phosphorylated IκB to the parasitophorous vacuole membrane. J. Cell Sci. 116, 4359–4371 (2003). 41. Saeij, J. P. et al. Toxoplasma co-opts host gene expression by injection of a polymorphic kinase homologue. Nature 445, 324–327 (2007). 42. Suss-Toby, E., Zimmerberg, J. & Ward, G. E. Toxoplasma invasion: the parasitophorous vacuole is formed from host cell plasma membrane and pinches off via a fission pore. Proc. Natl Acad. Sci. USA 93, 8413–8418 (1996). 43. Gilbert, R. E. et al. Screening for congenital toxoplasmosis: accuracy of immunoglobulin M and immunoglobulin A tests after birth. J. Med. Screen. 14, 8–13 (2007).
44. Ghosh, P. Process of protein transport by the type III secretion system. Microbiol. Mol. Biol. Rev. 68, 771–795 (2004). 45. Saeij, J. P., Boyle, J. P. & Boothroyd, J. C. Differences among the three major strains of Toxoplasma gondii and their specific interactions with the infected host. Trends Parasitol. 21, 476–481 (2005). 46. Saeij, J. P. et al. Polymorphic secreted kinases are key virulence factors in toxoplasmosis. Science 314, 1780–1783 (2006). Demonstrated the major role of two rhoptry proteins (ROP16 and ROP18) in T. gondii virulence in mice. 47. Taylor, S. et al. A secreted serine-threonine kinase determines virulence in the eukaryotic pathogen Toxoplasma gondii. Science 314, 1776–1780 (2006). Identified the major role that ROP18 has in T. gondii proliferation in vitro and virulence in mice. 48. Hunter, C. A. & Remington, J. S. The role of IL12 in toxoplasmosis. Res. Immunol. 146, 546–552 (1995). 49. Scharton-Kersten, T., Denkers, E. Y., Gazzinelli, R. & Sher, A. Role of IL12 in induction of cell-mediated immunity to Toxoplasma gondii. Res. Immunol. 146, 539–545 (1995). 50. Jung, C., Lee, C. Y. & Grigg, M. E. The SRS superfamily of Toxoplasma surface proteins. Int. J. Parasitol. 34, 285–296 (2004). 51. Parrish, C. R. & Kawaoka, Y. The origins of new pandemic viruses: the acquisition of new host ranges by canine parvovirus and influenza A viruses. Annu. Rev. Microbiol. 59, 553–586 (2005). 52. Boyle, J. P. et al. Just one cross appears capable of dramatically altering the population biology of a eukaryotic pathogen like Toxoplasma gondii. Proc. Natl Acad. Sci. USA 103, 10514–10519 (2006). 53. Lehmann, T., Marcet, P. L., Graham, D. H., Dahl, E. R. & Dubey, J. P. Globalization and the population structure of Toxoplasma gondii. Proc. Natl Acad. Sci. USA 103, 11423–11428 (2006). 54. Holland, G. N. & Lewis, K. G. An update on current practices in the management of ocular toxoplasmosis. Am. J. Ophthalmol. 134, 102–114 (2002). 55. Montoya, J. G. & Liesenfeld, O. Toxoplasmosis. Lancet 363, 1965–1976 (2004). 56. Dubremetz, J. F. Host cell invasion by Toxoplasma gondii. Trends Microbiol. 6, 27–30 (1998). 57. Ferguson, D. J. P. & Dubremetz, J. F. in Toxoplasma gondii: the Model Apicomplexan. Perspectives and Methods (eds. Weiss, L. M. & Kim, K.) 19–48 (Elsevier, London, 2007). 58. Beckers, C. J., Wakefield, T. & Joiner, K. A. The expression of Toxoplasma proteins in Neospora caninum and the identification of a gene encoding a novel rhoptry protein. Mol. Biochem. Parasitol. 89, 209–223 (1997). 59. Leriche, M. A. & Dubremetz, J. F. Characterization of the protein contents of rhoptries and dense granules
88 | january 2008 | volume 6
60.
61.
62.
63.
64. 65.
of Toxoplasma gondii tachyzoites by subcellular fractionation and monoclonal antibodies. Mol. Biochem. Parasitol. 45, 249–259 (1991). Park, Y. K. & Nam, H. W. Early recognized antigen (p34) of Toxoplasma gondii after peroral ingestion of tissue cyst forming strain (Me49 strain) in mice. Korean J. Parasitol. 37, 157–162 (1999). Karasov, A. O., Boothroyd, J. C. & Arrizabalaga, G. Identification and disruption of a rhoptry-localized homologue of sodium hydrogen exchangers in Toxoplasma gondii. Int. J. Parasitol. 35, 285–291 (2005). Poupel, O., Boleti, H., Axisa, S., Couture-Tosi, E. & Tardieux, I. Toxofilin, a novel actin-binding protein from Toxoplasma gondii, sequesters actin monomers and caps actin filaments. Mol. Biol. Cell 11, 355–368 (2000). Schwarz, J. A., Fouts, A. E., Cummings, C. A., Ferguson, D. J. & Boothroyd, J. C. A novel rhoptry protein in Toxoplasma gondii bradyzoites and merozoites. Mol. Biochem. Parasitol. 144, 159–166 (2005). Kissinger, J. C., Gajria, B., Li, L., Paulsen, I. T. & Roos, D. S. ToxoDB: accessing the Toxoplasma gondii genome. Nucleic Acids Res. 31, 234–236 (2003). Ahn, H. J., Kim, S. & Nam, H. W. Molecular cloning of a rhoptry protein (ROP6) secreted from Toxoplasma gondii. Korean J. Parasitol. 44, 251–254 (2006).
Acknowledgements
We thank J. Boyle, P. Bradley, M. Lebrun, M. Reese and J. Saeij for critical reading of the manuscript and access to unpublished information. J.C.B. is supported by grants from the National Institutes of Health (RO1 AI21423 and AI75473) and the Ellison Medical Foundation and J.F.D. is supported by grants from the Centre National de la Recherche Scientifique (UMR5539 and ANR 06-MIME‑024‑01).
DATABASES Entrez Genome Project: http://www.ncbi.nlm.nih.gov/ entrez/query.fcgi?db=geneomeprj Escherichia coli | Plasmodium falciparum | Toxoplasma gondii GenBank: http://www.ncbi.nlm.nih.gov/Genbank/index.html RON1 | RON2 | RON3 | RON4 | ROP1 | ROP2A | ROP4 | ROP5 | ROP6 | ROP7 | ROP8 | ROP9 | ROP10 | ROP11 | ROP12 | ROP13 | ROP14 | ROP15 | ROP16 | ROP17 | ROP18 | TgNHE2 | TgPP2C-hn |TgSUB2 | Toxofilin | Toxopain1
FURTHER INFORMATION John C. Boothroyd’s homepage: http://med.stanford.edu/ profiles/John_Boothroyd/ ToxoDB4.2: http://www.toxodb.org/toxo/home.jsp
SUPPLEMENTARY INFORMATION See online article: S1 (movie) All links are active in the online pdf
www.nature.com/reviews/micro © 2008 Nature Publishing Group
Correspondence
L i n k t o O r i g i n a l a rt i c l e L i n k t o A u t h o r ’ s r e p ly
From ‘perfect mix’ to ‘potion magique’ — regulatory T cells and anti-inflammatory cytokines as adjuvant targets Jagadeesh Bayry*‡§, Darren R. Flower||, David F Tough¶ and Srini V. Kaveri*‡§ The efficacy of vaccines can be greatly improved by the addition of adjuvants, which enhance and modify immune responses. Historically, adjuvants have been discovered empirically by using experimental models. Unfortunately, many adjuvants are associated with side effects that make them unsuitable for use in humans. At present, few adjuvants are available, and of these the adjuvant that is most commonly used, alum, induces immune responses that are suboptimal in several respects. There is an obvious need for the development of improved adjuvants that can induce cell-mediated responses in addition to antibodies. Recent advances in basic immunology have revealed the importance of innate pathogenrecognition receptors in shaping the adaptive immune response. The Review by Bruno Guy, in the July issue of Nature Reviews Microbiology 1, eloquently discusses how Toll-like receptor (TLR) and non-TLR innate-receptor signalling by antigen-presenting cells (APCs) can be used to enhance the immunogenicity of vaccines. Although the activation of APCs by innate receptors represents a promising approach to adjuvant development, we wish to draw attention to an alternative strategy, which could be used in conjunction with TLR agonists to optimize adjuvant activity. Specifically, we propose that selective interference with the activity of regulatory T (TReg) cells and suppressive cytokines could be used to boost responses to poorly immunogenic vaccines. Naturally occurring CD4+CD25+ TReg cells are crucial for the induction and maintenance of self-tolerance and are present in peripheral tissues, such as the skin and gut, under normal, non-inflamed conditions2. TReg cells not only suppress self-antigen-specific immune responses, but also negatively regulate immune responses against foreign antigens, including those driven by TLR-stimulated APCs3–6. TReg cells can inhibit initial T‑cell activation and downregulate ongoing immune responses, which suggests that they act both at the site of initiation of immune responses (secondary
lymphoid organs) and at sites of inflammation; part of the regulatory activity of TReg cells is related to their ability to modify APCs such as dendritic cells (DCs)3–6. Notably, TReg cells themselves have been reported to express several TLRs, including TLR4, TLR5, TLR7 and TLR8, and stimulation by TLR ligands can enhance the suppressive functions of TReg cells7–9. These results suggest that although TLR stimuli can activate APCs, such signals might also enhance the suppressive functions of TReg cells. Although this mechanism is probably important in controlling inflammatory responses in the context of infection, such concomitant suppression might limit the generation of effective immune responses after administration of poorly immunogenic vaccines. The idea that limiting TReg cell activity at the time of vaccination is an effective way of enhancing immune responses has been confirmed experimentally. Evidence for this has typically come from studies in which TReg cells are depleted using anti-CD25 monoclonal antibodies. Animals that are depleted of TReg cells show markedly superior primary and memory CD8+ T‑cell responses to both virus infection and vaccines10,11. Although such experiments provide a proof of principle, it is unlikely that this approach of TReg cell depletion would be of clinical use, as it has been associated with adverse consequences such as localized autoimmune disease12. Conversely, the transient inhibition of TReg cell function might be ideal for enhancing the immune response to vaccines. One strategy would be to block the trafficking of TReg cells to the site of vaccination. TReg cells express the chemokine receptors CCR4 and CCR8 and migrate in response to the chemokines CCL17 and CCL22. There is evidence to suggest that the secretion of these chemokines by DCs can attract TReg cells to the site of inflammation. Further, blocking CCL22 by monoclonal antibodies in vivo has been shown to significantly reduce TReg cell trafficking13. Similarly, the blockade of vascular endothelial
nature reviews | microbiology
growth factor has also been shown to reduce the migration of TReg cells14. Consequently, the inhibition of TReg cell migration at the time of vaccination — by blocking chemokines or growth factors through the use of antibodies or small-molecule antagonists — would alleviate the suppressive effects of TReg cells during the initiation of the immune response and allow for a greater response to the vaccine15. Cytokines are key regulators of the immune system that shape innate and adaptive immune responses. Because cytokines might have an adjuvant-like effect, researchers have attempted to use them to manipulate the immune response to vaccination16. Notably, the maturation-associated signalling of DCs (including that induced by TLR ligands) also leads to the production of anti-inflammatory cytokines such as interleukin-10 (IL-10). Importantly, suppression by IL-10 can have a significant effect on the outcome of the adaptive immune response to vaccination. Fms-like tyrosine-kinase‑3based immunoprophylaxis against infection has been shown to be improved by adjuvant treatment with anti-IL-10 antibody 17. In this study, a single injection of anti-IL-10 antibody was associated with an increase in early IL-12 and interferon-γ production from innate and adaptive immune cells. This experimental model suggests that neutralizing monoclonal antibodies to IL-10 also have a potential application in vaccinology. Together, these findings suggest that the use of combination adjuvants, which aim to both activate innate immune mechanisms and interfere with suppressive mechanisms, will achieve effective APC activation and generate robust responses to vaccines. *Centre de Recherche des Cordeliers, Equipe 16, Immunopathology and Therapeutic Immunointervention, Université Pierre et Marie Curie–Paris 6, UMRS872, 15 rue de l’Ecole de Médicine, Paris F‑75006, France. Université Paris Descartes, UMRS872 Paris F‑75006, France.
‡
INSERM, U872, Paris F‑75006, France.
§
The Jenner Institute, University of Oxford, Compton, Berkshire RG20 7NN, UK.
||
Target Discovery, RI CEDD, GlaxoSmithKline, Gunnels Wood Road, Stevenage SG1 2NY, UK.
¶
Correspondence to J.B. e-mail:
[email protected] doi:10.1038/nrmicro1681-c1 1. 2. 3.
Guy, B. The perfect mix: recent progress in adjuvant research. Nature Rev. Microbiol. 5, 505–517 (2007). Miyara, M. & Sakaguchi, S. Natural regulatory T cells: mechanisms of suppression. Trends Mol. Med. 13, 108–116 (2007). Oldenhove, G. et al. CD4+CD25+ regulatory T cells control T helper cell type 1 responses to foreign antigens induced by mature dendritic cells in vivo. J. Exp. Med. 198, 259–266 (2003).
www.nature.com/reviews/micro © 2008 Nature Publishing Group
Correspondence 4. 5.
6.
7.
8.
Tang, Q. et al. Visualizing regulatory T cell control of autoimmune responses in nonobese diabetic mice. Nature Immunol. 7, 83–92 (2006). Houot, R., Perrot, I., Garcia, E., Durand, I. & Lebecque, S. Human CD4+CD25high regulatory T cells modulate myeloid but not plasmacytoid dendritic cells activation. J. Immunol. 176, 5293–5298 (2006). Bayry, J., Triebel, F., Kaveri, S. V. & Tough, D. F. Human dendritic cells acquire a semimature phenotype and lymph node homing potential through interaction with CD4+CD25+ regulatory T cells. J. Immunol. 178, 4184–4193 (2007). Caramalho, I. et al. Regulatory T cells selectively express toll-like receptors and are activated by lipopolysaccharide. J. Exp. Med. 197, 403–411 (2003). Crellin, N. K. et al. Human CD4+ T cells express TLR5 and its ligand flagellin enhances the suppressive capacity and expression of FOXP3 in CD4+CD25+ T regulatory cells. J. Immunol. 175, 8051–8059 (2005).
Lewkowicz, P., Lewkowicz, N., Sasiak, A. & Tchorzewski, H. Lipopolysaccharide-activated CD4+CD25+ T regulatory cells inhibit neutrophil function and promote their apoptosis and death. J. Immunol. 177, 7155–7163 (2006). 10. Suvas, S., Kumaraguru, U., Pack, C. D., Lee, S. & Rouse, B. T. CD4+CD25+ T cells regulate virus-specific primary and memory CD8+ T cell responses. J. Exp. Med. 198, 889–901 (2003). 11. Toka, F. N., Suvas, S. & Rouse, B. T. CD4+ CD25+ T cells regulate vaccine-generated primary and memory CD8+ T‑cell responses against herpes simplex virus type 1. J. Virol. 78, 13082–13089 (2004). 12. Sutmuller, R. P. et al. Synergism of cytotoxic T lymphocyte-associated antigen 4 blockade and depletion of CD25+ regulatory T cells in antitumor therapy reveals alternative pathways for suppression of autoreactive cytotoxic T lymphocyte responses. J. Exp. Med. 194, 823–832 (2001). 9.
nature reviews | microbiology
13. Curiel, T. J. et al. Specific recruitment of regulatory T cells in ovarian carcinoma fosters immune privilege and predicts reduced survival. Nature Med. 10, 942–949 (2004). 14. Li, B. et al. Vascular endothelial growth factor blockade reduces intratumoral regulatory T cells and enhances the efficacy of a GM‑CSF‑secreting cancer immunotherapy. Clin. Cancer Res. 12, 6808–6816 (2006). 15. Johnson, Z., Schwarz, M., Power, C. A., Wells, T. N. & Proudfoot, A. E. Multi-faceted strategies to combat disease by interference with the chemokine system. Trends Immunol. 26, 268–274 (2005). 16. Salerno-Goncalves, R. & Sztein, M. B. Cell-mediated immunity and the challenges for vaccine development. Trends Microbiol. 14, 536–542 (2006). 17. Das, L., DeVecchio, J. & Heinzel, F. P. Fms-like tyrosine kinase 3‑based immunoprophylaxis against infection is improved by adjuvant treatment with anti‑interleukin‑10 antibody. J. Infect. Dis. 192, 693–702 (2005).
www.nature.com/reviews/micro © 2008 Nature Publishing Group
Correspondence
L i n k t o O r i g i n a l a rt i c l e L i n k to I n i t i a l c o r r e s p o n d e n c e
From ‘perfect mix’ to ‘potion magique’ — regulatory T cells and anti-inflammatory cytokines as adjuvant targets: reply from Guy Bruno Guy The proposal of Bayry and colleagues to replace adjuvant approaches, such as Toll-like receptor (TLR) stimulation, with the inhibition of regulatory T (TReg) cells, or to combine both strategies, is an attractive one. I would nevertheless suggest that this strategy would be best reserved for therapeutic vaccines, in situations where the negative role of TReg cells on protective immunity is well established (for example, in some chronic diseases such as parasitic diseases or tuberculosis1,2). As vaccine developers, we can try to mimic pathogens in their initiation of immune responses, but we must be cautious of interfering too much with natural regulatory mechanisms, and in particular the feedback mechanisms that prevent over-reactions. For example, TReg cells are essential for preventing autoimmunity; the role of some TLR agonists in inducing these disorders has been discussed3, and any potential risk of this occuring should be limited. Bayry and colleagues also propose the use of interleukin-10 (IL-10) suppression, but I feel this should also be reserved for therapeutic settings. Again, nature itself usually does things properly, and inflammatory T helper 1 cells have recently been shown to secrete IL‑10 as a feedback mechanism4. Preventing this
feedback could lead to over-reactive inflammatory responses. Although the goal, target and mechanism were different, the TGN1412 ‘disaster’ showed us that interfering directly with regulatory mechanisms can be dangerous. In conclusion, the approaches proposed by Bayry and colleagues could be powerful but the outcomes are difficult to predict precisely. For this reason, I would be cautious about using them in the routine vaccination of healthy infants (where even rare side-effects are not acceptable) and would reserve them, at least initially, for use in therapeutic vaccination against cancer5, or chronic infections, for which the benefits might outweigh the risks. Research Department, sanofi pasteur, Campus Merieux, Marcy l’Etoile 69280, France. e-mail:
[email protected] doi:10.1038/nrmicro1681-c2 1.
2.
3.
4.
5.
Belkaid, Y., Sun, C. M. & Bouladoux, N. et al. Parasites and immunoregulatory T cells. Curr. Opin. Immunol. 18, 406–412 (2006). Scott-Browne, J. P. et al. Expansion and function of Foxp3-expressing T regulatory cells during tuberculosis. J. Exp. Med. 204, 2159–2169 (2007). Marshak-Rothstein, A. Toll-like receptors in systemic autoimmune disease. Nature Rev. Immunol. 6, 823–835 (2006). Trinchieri, G. Interleukin‑10 production by effector T cells: Th1 cells show self control. J. Exp. Med. 204, 239–243 (2007). Curiel, T. J. Tregs and rethinking cancer immunotherapy. J. Clin. Invest. 117, 1167–1174 (2007).
nature reviews | microbiology
www.nature.com/reviews/micro © 2008 Nature Publishing Group