Chemical Biology Applied to the Study of Bacterial Pathogens

Virstatin.One early example of such a molecule is virstatin (see Table 1). Virstatin targets Vibrio cholerae specifically by inhibiting its ability to produce cholera toxin (CT) (step 1 in Fig. 1) (17). Virstatin was identified using a live-cell-based reporter assay in which the promoter for the genes encoding cholera toxin (ctxAB) directed expression of the tetracycline resistance gene tetA. In this screen, compounds that inhibited expression of the cholera toxin genes resulted in poor growth of the reporter strain in media containing tetracycline. Such cell-based assays are particularly useful to basic biology because the identified compounds may inhibit a protein or pathway not yet known to play a role in virulence, thus allowing new discoveries about biological systems to be made.

In contrast to screens against purified proteins, a cell-based screen requires postscreen target identification. This process can be extremely challenging because the target (i) may have an unspecified role in virulence, (ii) may be expressed only under certain conditions, and/or (iii) may not be a single protein but may rather be a protein complex or (iv) not a protein at all but rather DNA, RNA, lipids, or the redox state of the cell, to name a few of the possibilities.

In the case of virstatin, target identification was accomplished by taking advantage of the fact that the virulence regulatory pathway in V. cholerae is well defined (20–26). Expression profiles of genes expressed upstream of ctxAB transcription in the regulatory cascade were examined by quantitative reverse transcription-PCR (qRT-PCR), and none was affected by virstatin, suggesting that its target is late in the transcription cascade leading to ctxAB expression. Virstatin could still inhibit ctxAB transcription even when toxT, encoding the direct activator of ctxAB transcription, was expressed ectopically, indicating that virstatin inhibits ToxT activity (17).

To examine further whether ToxT is the likely target of virstatin, a library of toxT mutant alleles was generated and screened for resistance to the compound; one allele, toxT^L113P, was identified (17). The protein expressed from this mutant allele resembled wild-type ToxT in that it was found predominantly in the multimeric form; however, in the presence of virstatin, wild-type ToxT is largely in the monomeric state, while ToxT^L113P remains predominantly multimeric (18, 19). This supported a long-held hypothesis, based on the arrangement of the binding sites on ToxT-activated promoters known as toxboxes (27), that ToxT is active as a dimer. Virstatin was then used to probe ToxT activity at other ToxT-dependent promoters, which led to the discovery that some promoters favor the monomeric form whereas others favor dimerized ToxT (28).

To assess its in vivo efficacy, virstatin was tested in an infant mouse model of V. cholerae colonization (17). Mice inoculated with wild-type V. cholerae and given virstatin had a 4-log decrease in colonization relative to mice given the dimethyl sulfoxide (DMSO) control, suggesting that virstatin may have therapeutic potential. To determine whether the decreased colonization was ToxT dependent, an atypical strain of V. cholerae that colonizes the mouse in a ToxT-independent manner was used. This strain colonized mice treated with the DMSO control and mice treated with virstatin equally well, indicating that virstatin reduces colonization of the typical endemic strains of V. cholerae by inhibiting ToxT activity in vivo. Although it remains to be directly tested, that colonization by ToxT-independent strains of V. cholerae is not affected by virstatin suggests that other bacteria lacking ToxT, including the host microbiota, would be unaffected by virstatin treatment, making virstatin an attractive therapeutic lead.

The cell-based screening approach leading to the discovery of virstatin and the subsequent work characterizing its mechanism of action advanced the field of V. cholerae research in several ways. First, this approach provided further strong evidence for the importance of the dimerization state of ToxT. Second, it proved useful for determining whether a V. cholerae isolate colonizes in a ToxT-dependent or -independent manner. Finally, as virstatin was shown to be a compound that significantly reduces colonization of a pathogen by specifically inhibiting its virulence factors in vivo, it provided a proof of principle for the concept of antivirulence drugs.

Toxtazins.The toxtazins (step 1 in Fig. 1) constitute another class of antivirulence compounds that inhibit expression of virulence factors in V. cholerae (15). Toxtazins A and B were identified in a cell-based screen using a toxT-gfp reporter strain to probe the ToxT pathway directly. As indicated in Fig. 2, assays can be designed to target a phenotype or a molecular pathway, and each approach has unique advantages. Targeting a pathway restricts hits to those that affect only that pathway, while targeting a phenotype, CT production, for example, would include hits that target the ToxT pathway as well as inhibitors of CT folding and secretion.

The toxtazins significantly reduce expression of both cholera toxin and the toxin-coregulated pilus (15). Furthermore, toxtazin B reduced the level of V. cholerae colonization of infant mice by 2 logs, suggesting that it may have therapeutic potential in the treatment or prevention of cholera. Further analysis demonstrated that toxtazin A does not affect expression of ToxR or TcpP, the transcriptional activators of toxT, while toxtazin B reduced expression of both tcpP transcript and TcpP protein levels. TcpP is transcriptionally activated by AphA and AphB, neither of which is affected by toxtazin B treatment. In fact, ectopic expression of ToxT in cells treated with either toxtazin A or B restored expression of cholera toxin, confirming that both of these compounds act upstream of ToxT expression. From these results, it was determined that toxtazin A targets toxT transcription whereas toxtazin B targets tcpP transcription (15).

The exact mechanisms of action for these compounds have not been identified, but preliminary evidence showing that cells treated with toxtazin A mount a stress response indicates the presence of a pathway which was not previously known to influence regulation of toxT expression but which affects the virulence potential of V. cholerae.

FPSS.Unlike virstatin and the toxtazins, which target species-specific virulence regulators, fluoro-phenyl-styrene-sulfonamide (FPSS; see Table 1) targets a general microbial regulatory pathway—the sigma B (σB) regulon (step 2 in Fig. 1) (9). Sigma factors are dissociable subunits of RNA polymerase (RNAP) that associate with RNAP in response to certain environmental signals and directly activate or repress subsets of genes, resulting in a rapid change in global transcription that is appropriate for the given environment. Different sigma factors respond to different sets of environmental signals, allowing bacteria to quickly respond to specific environments.

Different sigma factors can respond to the same signal, and different sigma factors can regulate the same genes in response to distinct signals, creating a complex web of regulation (29). In addition, some sigma factors are essential in some or all growth conditions, making them difficult to study using classical genetic approaches. Small-molecule inhibitors of specific sigma factors enable the study of individual sigma factors in isolation, so the contributions of individual sigma factors and of different environmental signals can be explored. This class of inhibitor may also have therapeutic potential because sigma factors regulate virulence gene expression in many pathogens (30–32), as reviewed in reference 33.

In Listeria monocytogenes, σB responds to environmental stress (i.e., acidic conditions, ethanol, or high salt concentrations), energy stress, and growth at low temperatures (34). In response to these signals, σB activates or represses genes involved in central metabolism and activates PrfA, the master regulator of virulence (35). FPSS was identified as an inhibitor of σB by screening a diverse library of synthetic compounds using a cell-based reporter assay (9).

FPSS was discovered using a screening strain containing a σB-dependent promoter, opuCA, fused to the gene for glucuronidase, gus, such that glucuronidase activity could be used as a readout for σB activity. Inhibition of σB activity by FPSS was confirmed by qRT-PCR. Transcription of σB-dependent promoters was at the level of a σB mutant after FPSS treatment, indicating that FPSS completely inhibits σB activity at these promoters. Microarray analysis showed that L. monocytogenes cells treated with FPSS phenocopy an L. monocytogenes σB mutant, affecting 91% of previously defined σB-regulated genes (9). FPSS also affected 83 other genes, which, excluding side effects, could potentially expand the σB regulon. Gene-set enrichment analysis (GSEA) determined that genes specifically regulated by σH or by σL are not significantly enriched among genes differentially transcribed in FPSS-treated cells, indicating that FPSS inhibits σB specifically.

Bacillus subtilis was used to determine the mechanism by which FPSS inhibits σB activity, because its σB activity is also inhibited by FPSS (9) and the components of its well-characterized σB regulon are highly conserved with the components of the insufficiently studied σB regulon of L. monocytogenes. In B. subtilis, σB activity is regulated by three distinct branches (shown in Fig. 3 and reviewed in reference 36) that convey conditions of environmental stress (i.e., acidic conditions, ethanol, or high salt concentrations), energy stress (i.e., limitation of glucose, ATP, GTP, or phosphate), and growth at low temperatures.

• Open in new tab
• Download powerpoint

FIG 3

FPSS targets sigma B. Multiple stresses activate σ^B activity in B. subtilis. Energy stress is relayed via RsbPQ, environmental stress is relayed via the “stressosome” and RsbU, and low-temperature stress is relayed via an as-yet-unknown mechanism. Sensing of stress results in dephosphorylation of the RsbV anti-anti-sigma factor, allowing it to bind the RsbW anti-sigma factor, which in turn releases σ^B, to interact with RNAP and activate the σ^B regulon. Phosphorylated RsbV interacts poorly with RsbW, which is then free to bind σ^B, leading to low-level expression of the regulon. FPSS inhibits σ^B activity by driving equilibrium toward the state in which σ^B is bound to RsbW. SNP, single nucleotide polymorphism.

In B. subtilis, σB is kept in the off state by an anti-sigma factor, RsbW. The ability of RsbW to sequester σB is controlled by the phosphorylation state of another protein, RsbV; phosphorylated RsbV cannot bind RsbW, leaving RsbW free to sequester σB, thereby keeping the regulon from being expressed. In contrast, unphosphorylated RsbV binds RsbW, liberating σB and leading to activation of the σB regulon (37). RsbV phosphorylation is controlled by energy and environmental stresses (see Fig. 3) (38), and FPSS inhibited σB activation by several such stresses, indicating that it works on a factor common to these pathways (9). The possibility that FPSS could bind or interact with σB was ruled out by in vitro and in vivo experiments (10). Taken together, these results support a model whereby FPSS targets the partner-switching mechanism between RNA polymerase, σB, and its anti-sigma factor, RsbW.

FPSS provides a unique tool to study σB activity in multiple organisms. Microarray analysis of FPSS-treated cultures (9) suggests that the σB regulon is broader than currently understood, showing another value of using this compound to study σB. While animal studies with FPSS have not been reported, σB is an attractive therapeutic target because it is conserved and important in the activation of virulence genes in several pathogenic bacteria (39, 40). One potential drawback to the therapeutic potential of this molecule, however, is that FPSS may also inhibit σB in commensal strains of the microbiota, which could lead to dysbiosis, although this has not been experimentally assessed.

TSS29.Secretion mechanisms have been targeted in small-molecule screens for antivirulence drugs. In particular, the type 3 secretion system (T3SS) injectosome has been attractive because of its broad conservation in Gram-negative pathogens (step 3 in Fig. 1). Inhibitors of the Salmonella enterica serovar Typhimurium T3SS were identified using a reporter strain that secretes phospholipase in a T3SS-dependent manner (16). This phenotypic approach to cell-based screening can result in hits that affect any of several steps in production of a functional phospholipase, including transcription, translation, protein folding, protein-protein interaction, and translocation. Eighty-nine compounds remained after the primary and secondary screens. Tertiary screens are sometimes performed to select compounds with particular characteristics (Fig. 2). In that study, the authors performed two tertiary screens to eliminate compounds that affect bacterial growth and those that are not specific (i.e., that inhibit bacterial translation, sec-dependent secretion, or disulfide bond isomerization). Seven compounds passed the tertiary screens, and one of these, a 2-imino-5-arylidene thiazolidinone termed TTS29, was investigated further (16). Cultures grown with TTS29 displayed an overall reduction in the levels of type 3 secretion system needle complexes. Levels of needle complex protein constituents were not reduced in whole-cell lysates, indicating that the proteins were being produced but that their assembly into the needle complex was inhibited by TTS29 (16).

Because components of the T3SS are conserved in other bacteria, TTS29 has the potential to work against other T3S-encoding bacteria. Yersinia species express two types of T3SS: (i) the plasmid-encoded Ysc system in Yersinia pestis, Yersinia enterocolitica, and Yersinia pseudotuberculosis, which secretes Yop (Yersinia outer protein) into the cytosol of target cells, and (ii) the chromosomally encoded Ysa system in Y. enterocolitica, which secrets Ysp (Yersinia secreted protein) (41, 42). TSS29 inhibited secretion of both Yop and Ysp into Y. enterocolitica culture supernatants, indicating its potential utility as a broad inhibitor of T3SS (16). In contrast, TSS29 did not alter flagellar motility or decrease the levels of flagellar components in either S. Typhimurium or Pseudomonas aeruginosa, bacteria that depend on a flagellum-specific T3SS for motility (16). This suggests that TSS29 targets a component of the T3SS that is not conserved with the evolutionarily related flagellum-specific T3SS (42).

Because one component of the T3SS, secretin, is conserved in the type 2 secretion system (T2SS) that delivers enzymes and other proteins across the Gram-negative envelope (43), TSS29 was tested for its ability to inhibit such systems. Secretion of elastase (44) by the T2SS in P. aeruginosa was inhibited by TSS29, as was twitching motility, which is controlled by a pilus whose assembly includes components similar to those of the T2SS (16). These results demonstrate that TSS29 is a broad inhibitor of secretion affecting different secretion systems in multiple bacterial pathogens.

The assessment of the in vivo effectiveness of TSS29 was supported by demonstrating its ability to reduce killing of bone marrow-derived macrophages (BMMs) in a tissue culture model of infection (16). It also inhibited Pseudomonas syringae pv. tomato DC3000 from inducing a hypersensitivity response in tobacco plants (16). Thus, TSS29 has broad therapeutic potential.

Pilicides.Pili are recognized as important virulence factors because of their role in colonization of many pathogens in their respective hosts (step 4 in Fig. 1). While some microbes secrete pilus adherence structures through type 2-like secretion systems that can be inhibited by TSS29 as described above, others assemble pili using pathways that rely on a periplasmic chaperone (45–47) and an outer membrane usher (48, 49) (see references 50 and 51 for reviews). The chaperone mediates the folding, stabilization, and transport of pilus subunits, while the usher aids in incorporating subunits into the growing pilus (51).

Because the chaperone-usher pathway of pilus assembly is conserved in a wide range of pathogens (51), inhibitors of these systems would theoretically be effective against a broad range of bacterial species, making them attractive targets for therapeutic development. Substituted bicyclic 2-pyridones, called pilicides (pharmacophore shown in Table 1), are a well-studied group of synthetic small-molecule inhibitors that prevent formation of pili in uropathogenic Escherichia coli (UPEC) (51).

UPEC bacteria produce different pathogenicity-associated pili, the most representative of which are P pili and type I pili; these adhesion structures and others like them are generally termed chaperone-usher pili for the key proteins associated with their assembly (Fig. 4) (51). P pili are made up of subunits called PapA (the major subunit), PapE, PapF, PapG, PapH, and PapK. These are assembled by the PapD chaperone and the usher PapC. Type 1 pili are made up of FimA (the major subunit), FimG, and FimH and are assembled by the chaperone FimC and the usher FimD (51).

• Open in new tab
• Download powerpoint

FIG 4

Pilicides affect both P pili and the type 1 pili, which have similar structures. Pili consist of several repeating subunits arranged in a helical structure. Subunits are translocated from the cytoplasm to the periplasm, where a chaperone (PapD or FimC) folds the protein, stabilizes it, and transfers it to the usher protein (PapC or FimD), which secretes the protein and incorporates it into the pilus structure. Pilicides inhibit formation of pili by preventing the chaperone from passing the subunit to the usher protein.

In contrast to inhibitors discovered by screening compound libraries, pilicides were developed using a rational design approach (12). Pilus biogenesis requires the chaperone protein to bind its natural ligands (52), so the solved crystal structure of the PapD-PapG complex was used to chemically design molecular mimetics that would bind within the active site of periplasmic chaperones PapD and FimC (12). A library of pilicides was synthesized and demonstrated to be effective at inhibiting pilus formation (13).

PiIlicides do not affect cell growth or viability but inhibit mannose-sensitive hemagglutination (MSHA) and hemagglutination (HA), traits of type 1 pili and P pili, respectively (13). Pilicides also reduce biofilm formation and bacterial attachment to host cells, both of which require type 1 pili (13). Pilicide 2c in particular reduced the ability of E. coli strains to adhere to cultured bladder cells by 90% (13).

Cocrystallization of pilicide 2c with the PapD chaperone (13) demonstrated close contact with a hydrophobic patch that runs across the back of the F1-C1-D1 beta sheet on PapD, a region highly conserved in all pilus periplasmic chaperones, which may mediate interactions between the chaperone and the N-terminal domain of the usher protein (13, 53). Surface plasmon resonance showed that pilicide 2c also prevents FimC-FimH (chaperone-adhesin) from binding to FimD (usher) (13), suggesting that pilicide 2c inhibits the biogenesis of pili by preventing the chaperone from passing the subunit to the usher.

Studies of pilicides have uncovered knowledge about molecular mechanisms involved in the biogenesis of pili, and their discovery and characterization demonstrate the power of crystallography to guide chemists to create a small molecule with a predictable activity. This technique, called structure-based drug design (SBDD), can be extended by cocrystallizing the compound and target to provide information useful for improving compound activity.

LED209.Virulence gene expression is often induced by signals from the host environment, making sensory transduction mechanisms potential therapeutic targets (step 5 in Fig. 1). For example, E. coli, S. Typhimurium, and Francisella tularensis express a sensor histidine kinase called QseC that detects host-derived adrenergic signals (epinephrine and norepinephrine) as well as quorum-sensing autoinducer-3 (AI-3) (54, 55). In response to either of these signals, QseC autophosphorylates and then transfers the phosphate to transcription factor QseB, leading to transcription of key virulence genes, including the LEE1 operon in enterohemorrhagic E. coli (EHEC) (56).

A diverse chemical library of 150,000 compounds was screened for those that block expression of a LEE1-lacZ reporter strain of EHEC (11). Compounds were screened in an assay using spent medium (which contains AI-3) to activate QseC and induce reporter gene expression. The most potent inhibitor identified was improved by structure-activity relationship (SAR) studies and named LED209 (see Table 1 for structure) (11).

LED209 did not inhibit bacterial growth but rather selectively inhibited virulence gene expression by inhibiting QseC autophosphorylation (11). While LED209 was ineffective at reducing colonization loads in an infant rabbit model of EHEC infection (perhaps due to rapid absorption from the gastrointestinal tract), it significantly reduced the mouse pathogenicity of both S. Typhimurium and F. tularensis, which express QseC homologs of 87% and 57% similarity to EHEC QseC, respectively (11). QseC is important for motility in S. Typhimurium (41) and for systemic infection in F. tularensis (57).

Similar to the ToxT inhibitors virstatin and the toxtazins, LED209 inhibits virulence by targeting a specific virulence regulator without affecting growth. Its target, QseC, is broader, however, being conserved in over 25 pathogenic bacteria but absent in mammals (11), giving this molecule a bacterium-specific broad spectrum of activity and making it an attractive lead compound.

CCG-2979.In addition to targeting sensory transduction pathways, screens can target the ability of pathogens to manipulate the host. Such a strategy identified CCG-2979 (see Table 1 for structure), which reduces production of streptokinase (SK) by Streptococcus pyogenes (see step 6 in Fig. 1). CCG-2979 targets the promoter of SK (7), which is secreted by S. pyogenes and other group A streptococci (GAS) and activates the host zymogen plasminogen to form plasmin, the central protease of the fibrinolytic system critical for regulating blood clots (58, 59). CCG-2979 was identified using a cell-based assay in which ska, the SK gene promoter, controlled expression of the kanamycin resistance gene; positive hits were those that decreased the growth of the reporter strain but not growth of a constitutive kanamycin-resistant strain (7).

CCG-2979 reduces SK activity in a dose-dependent manner without inhibiting bacterial growth. Furthermore, GAS treated with 5 or 50 μM CCG-2979 were more susceptible to phagocytosis by host cells. A series of structural analogs of CCG-2979 was generated, and structure-activity relationship (SAR) studies were performed (8, 60). The effect of one chemical analog, CCG-102487, on global gene expression was examined by microarray analysis, which demonstrated that expression of 29% of GAS genes in addition to that of ska was altered, with the vast majority of them being reduced in expression. Included among these were other virulence-associated genes coding for adhesins and toxins, along with genes encoding some metabolic functions (7).

Transgenic mice expressing the human plasminogen gene were used to determine the therapeutic potential of CCG-2979. Using this well-established model (59), mice were subcutaneously injected with GAS, given a day to establish an infection, and then intraperitoneally treated with compound daily for 5 days. Mice treated with CCG-2979 showed a statistically significant improvement in survival, while those treated with the CCG-102487 analog were not protected from GAS-induced mortality (7).

Another analog, CCG-203592, inhibited SK at a level 35-fold greater than that seen with CCG-2979 (8). Subsequent work led to the observation that this and other analogs are also effective against another Gram-positive pathogen, Staphylococcus aureus. S. aureus is a major public health problem due in part to its ability to form biofilms on implantable devices (61). Biofilm formation of three different biofilm-producing S. aureus strains was reduced in the presence of 50 μM CCG-203592, both in laboratory microtiter plates and on silicone, the most common surface used for implantable medical devices (8). Thus, this class of compounds may have therapeutic applications against different traits expressed by two classes of important human pathogens, GAS and S. aureus, making them interesting therapeutic leads.

Pimozide.In contrast to screens using large, diverse chemical libraries whose functions are completely unknown, screens using compounds with known biological activities result in more readily identified mechanisms of action in pathogenic processes. Pimozide (see Table 1 for structure) was identified as an inhibitor of Listeria monocytogenes infection by screening a library of FDA-approved drugs (14).

L. monocytogenes infects macrophages (step 7 in Fig. 1) (62) and escapes the phagocytic vacuole by producing a hemolysin, listeriolysin O (LLO) (step 8 in Fig. 1). It replicates in the cytosol before spreading to neighboring cells by polymerizing host cell actin to propel itself into adjacent cells (step 9 in Fig. 1) (13, 63). While genetic and cellular biological studies have uncovered much about L. monocytogenes pathogenicity, the mechanisms regulating infection remain incompletely defined.

A novel chemical genetics approach was taken in which both bacterial cells and host cells were exposed to a library of compounds (14). Primary bone marrow-derived macrophages (BMMs) were infected with L. monocytogenes constitutively expressing green fluorescent protein (GFP) to enable visualization of internalized bacteria. A total of 480 compounds with diverse, known biological activities were screened; 21 of these altered L. monocytogenes infection in one of three ways (14). The majority of inhibitory compounds inhibited BMM infection, as observed by a decrease in GFP fluorescence within BMMs (i.e., internalized bacteria) relative to the DMSO control results. Others enhanced bacterial uptake or intracellular replication, causing an increase in GFP fluorescence per BMM cell. Lastly, some compounds inhibited cell-to-cell spread, observed as an increase in GFP fluorescence per BMM, but only in very few cells.

Compounds that answered the screen fell into four categories based on their previously known activities, and each category engenders testable hypotheses about the conditions required for L. monocytogenes infection. The first category of inhibitors includes compounds that disrupt actin, which is required for phagocytosis by BMMs and for cell-to-cell spread of L. monocytogenes (64). A second group targets kinases and phosphatases, important for actin rearrangement. These may provide new insights into the role of specific kinases and phosphatases in L. monocytogenes infection but were not further analyzed. The members of a third group of compounds were categorized together on the basis of lacking shared activities with other compounds from the screen; further studies on these may also yield new insights into L. monocytogenes infection. The fourth category, composed of nine compounds (43% of the hits), includes molecules that affect calcium pathways. Calcium is not only important in many cellular signaling pathways such as phagocytosis (65), it is also released in response to protein kinase C (PKC) activation, which occurs during L. monocytogenes infection and modulates bacterial uptake and escape from the vacuole (66). Pimozide falls into this fourth category and was chosen for further study; it should be emphasized, however, that whether the action of pimozide on calcium pathways is also what disrupts any given aspect of L. monocytogenes infection has yet to be resolved. This raises an issue that confounds design of any small-molecule inhibitor, whether used as a chemical probe for biological effects or as an actual therapeutic: inhibitors effective in specific screens may have other off-target effects.

Pimozide inhibited intracellular infection of BMMs by L. monocytogenes by an order of magnitude after a 10-h treatment (14). While the exact mechanism for this is unclear, pimozide was found to inhibit infection at three distinct steps. The most potent effect of pimozide was inhibition of macrophage phagocytosis, which was not limited to phagocytosis of L. monocytogenes but was seen with three other bacteria as well: B. subtilis, S. Typhimurium, and E. coli K-12. Pimozide inhibited internalization of bacteria by macrophages by 99% in a calcium-independent manner (14). Pimozide also reduced vacuolar escape of L. monocytogenes by 26%, and this was not due to inhibition of LLO. Finally, pimozide treatment decreased cell-to-cell spread by approximately 50% (14).

This screen generated many testable hypotheses regarding the cellular requirements of L. monocytogenes infection. Additionally, pimozide can be used to probe the molecular mechanisms underlying BMM phagocytosis of bacteria in general and of L. monocytogenes specifically. Pimozide blocks postsynaptic dopamine receptors and is an FDA-approved antipsychotic molecule used to treat severe Tourette's syndrome and schizophrenia (67). The molecular mechanism by which pimozide inhibits macrophage function leading to alterations in Listeria infection is not clear, but, given the class of compound into which it fell in this screen, its mechanism may involve calcium signaling. In addition, this compound was shown to affect this particular host-pathogen system in multiple and yet synergistic ways and highlights the connection between what otherwise appear to be different pathogenic mechanisms. Molecules such as pimozide have increased therapeutic appeal because of the lowered probability of bacteria overcoming their multiple effects by mutation. Additionally, an added benefit of an FDA-approved compound such as pimozide is that there is a wealth of information regarding safety, pharmacokinetics, and pharmacodynamics.

Type 4 secretion inhibitors.Compounds with known biological activities have also been used to probe the type 4 secretion system (T4SS; also called the Icm/Dot type IVB system) (step 10 in Fig. 1) in Legionella pneumophila, which infects and replicates in lung alveolar macrophages and causes Legionnaires' disease (68, 69). L. pneumophila avoids phagosome-lysosome fusion by using its T4SS to secrete effectors that interfere with vesicular trafficking, the host innate immune response, phosphoinositide metabolism, and ubiquitination (reviewed in reference 70). Given its importance in intracellular survival and replication (71), the T4SS is an attractive target for drug development. The T4SS can be activated by contact with the host cell (72), but the signals that trigger secretion of effectors are not well understood.

Compounds with known biological targets were screened to probe the mechanisms of type 4 secretion in L. pneumophila (6). A T4SS-secreted effector protein (LepA) was fused to the TEM-1 β-lactamase (BlaM), which cleaves a green substrate, CCF4, to produce a blue product. Host cells were incubated with compounds for 24 h and infected with the L. pneumophila lepA-blaM reporter strain, and host-bacterial cell contact was initiated by a low-speed centrifugation step. CCF4 was added 1 h later, and fluorescence was measured 2 h later to quantify the ratio of cleaved CCF4 to uncleaved CCF4 (6).

Of 2,640 compounds screened, 22 inhibited translocation with efficiencies ranging from 63% to 100% and were categorized into groups based on their previously known targets: ionophore or protonophores, calmodulin, cytoskeleton dynamics, NF-κB, serine proteases, kinases or phosphatases, and others (6). The majority of the inhibitors are molecules previously known to affect host cytoskeleton dynamics, including those affecting tubulin (step 11 in Fig. 1), actin (step 12 in Fig. 1), and phosphoinositol 3-kinase (PI3K). Ionophores and protonophores discharge membrane electric potential (Δψ) and collapse the proton gradient (ΔpH), both components of the proton motive force (PMF) (step 13 in Fig. 1) (73, 74). All three identified ionophores and protonophores inhibited Icm/Dot-dependent lysis of red blood cells by L. pneumophila (6), indicating that translocation of effectors is at least partially dependent on the PMF. Furthermore, one ionophore, carbonyl cyanide m-chlorophenylhydrazone (CCCP) (shown in Table 1), inhibited LepA translocation in L. pneumophila in a dose-dependent and reversible manner (6). The PMF was not previously known to play a role in T4SS-mediated translocation.

Nineteen of the 22 identified inhibitors significantly affected the ability of macrophages to phagocytose other bacteria (step 7 in Fig. 1) (6). To test whether phagocytosis is required for translocation of L. pneumophila Icm/Dot effector proteins, cytoskeleton inhibitors were used to inhibit coiling phagocytosis, the type of phagocytosis used by L. pneumophila to enter macrophages (75). Host-bacterium contact was instead initiated by opsonization with L. pneumophila-specific antibodies (76). Opsonization restored effector translocation in the absence of phagocytosis (6), indicating that host-cell binding, and not phagocytosis, is required for translocation by the Icm/Dot system. These results support a model where the T4SS in L. pneumophila is in a “locked and loaded” state, ready to inject effector proteins upon contact with a host cell (step 10 in Fig. 1).

Another important finding from this study came by further analyzing the RWJ-60475 translocation inhibitor (shown in Table 1), previously known to inhibit a receptor tyrosine phosphate phosphatase (step 14 in Fig. 1) (77). Work with this compound revealed that CD45 and CD148 are required to phagocytose L. pneumophila (6).

In this Legionella work, screening just 2,500 small molecules eventually generated many new testable hypotheses from the hits that arose. Additionally, new information was uncovered relating phagocytosis and effector translocation to the PMF, host cell contact, and CD45 or CD148, providing a great example of how chemical genetics deepens our understanding of pathogenicity mechanisms.