Plasmid Negative Regulation of CPAF Expression Is Pgp4 Independent and Restricted to Invasive Chlamydia trachomatis Biovars

ABSTRACT Chlamydia trachomatis is an obligate intracellular bacterial pathogen that causes blinding trachoma and sexually transmitted disease. C. trachomatis isolates are classified into 2 biovars—lymphogranuloma venereum (LGV) and trachoma—which are distinguished biologically by their natural host cell infection tropism. LGV biovars infect macrophages and are invasive, whereas trachoma biovars infect oculo-urogenital epithelial cells and are noninvasive. The C. trachomatis plasmid is an important virulence factor in the pathogenesis of these infections. Central to its pathogenic role is the transcriptional regulatory function of the plasmid protein Pgp4, which regulates the expression of plasmid and chromosomal virulence genes. As many gene regulatory functions are post-transcriptional, we employed a comparative proteomic study of cells infected with plasmid-cured C. trachomatis serovars A and D (trachoma biovar), a L2 serovar (LGV biovar), and the L2 serovar transformed with a plasmid containing a nonsense mutation in pgp4 to more completely elucidate the effects of the plasmid on chlamydial infection biology. Our results show that the Pgp4-dependent elevations in the levels of Pgp3 and a conserved core set of chromosomally encoded proteins are remarkably similar for serovars within both C. trachomatis biovars. Conversely, we found a plasmid-dependent, Pgp4-independent, negative regulation in the expression of the chlamydial protease-like activity factor (CPAF) for the L2 serovar but not the A and D serovars. The molecular mechanism of plasmid-dependent negative regulation of CPAF expression in the LGV serovar is not understood but is likely important to understanding its macrophage infection tropism and invasive infection nature.

IMPORTANCE The Chlamydia trachomatis plasmid is an important virulence factor in the pathogenesis of chlamydial infection. It is known that plasmid protein 4 (Pgp4) functions in the transcriptional regulation of the plasmid virulence protein 3 (Pgp3) and multiple chromosomal loci of unknown function. Since many gene regulatory functions can be post-transcriptional, we undertook a comparative proteomic analysis to better understand the plasmid's role in chlamydial and host protein expression. We report that Pgp4 is a potent and specific master positive regulator of a common core of plasmid and chromosomal virulence genes shared by multiple C. trachomatis serovars. Notably, we show that the plasmid is a negative regulator of the expression of the chlamydial virulence factor CPAF. The plasmid regulation of CPAF is independent of Pgp4 and restricted to a C. trachomatis macrophage-tropic strain. These findings are important because they define a previously unknown role for the plasmid in the pathophysiology of invasive chlamydial infection. KEYWORDS Chlamydia trachomatis, plasmid, proteomics, virulence factors C hlamydia trachomatis is an obligate intracellular bacterial pathogen of humans that causes blinding trachoma and sexually transmitted disease, both of which afflict hundreds of millions of people globally (1,2). C. trachomatis isolates are classified into 2 biovars-lymphogranuloma venereum (LGV) and trachoma-which are distinguished biologically by their host cell infection tropism. Macrophages and monocytes are the natural host for LGV biovars, whereas trachoma biovars infect oculo-urogenital epithelial cells (3). These distinct host cell infection tropisms produce invasive and noninvasive infections, respectively.
The chlamydial plasmid is an important virulence factor in both murine and nonhuman primate models of infection (4)(5)(6)(7)(8)(9). Plasmid-deficient organisms produce highly attenuated infections characterized by decreased organism loads of shorter duration that spontaneously resolve with reduced or no post-infection pathology. The C. trachomatis plasmid is highly conserved and carries eight open reading frames (ORFs) (4,10). All eight ORFs, designated Pgp1 to -8, are expressed in infected cells (11). Putative functions for several ORFs have been assigned based on homology to known proteins in the public databases. These are Pgp1, which is a DnaB-like helicase, Pgp7 and -8, which are integrase/recombinase homologues, and Pgp5, which is a homologue of plasmid-partitioning protein ParA. Pgp2 and -6 are chlamydia-specific proteins, showing little or no homology to proteins in the public databases. Pgp3 is highly conserved among chlamydiae and exits naturally as a homotrimer (12) that is secreted into the host cytosol of infected cells (13). The trimeric form is highly immunogenic (13) and is thought to function by neutralizing host antimicrobial peptides (14). Immunization with DNA expressing Pgp3 (15) or recombinant Pgp3 trimer (16) provides partial protection against chlamydial infection in mice. Lastly, a pgp3 null mutant is attenuated for mice, demonstrating its importance to chlamydial pathogenicity (17).
Transcriptional studies, comparing plasmid-positive and -negative C. trachomatis L2 isolates, have shown that the plasmid regulates the expression of several chromosomal loci; the most highly regulated genes include those coding for glycogen synthase, the Pmp-like proteins CT049 to -051, and putative type 3 secretion effectors CT142 to -144 (18). Importantly, the change in chromosomal gene expression in plasmid-free isolates could be reproduced by deletion of pgp4, indicating that Pgp4 is largely responsible for the plasmid-mediated regulation of chromosomal gene expression (18). In addition, these transcriptional studies indicated that Pgp4 also regulates the expression of Pgp3 (18).
To increase our understanding of Pgp4 gene regulation, we have employed a comparative proteomic approach using C. trachomatis plasmid-positive strains, plasmid-negative strains, and a strain with an inactivating single nucleotide polymorphism (SNP) mutation in pgp4. Our proteomic findings clearly show Pgp4 alone is responsible for the tight regulatory control of Pgp3 and a select core set of chromosomal genes. Moreover, the results implicate a hereto unrecognized role for the plasmid in the negative regulation of expression of CPAF (chlamydial protease-like activity factor), a documented chlamydial virulence factor. Notably, the plasmid-dependent negative regulation of CPAF was limited to an LGV biovar, implicating this function in macrophage infection tropism and invasive infection.

RESULTS
Comparative proteomic analysis of chlamydial proteins from L2 plasmidpositive and -negative infected HeLa cells. HeLa cells were infected with plasmidpositive L2 (here referred to as L2 Pϩ) and mutant plasmid-negative L2R (here referred to as L2 PϪ) at a multiplicity of infection (MOI) of 3. Infected cells were harvested at 30 h post-infection (hpi), and lysates were prepared for mass spectrometry (MS). Lysates were labeled equivalently with tandem mass tags (TMTs) and analyzed by nanoscale liquid chromatography-tandem mass spectrometry (nano-LC-MS/MS). A total of 752 chlamydial proteins were identified (see Table S1 in the supplemental material). Em-ploying a Ն2-fold change in mean log 2 -normalized intensity across four replicates and an adjusted P value of Յ0.05 from differential expression testing revealed altered expression of a limited number of proteins in L2 Pϩ compared to L2 PϪ (Fig. 1A). Protein identification was matched to the C. trachomatis serovar D UW-3 gene locus identification for ease of identification (19,20). The corresponding gene locus homologues in the other C. trachomatis serovars discussed in this study are summarized in Table 1. As expected, all plasmid proteins, except for Pgp7, which carries a natural inactivating mutation on the L2 plasmid (21,22), were detected and downregulated in L2 PϪ (Fig. 1A).
We used genetic complementation to validate the specificity of our proteomic findings between L2 Pϩ-and PϪ-infected cells. This was done by transforming L2 PϪ organisms with a recombinant L2 plasmid (18) (designated L2 PϪ complement) and doing a proteomic comparison of L2 Pϩ and L2 PϪ complemented organisms. The results shown in Fig. 1B and Table S1 clearly demonstrate that the decreased expression of L2 PϪ proteins described above is reversed following plasmid transformation. As expected, ␤-lactamase was detected as upregulated in the L2 PϪ-complemented strain as it is the selectable marker on the recombinant L2 plasmid. We also found the CT649 gene, encoding formyltetrahydrofolate synthetase, to be upregulated following complementation. The significance of this is unknown. The increased expression of CPAF was reversed by complementation.
To determine if the plasmid-regulated proteins identified in L2 were similarly regulated in other C. trachomatis serovars, we performed a proteomic comparison of serovars D Pϩ and PϪ and A Pϩ and PϪ (Table S1). A global view of the samples using chlamydial protein expression shows the strong similarity of the replicates for each condition, as indicated by the hierarchical clustering on the heat map ( Fig. 2A). As shown in the C. trachomatis L2 Volcano plots (Fig. 1A), the proteins CT798, CT049 to -051, and CT142 to -144 are also plasmid regulated in C. trachomatis serovars A and D ( Fig. 2B and C). We did not see expression changes in CT702 or CPAF. This result suggests that these proteins are more specifically plasmid regulated in L2. CT412 (PmpA) showed a statistically significant change (1.28-fold, P ϭ 0.0042) in serovar A but not D. Interestingly, we did see small but statistically significant alterations in the expression in the majority of the nine Pmps, and some Pmp changes were serovar specific ( Table 2). The interpretation of this data is complicated by the fact that Pmp expression is phase variable (30); thus, it is difficult to distinguish if the small changes seen result from plasmid regulation, phase variation, or both. Taken together, we conclude that CT798, CT049 to -051, and CT142 to -144 are the "core set" of plasmidregulated proteins among C. trachomatis serovars.
In previous studies, we used plasmid gene deletion mutants (pgp1 to -8) to study the gene specificity of plasmid transcriptional regulation (18). We found that pgp4 was the primary plasmid gene that regulated transcription of pgp3 and the plasmid-regulated chromosomal loci. However, the use of the plasmid gene deletion mutants for transcriptional and proteomics studies is potentially confounded by the fact that the deletions can remove noncoding small RNAs (sRNAs) known to be encoded on the plasmid (10, 31, 32). As chromosomal sRNAs are known to regulate expression of chlamydial proteins (33, 34) we reasoned it was possible that the plasmid sRNA might similarly function in the regulation of protein expression. Moreover, it is likely that deletion mutations in plasmid genes that are cotranscribed can affect the expression of the downstream gene. Thus, the observed changes in protein profiles between Pϩ, PϪ, and pgp4 null strains may result from alterations in multiple plasmid regulators. To minimize the chance that sRNA expression would be altered or eliminated in the pgp4 mutant, we used a pgp4 SNP nonsense mutant, specifically A37T introducing a stop codon early in pgp4, for comparative proteomic studies. The pgp4 SNP proteome results showed that in addition to Pgp4, Pgp3, CT798, CT702, CT049 to -051, and CT142 to -144 were downregulated (Fig. 3). This result supports the conclusion that the aforementioned plasmid-regulated proteins are directly and tightly regulated by Pgp4. Of note, the upregulation of CT858 (CPAF) was reversed in the L2 complement strain and in the pgp4 SNP mutant, implicating that its expression could be regulated by plasmid components other than Pgp4 ( Fig. 1B and 3).
Western blot. To validate our proteomics findings, we performed Western blotting using infected HeLa cell lysates from L2 Pϩ, PϪ, PϪ complement, and pgp4 SNP strains (Fig. 4A). Monoclonal antibodies (MAbs) specific to Pgp3, CT142, CT143, CT144, and OmpA were used to detect plasmid-regulated proteins. The Western blot results validate and support our proteomic findings. Moreover, the results show that Pgp4 is a master regulator of these genes and their regulation is remarkably tight. Interestingly, in contrast to the downregulation seen for Pgp4-regulated proteins, the proteome data showed an increased level of CPAF expression in the absence of plasmid. Western blots were performed using monoclonal antibodies specific to CPAF and chlamydial 60-kDa heat shock protein (HSP60) to validate the proteomic results. Anti-HSP60 was used as a protein loading standard among lysates. In agreement with the proteomics findings, Western blotting showed an increased level of CPAF expression in the absence of plasmid (Fig. 4B). Importantly, unlike the Pgp4-dependent decreased expression of Ppg3 and chromosomally encoded proteins, the regulation of CPAF expression was plasmid dependent but Pgp4 independent. The L2 PϪ-complemented strain showed a  Table S1.
Plasmid-Regulated Chlamydia Virulence Factors ® greater decrease in CPAF expression than the L2 Pϩ strain, consistent with the fact that plasmid-complemented strains have increased plasmid copy numbers (17). The regulation of CPAF expression was not as tight as the Pgp4-regulated loci. Collectively these data support the conclusion that CPAF expression is regulated by the plasmid but by a mechanism different from the Pgp4-regulated loci.  Immunostaining. Indirect fluorescence antibody (IFA) microscopy was performed with MAbs against CT143 and CT144 to assess the location and relative abundance of the Pgp4-regulated proteins. In agreement with our proteomic and Western blot findings, we observed inclusion staining of CT143 and CT144 in cells infected with L2 Pϩ and L2 PϪ complement strains but not with PϪ or pgp4 SNP strains. CT143 and CT144 staining was restricted to the inclusion lumen and produced an atypical globular staining profile. The staining density of CT143 and CT144 was significantly less than that of OmpA ( Fig. 5A and B). Staining patterns varied, ranging from small punctate particles to larger particles with outer membrane-like morphologies. Similar staining patterns have been reported by others for CT143 (28) and for plasmid-regulated chromosomal loci CT049 and CT50 (26). It is not understood why plasmid-regulated proteins exhibit atypical inclusion staining properties. In summary, our findings are in complete agreement with previous transcriptomic analyses (5,18,23) and the current proteomic study showing that Pgp4 is the primary master regulator of Pgp3 and chromosomal gene loci (18).
Impact of the chlamydial plasmid on host protein expression. Being an obligate intracellular pathogen, chlamydiae have evolved intricate networks for interacting with their host cell. For example, successful survival requires that chlamydiae dispense with a myriad of host-induced antimicrobial effectors (35). While it is well established in various animal model systems that plasmid-negative isolates are attenuated (4), the precise molecular mechanisms underlying the attenuated phenotype is unknown. As an initial experiment, we determined the host response to infection by comparing C. trachomatis L2 Pϩ-infected HeLa cells to mock-infected cells (see Table 2 in the supplemental material). In total, 7,548 host proteins were detected. There was a robust host response to C. trachomatis L2 Pϩ infection. Employing a Ն2-fold change in mean   (Fig. 6A). We employed Gene Ontology (GO) enrichment analysis using the DAVID resource (36) to functionally profile up-and downregulated genes in response to Pϩ infection (Fig. 6B). Upregulated proteins in chlamydia-infected cells compared to mock-infected cells were enriched with host defense response proteins such as interferons, NF-B, and small GTPase signaling and major histocompatibility complex (MHC) class I antigen processing and presentation pathways. Proteins associated with DNA replication, cell cycle progression, and transcription were downregulated in Pϩ-infected cells compared to mock-infected cells.
A direct comparison between Pϩ and PϪ C. trachomatis L2-infected cells enabled us to assess the effect the plasmid has on host response to infection. In general, the C D mRNA splicing, via spliceosome rRNA processing DNA replication cell division DNA repair transcription elongation from RNA polymerase II promoter mitotic nuclear division viral process mRNA export from nucleus mRNA processing protein sumoylation RNA processing gene silencing by RNA G1/S transition of mitotic cell cycle cell proliferation covalent chromatin modification proteasome−mediated ubiquitin−dependent protein catabolic process mitotic nuclear envelope disassembly regulation of cellular response to heat cell cycle viral transcription histone H4 acetylation mRNA polyadenylation histone deacetylation cellular response to DNA damage stimulus ATP biosynthetic process response to cytokine hepatocyte apoptotic process apoptotic mitochondrial changes gluconeogenesis response to hypoxia positive regulation of cell migration protein transport signal transduction proteolysis involved in cellular protein catabolic process regulation of apoptotic process intracellular protein transport response to virus cell redox homeostasis small GTPase mediated signal transduction regulation of cell shape oxidation−reduction process protein folding antigen processing and presentation of peptide antigen via MHC class I apoptotic process negative regulation of viral genome replication positive regulation of I−kappaB kinase/NF−kappaB signaling interferon−gamma−mediated signaling pathway defense response to virus type I interferon signaling pathway −log10(adj p−value) ® difference in host responses was quite minimal (Fig. 6C). Although there were numerous proteins differentially up-and downregulated, as assessed by statistical significance (P Յ 0.05), only one was greater than 2-fold: tissue-type plasminogen activator (TPA_ HUMAN). Gene Ontology enrichment analysis of statistically significant up-and downregulated proteins was performed (Fig. 6D). This analysis showed enrichment of proteins associated with host inflammatory and defense response upregulated in cells infected with L2 Pϩ compared to L2 PϪ, findings consistent with previous host transcriptional studies of L2 Pϩ-and PϪ-infected cells (37). Downregulated proteins included those involved in extracellular matrix organization and disassembly and cell adhesion. These results show that although chlamydial infection significantly induces a plethora of host cell responses, the chlamydial plasmid does not appear to have a large impact in this regard.

DISCUSSION
Results from our proteomics experiments support and extend the conclusions of C. trachomatis transcriptional studies on the role of the plasmid and Pgp4 in regulating the expression of pgp3 and several chromosomal loci, including those of the glycogen synthase, the Pmp-like proteins CT049 to -051, and the putative type 3 secretion effectors CT142 to -144 (5,18). Comparative proteomic analysis of plasmid-positive and -negative C. trachomatis serovars A and D indicate that while the number of plasmidregulated proteins is reduced compared to L2, the glycogen synthase, CT049 to -051, and CT142 to -144 loci remain highly regulated chromosomal loci in these serovars. We consider the genes coding for these proteins to be the core set of plasmid-regulated chromosomal loci in C. trachomatis. Results from Western blotting and IFA experiments indicate that the levels of expression of both plasmid (Pgp3) and chromosomal (CT142 and -143) proteins are highly Pgp4 dependent and the regulation is extremely tight. The mechanism responsible for Pgp4's positive regulation of transcription remains to be elucidated.
The positive regulatory effects of Pgp4 were conserved among all C. trachomatis serovars. In contrast, there was a clear plasmid-dependent, but Pgp4-independent, downregulation of expression of CPAF that was specific to the L2 serovar. Unlike the tight regulation of CT142 to -144 and Pgp3 expression shown by Pgp4, the effects of the plasmid on the expression of CPAF are less pronounced. The level of CPAF mRNA was not altered in previous C. trachomatis L2 transcriptional studies comparing a plasmid-positive strain and a pgp4 deletion mutant (18). This implies the plasmiddependent Pgp4-independent alterations in CPAF expression are regulated posttranscriptionally and are biologically linked to macrophage infection tropism and invasive infections.
In addition to the eight ORFs, the C. trachomatis plasmid also carries several small RNAs (sRNAs) (4,10,31,32). Bacterial sRNAs are accepted as the major post-transcriptional regulators involved in key processes such as virulence, quorum sensing, survival, plasmid maintenance, and primary and secondary metabolism (38)(39)(40). Recent studies indicate that sRNAs can both repress and activate translation (38).
There is precedent for sRNA regulation of gene expression in chlamydiae. Grieshaber et al. showed that the expression of the DNA condensing elementary body (EB)-specific histone-like protein Hc1 is regulated by a transacting sRNA (34). In addition, they showed that the inhibition of HctA translation by lhtA is a conserved function across chlamydiae (33). Intriguingly, in addition to HctA, lhtA also regulates the expression of CTL0322 (CT066) a highly conserved chlamydia-specific hypothetical protein (41).
Transcriptional studies indicate that the two primary plasmid antisense sRNAs, sRNA-2 and sRNA-7, encoded within pgp8 and pgp5, respectively, are highly expressed (32,42). Furthermore, there are biovar-specific differences in the level of expression of the eight ORFs and sRNAs (42). While the precise regulatory role of the abundant plasmidencoded sRNAs in chlamydiae remains to be elucidated, it has been suggested that sRNA-2 regulates expression of components besides pgp8 (4). It has recently been reported that the expression of the serine protease subtilisin is sRNA regulated in Bacillus (43). In addition to the possibility that sRNAs play a role in downregulating CPAF, it has been shown that Pgp5 selectively suppresses the expression of plasmiddependent genes (44). Clearly more work, such as targeted loss-of-function mutations in the various plasmid sRNAs and pgp5, needs to be done to determine the precise mechanism behind plasmid regulation of CPAF expression in C. trachomatis L2.
Given that CPAF is a critical virulence factor; its plasmid-dependent regulation is intriguing. CPAF is conserved among all serovars, so why would LGV strains regulate CPAF expression differently than non-LGV strains? As discussed above, LGV biovars express significantly higher levels (12-to 100-fold) of sRNA-2 and sRNA-7 than trachoma biovars (42). This increased sRNA expression would intuitively result in a more profound negative regulation of CPAF.
LGV strains naturally infect macrophages not epithelial cells (3). Macrophages have a well-characterized family of pattern recognition receptors (PRRs) that activate the inflammasome and exhibit antibacterial activities (45). A possible explanation is CPAF, or CPAF proteolytically modified type 3 secretion system (T3SS) effector(s) (46), might function as pathogen-associated molecular pattern (PAMP) ligands that engage cytosolic PRR pathways in macrophages, which may be absent or expressed at a lower level in nonphagocytic cells. It is therefore reasonable to propose that LGV strains evolved a unique plasmid sRNA regulatory mechanism to avoid PRRs that activate the inflammasome in a macrophage host as a survival mechanism. This pathogenic strategy would not be expected to function in non-LGV strains as they are epithelial cell tropic and do not naturally infect macrophages. Consistent with this hypothesis are the findings of Webster et al. (47), who showed that the LGV strain expressing CPAF is required for activation of the inflammasome in macrophages. This LGV-specific CPAF function could be a pathogenic characteristic shared with other more conserved CPAF functions common to all chlamydiae that have been described using epithelial cells as a model, which involve the inhibition of antimicrobial peptides (48) and p65 nuclear translocation that suppresses the production of proinflammatory cytokines (46).
A key factor in Pgp4 regulation of chlamydial virulence is no doubt its ability to tightly regulate the expression of Pgp3. By Western blot, there was limited Pgp3 detectable in both plasmidless isolates and the pgp4 SNP mutant. Immunolocalization studies indicate that Pgp3 is detectable in the cytosol of chlamydia-infected cells; Pgp3 is an outer membrane protein (13) and is one of the few chlamydial virulence factors that has been characterized biochemically and shown to be important to chlamydial pathogenesis in animal models (4). Although Pgp3 has been shown to neutralize host antimicrobial peptides (14), it is likely to confer additional functions important to chlamydial pathogenesis that remain to be defined.
The association between plasmid and the ability to deposit glycogen in the chlamydial inclusion has been known for many years (3,25), and the glycogen-negative phenotype is often used as marker to screen for plasmid-cured isolates. In all three plasmidless C. trachomatis serovars tested, glycogen synthase expression is highly downregulated. The decreased expression of glycogen synthase, the penultimate enzyme in the glycogen synthesis pathway, can explain the glycogen-negative phenotype.
Less is known about Pmp-like CT049 to -051 and putative type 3-secreted CT142 to -144. The CT049 to -051 genes are divergently expressed; thus it is likely Pgp4 regulates each of their expressions independently. CT049 and CT050 lack classical Sec-dependent secretion signals, and it has been proposed that they are secreted into the inclusion lumen by a novel mechanism to regulate events important for chlamydial replication and inclusion expansion (26). While the Pgp4-regulated expression of these proteins is conserved, there is substantial sequence variation for all three proteins between C. trachomatis serovars. Thus, these Pmp-like proteins are either under immune selection or have evolved strain-specific virulence functions.
CT142 to -144 are encoded in an operon (28). It has been suggested that CT142, -143, and -144 might form a protein complex (28). In contrast to CT049 to -051, CT142 and CT143 are highly conserved among C. trachomatis isolates. While the precise function of these proteins is currently unknown, the fact that they are all localized to the inclusion lumen (28) brings up the intriguing possibility that their absence could be responsible for the atypical "donut" inclusion morphology associated with plasmidless C. trachomatis isolates.
In keeping with transcriptional studies (37,49), our proteomic data show a substantial host reaction to chlamydial infection. In response to infection with plasmid containing C. trachomatis L2, there were 221 and 48 host genes up-and downregulated with Ն2-fold change, respectively. An overrepresented Gene Ontology (GO) term for all differentially expressed host proteins with annotation reveals a wide variety of functions. Upregulated proteins are those generally categorized as host response to infection, and downregulated proteins are associated with DNA replication and RNA processing. These findings are consistent with our current understanding of chlamydiahost cell interactions (50). Notably however, the host proteomic profiles of cells infected with L2 Pϩ and L2 PϪ were similar, implying that the plasmid has little effect on host response to infection. GO enrichment analysis of statistically significant up-and downregulated proteins showed enrichment of proteins associated with host inflammatory and defense response as upregulated and those involved in extracellular matrix organization as downregulated when cells were infected with Pϩ compared to PϪ isolates. These findings are in general agreement with previous host transcriptional studies of Pϩ-and PϪ-infected cells (37). More importantly, these findings are consistent with PϪ isolates being attenuated in animal models (4,5,7) and displaying altered plaque formation and host cell lysis in vitro (4,5,24,51).
In conclusion, we employed a comparative proteomics approach, using HeLa cells infected with different plasmid-cured C. trachomatis serovars and the L2 serovar transformed with a plasmid containing a nonsense mutation in pgp4, to elucidate the effects of the plasmid on both the chlamydial and host proteome. Our results indicate that Pgp4 exhibits tight regulation over Pgp3, glycogen synthase, and a conserved set of Chlamydiaspecific hypothetical proteins. In addition, we show that there is plasmid-dependent Pgp4-independent regulation of CPAF expression and suggest that this regulation could be dependent on plasmid-encoded sRNAs. Finally, while chlamydial infection has a substantial impact on the host proteome, the plasmid has a limited role in this response.
Proteomic sample preparation. HeLa cells grown in 6-well tissue culture (TC) plates (2 ϫ 10 6 cells/well) were infected in quadruplicate with L2 Pϩ, L2 PϪ, L2 PϪ complement, L2 pgp3 SNP, and L2 pgp4 SNP strains in sucrose-phosphate-glutamic acid (SPG) medium with a mock SPG-only control and rocked for 1.5 h at 37C°in a 5% CO 2 humidified atmosphere. Infections with C. trachomatis serovars APϩ, APϪ, DPϩ, and DPϪ were performed in triplicate in 150-cm 2 TC flasks (3 ϫ 10 7 cells/flask) and rocked for 1.5 h at 37°C in a 5% CO 2 humidified atmosphere. All infected and mock-infected cells were refed with DMEM-10 at 1.5 h post-infection (hpi). A multiplicity of infection (MOI) of 3 was used to ensure that monolayers had Ͼ95% infection. At 30 hpi (L2 and L2 mutants) and 42 hpi (A strain and D strain), medium was aspirated, and cells were washed once with warm Hanks balanced salt solution (HBSS). Cells were treated with 250 l of hot (~100°C) 2% SDS in 50 mM HEPES buffer (pH 8.2), immediately boiled for 10 min, and frozen at Ϫ80°C. Mass spectrometry and differential protein expression data analysis. We used three technical replicates for each condition for serovars A and D and 4 replicates for serovar L2. Mascot search was conducted on all MS/MS samples against the Homo sapiens sequences from Swiss-Prot, Chlamydia trachomatis sequences from serovars D (19), A (7), and L2 (21) with a fragment ion mass tolerance of 0.50 Da and a parent ion tolerance of 10.0 ppm. Carbamidomethyl of cysteine and TMT6plex of lysine and the N terminus were specified in Mascot as fixed modifications and oxidation of methionine as a variable modification. Scaffold Qϩ (version Scaffold_4.7.3; Proteome Software, Inc.) was used for label-based quantitation of TMT peptide and protein identifications. Peptide identifications were accepted at a false-discovery rate (FDR) of Ͻ0.1%. Protein identifications were accepted at less than Ͻ1.0 and with at least 2 identified peptides. Protein probabilities were assigned by the Protein Prophet algorithm (54).
Proteins that contained similar peptides and could not be differentiated based on MS/MS analysis alone were grouped. Proteins sharing significant peptide evidence were grouped into clusters. Channels were corrected in all samples according to the algorithm described in i-Tracker (55). Normalization was performed iteratively (across samples and spectra) on intensities, as previously described (56). Medians were used for averaging. Spectrum data were log 2 transformed, pruned of those matched to multiple proteins, and weighted by an adaptive intensity weighting algorithm. Additional filtering was done to remove proteins with any 0 entries. To account for proportions of chlamydiae within host cells, an additional cyclic LOWESS normalization (57) was done separately for chlamydial and human proteins and was chosen as the approach resulting in the highest reduction in coefficient of variation distribution compared to median, scale, quantile, and variance stabilizing normalization approaches. Differential expression analysis on log 2 -normalized intensity was conducted using Linear Models for Microarray Data (LIMMA) in R with a multigroup comparison design (58,59). To correct for multiple hypothesis testing, the P values for each pairwise comparison were adjusted using the Benjamini-Hochberg method (60). Proteins with adjusted P values of Յ0.05 were considered statistically significant, while an additional criterion of 2-fold change was used to highlight potential biological importance.
Bioinformatics analyses. To enable the comparison between three serovars, pairwise BLASTP was done using the amino acid sequences for each pair of serovars to identify protein pairs (20). The BLOSUM45 scoring matrix was used, and proteins that have Ͼ80% identity and are one-to-one were paired. Eighty percent identity was selected from examining the distribution of the percentage of identity of conducting BLASTP against the same serovar. Functional enrichment of Gene Ontology terms was tested using the DAVID resource (36) by comparing significantly upregulated or downregulated human proteins (adjusted P values of Յ0.05) to all human proteins identified from the mass spectrometry data. The Benjamini-Hochberg (60) adjusted P values from the GO enrichment test were reported to account for multiple hypothesis testing. The GO enrichment plot was generated with modified R scripts from the GO plot R package for biological process terms (61).
The cloned ORFs were expressed as fusion proteins with a C-terminal His tag. Expression of the fusion proteins was induced with isopropyl-␤-D-thiogalactoside for 2 to 3 h during mid-log growth. Fusion proteins were extracted by lysing the bacteria with a French press in Triton X-100 lysis buffer (1% Triton X-100, 1 mM phenylmethylsulfonyl fluoride, 75 U/ml aprotinin, 20 M leupeptin, and 1.6 M pepstatin). After high-speed centrifugation to remove debris, the fusion protein-containing supernatants were passed over His GraviTrap columns (GE Healthcare Life Sciences), and the purified proteins were used to immunize mice for production of monoclonal antibodies (MAbs) (62).