ABSTRACT
Evident in its name, the gastric pathogen Helicobacter pylori has a helical cell morphology which facilitates efficient colonization of the human stomach. An improved light-focusing strategy allowed us to robustly distinguish even subtle perturbations of H. pylori cell morphology by deviations in light-scattering properties measured by flow cytometry. Profiling of an arrayed genome-wide deletion library identified 28 genes that influence different aspects of cell shape, including properties of the helix, cell length or width, cell filament formation, cell shape heterogeneity, and cell branching. Included in this mutant collection were two that failed to form any helical cells, a soluble lytic transglycosylase and a previously uncharacterized putative multipass inner membrane protein HPG27_0728, renamed Csd7. A combination of cell fractionation, mutational, and immunoprecipitation experiments show that Csd7 and Csd2 collaborate to stabilize the Csd1 peptidoglycan (PG) endopeptidase. Thus, both csd2 and csd7 mutants show the same enhancement of PG tetra-pentapeptide cross-linking as csd1 mutants. Csd7 also links Csd1 with the bactofilin CcmA via protein-protein interactions. Although Csd1 is stable in ccmA mutants, these mutants show altered PG tetra-pentapeptide cross-linking, suggesting that Csd7 may directly or indirectly activate as well as stabilize Csd1. These data begin to illuminate a highly orchestrated program to regulate PG modifications that promote helical shape, which includes nine nonessential nonredundant genes required for helical shape and 26 additional genes that further modify H. pylori’s cell morphology.
IMPORTANCE The stomach ulcer and cancer-causing pathogen Helicobacter pylori has a helical cell shape which facilitates stomach infection. Using light scattering to measure perturbations of cell morphology, we identified 28 genes that influence different aspects of cell shape. A mutant in a previously uncharacterized protein renamed Csd7 failed to form any helical cells. Biochemical analyses showed that Csd7 collaborates with other proteins to stabilize the cell wall-degrading enzyme Csd1. Csd7 also links Csd1 with a putative filament-forming protein via protein-protein interactions. These data suggest that helical cell shape arises from a highly orchestrated program to regulate cell wall modifications. Targeting of this helical cell shape-promoting program could offer new ways to block infectivity of this important human pathogen.
INTRODUCTION
As its name implies, Helicobacter pylori is a helical-shaped, Gram-negative bacterial pathogen that colonizes the human stomach of approximately half of the worldwide population. Infection with H. pylori can lead to serious disease, including ulcers and gastric cancer (1), which is the third leading cause of cancer deaths globally (2). Interestingly, H. pylori’s eponymous cell shape is conserved in clinical isolates (3), supporting the theory that this bacterium’s morphology plays an important role during infection (4, 5). For H. pylori, helical shape has been proposed to enhance motility through the viscous gastric mucus, via a corkscrew-like manner (6), to reach its niche on and adjacent to the stomach epithelia. However, experimental evidence to support this theory has only recently been established with the isolation of stable, non-helical-shape mutants that show defects in stomach colonization and motility through gastric mucin medium (7–11). These mutants are the result of loss of function mutations in csd, or cell shape determining, genes (7–10). How H. pylori generates its helical cell shape is still poorly understood, but identification of the csd genes suggests that shape generation is distinct from processes in traditional model organisms used to study cell shape.
In bacteria, the peptidoglycan (PG) sacculus is the macromolecular structure that maintains the shape of the cell (12). PG is made up of glycan strands with repeating disaccharides of N-acetylglucosamine (GlcNAc) and N-acetylmuramic acid (MurNAc) covalently cross-linked by short peptide stems. The PG layer forms a contiguous meshwork that encases the cell and is found in the periplasm of Gram-negative bacteria. The penicillin-binding proteins (PBPs) possess the enzymatic activities to polymerize new PG, while PG hydrolases are responsible for cleaving PG bonds to allow insertion of new PG material, cell wall modifications, and/or separation during cell division (13). In other well-studied bacteria, targeted PG synthesis is the driving force behind cell shape. For example, the characteristic curved rod morphology of Caulobacter crescentus is achieved by directed PG synthesis of the cytoskeletal element CreS into a wedge-like pattern (14, 15). In Vibrio cholerae, the recently identified periskeletal protein CrvA is responsible for directing asymmetric PG synthesis such that more cell wall material is inserted into the outer cell face than the inner cell face, resulting in cell curvature that gives this bacterium its vibroid shape (16). At least four of the csd genes identified to date encode PG hydrolases (7–10). This suggests that structural modifications to the cell wall, apart from PG synthesis, help to drive cell shape generation in H. pylori.
Although previous screens to identify csd genes in H. pylori have been foundational, these efforts were not saturating, and it is likely that additional factors have yet to be identified (7–10). To identify all nonessential csd genes, we used flow cytometry to screen a nonredundant mutant library of H. pylori for cell shape perturbations. Using the differential light-scattering properties of wild-type H. pylori compared to those of known csd mutants, we identified new mutants with a range of cell shape phenotypes, including two new curved rod mutants, slt and csd7. Csd7 is conserved within the Helicobacter genus, and topology modeling predicts that it is a polytopic inner membrane protein. We show that Csd7 is required for the stability of Csd1 and Csd2, two previously characterized Csd factors (9, 17), and directly interacts with these two periplasmic proteins as part of a larger helical cell shape-promoting complex linking periplasmic and cytosolic factors.
RESULTS
Flow cytometry screen of a nonredundant mutant library identifies new cell shape genes with diverse phenotypes.To identify new csd genes, we leveraged prior observations that cell shape mutants show altered light-scattering properties (Fig. 1) (7) and used flow cytometry to evaluate the light scattering profiles of an arrayed genome-wide H. pylori deletion mutant library collection. This collection covers 1,021 of the 1,504 predicted protein-coding genes in H. pylori strain G27 (see File S1 in the supplemental material). Mutants not present in the library are enriched for genes presumed or known to be essential in H. pylori and other organisms (18). Candidate shape mutants were identified by flow cytometry profiles that differed from a wild-type control. As shown in Fig. 1, in addition to wild-type cells, two genetically defined mutant populations of known shape classes—curved rods (ccmA) and variable, “c”-shaped cells (csd3)—were also included as controls in this screen to further define the light-scattering properties of curved or variable shaped mutants (Fig. 1B and C) (7–9). To identify new straight or curved mutants, we looked for clones that displayed a decrease in forward side scatter (FSC), characteristic of the curved rod mutant phenotype (Fig. 1B). Mutants with broader or differently shaped light scattering profiles from wild-type (e.g., csd3) (Fig. 1C), where the fraction of the events in the wild-type control gate was <60% of those captured for wild-type in the same flow cytometry session, were also identified as a preliminary screen hit (Fig. 1D to I). Based on these criteria, 96 mutant hits were selected for a secondary light microscopy screen. Of the initial 96 mutants with altered light-scattering properties, 31 mutants showed altered cell shape by visual inspection (Table 1). Known csd mutants, present in the library, were also excluded from this list. PCR verification for each mutant insertion site revealed that 28 of the 31 mutants had the correctly annotated gene disrupted. The remaining 3 mutants did not contain insertions in the expected loci and were not pursued further. Table 1 summarizes the observed shape phenotypes seen by light microscopy for the final 28 screen hits of interest. The screen identified a range of different mutants with various shape phenotypes, including altered helicity, altered length/width, increased shape heterogeneity, bulbous/branched cells, and filamentous mutants (Fig. 1; Table 1; see also Table S1). Of the new shape mutants, only two lacked helical shape, HPG27_0728 and slt (HPG27_0607) mutants, and both displayed a curved rod phenotype. Based on these findings, we have likely identified all nonessential and nonredundant genes required for helical shape.
Light scattering profiles distinguish diverse shape phenotypes in an ordered H. pylori deletion library. (A to I) Forward versus side scatter contour plots measured by flow cytometry showing the wild-type and the indicated mutant populations. Drawn gates represent 64% of the wild-type population. For cell shape mutants, the numbers listed refer to the percentages of the populations in this wild-type gate. Insets, light micrographs (×1,000 magnification, phase-contrast) of strains with indicated shapes. For the contour plots: blue and green represent areas of lower cell density, red and orange correspond to areas of high cell density, and yellow is midrange. Scale bars, 2 μm. Strains used were wild type (WT; LSH100), ccmA (LSH142), csd3 (NSH152a), slt (MHH9), HPG27_1093 (DCY8a), HPG27_1325 (DCY113), prc (HPG27_1298; DCY114), HPG27_0119 (DCY115), and HPG27_0750 (DCY6a).
Summary of the flow cytometry screen for new cell shape mutants in H. pylori
Two curved rod mutants map to slt and a hypothetical protein.Two mutants identified in the flow cytometry screen showed no helical cells and displayed profiles with lower forward side scatter, consistent with the ccmA (HPG27_1480) curved rod phenotype. These clones contained insertions in HPG27_0607 (a soluble lytic transglycosylase, slt) and HPG27_0728 (a hypothetical protein). Phase-contrast images show that these mutants are indeed curved rods and have significantly less side curvature than the wild type (Fig. 1 and 2). Helical shape was restored in both cases by expressing a wild-type copy of the respective gene at the rdxA locus, indicating that slt and HPG27_0728 are required for proper helical cell shape in H. pylori (Fig. 2). HPG27_0728 encodes a hypothetical protein that has not been previously implicated in cell shape generation or maintenance in H. pylori; therefore, we have designated this gene csd7. Slt has been characterized in H. pylori as having lytic transglycosylase activity (19). Interestingly, in contrast to the curved rod phenotype we observed for the slt mutant in G27, disruption of slt in the H. pylori 26695 strain background is reported to retain the wild-type shape (19). However, muropeptide analysis of the G27 slt mutant mirrors the PG profile previously reported by Chaput et al. (19) (Table 2). In both the 26695 and G27 strain backgrounds, the slt mutant showed a decrease in tetrapeptide monomers, dimers, anhydromuropeptides, and cross-linking but an increase in tripeptides compared to that in the wild type (Table 2) (19). Although the shape phenotype was not consistent between the two strain backgrounds, the PG profiles of the slt mutants were congruous.
csd7 and slt are essential for normal helical shape. Complementation strains have the native locus disrupted by deletion and insertion of chloramphenicol acetyltransferase cassette (cat) and the indicated gene expressed from the rdxA locus. (A) Light micrographs (×1,000 magnification, phase-contrast) of various strains (genotypes listed below). Scale bars, 5 μm. (B) Forward versus side scatter contour plot measured by flow cytometry showing the csd7 mutant population. The drawn red gate encompasses 64% of the control wild-type population. For csd7, 48% of the mutant population falls within this control gate. (C) Scatter plot arraying wild-type, csd7, and the csd7 complemented (csd7 rdxA::csd7) populations by cell length (x axis) versus side curvature (y axis). Quantitative analyses of phase-contrast images of bacteria to measure side curvature and central axis length were performed with the CellTool software package as described previously (9). (D) Smooth histogram displaying population cell curvature (x axis) as function of density (y axis). (E) Scatter plot arraying wild-type, slt, and the slt complemented (slt rdxA::slt) populations by cell length (x axis) versus cell curvature (y axis). (F) Smooth histogram displaying population cell curvature (x axis) as function of density (y axis). Strains used were WT (LSH100), slt (MHH9), slt rdxA::slt (AC1), csd7:cat rdxA::csd7 (DCY28), and csd7:cat (DCY7a).
Summary of muropeptide composition of PG in mutant strains
Csd7 is a predicted polytopic integral membrane protein.Recognizable homologous csd7 genes appear restricted to the Helicobacter genus. csd7 is encoded directly downstream of amiA (HPG27_0729), an N-acetylmuramoyl-l-alanyl amidase required for proper cell division in H. pylori (Fig. 3A) (20, 21). Topology modeling using jackhmmer (22) predicts that Csd7 is an inner membrane protein with 8 transmembrane helices (Fig. 3B). To further study Csd7, we constructed a strain with a 3×FLAG-tagged version of csd7 (csd7-FLAG) expressed at the native locus as well as a complemented csd7 mutant with csd7-FLAG expressed at the rdxA locus. For studies with other proteins already FLAG epitope tagged, we engineered a 3×VSV-G-tagged version of csd7 (csd7-VSV-G) at the native locus. All epitope-tagged strains displayed wild-type helical morphology, indicating that the FLAG or VSV-G tags do not disrupt normal Csd7 function (see Fig. S1).
Csd7 is an integral membrane protein. (A) Schematic of the csd7 gene locus identified in the flow cytometry screen. The disruption site of the chloramphenicol acetyltransferase cassette (cat) was inserted 347 bp in the same orientation as the HPG27_0728 open reading frame and results in a 60-bp internal deletion. (B) Predicted topology for Csd7. IM, inner membrane. (C) Cell fractionation and localization of Csd7-FLAG with and without detergent (0.5% Triton X-100). The whole-cell extract (WCE), pellet (PEL), and supernatant (SUP) fractions were analyzed by SDS-PAGE and immunoblotted for Csd7-FLAG using anti-FLAG antibodies. The molecular weight marker (kDa) is on the left. Strains used: LSH100 and DCY71.
Because Csd7 is predicted to be a polytopic membrane protein, we examined Csd7-FLAG solubility during cell lysis and fractionation in the presence and absence of detergent. In whole-cell extracts, Csd7-FLAG was found predominantly in the insoluble pellet fraction without the addition of detergent (Fig. 3C). However, addition of detergent partially solubilized Csd7-FLAG (Fig. 3C), consistent with the predicted behavior of an integral membrane protein.
Muropeptide analysis of the csd7 mutant resembles the csd1, csd2, and ccmA shape mutants.As with most other bacteria, the PG sacculus is the structural determinant of cell shape in H. pylori (9). Therefore, to better understand how Csd7 might contribute to proper H. pylori shape generation, we analyzed the global muropeptide composition of the csd7 mutant, wild-type, and representative strains with similar curved rod phenotypes (Table 2). Interestingly, the muropeptide analysis revealed that the csd7 mutant bears a strikingly similar pattern to those of the csd1, csd2, and ccmA mutants (Table 2) (9). Reminiscent of these three mutants, compared to the wild type, the csd7 mutant showed an increase in tetrapentapeptide dimers (35%) and a decrease in both tetrapeptide (31%) and tripeptide monomers (34%) (Table 2) (9). This shared muropeptide phenotype suggests a connection among all four gene products. Although csd7 is not genetically linked to csd1, csd2, or ccmA, the latter three genes are found in the same locus and reside directly adjacent to each other (9). As previously described, csd1, csd2, and ccmA are all individually required for proper wild-type helical shape in H. pylori, and the null mutants display a curved rod morphology comparable to that of the csd7 mutant.
Csd7 and Csd2 promote shape by stabilizing Csd1.To further investigate a possible link(s) between Csd7, Csd1, Csd2, and CcmA, we performed Western blot analyses on whole-cell extracts from the null and complemented mutant strains to determine if there were any protein stability codependencies. Interestingly, absence of csd2 or csd7, but not ccmA, resulted in a complete loss of Csd1 similar to that in a csd1 null mutant (Fig. 4A). Similarly, wild-type expression of a FLAG-tagged version of Csd2 at the native locus (Csd2-FLAG) required csd1 and csd7 but not ccmA (Fig. 4B). In both cases, expression was rescued by complementation of the appropriate gene at the rdxA locus (Fig. 4A and B). Csd7-FLAG expression was not dependent on the presence of csd1, csd2, or ccmA, and CcmA did not require csd1, csd2, or csd7 for its expression (Fig. 4C and D). The stability codependency (Fig. 5H) for some of these proteins on each other and the shared muropeptide profiles provide a partial explanation for why these mutants have a shared shape phenotype. Based on these results, we suggest that the csd7 and csd2 mutants have a csd1 shape phenotype because they are essentially csd1 null mutants. How the ccmA mutant displays the same morphological perturbation and PG profile, independent of affecting Csd1, Csd2, or Csd7 protein stability, is unclear and will require further investigation.
Csd1, Csd2, and Csd7 show costability dependency, and catalytically inactive variants of Csd1 and Csd2 do not affect protein stability. Whole-cell extracts of the indicated strains, normalized by OD600, were analyzed by SDS-PAGE and immunoblotted with anti-Csd1 (A, E, and F), anti-FLAG (B and D), anti-CcmA (C), or anti-VSV-G (G) antibodies. Positions of the molecular weight markers (in kilodaltons) are on the left. Complemented strains are indicated with “C.” LS, unmarked LSH100 strain; csd2-F, Csd2-FLAG; csd7-F, Csd7-FLAG; csd2-V, Csd2-VSV-G strain. The gene listed above each band indicates the relevant deletion or mutation made at the native loci. Strains used: ACH1, DCY7a, DCY28, DCY72, DCY73, DCY74, DCY75, DCY76, DCY77, DCY89, DCY102, DCY106, DCY105, DCY110, DCY111, DCY112, JTH4, LSH100, LSH113, LSH120, LSH121, LSH140, LSH141, LSH142, LSH120, and NAH1.
Csd1, Csd2, and Csd7 directly interact with each other in bacterial two hybrid (BATCH) and co-immunoprecipitation (co-IP) assay blots. (A to D) Plasmid pairs encoding indicated T18 and T25 fusion proteins were cotransformed into E. coli BTH101 (cya-99). Individual colonies were patched on M9-glucose supplemented with Amp, Kan, X-Gal, and 1 mM IPTG. Plates were incubated at room temperature and photographed after 72 h. Interacting partners bring together T18 and T25 to reconstitute adenylate cyclase activity that is detected using lacZ induction (blue) as a reporter. (E) FLAG co-IP of Csd1-FLAG and Western blotted for FLAG-tagged Csd1 and VSV-G-tagged Csd2. (F) FLAG co-IP of Csd2-FLAG and Western blotted for FLAG-tagged Csd2 and Csd1. (G) VSV-G co-IP of Csd7-VSV-G and Western blotted for VSV-G-tagged Csd7 and FLAG-tagged Csd2. (H) Protein stability codependencies. The arrows represent promotion of protein stability, black dashed lines indicate protein interactions validated by pulldown under native conditions, and the gray dashed line indicates a putative protein interaction validated by pulldown using formaldehyde cross-linking.
As previously reported, Csd1 and Csd2 are predicted M23 family endopeptidases, possessing characteristic LytM domains with active site motifs HXXXD and HXH (9, 17). In Csd2, the first two histidines of these motifs are not conserved and are replaced with Glu165 and Lys246, respectively, suggesting that Csd2 may not possess PG hydrolase activity (9, 17). In further agreement with this observation, when we mutated the Csd2 aspartic acid residue in the HXXXD motif to either a cysteine (Csd2D169C) or alanine (Csd2D169A), cells were motile and displayed wild-type helical morphology (Fig. S1B). Moreover, for both Csd2 variants, Csd1 and Csd2 protein levels were comparable to that of the wild type (Fig. 4E to G). Taken together, these results bolster the idea that Csd2 is not acting as an enzyme itself but functions to help stabilize Csd1 in H. pylori.
Csd7 directly interacts with Csd1, Csd2, and CcmA.Csd7 is predicted to be a polytopic inner membrane protein, and both Csd1 and Csd2 contain N-terminal signal sequences, suggesting that they are secreted into the periplasm (8). CcmA, on the other hand, is a bactofilin-like protein thought to reside in the cytoplasm (8). Based on the predicted cellular localization of these proteins, we wondered if Csd7 could link Csd1 and Csd2 in the periplasm and CcmA in the cytoplasm. Three lines of evidence support this hypothesis. We first used a bacterial two-hybrid (BACTH) assay that takes advantage of the split adenylate cyclase fragments, T18 and T25, from Bordetella pertussis to reconstitute cAMP production in an E. coli cya strain (23). T18/T25Csd7 showed a strong interaction signal when paired with T25/T18Csd1 as well as T25/T18Csd2 (Fig. 5B and C). Additionally, T18/T25Csd1 displayed a robust signal when expressed with T25/T18Csd2 (Fig. 5A). Evidence that Csd2 and Csd7 may form self-dimers or self-oligomers was also indicated by the BACTH assay (Fig. 5A and C). The Csd1/Csd2 results further support published structural findings by An et al., who reported that Csd2 exists as both a monomer and dimer and that Csd1 and Csd2 form a heterodimer (17). An interaction signal was not detected between Csd7 and CcmA by BACTH (Fig. 5D). Additionally, we saw no evidence of CcmA directly interacting with Csd2 by BACTH (see Fig. S2); however, since these proteins are not predicted to reside in the same cellular compartment, a positive interaction signal was not expected. CcmA did show a strong interaction signal with itself (Fig. 5D), suggesting that this protein forms oligomers, which would be consistent with its role as a putative cytoskeletal element.
To further validate the BACTH interaction data, we performed a series of coimmunoprecipitation (co-IP) experiments from detergent-solubilized H. pylori cell extracts. Using a strain expressing csd1-FLAG and csd2-VSV-G, Csd1-FLAG was readily immunoprecipitated using anti-FLAG magnetic beads (Fig. 5E). Immunoblotting for Csd2-VSV-G revealed that it was also present in the input and was pulled down in the Csd1-FLAG IP sample (Fig. 5E). A reciprocal pulldown experiment using a strain expressing csd2-FLAG showed that Csd2-FLAG also immunoprecipitated with anti-FLAG magnetic beads and that Csd1, immunoblotted with an anti-Csd1 polyclonal antibody, readily coimmunoprecipitated as well (Fig. 5F). A control IP using a RecA-FLAG strain failed to coimmunoprecipitate Csd1, demonstrating that the Csd1-Csd2 interaction is specific (see Fig. S3). Pulldown experiments using a strain expressing csd2-FLAG and csd7-VSV-G and immunoprecipitating Csd7-VSV-G with an anti-VSV-G resin revealed that Csd2-FLAG coimmunoprecipitated only with the addition of a formaldehyde cross-linker (Fig. 5G).
In addition to performing the immunoprecipitation experiments to probe for specific interactions by immunoblotting, we submitted these samples for mass spectrometry analysis to further validate these findings as well as potentially identify new interaction partners. Wild-type (no tag) and RecA-FLAG extracts were used as negative controls (see Table S5). In accordance with published findings and our interaction data, mass spectrometry identified both Csd1 and Csd2 as top hits for both the Csd1-FLAG and Csd2-FLAG IP samples (Table 3; see Table S6 to S8) (17). For Csd7-FLAG, Csd1, Csd2, and CcmA were among the top scoring protein hits (Table 3). Csd7 came up as the 5th place hit by mass spectrometry most likely due to the hydrophobicity of this protein, since it is predicted to be a polytopic transmembrane protein. The mass spectrometry, co-IP, and BACTH data all support a multiprotein complex containing Csd1, Csd2, and Csd7. It is possible that CcmA is part of this complex, but this interaction may be through indirect binding with Csd7.
Protein interactions identified by immunoprecipitation and mass spectrometry of Csd1, Csd2, and Csd7
csd7 mutation causes strain-dependent morphology perturbations and stomach colonization deficits.Previous studies have shown that csd1 (curved rods), csd3 (variable curved rods), and csd4 (straight rods) mutants are all attenuated for stomach colonization in a mouse model of colonization (8–10). Based on the shared protein stability and interaction data of Csd1 and Csd2 with Csd7, we predicted that a csd7 mutant would display a similar colonization defect to that of the csd1 curved rod mutant. We thus created csd7 mutant and complemented strains in a mouse-adapted derivative of strain G27 (MSD132 [24]) and an unrelated mouse-colonizing strain PMSS1 (25, 26). In the PMSS1 strain, deletion of csd7 led to both straight and highly curved cells (Fig. 6A) and a much broader curvature distribution than the wild type, which was restored by expression of csd7 at the rdxA locus (Fig. 6B). Deletion of csd1 in the PMSS1 strain resulted in a very similar morphology phenotype as that in PMSS1 csd7 and distinct from that previously observed in strain G27 derivatives (9). In both strain backgrounds, the csd7 mutant and complemented strains retained motility.
csd7 alters cell morphology in strain PMSS1. (A) Scatter plot arraying wild-type, csd1, csd7, and the csd7 complemented (csd7 rdxA::csd7) populations by cell length (x axis) versus side curvature (y axis). Quantitative analysis of phase-contrast images of bacteria to measure side curvature and central axis length was performed with the CellTool software package as described previously (9). (B) Smooth histogram displaying population cell curvature (x axis) as function of density (y axis). Strains used: DCY119, DCY120, LMH1, and PMSS1.
In the MSD132 strain, the csd7 mutant was outcompeted by the wild type in the C57BL/6 mouse model (Fig. 7). Interestingly, the csd7 complemented strain did not fully restore colonization and did not compete as well as the wild type (Fig. 7). This is despite an ability of both the csd7 and complemented strains to maintain a similar ratio during 72 h of continuous coculture with the wild type in broth culture (see Fig. S4). In the PMSS1 strain, we also saw partial rescue compared to the wild-type for colonization; although in this case, the complemented strain colonized significantly better than the csd7 strain (Fig. 7). Thus, although shape was fully complemented (Fig. 2 and 6), the rescue of colonization by the null mutant was only partial when csd7 was expressed at the rdxA locus. These results suggest that the lack of full complementation of the csd7 mutant is specific to infection and not manifested under in vitro growth conditions. Western blotting of Csd7-FLAG levels indicated that there was more expression of Csd7 from rdxA than from the native locus (see Fig. S5). It is possible that this difference in Csd7 expression accounts for the complementation discrepancy results; however, further studies would be necessary to make this determination.
csd7 promotes stomach colonization. Stomach colonization loads from 1-week infections of csd7 mutant (Δcsd7), complemented (cmpl), and wild-type strains in pairwise competition as indicated in the MSD132 (left) and PMSS1 (right) strain backgrounds. Each data point represents the CFU per gram of stomach tissue for each genotype from a single mouse. Lines connect the data points of each genotype from one animal and the dotted line indicates the limit of detection of infection. Numbers below the x axis indicate the fractions of mice infected with each genotype. Displayed data were pooled from two separate experiments, and each replicate is shown in black or gray. Differences in colonization loads among genotype in competition were assessed by Wilcoxon matched pairs signed-rank test. *, P < 0.05; **, P < 0.01. Strains used: DCY27, DCY119, DCY120, DCY121, MSD132, and PMSS1.
DISCUSSION
As the mechanistic details driving bacterial shape generation become further elucidated, it has become increasingly apparent that commonalities exist along with unique differences among cell shape systems. In almost all bacteria, PG remains the underlying unifier when it comes to defining a bacterium’s shape, with its biosynthesis and modification directly influencing the overall gross morphology of the cell. This is made evident by the observation that intact, purified PG sacculi maintain the overall morphology of the bacterial cell (12). However, the field has begun to appreciate the diversity of specific PG modifications and molecular mechanisms employed by different bacteria to target these modifications to the appropriate site(s) within the cell wall. This includes the previously mentioned nonenzymatic scaffolding proteins CreS and CrvA that drive differential PG synthesis along one side of the cell, resulting in a curved rod morphology in Caulobacter crescentus and Vibrio cholerae, respectively (14–16).
Based on our previous observations that nonhelical mutants of H. pylori have altered light-scattering properties, we undertook a genome-wide screen to identify more helical cell shape-regulating genes. In addition to identifying two new genes required for helical shape, we identified an additional 26 genes that modify cell shape in diverse ways, including altering the properties of the helix, altering cell length or width, forming cell filaments, and increasing cell shape heterogeneity. Many genes identified in our screen have not previously been linked to cell wall metabolism and include six hypothetical proteins (see Table S1 in the supplemental material). The largest class of new shape mutants was helical but produced filaments of various lengths. The filamentous mutants include the H. pylori homologue of FtsE, which participates in the activation of amidases that remove stem peptides from the sugar chains to cleave septal PG, allowing for daughter cell separation in E. coli (27–29). We also identified mutations yielding filamentous cells in all but one gene in the H. pylori tol-pal operon (a pal mutation in HPG27_1071 was not present in our library). The Tol-Pal complex spans the cell envelope with components in the outer membrane (Pal lipoprotein), periplasm (TolB), and inner membrane (TolA, TolR, and TolQ). This complex may facilitate the invagination of the outer membrane during cell division, as tol-pal mutants show increased formation of outer membrane vesicles (OMVs) and cell division defects under low osmotic conditions in E. coli (30). In addition, Tol proteins modulate the activity of a PG synthase-OM regulator pair active in cell division, PBP1B-LpoB (31). Recently, enhanced OMV accumulation was also noted for H. pylori tolB but not pal mutants (32). Another mutation mapped to rare lipoprotein A (rlpA). RlpA localizes to the division septum in E. coli, but mutants do not show defects in cell division (33). In Pseudomonas aeruginosa, however, rlpA mutants filament, and recent work demonstrated RlpA has lytic transglycosylase activity on PG sugar chains lacking stem peptides (34). One additional mutant we identified participates in turnover and repair of damaged proteins (Pcm), and two mutants map to uncharacterized predicted outer membrane proteins. Only one of the other gene hits that had more subtle cell shape phenotypes had annotations previously linked to cell wall modification or cell division. A mutant with altered helicity is annotated as periplasmic carboxyterminal protease Prc, which in E. coli, processes the cell division-associated lipoprotein NlpI and penicillin binding protein 3 (35, 36) and digests the cell elongation-specific PG hydrolase MepS in an NlpI-dependent manner (37, 38). Further exploration of these genes and phenotypes promises to yield new information on unique features of the cell envelope of H. pylori and links between cell shape and other cellular processes.
Our screen yielded two new genes required for helical cell shape in strain G27: the PG hydrolase slt and a previously uncharacterized protein csd7. In H. pylori, four Csd PG hydrolases (Csd1, Csd3, Csd4, and Csd6) have been linked to helical cell shape maintenance, but how cleaving of PG bonds results in helical shape generation is still unknown. What is becoming more apparent is that these enzymes are not working alone but in concert with other proteins for stability and proper functioning. In addition to Csd7, Csd5 and the bactofilin CcmA are also required for helical shape but lack predicted enzymatic function. As our screen covered nearly all nonessential genes in H. pylori strain G27, it appears that these nine genes likely represent the full complement of nonessential, nonredundant helical cell shape-determining genes for this strain.
We show here that Csd1, Csd2, and Csd7 directly interact and that Csd1 and Csd2 require each other and Csd7 for stability. Like csd1 mutants, csd7 mutants show attenuated stomach colonization. Analysis of csd7 mutation in a second strain (PMSS1) revealed a distinct morphologic perturbation. We confirmed that a PMSS1 csd1 mutant showed the same broad curvature distribution as the csd7 mutant and distinct from the slightly curved phenotype observed for both mutants in G27 strain derivatives. These results further support a connection between csd1 and csd7 across strains. In both strain backgrounds, shape in broth culture was fully restored by complementation at the rdxA locus but full stomach infectivity was not. This might result from altered Csd7 expression compared to that in broth at the nonnative locus or from polar effects of the mutation cassette on neighboring genes that only become evident under infection conditions. Further study will be required to resolve these possibilities.
We also provide evidence to suggest that CcmA may indirectly interact with Csd1 through Csd7, which would support a multiprotein transmembrane complex that provides a link from the cytoplasm to the periplasm via Csd7. In a separate study, we showed that Csd5 interacts with CcmA and the cytoplasmic PG precursor enzyme MurF and directly binds PG in the periplasm (39). Interestingly, among the top 20 hits in the Csd7 IP was MurF. This implicates the existence of a putative “helical shapesome” that includes at least one PG-modifying enzyme (Csd1) and additional proteins (CcmA, Csd2, Csd7, and Csd5) that serve to stabilize and/or localize their biochemical activity as well as a PG precursor synthesis enzyme (MurF) to ensure proper helical shape generation in H. pylori (Fig. 8). It is also possible that along with stabilizing Csd1, Csd2 and Csd7 stimulate Csd1 enzymatic activity to further regulate this PG hydrolase. Understanding how PG bond cleavage can lead to a helically shaped cell wall is still unclear and is under active investigation.
Model of the Csd7 protein interaction network. Shown is a schematic diagram of a putative “helical shapesome” which includes the PG hydrolase Csd1, nonenzymatic proteins (CcmA, Csd2, Csd7, and Csd5) that serve to stabilize and/or localize this complex, and the PG precursor synthesis enzyme MurF. See the text for details. OM, outer membrane; IM, inner membrane; PG, peptidoglycan.
MATERIALS AND METHODS
Bacterial strains and growth.Strains used in this work are described in Table S2 and strain construction details can be found in the supplemental material. LSH100, a derivative of the sequenced human clinical isolate G27 (40, 41), or isogenic mutants were grown at 37°C in a tri-gas incubator with 10% oxygen, 10% carbon dioxide, and 80% nitrogen on solid medium containing horse blood (HB) agar or in liquid culture containing Brucella broth (BD Biosciences) with 10% heat-inactivated fetal bovine serum (FBS; Gemini-Benchmark) (BB10), as previously described (18). Isogenic knockout mutants were constructed using a vector-free allelic replacement strategy where a portion of the coding sequence was replaced by a chloramphenicol (vast majority of the mutants) or kanamycin (minor proportion of the mutants) resistance cassette, and mutants were selected with 15 μg/ml chloramphenicol or 25 μg/ml kanamycin, respectively, as previously described (42). To generate complementation mutants, we integrated a wild-type copy of the deleted gene at a neutral locus: either rdxA with metronidazole selection (43) or an intragenic region between HPG27_0186 and HPG27_0187 (McGee locus) with kanamycin selection (44). The primers used for generating H. pylori mutants are listed in Table S3.
Mutant library preparation and flow cytometry screening.Included in the mutant library array are 311 transposon insertions (18), 682 stitch PCR integrants containing a chloramphenicol acetyltransferase cassette (cat) genomic DNA (gDNA) from 55 mutant strains in our strain collection, and 5 mutants generously provided by Karen Ottemann (HPG27_1006-cheW, HPG27_0772-motB, HPG27_0543-fliN, HPG27_0576-cheV, and HPG27_1125) (File S1). To conduct the flow cytometry screen, the arrayed mutant genomic DNA library was first transformed into the LSH100 strain background. Transformations were performed in 96-well-plate format using 100 μl of liquid LSH100 grown overnight to an optical density at 600 nm (OD600) of ∼1.0. To each well, 5 μl of gDNA (DNA concentration range between ∼20 and 300 ng/μl) was added, and the transformations were allowed to go 6 h without agitation in the tri-gas incubator. Cells were then pin transferred onto selective medium and allowed to incubate for 2 days. After 2 days, transformants were pin transferred onto fresh selective medium and grown overnight. Transformants were then individually inoculated in 200 μl of BB10 medium with selection and grown shaking overnight. The following morning, the overnight minicultures were back diluted 1:10 into fresh selective medium (200 μl final volume) and grown shaking for 6 h. Cells were then fixed in 2× fix solution (8% formaldehyde in 1× phosphate-buffered saline [PBS]) and diluted 1:40 into 1 ml of 1× PBS for flow cytometry analysis. As indicated in File S1, we were not able to collect data for 15 clones present in the library due to an inability to transform or poor growth.
Flow cytometry was performed using an Influx high-speed cell sorter (BD Biosciences), and the results were analyzed with FlowJo (Tree Star). For optimal detection of small cells, we improved the sensitivity in the forward scatter direction by effectively increasing the magnification of the detection system, a technique that has been used with high-speed cell sorters (45). Rather than using a simple lens to collect forward-scattered laser light, we employed a high numerical aperture (NA) microscope objective (Mitutoyo 20×; Edmund Optics, Barrington, NJ) and a second lens to focus the light from the objective onto a mirrored surface. A 0.7-mm pinhole located on the central axis of the mirror acts as a field stop to remove light that is not in the focal image plane. This system effectively separates scattered light from stray laser light, allowing the detection of particles as small as 100 nm and increases its sensitivity to small changes in the amount of forward scattered light caused by morphological perturbations.
Phase-contrast microscopy of H. pylori cells and quantitative morphology analyses.Phase-contrast microscopy was performed as previously described (9). In brief, cells were grown in shaken liquid culture to mid-log phase (OD600 of 0.2 to 1.0), fixed in a 4% paraformaldehyde-PBS solution containing 10% glycerol, and mounted on glass slides. Cells were imaged with a Nikon TE 200 microscope equipped with a 100× oil immersion lens objective and Nikon CoolSNAP HQ charge-coupled-device (CCD) camera controlled by MetaMorph software (MDS Analytical Technologies). Quantitative analysis of phase-contrast images of bacteria was performed with the CellTool software package as previously described (9).
Muropeptide analysis.PG was harvested from 300 OD600 of H. pylori cells grown in liquid culture and prepared as previously described (9).
Mouse colonization.Female C57BL/6 mice 35 to 49 days old were obtained from Jackson Laboratory and certified free of endogenous Helicobacter infection by the vendor. Mice were housed in an Association for the Assessment and Accreditation of Laboratory Animal Care-accredited facility in sterilized microisolator caging and provided with irradiated PMI 5053 rodent chow and acidified reverse-osmosis-purified water ad libitum. The FHCRC Institutional Animal Care and Use Committee approved all manipulations. H. pylori infection and recovery from the stomach were performed as described previously (46) using 5 × 107 cells/strain in the inoculum for competition experiments. After 1 week, the mice were euthanized by inhalation of CO2, and the glandular stomachs were removed and opened to remove any food. Half of each stomach was homogenized in 500 μl BB10 medium. Dilutions of homogenate were plated on nonselective and selective HB plates to enumerate bacteria of each genotype. If no bacteria were recovered, we set the number of colonies on the lowest dilution plated to 1 to calculate the competitive index.
Immunoblotting.H. pylori whole-cell extracts were prepared by harvesting log-phase bacteria at an OD600 of 1.0 by centrifugation for 2 min at maximum speed in a microcentrifuge, suspending the pellets in 2× sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) sample buffer at an OD600 of 10.0, and then boiling them for 10 min. Proteins were loaded into Mini-Protean TGX polyacrylamide gels (Bio-Rad) and then transferred onto polyvinylidene difluoride (PVDF) membranes using the TurboTransfer (Bio-Rad) system. Membranes were then blocked for 2 h in 5% nonfat milk–Tris-buffered saline containing 0.05% Tween 20 (TBS-T) and then incubated with primary antibody overnight at 4°C. Membranes were then washed with TBS-T and then incubated with secondary antibody overnight at 4°C. Proteins were detected using a chemiluminescent substrate according to the manufacturer’s directions (Pierce ECL) and imaged directly with the Bio-Rad gel documentation system.
H. pylori cell fractionation.A final total of 20 OD600 of H. pylori bacteria were grown in liquid culture to mid-log phase (OD600 of ∼0.45) and harvested by centrifugation at 8,000 rpm for 10 min at 4°C. Each cell pellet was resuspended in 2 ml of buffer A (50 mM Tris [pH 7.0], 200 mM NaCl, 75 mM potassium glutamate, 10 mM magnesium acetate, 5% glycerol, 20 mM phenylmethylsulfonyl fluoride [PMSF]) in the presence or absence of 0.5% Triton X-100. The suspension was sonicated at 10% power with a microtip (Sonic Dismembrator model 500; Branson) for a total of 90 s using 3 pulses of 30 s (with 60 s on ice between each pulse). Normalized volumes of whole-cell lysate and soluble and pellet fractions were prepared by centrifugation at 14,000 rpm for 10 min at 4°C. Equal volumes of each sample were analyzed by 10% SDS-PAGE followed by Western blotting using PVDF membranes and anti-FLAG M2 (Sigma) primary antibody as described above.
Cross-linking, immunoprecipitations, and mass spectrometry analysis.Cells freshly grown on HB plates were resuspended in ice-cold PBS and washed twice. A total of 3 × 1010 cells were resuspended in 2 ml ice-cold PBS, and formaldehyde was added to a 1% final concentration. The cross-linking reaction mixture was incubated for 30 min on ice. The reaction was stopped by the addition of 0.125 M glycine (pH 2.0) and incubated for an additional 10 min on ice. After one wash with ice-cold PBS, the cells were pelleted and stored at −80°C. Cross-linked pellets were lysed in 2 ml of lysis buffer (20 mM Tris [pH 8.0], 150 mM NaCl, 2.0% glycerol, 1.0% Triton X-100) plus 1× Complete protease inhibitors (Roche), 2 μl of Benzonase (EMD), and 20 μl 100 mM PMSF (Sigma). Cells were briefly sonicated and centrifuged at 20,000 × g for 30 min at 4°C. To the supernatants, 40 μl of anti-FLAG M2 magnetic beads (Sigma-Aldrich) was added and the mixture was rotated at 4°C for 90 min. Beads were washed four times with wash buffer (20 mM Tris [pH 8.0], 150 mM NaCl, 2.0% glycerol, 0.1% Triton X-100). After the final wash, the beads were resuspended in 40 μl of 2× SDS-PAGE loading buffer and run on an SDS-PAGE gel for Western blotting and/or mass spectrometry analysis.
For mass spectrometry, samples were analyzed as previously reported (39). Briefly, the immunoprecipitated fractions were run for approximately 5 min into the wells of an SDS-PAGE gel (Bio-Rad TGX), excised as a gel slice, and submitted for analysis to the Proteomics Shared Resource of the Fred Hutchinson Cancer Research Center. The data were transferred through automated pipelines to systems supported by the FHCRC Computing Support Shared Resource. Data were analyzed using the Proteome Discoverer (Fisher) software package.
Bacterial two-hybrid assay.The BACTH assay was carried out as described in reference 23. Briefly, plasmid pairs encoding the indicated T18 or T25 fusion proteins were cotransformed into BTH101 (cya-99). Individual colonies were patched on M9-glucose supplemented with ampicillin (100 μg/ml), kanamycin (50 μg/ml), X-Gal (5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside; 40 μg/ml), and 1 mM IPTG (isopropyl-β-d-thiogalactopyranoside). Plates were incubated at room temperature and photographed after 72 h. For this particular BACTH assay, interacting partners bring together T18 and T25 to reconstitute adenylate cyclase activity. This activity is detected using lacZ induction as a reporter.
Growth experiments.Growth testing was accomplished using 200 μl of BB10 minicultures grown in a 96-well plate. For growth experiments, liquid cultures were grown overnight to an OD600 of <1 and diluted to 0.005 OD600/ml (0.0025 OD600/ml for each strain in coculture experiments). At desired intervals, cell aliquots were diluted serially and plated on nonselective and (for coculture experiments) selective HB plates to enumerate total and mutant CFU.
ACKNOWLEDGMENTS
We thank Karen Ottemann (University of California, Santa Cruz) for providing mutants for the sequence-defined mutant library, Ashley Roarty (Molecular and Cellular Biology Graduate Program, University of Washington) for constructing the csd7 BACTH plasmid strains, and Laura Borkenhagen, Allison Chistiansen, April Dennison, Nathan Ferrey, Chloe Hanson, Laura Lhotka, and Aubrey Thyen (University of Minnesota, Morris) for csd2 mutant construction.
This work was supported by the National Institutes of Health under award numbers R01 AI094839, R01 AI136936 (to N.R.S.), and T32 CA009657 (to K.M.B. and D.C.Y.), National Science Foundation Graduate Research Fellowship grants DGE-1256082 (to K.M.B. and J.A.T.) and DGE-0718124 (to J.A.T.), a Department of Defense (DoD) National Defense Science & Engineering Graduate Fellowship (to J.A.T.), an American Cancer Society–Lakeshore Division postdoctoral fellowship (PF-13-391-01-MPC to D.C.Y.), and the Proteomics and Genomics Shared Resources of NIH/NCI Cancer Center support grant P30 CA015704. W.V. received support from the Wellcome Trust (101824/Z/13/Z). T.J.W. received support from the University of Minnesota, Morris Faculty Research Enhancement Funds (FREF), the University of Minnesota Undergraduate Research Opportunities Program (UROP), and the Howard Hughes Medical Institute Precollege and Undergraduate Science Education Program award to UMM.
The funders had no role in study design, data collection, or interpretation.
Conceptualization of the study was by D.C.Y. and N.R.S.; methodology was by T.W.P. and W.V.; investigation was by D.C.Y., K.M.B., J.A.T., T.S., C.M.T., C.K.L., A.L.C., T.J.W., and J.B. D.C.Y., K.M.B., and N.R.S. wrote the original draft of the manuscript.
FOOTNOTES
- Received 21 November 2018.
- Accepted 26 April 2019.
- Accepted manuscript posted online 29 April 2019.
Supplemental material for this article may be found at https://doi.org/10.1128/JB.00724-18.
- Copyright © 2019 American Society for Microbiology.
