Control of the flagellation pattern in Helicobacter pylori by FlhF and FlhG

ABSTRACT

FlhF and FlhG control the location and number of flagella, respectively, in many polar-flagellated bacteria. The roles of FlhF and FlhG are not well characterized in bacteria that have multiple polar flagella, such as Helicobacter pylori. Deleting flhG in H. pylori shifted the flagellation pattern where most cells had approximately four flagella to a wider and more even distribution in flagellar number. As reported in other bacteria, deleting flhF in H. pylori resulted in reduced motility, hypoflagellation, and the improper localization of flagella to nonpolar sites. Motile variants of H. pylori ∆flhF mutants that had a higher proportion of flagella localizing correctly to the cell pole were isolated, but we were unable to identify the genetic determinants responsible for the increased localization of flagella to the cell pole. One motile variant though produced more flagella than the ΔflhF parental strain, which apparently resulted from a missense mutation in fliF (encodes the MS ring protein), which changed Asn-255 to aspartate. Recombinant FliF^N255D, but not recombinant wild-type FliF, formed ordered ring-like assemblies in vitro that were ~50 nm wide and displayed the MS ring architecture. We infer from these findings that the FliF^N225D variant forms the MS ring more effectively in vivo in the absence of FlhF than wild-type FliF.

IMPORTANCE

Helicobacter pylori colonizes the human stomach where it can cause a variety of diseases, including peptic ulcer disease and gastric cancer. H. pylori uses flagella for motility, which is required for host colonization. FlhG and FlhF control the flagellation patterns in many bacteria. We found that in H. pylori, FlhG ensures that cells have approximately equal number of flagella and FlhF is needed for flagellum assembly and localization. FlhF is proposed to facilitate the assembly of FliF into the MS ring, which is one of the earliest structures formed in flagellum assembly. We identified a FliF variant that assembles the MS ring in the absence of FlhF, which supports the proposed role of FlhF in facilitating MS ring assembly.

INTRODUCTION

The bacterial flagellum is a complex nanomachine that is organized into three basic structures: the basal body, hook, and filament (1 - 3). The basal body anchors the flagellum in the cell envelope and includes a rotary motor and a flagellar type III secretion system (T3SS) that transports axial components of the flagellum (e.g., rod, hook, and filament subunits) across the cell membrane. The core bacterial motor consists of the C ring (rotor and switch complex), MS ring (rotor base plate), rod (driveshaft), P and L rings (bushing), and stator units (torque generator) (1 - 3). The stator unit consists of five subunits of MotA and two subunits of MotB (4, 5). MotB has a transmembrane region and a C-terminal periplasmic region that anchors the stator to the peptidoglycan layer through interactions with the P ring protein (FlgI) and the peptidoglycan layer (6, 7). MotA is an integral membrane protein that forms a proton channel with MotB through which protons flow to rotate the MotA pentamer (4, 5). Engagement of MotA with the C ring protein FliG results in the rotation of the C ring/MS ring complex, and the resulting torque is transmitted through the rod and hook to the filament, which acts as a propeller (8, 9). In addition to these core structures, the H. pylori motor has additional features that are not present in the archetypical Escherichia coli and Salmonella enterica motors (10).

The MS ring is formed by FliF and is one of the earliest flagellar structures assembled (11). FliF is embedded in the inner membrane by two transmembrane helices that flank a large periplasmic region that contains the ring-building motif (RBM) domains RBM1, RMB2, RBM3a, and RBM3b, plus the β-collar domain (12 - 15). The MS ring interfaces with the C ring through interactions between the C-terminal, cytoplasmic domain of FliF, and the N-terminal domain of the C ring protein FliG (FliG_N) (16, 17). The MS ring also houses the export gate of the flagellar T3SS. S. enterica FliF forms an inner ring with 23-fold symmetry that interacts with the export gate of the T3SS and an outer ring with 34-fold symmetry that matches the symmetry of the C ring. In E. coli, the flagellar T3SS protein FlhA initially assembles in the membrane and subsequently recruits FliF, which is further aided by FliG (18). FliF self-assembles in S. enterica to form MS ring structures, although efficient MS ring formation is dependent on FliG (19). Moreover, S. enterica FliF proteins oligomerize in vitro to form ring structures (20, 21).

The human gastric pathogen Helicobacter pylori uses a cluster of polar flagella for motility, which is required for colonization in animal models for infection (22, 23). H. pylori possesses FlhF and FlhG, which control the flagellation patterns in several bacterial species with polar flagella, as well as in Bacillus subtilis, which has a peritrichous arrangement of flagella (24). Genetic, biochemical, and structural studies of FlhF and FlhG have provided valuable information on how these proteins control flagellation patterns, but the mechanisms these proteins use to affect the location and number of flagella remain obscure.

FlhF belongs to the GTP-binding signal recognition particle family, which includes Ffh and FtsY, proteins required for targeting various secretory and membrane proteins to the cell membrane (25). FlhF is a GTPase that cycles between a GTP-bound form that facilitates flagellum assembly at the cell pole and an inactive GDP-bound or apo-form (24). Green and co-workers reported Vibrio cholerae FlhF localizes to the cell pole where it promotes oligomerization of a FliF-GFP fusion into distinct fluorescent foci (26). Terashima and co-workers subsequently showed that FlhF is required for the efficient formation of the MS ring in marine Vibrio species (27). In many polar-flagellated bacteria, deleting flhF results in hypoflagellation and the mislocalization of flagella to nonpolar sites (26, 28 - 30). In Campylobacter jejuni, cells expressing FlhF variants with reduced GTPase activity display various aberrant flagellar phenotypes, including lack of flagella, multiple flagella at one cell pole, flagella at nonpolar sites, and significantly shorter flagella (28).

FlhG belongs to the MinD/ParA ATPase family, members of which act in cell division site determination and plasmid/chromosome partitioning (31). Deleting flhG in bacteria that have a single polar flagellum often results in hyperflagellation (32 - 36). The N-terminal region of B. subtilis FlhG interacts with FlhF to stimulate its GTPase activity, and a conserved glutamine residue in FlhG is required for the stimulation (37). Substituting the conserved glutamine with alanine in C. jejuni FlhG results in hyperflagellation (38), suggesting that FlhG controls flagella number by modulating FlhF GTPase activity. Upon binding ATP, Geobacillus thermodenitrificans FlhG forms a dimer, and an amphipathic helix at the C-terminus, known as the membrane targeting sequence, is exposed to allow FlhG to interact with the inner membrane (39). After hydrolyzing ATP, FlhG undergoes a conformation change that converts the protein to a monomer and buries the membrane-targeting sequence in the protein (24, 39).

Most studies on FlhF and FlhG have been done in bacteria that have a single polar flagellum, and little is known about the function of these proteins in bacteria with a lophotrichous arrangement of flagella. To address this gap in our knowledge, we deleted flhF and flhG in a H. pylori G27-derived strain with enhanced motility (referred to as strain G27M) and examined the motility and flagellation phenotypes of the resulting mutants. In contrast to ΔflhG mutants in bacteria that have a single polar flagellum, the motility of the H. pylori G27M ΔflhG mutant in a soft agar medium was unimpaired, although the ΔflhG mutant had a greater proportion of cells that were hypoflagellated (i.e., aflagellated or possessed a single flagellum) compared to the G27M parental strain. Similar to reports on flhF mutants in other bacteria (26, 28 - 30), deleting flhF in H. pylori strains G27M and B128 resulted in reduced motility, hypoflagellation, and mislocalization of flagella to nonpolar sites. Variants of the H. pylori G27M ∆flhF mutant with enhanced motility had a significantly higher proportion of flagella that localized correctly to the cell pole or produced more flagella compared to the parental ∆flhF strain. A fliF allele where Asn-255 was switched to aspartate was identified in one of the motile variants, and introducing the fliF allele into the H. pylori G27M ∆flhF mutant resulted in enhanced motility and an increased number of flagella. In contrast to the wild-type H. pylori FliF, recombinant FliF^N255D formed ordered ring-like assemblies in vitro that displayed the MS ring architecture. Taken together, these findings suggest that the FliF^N255D variant is more proficient than wild-type FliF in forming the MS ring in the absence of FlhF.

RESULTS

Deletion of flhG in H. pylori G27M influences the number of flagella per cell

H. pylori G27M is an isolate of strain G27 with enhanced motility in a soft agar medium that was isolated in our laboratory. H. pylori G27M has a nonsense mutation at codon 78 of fliL, which encodes a protein associated with the MotA/MotB stator units. As expected, the in situ structure of the flagellar motor of H. pylori G27M determined by cryo-electron tomography and subtomogram averaging indicates the motor lacks the FliL ring (40).

In contrast to reports on disrupting flhG in other polar flagellated bacteria (32 - 36), deletion of flhG in H. pylori G27M did not significantly affect the motility of the strain in a soft agar medium (Fig. 1A). Examination of the ΔflhG mutant by transmission electron microscopy (TEM), however, showed that the flagellation pattern of the mutant was altered. A histogram of the number of flagella per cell for H. pylori G27M was symmetrical and unimodal (Fig. 1B). Four flagella per cell were both the mode and mean for H. pylori G27M, and >68% of the cells (82%) were within one standard deviation of the mean, consistent with a Gaussian distribution in the number of flagella per cell. In contrast, cells of the ∆flhG mutant did not display a Gaussian distribution in flagellar number but instead had a wider and more even distribution in the number of flagella per cell (Fig. 1B). In contrast to the hyperflagellated phenotype observed for flhG mutants in bacteria that normally have a single polar flagellum (32 - 36), only a small fraction of the ∆flhG mutant cells were hyperflagellated (≥8 flagella per cell), although we occasionally observed cells with as many as 12 flagella. Compared to the parental G27M strain, there were a greater proportion of cells with no flagella or a single flagellum in the ∆flhG mutant population (18% for the ΔflhG mutant versus 1% for wild type; Fig. 1B).

Fig 1.

Motility and flagellation phenotypes of the H. pylori G27M ∆flhG mutant. (A) The motility of the H. pylori G27M wild type (WT) and ∆flhG mutant was assessed in a soft agar medium by measuring the diameters of swim halos 7 d post-inoculation. Bars indicate mean values for the swim halo diameter (n = 8 for wild type; n = 9 for ∆flhG mutant). The motility of the ∆flhG mutant did not differ significantly from that of the wild type. Error bars indicate sample standard deviation. Statistical analysis of the data was done using a two-sample t-test. (B) Comparison of the motility of H. pylori G27M and ΔflhG mutant in a soft agar medium. Swim halos are smaller than those measured for the data presented in panel A due to inoculating multiple strains on the same plate. (C) Flagella were counted for 100 cells of wild type and ΔflhG mutant. Data for wild type are shown with black bars, and data for ∆flhG mutant are indicated with gray bars. One of the ∆flhG mutant cells that was counted had 12 flagella, which is not shown in the histogram. For the wild-type flagella counts, the mode, mean, and sample standard deviation were 4, 4.0, and 1.2, respectively. The mode, mean, and sample standard deviation for the ∆flhG mutant flagella counts were 5, 3.5, and 2.1, respectively. The distribution of the number of flagella per cell for the ∆flhG mutant differed from that of the wild type (P value = 0.035). Statistical significance was determined using a Mann-Whitney U test. (D) TEM images of flagellated H. pylori G27M (WT) and ΔflhG mutant cells. The scale bars correspond to 1 µm.

Deletion of H. pylori flhF results in hypoflagellation and mislocalization of flagella to nonpolar sites

Similar to the results from previous studies with mono-flagellated bacteria, deleting flhF in H. pylori G27M resulted in reduced motility (Fig. 2A), a significant reduction in the number of flagella per cell (Fig. 2B) and the improper localization of flagella to nonpolar sites (Fig. 2C). Approximately 30% of the H. pylori G27M ∆flhF mutant cells lacked flagella, and ~45% of the cells had a single flagellum, whereas the vast majority of the cells of the G27M parental strain possessed multiple flagella (Fig. 2B). Since G27M flagellar motor lacks the FliL_C ring, we wished to confirm that the phenotype of the ΔflhF mutant was not related to the absence of the FliL_C ring. To address this issue, we deleted flhF in H. pylori B128, which expresses full-length FliL and has an intact FliL_C ring in its motor. Deletion of flhF in H. pylori B128 had a profound effect on flagellation as ~48% of the B128 ΔflhF mutant cells lacked flagella and ~42% of the cells had a single flagellum, while the vast majority of the B128 parental cells possessed multiple flagella.

Fig 2.

Motility and flagellation phenotypes of H. pylori G27M ΔflhF mutant. (A)The motility of the H. pylori G27M wild type (WT) and ∆flhF mutant was assessed in a soft agar medium by measuring the diameters of swim halos 7 d post-inoculation. Bars indicate mean values for the swim halo diameter (n = 8 for wild type; n = 9 for ∆flhF mutant). The motility of the ∆flhF mutant differed significantly from that of the wild type (P value <0.00001). Error bars indicate sample standard deviation. Statistical analysis of the data was done using a two-sample t-test. (B) Comparison of H. pylori G27M and ΔflhF mutant in a soft agar medium. Swim halos are smaller than those measured for the data presented in panel A due to inoculating multiple strains on the same plate. (C) Flagella were counted for 100 cells of wild type and ΔflhF mutant. Data for wild type and ΔflhF mutant are shown with black bars and gray bars, respectively. The number of flagella per cell for the ∆flhF mutant differed from the wild type (P value <0.00001). Statistical significance was determined using a Mann-Whitney U test. (D) TEM images of H. pylori G27M (WT) with polar flagella and ∆flhF mutant cells with lateral flagella indicated by arrows.

We rarely observed flagella at nonpolar sites in the H. pylori G27M or H. pylori B128 parental strains. In contrast, ~37% of the flagella were located on the side of the cell (lateral location), and ~25% of the flagella had a subpolar location (which we defined as the curved region located between the apex of the cell pole and the side of the cell; see Fig. S1) in the G27M ΔflhF mutant, and ~53% of the flagella located laterally and ~25% had a subpolar position in the B128 ΔflhF mutant (Table 1).

TABLE 1.

Distribution of flagella at various sites in H. pylori ∆flhF mutants and isolated variants of the strain with enhanced motility

	Cellular locationa
Strain	Polar (%)	Subpolar (%)	Lateral (%)	P valueb
H. pylori G27M (wild type)	100	0	0
G27M ∆flhF parent (KG43)	37	36	27
G27MV1 (KG43 motile variant)	80	12	8	<0.00001
G27MV2 (KG43 motile variant)	74	12	14	0.0002
G27MV3 (KG3 motile variant)	58	11	30	0.156
Wild-type B128	100	0	0
B128 ΔflhF parent (KG61)	22	25	53
B128MV1 (KG61 motile variant)	33	18	48	0.390
B128MV3 (KG61 motile variant)	34	15	51	0.478
B128MV6 (KG61 motile variant)	35	7	57	0.802

^a Polar flagella were defined as ones at the apex of the cell pole, subpolar flagellar were defined as ones in the curved region between the apex and side of the cell, and lateral flagella were defined as ones on the side of the cell.

^b Statistical significance determines using a Mann-Whitney U test.

Isolation of H. pylori ∆flhF variants with enhanced motility in a soft agar medium

To identify mutations that bypass the requirement of FlhF for flagellum assembly or localization, we enriched for variants of the G27M and B128 ∆flhF mutants with enhanced motility by repeated passage of the mutants on a soft agar medium. Clonal isolates were obtained from three independent enrichments of the G27M ΔflhF mutant (designated as G27MV1 through G27MV3) and three independent enrichments of the B128 ΔflhF mutant (total of nine isolates; designated as B128MV1 through B128MV9). Motilities of G27MV1, G27MV2, and G27MV3 in a soft agar medium were significantly improved compared to the ∆flhF parental strain and approached wild-type motility (Fig. 3A). Although the swim halo diameters of the B128 ΔflhF mutant increased from 6–10 mm to 18–21 mm following multiple passages on a soft agar medium, only four clonal isolates from the motility enrichment produced swim halos that were significantly larger than that of the ΔflhF parental strain (Fig. 3B), which suggested that the suppressors that rescued motility in many of the B128 ΔflhF isolates were not stably maintained during the segregation and propagation of the motile variant clonal isolates.

Fig 3.

Motility phenotypes for the motile variants of the H. pylori G27M and B128 ∆flhF mutants. (A) Motilities of H. pylori G27M (WT), original G27M ΔflhF mutant (ΔflhF), and G27M ∆flhF motile variants (G27MV1, G27MV2, and G27MV3) in a soft agar medium. Mean values for the swim halo diameter with 3–9 replicates are indicated. Error bars indicate sample standard deviation. Asterisk (*) indicates swim halo diameters that were significantly greater than that of the original ∆flhF mutant (P value <0.00001). Statistical analysis of the data was done using a two-sample t-test. (B) Comparison of H. pylori G27M, ΔflhF mutant, and motile variant G27MV1 in a soft agar medium. Swim halos are smaller than those measured for the data presented in panel A due to inoculating multiple strains on the same plate. (C) Motilities of H. pylori B128 (WT), original B128 ΔflhF mutant (ΔflhF), and B128 ∆flhF motile variants (B128MV1 through B128MV9) in a soft agar medium. Three replicates were done for each strain. Asterisk (*) indicates swim halo diameters that were significantly greater than that of the original ∆flhF strain (P value <0.002). Statistical analysis of the data was done using a two-sample t-test. (D) Flagella counts for G27M, G27M ΔflhF mutant, and G27M ΔflhF motile variants. The number of flagella per cell was counted for 100 cells for each strain. The number of flagella per cell for G27MV3 differed from that of the ∆flhF parental strain (P value <0.0005), while the number of flagella per cell for G27MV1 and G27MV2 did not differ significantly (ns) from that of the ∆flhF parental strain. (E) Flagella counts for B128 ΔflhF parental strain and selected B128 ΔflhF motile variants. The number of flagella per cell was counted for >100 cells for each strain. The number of flagella per cell for the motile variants did not differ significantly (ns) from that of the ∆flhF parental strain. Statistical analysis of flagella numbers was determined using a Mann-Whitney U test.

Examination of the G27M ΔflhF motile variants by TEM revealed that the flagella number for G27MV1 and G27MV2 did not differ significantly from that of the original H. pylori G27M ∆flhF mutant (Fig. 3C). By contrast, the motile variant G27MV3 had more flagella per cell than the parental strain, having a much lower proportion of aflagellated cells and a higher proportion of cells with multiple flagella (Fig. 3D). G27MV1 and G27MV2 localized a significantly higher proportion of flagella correctly to the cell pole compared to the parental strain, with 70%–80% of the flagella positioned at the cell pole in the motile variants versus ~37% for the parental strain (Table 1). The higher proportion of flagella localizing to the cell pole in G27MV1 and G27MV2 presumably accounted for the robust motility of these strains in a soft agar medium (Fig. 3A). The proportion of flagella that localized to the cell pole appeared to have increased in G27MV3 (from ~37% in the parental strain to ~58% in G27MV3), but the difference was not statistically significant. The combination of increased flagella number (Fig. 3D) and the slight increase in the proportion of flagella that localized to the cell pole in G27MV3 (Table 1) presumably accounted for the robust motility of the motile variant in a soft agar medium (Fig. 3A).

Flagella numbers for three isolates from the motility enrichment with the B128 ΔflhF mutant that were more motile than the parental strain (B128MV1, B128MV3, and B128MV6) did not differ significantly from that of the parental strain (Fig. 3D). The percentage of flagella that localized correctly to the cell pole in the motile variants increased about 1.5-fold (from ~22% to ~33%; Table 1), but the increase was not statistically significant. The slight increase in the proportion of flagella that localize to the cell pole in B128 ΔflhF motile variants may have been sufficient to support the enhanced motility of the isolates in a soft agar medium (Fig. 3C). Alternatively, the suppressor mutations responsible for enhanced localization of flagella to the cell pole in the B128 ΔflhF isolates may not have been maintained during the propagation of the isolates for TEM examination.

Whole-genome sequencing of the three G27M ΔflhF motile variants and five B128 ΔflhF variants identified several single-nucleotide mutations and insertion/deletions within coding sequences and intergenic regions (Tables 2 and 3; Tables S1 and S2). Interestingly, all eight isolates had a frameshift mutation in faaA (locus tags HPG27_RS02965 and CV725_RS05915 in H. pylori strains G27 and B128, respectively; Table 2), which encodes a type V secretion system (T5SS) that localizes to the flagellar sheath (41). Although all of the variants examined had a frameshift mutation in faaA, replacing the wild-type faaA allele in the H. pylori G27M ΔflhF mutant with the faaA allele from G27MV1 and G27MV2 (designated as strain JB08) had a modest impact on motility and no impact on the flagellar number or localization of flagella (Fig. 4).

TABLE 2.

Relevant mutations in motile variants of G27M ΔflhF and B128 ΔflhF

Isolate	Gene description	Locus taga	Sequenceb	Impactc	Freq (%)d
G27MV1	faaA	HPG27_RS02965 (HP0609/HP0610)	Δ1 bp (5028/9579 nt)	L1676fs	98.8
G27MV2	faaA	HPG27_RS02965 (HP0609/HP0610)	Δ1 bp (5028/9579 nt)	L1676fs	99.6
G27MV3	faaA	HPG27_RS02965 (HP0609/HP0610)	(G)7→6 (7908/9579 nt)	G2636fs	99
	fliF	HPG27_RS01760 (HP0351)	AAT→GAT	N255D	99.5
G27MV4	faaA	HPG27_RS02965	(A)5→4 coding (3756/9579 nt)	K122fs	94.8
G27MV5	faaA	HPG27_RS02965	(A)5→4 coding (3756/9579 nt)	K1252fs	99.1
B128MV1	faaA	CV725_RS05915 (HP0609/HP0610)	(C)7→8 (502/9561 nt)	G168fs	90
B128MV2	faaA	CV725_RS05915 (HP0609/HP0610)	(C)7→8 (502/9561 nt)	G168fs	95.8
B128MV3	faaA	CV725_RS05915 (HP0609/HP0610)	(C)7→8 (502/9561 nt)	G168fs	95.7
B128MV4	faaA	CV725_RS05915 (HP0609/HP0610)	(C)7→8 (502/9561 nt)	G168fs	96.7
B128MV5	faaA	CV725_RS05915 (HP0609/HP0610)	(C)7→8 (502/9561 nt)	G168fs	95.7

^a Locus tag for homolog in H. pylori 26695 is indicated in parentheses.

^b The numbers in parentheses indicate the position of the mutation (first number) within the entire length of the open reading frame (second number).

^c Indicates the amino acid position where a different amino acid was introduced, where a stop codon was introduced (*), or site where a frameshift mutation occurred (fs).

^d Indicates the percentage of reads at the position that had the particular single-nucleotide polymorphism (SNP).

TABLE 3.

Distribution of flagella at various sites in H. pylori strain JB06 (G27M ΔflhF fliF255) and variants of JB06 with enhanced motility

	Cellular location
Strain	Polar (%)	Subpolar (%)	Lateral (%)	P valuea
G27M ΔflhF parent (KG43)	37	36	27
JB06 (KG43 fliF255)	39	21	39	0.43
G27MV4 (JB06 motile variant)	58	15	27	0.021
G27MV5 (JB06 motile variant)	56	19	25	0.011

^a Statistical significance determines using a Mann-Whitney U test.

Fig 4.

Flagellation patterns of H. pylori G27M ΔflhF mutant containing mutant faaA alleles or fliF alleles. (A) Motilities in a soft agar medium of H. pylori G27M (WT), G27M ΔflhF mutant (KG43), G27M ΔflhF mutant in which wild-type faaA allele was replaced with faaA5028 (JB08), and G27M ΔflhF mutant in which wild-type fliF allele was replaced with fliF255 (JB06). Bars show mean values for the swim halo diameter (n = 3 or 4 for each strain). Error bars indicate sample standard deviation. Asterisks indicate swim halo diameters that were significantly greater than that of the parental ∆flhF strain (*P value <0.05; ***P value <0.0005). Statistical analysis of the data was done using a two-sample t-test. (B) Flagella counts for H. pylori G27M, KG43, JB08, and JB06. The number of flagella per cell was counted for ≥91 cells for each strain. The distribution of the number of flagella per cell for JB08 and JB06 did not differ significantly (ns) from that of the ∆flhF parental strain (KG43). All of the strains had fewer flagella per cell than wild-type G27M (****P value <0.00001). (C) When only the flagellated cells were considered, the average number of flagella per cell for JB06 was significantly higher than that of KG43 but less than that for G27M (****P value <0.00001). Statistical significance was determined using a Mann-Whitney U test.

A fliF allele rescues motility in the H. pylori G27M ∆flhF mutant

G27MV3 contained a mutation in fliF that changed Asn-255 to Asp in the coding sequence (Table 2). The wild-type fliF allele in the H. pylori G27M ΔflhF mutant was replaced with the fliF allele from G27MV3 (designated allele as fliF255) to generate strain JB06. JB06 was significantly more motile in a soft agar medium compared to the ΔflhF parental strain (Fig. 4A). JB06 and the ΔflhF parental strain did not differ significantly with regard to distribution of the number of flagella per cell (Fig. 4B). When only the flagellated cells of the two strains were compared, however, JB06 had significantly more flagella per cell than the ΔflhF parental strain (Fig. 4C). The fliF255 allele did not influence the localization of flagella in the G27M ΔflhF mutant (Table 3; P-value = 0.43).

Following the procedure for isolating the G27M ΔflhF motile variants, two variants of JB06 with enhanced motility in a soft agar medium (designated G27MV4 and G27MV5) were isolated from the same enrichment. Compared to the JB06 parental strain, both G27MV4 and G27MV5 had a significantly higher proportion of flagella that localized to the cell pole (Table 4). The number of flagella per cell for G27MV4 and G27MV5 though did not differ significantly from that of the JB06 parental strain (P values = 0.337 and 0.535, respectively). Whole-genome sequencing of G27MV4 and G27MV5 revealed that the isolates had almost identical genotypes (Tables S1 and S2), suggesting the isolates were siblings. As with all the previous ΔflhF motile variants, G27MV4 and G27MV5 had a frameshift mutation in faaA (Table 2).

TABLE 4.

Relevant H. pylori strains used in this study

Strain	Relevant characteristics	Source
H. pylori G27	Wild type	D.Scott Merrell
H. pylori G27M	Derived from G27; nonsense mutation in fliL (Glu-78 to stop), missense mutation in flgH (Gly-178 to Cys)	This study
H. pylori B128	Wild type	Richard M. Peek, Jr.
KG43	G27M ΔflhF	This study
KG47	G27M ΔflhG	This study
KG61	B128 ΔflhF	This study
JB06	KG43 fliF255	This study
JB08	KG43 faaA5028	This study
G27MV1	Motile variant of KG43	This study
G27MV2	Motile variant of KG43	This study
G27MV3	Motile variant of KG43	This study
G27MV4	Motile variant of JB06	This study
G27MV5	Motile variant of JB06	This study
B128MV1	Isolate from enrichment #1 for motile variants of KG61	This study
B128MV2	Isolate from enrichment #2 for motile variants of KG61	This study
B128MV3	Isolate from enrichment #3 for motile variants of KG61	This study
B128MV4	Isolate from enrichment #1 for motile variants of KG61	This study
B128MV5	Isolate from enrichment #2 for motile variants of KG61	This study
B128MV6	Isolate from enrichment #3 for motile variants of KG61	This study
B128MV7	Isolate from enrichment #1 for motile variants of KG61	This study
B128MV8	Isolate from enrichment #2 for motile variants of KG61	This study
B128MV9	Isolate from enrichment #3 for motile variants of KG61	This study

H. pylori FliF^N225D displays an increased tendency for in vitro self-assembly

The observation that the introduction of the fliF255 allele into the G27M ΔflhF mutant rescued motility and resulted in a higher proportion of cells with multiple flagella suggested FliF^N255D assembles the MS ring more effectively than wild-type FliF (FliF^WT) in the absence of FlhF. Based on the previously reported structure of the MS ring (13, 14, 20), Asn-255 is located in the middle of a helix in RBM3a, which forms the C-terminal globular domain together with RBM3b. When assembled into the full MS ring, Asn-255 is located on an interface that appears to interact with the alternative positioning of the central globular domain (RBM2) of adjacent subunits (Fig. 5A). Mariano and co-workers previously reported that RBM3 of S. enterica FliF adopts a native 34-mer state on its own, which is sufficient to promote the overall oligomerization of the protein (21). By contrast, no oligomerization was reported for the H. pylori or Vibrio FliF orthologs when purified in isolation (21, 27). This prompted us to hypothesize that in H. pylori FliF, the Asn-255 to Asp substitution increases the propensity of the protein to oligomerize in the absence of FlhF by promoting interface formation between RBM3 and RBM2.

Fig 5.

Purification and precipitation profiles of FliF^WT and FliF^N255D recombinant proteins. (A) Hypothetical location of the N255D mutation modeled onto a S. enterica FliF structure (PDB:6SD3) with a monomer of FliF on the far left and 34-mer of FliF in both side (middle) and top views (right). The mutation is in red, RBM3 domains in blue, and RBM2 domains in green. (B) Normalized UV traces from gel filtration chromatography of FliF^WT (blue) and FliF^N255D (red). (C) SDS-PAGE gel of the precipitation assay with both FliF^WT (WT) and FliF^N255D (M) at each concentration separated by pellet (P) and supernatant (S).

To examine this hypothesis, we compared the ability of the H. pylori FliF wild-type protein (FliF^WT) and FliF^N255D to oligomerize in vitro. Truncated forms of the two FliF proteins consisting of amino acid residues 52–427 and lacking the transmembrane helices were expressed and purified. Both proteins were soluble and eluted with a similar profile on the gel filtration column (Fig. 5B) with a small peak at the void volume (9–10 mL), a shoulder corresponding to proteins at ~700 kDa (10–15 mL), and a large peak corresponding to proteins at ~45 kDa (17 mL), which was very close to the predicted molecular weight of a single monomer (42 kDa). We noted that the solubility of the proteins varied significantly between FliF^WT and FlilF^N255D. Specifically, FliF^WT is soluble to a concentration >10 mg/mL, whereas FliF^N255D precipitates at ~2 mg/mL (Fig. 5C). This suggested that FliF^N255D might oligomerize at high concentrations, although it could also indicate reduced stability induced by the Asn-255 to Asp substitution. To investigate this further, we used negative-stain electron microscopy to characterize the void volume fractions of the FliF^WT and FliF^N255D preparations. As shown in Fig. 6, FliF^WT consisted mainly of aggregated protein. In contrast, for FliF^N255D, we observed the presence of ordered ring-like assemblies that were ~50 nm wide and displayed the clear MS ring architecture (Fig. 6B and 2), consistent with the hypothesis that FliF^N255D variant assembles the MS ring in the absence of FlhF. Although the results from the size exclusion chromatography indicated that even in the context of the Asn-255 to Asp substitution the majority of the FliF protein remains monomeric, the fact that we observe some ordered ring-like assemblies with the FliF^N255D variant, but not with FliF^WT, is consistent with our hypothesis and our interpretation of this rescue mutation.

Fig 6.

Negative-stain images of FliF^WT and FliF^N255D recombinant proteins. (A) Negative-stain image of FliF^WT at ×40,000 magnification showing aggregate. (B) Negative-stain images of FliF^N255D at ×25,000 (left) and ×40,000 (right) magnifications showing spontaneously oligomerized complexes of about 30 nm (300 Å) in width. Scale bars are equal to 200 nm. Additional images of FliF^N255D recombinant protein are shown in Fig. S2.

DISCUSSION

How bacteria control their flagellation patterns is a fascinating but poorly understood area. Studies on FlhF and FlhG from various bacterial species have provided valuable insights into structure-function relationships in these proteins (24, 26, 28, 32, 38, 39), but elucidating the mechanisms by which FlhF and FlhG control the location and number of flagella has remained elusive. In examining the roles of FlhF and FlhG in regulating the flagellation pattern in H. pylori, we found the phenotypes of H. pylori flhF and flhG mutants to be both similar and different from those reported for bacteria that have a single polar flagellum.

In contrast to the hyperflagellation and reduced motility phenotypes observed for flhG mutants in other bacteria that normally have a single flagellum at one or both cell poles (32 - 36), deleting flhG in H. pylori G27M did not result in a significant proportion of the cells being hyperflagellated (Fig. 1C), nor did it inhibit motility in a soft agar medium (Fig. 1A and B). The lack of a motility defect in a soft agar medium for the H. pylori G27M ΔflhG mutant was not particularly surprising since H. pylori has evolved to coordinate the rotation of multiple, polar flagella to swim effectively, while bacteria that normally have a single polar flagellum are not equipped to navigate with multiple flagella. The H. pylori G27M parental strain exhibited a Gaussian distribution for the number of flagella per cell, which was lost in the ∆flhG mutant (Fig. 1C). These findings indicate that FlhG controls the flagellation pattern in H. pylori so that most of the cells in the population have approximately the same number of flagella. Compared to the H. pylori G27M parental strain, a higher proportion of the ∆flhG mutant cells were either aflagellated or had one or two flagella (Fig. 1C). The high proportion of hypoflagellated cells for the ∆flhG mutant may have resulted from interference with C ring assembly as B. subtilis and Shewanella putrefaciens FlhG proteins are reported to promote assembly of the C ring in vitro (39). Since the C ring has a role in protein secretion by the flagellar T3SS (42, 43), inhibiting C ring formation in H. pylori would have likely interfered with flagellum assembly. Alternatively, the high proportion of hypoflagellated cells for the H. pylori ∆flhG mutant may be due to unregulated FlhF activity that results in the cells initiating assembly of more nascent flagella than can be completed with the levels of flagellar protein subunits synthesized. Insufficient levels of flagellar protein subunits might also explain the relatively low number of hyperflagellated cells observed for the H. pylori ∆flhG mutant.

FlhF generally functions as a positive factor for flagellar biosynthesis in polar-flagellated bacteria. Depending on the bacterial species, however, disrupting flhF results in a range of phenotypes, including decreased or complete loss of motility, abnormal flagella number, mislocalization of flagella, reduced flagellar gene expression, and decreased virulence (44). For example, the deletion of flhF in various Vibrio species results in the absence of flagella in virtually all of the cells in the population (34, 45, 46). In contrast, Pseudomonas putida and Pseudomonas aeruginosa flhF mutants continue to produce a single flagellum, but the flagellum incorrectly localizes to a nonpolar site (29, 30). In the peritrichous bacterium B. subtilis, loss of flhF results in the positioning of flagella more closely to the cell poles (47). Deleting flhF in C. jejuni abolishes motility and results in the vast majority of cells being aflagellated (28). Missense mutations in C. jejuni flhF that interfere with the GTPase activity of FlhF, however, result in a range of flagellation phenotypes, including increased frequency of aflagellation, polar hyperflagellation, and lateral flagella, indicating that FlhF GTPase activity is required for normal flagellation in C. jejuni (28). Expression of σ⁵⁴-dependent flagellar genes is dramatically inhibited in the C. jejuni ΔflhF mutant but not in the strains harboring the flhF missense mutations, which accounts at least partially for the differences in flagellar biosynthesis between the ΔflhF and flhF missense mutants (28). The phenotypes of the H. pylori G27M and H. pylori B128 ∆flhF mutants were similar to those reported for other bacteria, including reduced motility, hypoflagellation, and increased frequency of lateral flagella (Fig. 2 and 3). In contrast to the C. jejuni ΔflhF mutant though, most of the H. pylori ΔflhF mutant cells were flagellated (Fig. 2C, 3D, and E). There is considerable overlap between H. pylori and C. jejuni in the flagellar genes that are dependent on σ⁵⁴ for their expression and the mechanisms that control the transcription of these genes (48 - 54). The differences in the flagellation patterns of the H. pylori ΔflhF and C. jejuni ΔflhF mutants suggest that there is a significant distinction in how FlhF influences transcription of σ⁵⁴-dependent flagellar genes in the two bacterial species.

Although we isolated motile variants of the G27M ΔflhF mutant that had higher proportions of flagella that localized correctly to the cell pole compared to the parental strain (Tables 1 and 3), we failed to identify a genetic determinant responsible for suppressing the flagellar localization defect in the ΔflhF mutants. It is possible that the suppression of the flagellar localization defect in the ΔflhF mutants involved a rapid phenotypic switching. Epigenetic events, such as DNA methylation or gene silencing mediated by noncoding RNAs, are one means for accomplishing phenotypic switching (55). Alternatively, phenotypic switching can be achieved through phase-variation-associated genetic modifications, including site-specific recombination, inversion of promoter elements, and slippage of DNA polymerase during replication in homopolymeric DNA tracts (55). Phase variation resulting from changes in homopolymeric tract length is common in H. pylori, and when they occur within coding regions they generate frameshift mutations that result in expression of truncated and often nonfunctional proteins (56 - 58). Alternatively, changes in homopolymeric tract length within intergenic regions can alter transcript levels to fine-tune protein expression, which was shown to be the case for H. pylori sabA (59). Since all the ΔflhF motile variants that we examined had a frameshift mutation in faaA, it seems likely that disrupting faaA has a role in suppressing the flagellar localization defect. It may be that disruption of faaA acted in concert with a phase variable mutation to suppress the flagellar localization defect.

While we did not identify genetic determinants that suppressed the flagellar localization defect in the ΔflhF mutants, we did find that the FliF^N255D variant partially suppressed the motility and reduced flagellation defects of the G27M ΔflhF mutant. FliF contains the RBMs that are found in ring-forming proteins of other protein secretion systems (12). RBM3 is formed by two stretches of sequence denoted as RBM3a and RBM3b, which are separated by a stretch of β-strands that form a structure referred to as the β-collar that constitutes the driveshaft housing (20). Asn-255 of H. pylori FliF corresponds to Ala-251 of S. enterica FliF, which is located in an α-helix within RBM3a. In the S. enterica MS ring, RBM3 forms a ring with 34-fold symmetry, and Ala-251 is located at the interface between RBM3s of adjacent FliF monomers, as well as the occasional conformation of the RBM2 present in the full 34-mer (20) (Fig. 6A). The region near Ala-251 is important for S. enterica FliF function as alanine substitutions at Ile-252 and Leu-253 result in reduced numbers of flagella and lower levels of hook-basal body complexes compared to wild type (60). Examination of purified MS rings from S. enterica strains expressing the FliF^I252A or FliF^L253A variants by electron microscopy revealed that most of the MS rings were not completely closed, indicating that the substitutions in FliF destabilized the MS ring (60).

The predicted location of Asn-255 in H. pylori MS ring suggests the aspartate substitution at this position renders the complex more stable, which may account for the increased number of flagella in the ΔflhF motile variant G27MV3. This prompted us to hypothesize that purified FliF^N255D would be more prone to oligomerization than FliF^WT, which presumably requires FlhF to assemble as shown in Vibrio species (26, 27). Consistent with this hypothesis, recombinant FliF^N255D displayed an increased tendency to oligomerize compared to FliF^WT (Fig. 6; Fig. S2). This supports the observation of rescue in motility in the G27MV3 motile variant and indicates that Asn-255 has a key role in inducing FliF oligomerization to initiate flagellum formation. It also suggests a mechanism for the role of FlhF in promoting the polar localization of the flagellum, by suggesting that this may be at least in part mediated by its inducing of FliF oligomerization. However, further experimental characterization will be required to demonstrate this. It remains to be seen if Asn-255 prevents FliF oligomerization, which is lifted in the presence of FlhF, or if the N255D mutation is a gain-of-function mutation that induces excessive FliF oligomerization to compensate for the inhibitory effect resulting from the loss of FlhF.

MATERIALS AND METHODS

Bacterial strains and growth conditions

A list of relevant H. pylori strains generated and used in the study is indicated in Table 4. E. coli strains DH5α and NEB Turbo (New England Biolabs) were used for cloning and plasmid construction. E. coli strain BL21 (New England Biolabs) was used for overexpression of the recombinant H. pylori FliF^N255D and FliF^WT. E. coli strains were grown in lysogeny broth (LB) or LB agar medium supplemented with ampicillin (100 µg/mL) or kanamycin (30 µg/mL) when appropriate. H. pylori strains were grown under an atmosphere consisting of 10% CO₂, 4% O₂, and 86% N₂ at 37°C on tryptic soy agar (TSA) supplemented with 5% horse serum (TSA-HS). Liquid cultures of H. pylori were grown in brain heart infusion (BHI) broth supplemented with 5% heat-inactivated horse serum (BHI-HS) with shaking in serum bottles under an atmosphere consisting of 5% CO₂, 10% H₂, 10% O₂, and 75% N₂. H. pylori growth medium was supplemented with kanamycin (30 µg/mL) or sucrose (5%) where appropriate.

Construction of H. pylori ΔflhF and ΔflhG mutants

PCR primers and plasmids used in the construction of H. pylori mutants are listed in Tables S3 and S4. Genomic DNA from H. pylori G27M was purified using the Wizard Genomic DNA Purification Kit (Promega) and used as the PCR template to construct the deletion mutants. Primers 58 and 59 were used to amplify a 607-bp region corresponding to the sequence directly upstream of flhF. Primers 60 and 61 were used to amplify a 617-bp region corresponding to the sequence directly downstream of flhF. The resulting amplicons were joined by PCR SOEing and were incubated with Taq polymerase (Promega) at 72°C for 1 h to add 3′-A overhangs, which facilitated cloning of the amplicon into the pGEM-T Easy vector (Promega) to generate plasmid pKHG27. Plasmid pJC038 carries a cassette bearing a kanamycin resistance gene (kan^R) and Bacillus subtilis sacB under the control of the H. pylori ureA promoter (61). The kan^R-sacB cassette from pJC038 was introduced into unique NheI and XhoI sites in plasmid pKHG27 to generate suicide plasmid pKHG30. The suicide plasmid was introduced by natural transformation into H. pylori strains G27M and B128 to replace flhF with the kan^R-sacB cassette through homologous recombination and was screened by plating on TSA supplemented with horse serum and kanamycin. The presence of the kan^R-sacB cassette in the resulting transformants (KG38 and KG41) was verified by PCR. The kan^R-sacB cassette was removed by transforming strains KG38 and KG4 with plasmid pKHG27 and then using a sucrose-based counter-selection to enrich for the desired recombinants with an unmarked deletion of flhF (62). The genotypes of the resulting G27M ∆flhF mutant (KG43) and B128 ΔflhF mutant (KG61) were confirmed by PCR.

A similar method was used to generate an unmarked deletion of flhG. Primers 143 and 144 were used to amplify a 611-bp region corresponding to the upstream sequence of flhG. Primers 145 and 146 were used to amplify 720 bp corresponding to the downstream sequence of flhG. The resulting amplicons were joined by PCR SOEing. The resulting amplicon was incubated with Taq polymerase (Promega) and then ligated into pGEM-T Easy to generate pKHG31. Plasmid pKHG31 was cut at unique XhoI and NheI restriction sites and ligated with the kan-sacB cassette to generate plasmid pKHG33, which was introduced into H. pylori G27M generating KG46. Plasmid pKHG31 was introduced into KG46, and transformants were screened on TSA supplemented with sucrose to enrich for strains that had an unmarked deletion of flhG. The genotype of the resulting H. pylori G27M ∆flhG mutant (KG47) was confirmed by PCR.

Introduction of fliF and faaA alleles into H. pylori G27M ΔflhF

Primer pairs 150 and 151 and 152 and 153 were used to amplify a 513-bp region upstream of fliF and a 506-bp region downstream of fliF, respectively. The resulting amplicons were joined by PCR SOEing, and the joined fragment was cloned into pGEM-T Easy to generate plasmid pJMB10. The kan^R-sacB cassette from pJC038 was introduced into unique NheI and XhoI sites in plasmid pJMB10 to generate suicide plasmid pJMB11, which was introduced into KG43 to generate JB05. Primers 150 and 153 were used to amplify a 2,415-bp region from G27MV3 genomic DNA that included the fliF allele from this ΔflhF motile variant. The resulting amplicon was cloned into pGEM-T Easy to generate the suicide plasmid pJMB12, which was transformed into JB05 to replace the kan^R-sacB insertion in fliF with the fliF allele from G27MV3. PCR and sequencing of the resulting amplicon (Eton Biosciences) was done to confirm that the kan^R-sacB cassette had been replaced with the fliF allele from G27M3. The resulting strain was designated JB06.

Using H. pylori G27M genomic DNA as a template, primers 154 and 155 were used to amplify a portion of faaA corresponding to nucleotide positions 3,914 to 4,439, and primers 156 and 157 were used to amplify a portion of faaA corresponding to nucleotide positions 5,039 to 5,573. The primers introduced unique NheI and XhoI sites in the amplified region of faaA. The amplicons were joined by PCR SOEing, and the PCR product was cloned into pGEM-T Easy to generate plasmid pJMB13. The kan^R-sacB cassette from pJC038 was introduced into the NheI and XhoI sites in plasmid pJMB13 to generate suicide vector pJMB14, which was introduced into KG43 to create strain JB07. Using genomic DNA from G27MV1 as a template, primers 154 and 157 were used to amplify a 1,658-bp fragment of faaA that included the SNP identified in G27MV1 (faaA allele designated at faaA5028) and G27MV2. The resulting amplicon was cloned into pGEM-T Easy to generate the suicide vector pJMB15, which was transformed into JB07 to replace the kan^R-sacB insertion in faaA with the faaA allele from G27MV1. PCR and sequencing of the resulting amplicon was done to confirm that the kan^R-sacB cassette had been replaced with the faaA allele from G27M1. The resulting strain was designated JB08.

Motility assay

Motility was assessed using a soft agar medium consisting of Mueller-Hinton broth, 10% heat-inactivated horse serum, 20 mM 2-(N-morpholino)ethanesulfonic acid (MES; pH 6.0), and 0.4% noble agar. The soft agar medium was stab inoculated with H. pylori cells grown on TSA-HS. The inoculated plates were incubated for 7 d at 37°C in an atmospheric condition consisting of 10% CO₂, 4% O₂, and 86% N₂, after which the diameters of the swim halos emanating from the point of inoculation were measured. At least six replicates were done for each strain, mean values for the diameters of the swim halos were calculated, and a two-sample t-test was used to determine statistical significance.

Enrichment and isolation of motile variants of ∆flhF mutants

Motile variants of the H. pylori ∆flhF mutants were isolated by inoculating the mutant strains in a soft agar medium and allowing the cells to migrate from the point of inoculation as described for the motility assay. Cells were picked from the edge of the swim halos, inoculated into a soft agar medium, and allowed to migrate from the point of inoculation. The process was repeated five more times, at which point, cells from the edge of the swim halo were streaked onto TSA-HS to obtain single colonies. Clonal isolates were propagated on TSA-HS for one passage and then stored frozen.

Protein expression and purification

The plasmid for the expression and purification of FliF^52-427 has been reported previously (21). The N255D mutation was constructed by site-directed mutagenesis, and primers for this protocol are listed in Table S3. FliF^52-427 plasmid was amplified by PCR with divergent primers containing targeted nucleotide substitutions in the forward and reverse primers. The remaining original methylated DNA was digested by the DpnI restriction enzyme. QIAquick PCR Purification Kit (QIAGEN) was used to extract DNA for further transformation.

FliF^WT and FliF^N255D were purified as described previously (21). Briefly, plasmids were transformed into E. coli BL21 cells, and a single colony was used to inoculate 1 L of LB. Cultures were grown at 37°C to the early log phase, and expression of the FliF proteins was induced by adding 1 mM isopropyl β-D-1-thiogalactpyranoside (IPTG), which was carried overnight at 18°C. Cells were pelleted through centrifugation, resuspended in 50 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES), 500 mM NaCl, 20 mM imidazole, pH 8, at room temperature (~25°C), lysed by sonication on ice, and the supernatant was isolated through centrifugation at 14,000 × g for 45 min at 4°C. The protein was purified by Ni-NTA affinity chromatography at room temperature using HisPure Ni-NTA (ThermoScientific) resin and eluted with 50 mM HEPES, 500 mM NaCl, 500 mM imidazole, and pH 8.0. Fractions containing the eluted protein were pooled and concentrated using Vivaspin Turbo 30 kDa concentrator (Sartorius) and purified further using Superose 6 Increase 10/300 Gl (GE Healthcare) gel filtration column in 50 mM HEPES 500 mM NaCl pH 8.

Precipitation assay

The gel filtration fractions containing FliF^WT and FliF^N255D, respectively, were pooled and concentrated in steps to 0.5, 1, 2, 3, 4, and 5 mg/mL using Vivaspin Turbo 30 kDa concentrators (Sartorius) and centrifuged at 15,000 rpm for 2 min to separate the precipitate from the soluble fraction with the results visualized on a 12% SDS-PAGE gel.

Transmission electron microscopy

H. pylori cultures were grown to late-log phase (A₆₀₀ = ~1.0) in BHI-HS. Cells from the cultures were collected by centrifugation, fixed with formaldehyde and glutaraldehyde, and stained with uranyl acetate as described (61). Cells were visualized using a JEOL JEM 1011 transmission electron microscope operated at 80 kV. Flagella counts were determined for at least 100 cells for each strain. A Mann-Whitney U test was used to determine the statistical significance of differences in the number of flagella per cell or location of flagella. For visualization of FliF complexes, purified proteins were characterized by applying 5 µL of sample at 3 mg/mL onto Cu 300 mesh carbon-coated copper grids and stained with 0.75% uranyl acetate. The grids were imaged using Jem-1400 Flash Electron Microscope (TEM) equipped with a sCMOS Matakaki Flash camera at 40K magnification.

DNA sequencing and analysis

Genomic DNA from H. pylori strains was purified using the Wizard Genomic DNA Purification Kit (Promega). For some strains, genomic libraries were prepared with the Illumina iTruSeq adaptor kit from 500 ng of gDNA, and the libraries were sequenced at the University of Georgia Genomics Facility by Illumina sequencing. The quality of the reads was assessed with FastQC and trimmed using Trimmomatic (63, 64). Alternatively, gDNA preparations from H. pylori strains were submitted to the Seqcenter (Pittsburgh, PA) for genomic library preparation and Illumina sequencing. Reads for H. pylori genomic DNA sequences were mapped using the breseq computational pipeline (65) with the published NCBI genomes for H. pylori G27 (Accession no.: NC_011333.1) or H. pylori B128 (Accession no.: NZ_CP024951.1). The genomic DNA sequences of the H. pylori mutants were aligned with that of H. pylori G27 or H. pylori B128, which served as the backbones for aligning the gDNA sequences.

DATA AVAILABILITY

All genomic data are available upon request.

SUPPLEMENTAL MATERIAL

The following material is available online at https://doi.org/10.1128/jb.00110-23.

Tables S1 to S4; Fig. S1; Fig. S2. jb.00110-23-s0001.docx.

Table S1: Identified intragenic mutations; Table S2: Identified intergenic mutations; Table S3: Primers used in study; Table S4: Strains and plasmids used in study; Fig. S1: Examples of polar, subpolar, and lateral flagella; Fig. S2: Electron micrographs of FliF assemblies.

ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.