Expression of the Helicobacter pylori adhesin SabA is controlled via phase variation and the ArsRS signal transduction system

Adaptation to the acidic microenvironment, and adherence to mucosal epithelium, are essential for persistent colonization of the human stomach by Helicobacter pylori. The expression of SabA, an adhesin implicated in the ability of H. pylori to adhere to the host gastric epithelium, can be modulated by phase variation via slipped-strand mispairing in repetitive nucleotide tracts located in both the promoter region and the coding region. This study demonstrates the occurrence of phase variation at the sabA locus within individual strains of H. pylori, and among multiple isolates from a single patient. In addition, transcription of sabA is repressed by the acid-responsive ArsRS two-component signal transduction system in vitro. Our results demonstrate that isogenic inactivation of the arsS (jhp0151/HP0165) histidine kinase locus results in a 10-fold SabA-dependent increase in adherence to gastric epithelial cells in strain J99 (contains an in-frame sabA allele), but not in strain 26695 (out-of-frame sabA allele). The combination of transcriptional regulation of the sabA locus by the ArsRS two-component signal-transduction system and the generation of subpopulations harbouring alternate sabA alleles by slipped-strand mispairing during chromosomal replication could permit H. pylori to rapidly adapt to varying microenvironments or host immune responses. As a pathogen with a paucity of regulatory proteins, this dual regulation indicates that SabA expression is a tightly regulated process in H. pylori infection.


INTRODUCTION
Helicobacter pylori is a Gram-negative bacterium that infects more than half the world's population, and it colonizes the human gastric epithelium. Colonization generally occurs in childhood and, without treatment, persists for the lifetime of the host. As one of the most genetically diverse bacterial species, H. pylori has coevolved with human hosts, and generated populations that productively colonize a particular gastric niche. Although many colonized individuals remain asymptomatic, H. pylori is a major aetiological agent of peptic ulcer disease, and a recognized risk factor for gastric cancer (Blaser & Berg, 2001;Merrell & Falkow, 2004;Kusters et al., 2006).
Adherence to host cell receptors protects H. pylori from clearance during mucus shedding, and ensures consistent access to nutrients released by damaged gastric epithelial cells, facilitating long-term colonization, and potentially contributing to disease onset (Gerhard et al., 1999;Odenbreit, 2005;Aspholm et al., 2006). H. pylori has several well-characterized adhesins, including BabA, which binds to the Lewis B (Le b ) antigen (Boren et al., 1993;Ilver et al., 1998), and SabA, which binds to glycosphingolipids displaying a sialyl-dimeric Lewis X (sialyl-Le x ) antigen (Mahdavi et al., 2002).
Glycoconjugates bearing the sialyl-Le x antigen are rarely expressed in healthy gastric epithelial cells (Madrid et al., 1990), but they are upregulated during inflammation, and serve as binding sites for host cell adhesins of the selectin family (Alper, 2001). Accordingly, SabA-mediated adherence is positively correlated with sialyl-Le x concentration in vitro (Linden et al., 2004), and colonization density is increased in patients who produce high levels of sialyl-Le x , or are infected with SabA-positive strains of H. pylori (Sheu et al., 2006). H. pylori shows a tropism for areas of reduced acidity in the stomach that contain gastric pit cells producing sialyl-Le x (Bjorkholm & Salama, 2003), and studies have shown that clearance of infection reduces production of sialyl-Le x receptors to pre-infection levels (Mahdavi et al., 2002;Acheson & Luccioli, 2004;Roche et al., 2004).
Expression of bacterial adhesins can be regulated both by reacting to changes in the environment using signal transduction, and by generating genetic changes that affect production of functional proteins. It has been demonstrated that some H. pylori genes are regulated by phase variation (Saunders et al., 1998;Salaun et al., 2004), a mechanism by which genes can be expressed in an all-ornothing manner. Phase variation can control gene expression at the transcriptional and the translational levels, and in other organisms it has been shown to mediate evasion of the host immune response, or to modify virulence properties (van der Woude & Baumler, 2004). At the level of chromosomal replication, a molecular mechanism of phase variation known as slipped-strand mispairing can insert or delete nucleotides within repetitive DNA tracts, usually near the 59 end of genes. This results in altered reading frames, alternatively yielding truncated or full-length proteins (de Vries et al., 2002).
The sabA locus contains a homopolymeric thymine (poly-T) tract in the promoter region, and a dinucleotide cytosine-thymine repeat (CT repeat) in the coding region. Studies have demonstrated that collections of H. pylori strains exhibit significant diversity in the presence of sabA, CT-repeat tract lengths, and resulting expression of SabA (Lehours et al., 2004;Sheu et al., 2006;Yamaoka et al., 2006). This regulatory mechanism may explain why 1 % of J99 colonies in one study spontaneously lost their ability to bind sialyl-Le x (Mahdavi et al., 2002). Yamaoka et al. (2002) demonstrated that adherence, colonization ability, bacterial density, and induction of inflammation were all decreased when sabA or sabB was switched off, indicating that this mechanism of sabA regulation has functional significance.
Aside from genetic changes, H. pylori also uses twocomponent signal transduction (TCST) systems to respond to environmental changes. Activation of a TCST system results in changes in the rates of histidine kinase and response regulator protein phosphorylation, leading to altered promoter-region DNA-binding activity of the response regulator, and either positive or negative regulation of gene transcription (Beier & Frank, 2000;Stock et al., 2000). A previous study in our laboratory utilized DNA macroarrays to define the set of genes regulated by the HP0165-HP0166 TCST system in Helicobacter pylori strain 26695 (Forsyth et al., 2002). That study, which compared genome-wide transcriptional profiles between wild-type H. pylori and an isogenic HP0165 histidine kinase mutant, found one gene to be repressed in the null mutant, while six genes, including sabA (HP0725), were derepressed in the mutant.
Additional studies have further characterized and expanded this regulon, and identified acidic pH as the key environmental signal for HP0165-HP0166, and this locus has thus been redesignated arsRS (acid-responsive signalling) (Dietz et al., 2002;Pflock et al., 2004;Sachs et al., 2006). arsRS has accordingly been demonstrated to be essential for the production of urease under acidic conditions (Panthel et al., 2003), and is required for virulence in a mouse model (Pflock et al., 2005). Transcription of the response regulator locus arsR (HP0166), an essential gene in H. pylori (Beier & Frank, 2000), is downregulated at pH 5.0 (Bury- Mone et al., 2004) and non-phosphorylated ArsR has additional, phosphorylation-independent, regulatory activity (Schar et al., 2005).
Recent studies have provided further insights into the relationship between pH and the expression of the genes regulated by ArsRS at the transcriptional and translational levels. Global gene expression analyses by Merrell et al. (2003) and Bury-Mone et al. (2004) have found that at pH 5.0 sabA and its paralogue sabB (de Jonge et al., 2004) are downregulated, while HP1188, a novel H. pylori adhesin (Rubinsztein-Dunlop et al., 2005), is induced. It has also been reported that SabA-positive status is associated with decreased acid secretion in patients, and that SabA protein levels are reduced at pH 5.0 (Yamaoka et al., 2006).
In the present study, we hypothesized that the role of SabA in H. pylori adherence to AGS gastric epithelial cells is governed by phase variation and transcriptional regulation of sabA via the ArsRS system. We demonstrate that derepression of sabA transcription in an ArsS isogenic knockout strain of H. pylori (Forsyth et al., 2002) results in a corresponding functional change in the ability of the bacterium to adhere to gastric epithelial cells. In addition, we demonstrate the existence of multiple alleles of sabA within a single H. pylori strain population, and among multiple isolates from a single patient, differing in the nucleotide-repeat tract lengths. Our findings provide new insights into the complex mechanisms regulating the expression of the SabA adhesin and may contribute to an improved understanding of persistent H. pylori infection, and thus have implications for development of therapeutics.

METHODS
Bacterial culture. The H. pylori strains used in this study are described in Table 1. H. pylori was cultured on Trypticase Soy Agar II plates with 5 % sheep blood (BBL) at 37 uC and 5 % CO 2 , or Brucella agar plates supplemented with 10 % newborn calf serum (Gibco/ Invitrogen). Escherichia coli DH5a was cultured in Luria-Bertani medium. When appropriate, media were supplemented with 100 mg ampicillin ml 21 , 20 mg kanamycin ml 21 , 25 mg chloramphenicol ml 21 (for E. coli), or 5 mg chloramphenicol ml 21 (for H. pylori).
Oligonucleotides were designed to amplify the regions containing the poly-T (sabAampF and sabArev; see Table 2 for oligonucleotide sequences) and CT-repeat (sabAForC and HP0725R) tracts, as well as a downstream control region (sabABCFw and sabABCon). A 6carboxyfluorescein (FAM) label was added to the 59 end of one primer from each oligonucleotide pair. PCRs for AFLP analysis were performed using Vent High-Fidelity DNA polymerase (New England Biolabs). Amplicons were purified (QiaQuick PCR purification kit; Qiagen), processed at the DNA Sequencing and Synthesis Facility (Iowa State University, Ames, IA, USA), and analysed using GENESCAN software (Applied Biosystems).
Sequencing of sabA clones. The sabA poly-T region was amplified by PCR from H. pylori strains 26695 and J99 with primers sabAampF and sabArev. A population of clones (denoted as pSabA-T) was generated by cloning the resulting PCR products in pGEM-T Easy. Similarly, the sabA CT-repeat tract region was amplified from H. pylori 26695 genomic DNA (gDNA) or cDNA with primers HP0725ForwII and HP0725Ext, and cloned to create a collection of clones denoted as pSabA-gDNA-CT and pSabA-cDNA-CT, respectively. Plasmid DNA was isolated using the QiaSpin Miniprep kit (Qiagen) or Wizard Plus Midiprep kit (Promega), and clones were sequenced as described below to determine the length of the poly-T or CT-repeat tracts.
Quantitative real-time PCR. RNA was extracted from exponentialphase H. pylori cultures using the Invitrogen RNA Extraction System, treated with Turbo DNase (Ambion), and assayed for gDNA contamination by PCR. cDNA was synthesized from 2 mg RNA using random hexamers (Applied Biosystems) and AMV reverse transcriptase (Promega), purified using the QIAquick Nucleotide Removal kit (Qiagen), and verified via PCR. Quantitative real-time PCR was performed using the iCycler iQ real-time PCR detection system and SYBR Green supermix reagents (Bio-Rad). Relative expression of sabA (primers HP0725fwd and HP0725rev), HP0218 (control gene; primers HP0218forw and HP0218rev) and gyrB (normalization gene; primers gyrB forw and gyrB rev) was calculated for J99 and the J99-arsS : : cat mutant strain. PCRs were performed in triplicate, and meltcurve analysis was used to ensure that a single product was amplified with each primer set. Differences in gene expression from three independent experiments were calculated by the DDC T method (Livak & Schmittgen, 2001), and evaluated by Student's t-test.
Creation of H. pylori sabA : : cat, arsS : : cat, arsS : : km and arsS complemented mutants. A 525 bp fragment was amplified by PCR from the 59 region of H. pylori J99 sabA using primers JCN725F and JCN725R, and cloned into pGEM-T Easy (Promega) to generate pJCN1. The chloramphenicol acetyltransferase gene (cat) was excised from pCM7 (a kind gift of Dr John Loh and Dr Timothy Cover, Vanderbilt University Medical Center, Nashville, TN, USA) and cloned into the pJCN1 sabA HindIII site, resulting in pJCN2. H. pylori strain J99 was naturally transformed with pJCN2, as previously described (Forsyth et al., 2002), yielding strain J99-sabA : : cat. The cat cassette was also inserted into the arsS BglII site to generate strain J99-arsS : : cat.
H. pylori adhesion to AGS cells. The AGS (human gastric epithelial cell adenocarcinoma; ATCC) cell line was cultured in F12 Kaighn's medium supplemented with 10 % fetal bovine serum (Gibco/ Invitrogen) at 37 uC and 5 % CO 2 . The cells were then used to seed 24-well plates with 2610 5 cells per well, and the plates were maintained at 37 uC and 5 % CO 2 overnight. Plate-grown H. pylori was resuspended in 3.5 ml F12/10 % FBS, concentration was Regulation of sabA gene expression estimated by OD 600 , and the suspension was added to cells at 5610 7 c.f.u. per well. Plates were then centrifuged at 480 g for 5 min to initiate contact between H. pylori and AGS cells, and incubated for 3 h at 37 uC. After incubation, cell monolayers were washed three times with cold 16 PBS (pH 7.0) to remove non-adherent bacteria, and 0.5 ml F12/10 % FBS/1 % saponin was added to lyse cells. Tenfold serial dilutions of these lysates were plated on blood agar plates and incubated at 37 uC/5 % CO 2 for 4 days, and titres of the original and post-infection cultures were determined. Relative differences in adherence between strains in multiple independent experiments were evaluated by using Student's t-test.

RESULTS AND DISCUSSION
Multiple alleles of sabA exist in H. pylori 26695 and J99 populations due to poly-CT and poly-T tract length variation The first annotated H. pylori genome (strain 26695) identified the HP0725 locus (subsequently named sabA) as containing putatively phase-variable repetitive nucleotide sequences (Tomb et al., 1997). Depending on the length of a CT-repeat tract near the 59 end of the sabA ORF, subsequent translation either encounters a premature termination codon (resulting in a truncated, non-functional gene product), or encodes a full-length functional SabA adhesin protein. Strain 26695 is predicted to contain a 14-base poly-T tract in the promoter region, and to encode an out-of-frame sabA allele containing six CT repeats (see schematic diagram, Fig. 1a). A recent study showed that 49 % of strains examined (n589) were predicted to be in-frame based on the number of CT repeats (de Jonge et al., 2004). However, to our knowledge, no studies have evaluated the degree of diversity at the sabA repetitive tracts within a single strain of H. pylori, or among multiple strains isolated from a single patient.
We first tested the hypothesis that phase variation could result in H. pylori populations containing multiple alleles at the sabA poly-T and CT-repeat tracts. Preliminary evidence that the length of the sabA CT-repeat region varies in strain 26695 was obtained by electrophoresis of PCR products of the repeat-containing region of sabA on a denaturing polyacrylamide gel. When the region was amplified from a plasmid template (pSabA), a single amplicon was observed, while the same amplification using H. pylori strain 26695 gDNA as template revealed the presence of a second visible band (data not shown). Next, to better demonstrate allelic variation in the poly-T and CT-repeat tracts of the sabA locus, we conducted AFLP analysis. Results of these analyses indicated the presence of alleles of multiple lengths in amplicons containing the sabA poly-T and CTrepeat tracts from strain 26695 (Fig. 1). Control amplicons consisting of a portion of the sabA locus lacking a repetitive sequence produced a single detectable allele. Analysis of the putatively variable regions amplified from a cloned fragment of sabA, pSabA, also revealed a single fragment length (data not shown), thus indicating that the variation seen in AFLP analysis was not an artefact of PCR.
To quantify allelic variation at the CT-repeat tract within the sabA coding region of H. pylori strain 26695, we sequenced pSabA-gDNA-CT clones (Fig. 2a, black bars). Our sequencing results indicated that while 97.6 % (40/41) of the clones confirmed the out-of-frame sabA status, 25 % of these clones contained a tract length of eight CT repeats rather than the predicted length of six repeats. A single clone possessed a sabA allele containing seven CT repeats; this CT tract length is predicted to express a functional SabA adhesin. These data support the hypothesis that slipped-strand mispairing could generate allelic variation at the sabA locus; subsequent environmental pressures could rapidly select for a subpopulation that does or does not express the SabA adhesin.
We next hypothesized that RNA polymerase could err during transcription instead of, or in addition to, mutations induced by slippage of DNA polymerase during replication. As a result of the coupling of transcription and translation in prokaryotes, such an event could have a significant impact on protein expression. This might allow for the potential synthesis of a full-length mRNA transcript, despite the presence of a gDNA sequence that would predict abortive translation. To investigate this possibility, we cloned and sequenced PCR products amplified from H. pylori 26695 cDNA synthesized using a sabA-specific oligonucleotide (pSabA-cDNA-CT; Fig. 2a, grey bars). As with those derived from a gDNA template, nearly all pSabA-cDNA-CT clones analysed (n523) harboured out-of-frame sabA sequences: 91.3 % contained six or eight CT repeats, while two clones possessed an inframe CT tract with seven repeats. No statistically significant difference in the distribution of sabA CT-repeat tract lengths was observed in clones derived from cDNA versus gDNA templates (2 d.f., x 2 51.76, P50.415). Thus, it seems unlikely that significant expression of SabA could result from the generation of in-frame mRNA transcripts with different CT-repeat tract lengths than the corresponding chromosomal sequence.
Variation in the sabA promoter-region poly-T tract was likewise characterized by sequencing pSabA-T clones containing amplicons derived from H. pylori 26695 or J99 gDNA (n519 and n515, respectively; Fig. 2b). Sequencing analysis of all clones containing sabA promoter region PCR amplicons from strain 26695 revealed a distribution of 13-16 thymine nucleotides, with a predominant length of 15 bases (68 % of clones). pSabA-T clones derived from strain J99 contained 16-19 thymine nucleotides; 40 % of these had a repeat tract length of 17 bases. Annotated genomic sequences indicate poly-T tracts of 14 nt in 26695 (Tomb et al., 1997), and 18 nt in J99 (Alm et al., 1999). These results provide strong evidence that the sabA promoter poly-T tract undergoes slippedstrand mispairing, resulting in a population of H. pylori with numerous alleles. Further studies will be needed to ascertain the potential effects of poly-T length variation on transcriptional control at the sabA locus. Allelic diversity at the sabA locus exists within a single patient In order to gain a better understanding of the actual sabA diversity within a single host, we analysed a collection of 12 low-passage strains re-isolated by gastric biopsy from the J99 source patient 6 years after the initial endoscopy (Israel et al., 2001). These strains were isolated from several regions of the stomach: antrum (n55), cardia (n51), corpus (n54), and foci of gastric metaplasia in the duodenum (n52). The sabA CT-repeat region of each isolate was amplified by PCR and directly sequenced without cloning in order to assess the presence or absence of variation in the length of the tract relative to the archival strain J99 (Fig. 2c). Five isolates had a tract of seven CT repeats (in-frame sabA allele), one isolate contained eight CT repeats (out-of-frame), and six isolates possessed 10 CT repeats (in-frame).
While the annotated genome sequence of archival strain J99 predicts an out-of-frame sabA locus with nine CT repeats (Alm et al., 1999), 11 of the 12 re-isolates analysed were predicted to be in-frame based on a CT-repeat tract length of 7 or 10 repeats This suggested the possibility that a selective pressure favouring expression of SabA may have developed in the host that was not present when the strain was originally isolated. However, we proceeded to sequence the repetitive region of archival J99, as well as two singlecolony isolates derived from that strain, and found 10 CT repeats, corresponding to a phase-on sabA allele in each case. Taken together, these sequencing results, along with published observations that J99 (but not 26695) binds the Lewis X antigen (Mahdavi et al., 2002), and adherence data in the current study (see below), suggest that J99 does in fact harbour an in-frame sabA locus. It is possible that a portion of the sabA variation observed in the re-isolates examined in the current study was present at the time strain J99 was initially isolated from an antral biopsy, but not reflected in the published genome sequence. However, as a recent study would suggest (Kuipers et al., 2000), selective pressures in the 6 years between the isolation of the original and novel J99 strains probably contributed to further genetic diversity at sabA and other loci.
One isolate (J99 C-6) was selected for additional study of variation in the CT-repeat tract as described above (Fig. 2d). Plasmid clones (n533) were generated from the amplicons generated from the sabA CT-repeat region of H. pylori J99 C-6, and sequenced to determine the initial diversity present at the sabA locus at the time of the biopsy ('low passage'). In addition, the effect of prolonged in vitro passage on the sabA CT-repeat tract was studied in a similar manner using 34 plasmid clones obtained from PCR amplicons of the same region of sabA after numerous in vitro passages ('high passage', .50 passages). In each case, the vast majority of clones (94 % of low-passage and 85 % of high-passage clones; Fisher's exact test, P50.43) contained in-frame sabA alleles with 7 or 10 CT repeats. However, a significant shift towards a longer CT-repeat tract was observed in high-passage clones (x 2 for trend519.7, d.f.51, P,0.0001). These results indicate that a functional SabA was favoured in the gastric niche from which the strain was isolated.
Taken together, the above results clearly demonstrate that phase variation via slipped-strand mispairing occurs at the H. pylori sabA locus, and results in the generation of diversity in the length of promoter region poly-T and coding region CT-repeat tracts within a single strain, and among multiple isolates from a single patient. As a result of variations in this repetitive nucleotide tract, populations of H. pylori may or may not express the functional SabA adhesin molecules needed to mediate BabA-independent binding to sialyl-Le x antigens on the host cell surface.
Sensory histidine kinase arsS null mutants express a sabA-dependent hyper-adherent phenotype We next sought to study the effects of ArsRS-mediated transcriptional control of sabA on adherence by H. pylori strains harbouring in-frame (J99) and out-of-frame (26695) sabA loci. Several studies have demonstrated that the ArsRS TCST system, in response to environmental changes in pH, regulates the transcription of sabA (Dietz et al., 2002;Forsyth et al., 2002;Pflock et al., 2004). We conducted a series of in vitro assays that quantified adherence of H. pylori strains to AGS cells to test the hypothesis that deletion of the arsS histidine kinase locus would result in derepression of sabA transcription and, in strains containing in-frame sabA CT-repeat tracts, increased adherence to gastric epithelial cells.
We first confirmed the transcriptional control of sabA by arsS (previously demonstrated only by genome-wide transcriptional profiling) by performing quantitative realtime PCR to compare sabA transcription in wild-type H. pylori and an isogenic arsS mutant strain (Fig. 3). Results showed a 3.75±0.25 fold increase in sabA cDNA in H.
pylori J99-arsS : : cat compared to wild-type J99 (mean±SEM from three independent experiments; P50.0026, Student's t test). These results are concordant with our earlier DNA macroarray study that indicated that the adhesin sabA is 3.34-fold derepressed in the absence of a functional allele of the ArsS histidine kinase (Forsyth et al., 2002). Expression of HP0218, a gene not under the transcriptional control of the ArsRS TCST system, was not significantly different in J99 versus J99-arsS : : cat (P50.62, Student's t test).
To examine the functional effect of increased transcription of sabA in the absence of arsS, we assayed the ability of wild-type H. pylori strain J99 and three J99-derived mutant strains (J99-arsS : : km, J99-arsS : : km-sabA : : cat and J99-arsS : : km-rdxA : : arsS) to bind to AGS cells in vitro (Fig. 4) Strain J99-arsS : : km-rdxA : : control, a control strain created by inactivating the rdxA gene with an unrelated sequence (Loh & Cover, 2006), retained the hyper-adherent phenotype (data not shown), confirming that the decreased binding seen in strain J99-arsS : : km-rdxA : : arsS was due to complementation of arsS rather than the inactivation of rdxA. A hyper-adherent phenotype similar to that of J99-arsS : : km was also seen in a J99-arsS : : cat strain, demonstrating that independently generated arsS-null strains possess the same phenotype (data not shown). In contrast, there was no increase in adherence of H. pylori strain 26695-arsS : : cat relative to wild-type 26695 [2.5(±0.3)610 5 vs 3.9(±0.5)610 5 , P50.3, n52]. There was also a trend towards lower basal adherence by wildtype strain 26695 than by J99, potentially due to the lack of SabA-mediated binding in 26695. Further studies will be necessary to elucidate what other factors, such as relative BabA expression, may impact the degree of adherence to AGS cells.
These results establish the functional significance of two distinct mechanisms regulating the expression of the SabA adhesin. The elimination of ArsS in J99 led to a greater than 10-fold increase in adherence that was SabAdependent, importantly demonstrating that the transcriptional regulation of sabA by the ArsRS TCST results in functional changes in H. pylori adherence to gastric epithelial cells. However, this elevated binding upon disruption of arsS only occurred in strain J99, which contains an in-frame sabA allele, but not in strain 26695, which possesses a predominantly out-of-frame sabA locus.
Persistent H. pylori colonization requires continual adaptation to variations in the gastric microenvironments it inhabits, and to robust host immune and inflammatory responses. Due to the fact that the H. pylori genome contains relatively few conserved transcriptional regulators, alternative mechanisms of gene regulation, such as variation in repetitive DNA sequences, play an important role in generating genetic diversity in H. pylori (Aras et al., 2003;Mrazek et al., 2007). The present study demonstrates an example of how TCST-mediated regulation and phase variation combine to regulate the transcription of sabA, and the subsequent ability of H. pylori to bind sialyl-Le x displayed on gastric epithelial cells. The combination of transcriptional regulation of the sabA locus by the ArsRS TCST system, and the generation of subpopulations harbouring alternate sabA alleles by slipped-strand mispairing during chromosomal replication, could permit H. pylori to rapidly adapt to varying microenvironments or host immune responses. The existence of multiple means by which the expression of SabA is controlled in H. pylori suggests that precise regulation of this adhesin may be crucial to the virulence of this important pathogen.
Helicobacter pylori adhesion to carbohydrates. Methods Enzymol 417, 293-339. Differences between strains in mean adherence were evaluated by using Student's t-test.