H. pylori is a spiral bacterium, and this helical shape is thought to enhance motility through the viscoelastic mucus layer by corkscrew propulsion. Helical bacteria such as H. pylori and Campylobacter can swim faster than straight-rod E. coli in liquid cultures presumably due to their helical body shape.28 Helical shape is established by the sum of cell elongation, curvature, and twist.29 The curvature structure is formed by peptidoglycan cross-linking relaxation associated with the gene products of ccd1, ccd2, ccmA, ccd3, ccd4, ccd5, and ccd6. The Ccd proteins play the role of DD-endopeptidases or carboxypeptidases. Csd1, Csd2, and CcmA may dissolve the tetrapentapeptide crosslinking of peptidoglycans to support helical twist and cell curvature. Csd3 induces changes in the level of cross-linked muropeptide content to maintain a helical form.29 Csd4 has DL-carboxypeptidase activity on tripeptide monomers, cleaving the terminal mDap residue to produce dipeptide monomers. Ccd6 has LD-carboxypeptidase activity and seems to act in concert with Csd4 to produce dipeptides.30 Csd4 and Csd6 proteins have opposite effects on the conversion from tripeptides to dipeptides, related to the induction of cell curvature.30 Csd5 might provide an activating signal for cell shape modulation, perhaps through recruitment of enzymes involved in peptidoglycan synthesis, such as Csd4 and Csd6. All deletion mutants for these genes showed loss of helical shape and demonstrated a decrease in the infectivity to murine gastric mucosa.29,30 In addition, H. pylori contains a complex cytoskeleton which affects cell morphology as well as motility.31 It possesses four coiled-coil rich proteins, Ccrp58, Ccrp89, Ccrp1142, and Ccrp1143, which extend filamentous structures and control cell shape and motility.31 The deletion of these Ccrp genes caused a loss of the helical shape and decrease of motility.31 Ccrp filaments may be incorporated with proteins which synthesize the cell wall, and Csd proteins form the helical shape during bacterial growth.

The mobility of H. pylori is conferred by flagella. Motile flagella are critical for the persistence of H. pylori in the gastric mucus layer, meaning they endow the locomotion ability of the organism in the mucus layer.32 H. pylori carries a polar tuff of three to five sheathed flagella, 4.3-9.5 µm long and 23.5-40 nm thick. A characteristic terminal bulb is observed, which may be an extension of a flagellar sheath. Under electron microscopic observation, the flagella appear to be 1.35-nm thick filaments surrounded with sheaths, showing the typical bilayer structure of a membrane containing lipopolysaccharide (LPS) and several proteins.33 H. pylori flagella are synthesized by the coordination of over 40 genes dissipated in the whole genome. Bacterial flagella consist of a basal body, hook, and flagellar filament. The basal body is composed of the LP ring (FlgH and FlgI), the MS-C ring (FliF, FliG, FliM, and FliN), and the rod (FliE, FlgB, FlgC, and FlgG).32 The LP ring is assembled at the level of the outer membrane and peptidoglycan. The MS ring (M, membrane; S, supramembrane) is built up by the self-assembly of FliF above and within the cytoplasmic membrane.34 The cytoplasmic face of the MS ring is mounted with the drum-shaped C-ring. The rod is a shaft which connects the LP ring and the MS-C ring. The center of the cytoplasmic face of the MS-C ring is located with the flagellar protein export apparatus. The flagellar filament is made of two flagellins: the major, FlaA, and the minor, FlaB.34 FlaA builds up an extracellular region of the flagella. H. pylori flagellins were identified to be modified post-translationally by glycosylation. The FlaA protein is modified with a total of seven O-linked pseudaminic acid (Pse5Ac7Ac) residues, while FlaB is modified with ten O-linked Pse5Ac7Ac residues. Pseudaminic acid is a nine-carbon sugar absent from humans and is biosynthesized by several genes. The deletion of any of these genes results in non-motile H. pylori that lack functional flagella, and subsequently, there is a loss of mouse infectivity, demonstrating that the pseudaminylation of flagella subunits is absolutely essential for H. pylori to synthesize functional flagella to colonize the host's stomach.35

Bacteria move by a rotating movement of the rotary motor at the base of the filament in both clockwise (CW) and counterclockwise (CCW) directions.34 The basal body imbedded in the cytoplasmic envelope works as a reversible rotary motor, while the hook and filaments extended extracellularly act as a propeller. The basal body also includes the flagellar protein export apparatus, which consists of six integral membrane proteins (FlhA, FlhB, FliO, FliP, FliQ, and FliR) and three soluble proteins (FliH, FliI, and FliJ). The export of LP ring is carried out by a Sec-dependent pathway. And the extracellular components of the flagellum may be transported by a central channel of the on-going structure by a type III secretion mechanism.32 FliI is associated with provision of the energy required for the export process.32

The rotary motor is powered by proton or sodium ion motive force across the cytoplasmic membrane. The motor is made of a rotor (MS-C ring), a drive shaft (rod), a bushing (LP ring), and about a dozen stators. The rotor (rotating part) is surrounded by a circular array of the stator complexes (nonrotating part) in the membrane. The stator complexes are formed by the MotA and MotB proteins, in which MotA is inserted into the cytoplasmic membrane and MotB contacts with the peptidoglycan. The C-ring is assembled with FliG, FliM, and FliN to be termed the switching complex, producing both rotational movement and CW-CCW switching.36 FliG is the rotor protein most directly involved in the generation of torque. The C-terminal domain of FliG forms the rotor portion for the site of energy transduction in the flagellar motor and interacts with the stator protein MotA.37 H. pylori demonstrates positive chemotaxis towards urea, bicarbonate, and body fluid. H. pylori senses chemotactic attractants through the membrane-bound methyl-accepting chemotaxis proteins (MCPs) TlpA, TlpB, TlpC, and HP0599, which transduce the signals to CheA. CheA is stimulated by the MCPs to become autophosphorylated via CheW, and transfers a phosphoryl group from a histidine residue to an aspartate of CheY. The phosphorylated CheY then interacts with FliM to change the direction of motor rotation with CW-CCW switching.38

After ingestion, the gastric mixture becomes weakly acidic, to which H. pylori is tolerant. H. pylori attaching to the gastric mucus layer uses urease to hydrolyze urea in the mucus layer for the production of NH₃. NH₃ rapidly diffuse to the mucus layer in the vicinity of H. pylori to change the pH, which modifies the rheologic properties of the mucin network. Thereafter, H. pylori achieves penetration into the gastric mucus layer by coupling the corkscrew propulsion originating from the helical shape with flagellar motility, which is obtainable due to the liquefied mucin in the vicinity of H. pylori.

The outer envelope components of H. pylori, LPS and outer membrane proteins (OMPs) are closely associated with its adherence to host cells. The most outer envelope of H. pylori contains LPS, as do those of other Gram-negative bacteria.43 Generally, LPS is the component of Gram-negative bacteria that contributes to immune modulation and biological toxicity associated with the pathogenesis of bacterial infections. In enterobacteria, LPS consists of lipid A, core oligosaccharide, and O-chains of repeating units. The typical lipid A backbone of enterobacteria is a β, 1'-6-linked disaccharide of glucosamine which is bisphosphorylated and multiply acylated with (R)-3-hydroxymyristate at the 2-, 3-, 2'-, and 3'-positions. Core oligosaccharide is a branched polysaccharide of 9-12 sugars, including the unusual sugar 2-keto-3-dexoxy-octanoate (KDO), that covalently binds to the glucosamine ring of lipid A. The O-side chain is a long, branched, or linear polysaccharide consisting of 50-100 repeating units, which bind the core oligosaccharide.43 The LPS of H. pylori is synthesized by the incorporation of at least 27 genes, which are scattered throughout the whole genome and are unlike other bacteria, with the exception of rfbM, rfbD, and wbcJ orthologs encoding proteins involved in GDP-fucose synthesis.44 Orthologs to all enzymes required for lipid A and core oligosaccharide synthesis have been identified in the H. pylori genome. The lipid A of H. pylori lacks a phosphate group at the 4'-hydroxyl position on the distal nonreducing glucosamine of the disaccharide, is substituted at the C-1 position on the proximal glucosamine with a phosphoethanolamine residue, and is tri-acylated or tetra-acylated with either (R)-3-hydroxystearate (C18) or (R)-3-hydroxypalmitate (C16).45 During the synthesis of H. pylori LPS, H. pylori WaaA transfers two KDO sugars to the distal glucosamine of the tetra-acylated lipid A precursor lipid IV_A, followed by the addition of the acyloxyacyl-linked fatty acyl chains, resulting in a minor lipid A.46 After constitutive biosynthesis, Kdo2-lipid A is modified by the removal of the 1-phosphate group from lipid A by lipid A-1 phosphatase (HP0021) on the periplasmic side of the inner membrane, after which LPS is transported across the inner membrane. The removal of the outer KDO sugar or the addition of a phosphoethanolamine residue to the 1 position is carried out by phosphoethanolamine transferase (HP0022) in the periplasmic space.46,47 Conclusively, H. pylori LPS shows significantly lower immunologic activity and endotoxicity compared with enterobacteria due to underphosphorylation, underacylation, and substitution with long chain fatty acids in its lipid A. In addition, modification at the 1 position of H. pylori lipid A confers resistance to cationic antimicrobial peptides like polymyxin B. A core region with an unusual branching patterns and O-chains of relatively constant chain length are found in H. pylori LPS.48 One of the characteristic features of H. pylori LPS is the abundance of D-glycero-D-manno-heptose (DD-heptose) residues in its core region, unlike other Gram-negative bacteria, which mainly contain L-glycero-D-manno-heptopyranose (LD-heptose). H. pylori HP0479 was identified to encode a D-glycero-D-manno-heptosyltransferase gene which adds a D-glycero-D-manno-heptose residue (DDHepII) to a distal DD-heptose of the core oligosaccharide backbone of H. pylori LPS.49

The H. pylori O-chains are assembled in the cytoplasm onto a polyisoprenoid membrane anchor. WecA, which plays the role of the UDP-GlcNAc:undecaprenyl-phosphate GlcNAc-1-phosphate transferase, assembles the glycan of Gal and GlcNAc residues, generating the linear O-chain backbone. Fucosyltransferases then decorate fucose residues to specific locations of the O-chain backbone, forming Lewis antigens.50 The flippase Wzk translocates the O polysaccharide to the periplasm, where the O antigen ligase WaaL connects it to the lipid A-core to assembe the LPS molecule. The LPS molecule can then be transferred to the outer membrane to appear in the outer side of the bacterium.50

The most striking observation about H. pylori LPS was the identification of Lewis X in its O-chain, followed by the discovery of the mimicry of type 1 and type 2 blood group antigenic determinants on the core oligosaccharide. Some strains express sialyl-Lewis X in the O-chain. Additionally, novel sugars of L-rhamnose, D-rhamnose, or 3-C-methyl-D-mannose, as well as heptoglycans composed of 2-linked and 3-linked D-glycero-a-D-manno-heptopyranose units, were found in the O-chain of some strains, revealing that not all H. pylori strains express Lewis antigen mimicry.51

The O-chains of H. pylori LPS are posttranslationally decorated with fucose residues to generate Lewis antigens. The precursor of type 2 Lewis antigens, LacNAc (N-acetyl-β-lactosamine, Gal β1-4GlcNAc), is synthesized by rfaJ homologues (HP0159) and HP0826, encoding β1,4-galactosy-ltransferase (β1,4GalT).51 There are three fucosyltransferases in the genome: FutA, FutB, and FutC. Of these, FutA and FutB (α1,3/4-fucosyltransferases, α1,3/4-FucT) catalyze the transfer of GDP-fucose on nonreducing ends and internal sites of O-antigen polymers of various sizes to produce the Lewis X epitope, a substrate for FutC (α1,2-FucT) that yields difucosylated Lewis Y.52 FutA and FutB, each of which contains a C-terminal heptad repeat region, are expressed under the regulation of on/off status via slipped-strand mispairing in the intragenic polyC tract regions of their genes.53 The C-terminal heptad repeat region consists of a variable number of DD/NLRV/INY tandem repeats in which variation correlates to the size of the O-antigen polymers, which are decorated by fucose residues.53 The expression of FucC is regulated by frame shifts in a polyC tract situated in the middle of the gene.53 Type 1 precursor, Gal β1-3GlcNc, is decorated by FutA and FutB to produce Lewis A antigen and by FutC to produce H type 1 antigen. Lewis A is a substrate of FutC, which produces Lewis B.53 Most H. pylori strains (>80%) have serologically been identified to express type 2 antigen Lewis X and/or Y, but a minor portion of the strains (<10%) were shown to carry type 1 antigen Lewis A and B.54 East Asian strains express both type 1 and type 2 antigens far more frequently compared to others.55 The Lewis antigens of H. pylori LPS show extensive diversity over the infection periods and in the colonizing regions of the host stomach in vivo, as well as over the duration of culture and according to culture conditions in vitro.56 H. pylori was observed to adapt to the host by phase variation of its LPS, in which the bacterial Lewis antigen evolves to resemble the host's gastric Lewis phenotype, demonstrating that the bacterial genetic locus is subjected to host-driven selective pressure.56 Such phase variation could be explained by regulation of the on/off status of fucosyltransferase genes.

H. pylori LPS plays several delicate and complicated pathogenic roles to allow persistent infections of the gastric mucosa, of which the foremost function is adherence to the gastric mucosa. Lewis antigen mimicry was suggested to be implicated initially in immune evasion and gastric adaptation by the bacterium. However, long-term H. pylori infection can induce the production of cross-reacting autoreactive anti-Lewis antibodies, which is reported to react to the β-chain of the gastric proton pump (H⁺,K⁺-ATPase) contained within parietal cell canaliculi, contributing in part to the development of gastric atrophy.57 The Lewis X antigen in the O-chain is involved in the adhesion of H. pylori to the human antral gastric mucosa through the gastric receptor of the galactoside binding lectin, galectin-3.58 In addition, H type 1 antigens are suggested to attach to a membrane mucin MUC1, through carbohydrate-carbohydrate interaction.59 Lewis antigen mimicry in the O-chain contributes directly to the development of bacterial pathogenesis, but not to bacterial colonization. To date, no investigations have been done on the circumstances under which Lewis antigen mimicry in the O-chain of LPS could mediate the adherence of bacteria to the gastric mucosa.

In addition to LPS, the H. pylori outer membrane mediates direct contact with the surrounding environment of the host, including the mucus and the epithelial surface of the gastric mucosa. Therefore, H. pylori OMPs have been postulated to be crucial for adaptation to and persistent colonization of the host. Of the 1600 ORFs in the H. pylori genome, approximately 4% have been annotated as OMPs, representing a significantly larger amount than in any other bacterial genome.44 Of the ORFs, 64 were predicted to be OMPs, classified into 5 paralogous families. Of these, 33 belong to the Hop (Helicobacter outer membrane porins) and Hor (Hop-related proteins) groups.60 The HOP subfamily includes several known H. pylori adhesins including BabA (HopS), SabA (HopP), AlpA (HopC), AlpB (HopB), HopZ, and OipA (HopH), which promote and enhance bacterial interaction with the epithelial surface of the gastric mucosa. These genes share high similarity or identical sequences at the amino and carboxy termini which could elevate the possibility of intragenomic or intergenomic recombination in the H. pylori population. The genes also contain dinucleotide (CT) repeats in their signal sequences, suggesting that their expression could be regulated by slipped-strand repair.

BabA, the H. pylori blood group antigen-binding adhesion, is the best-characterized of the H. pylori adhesins. It binds Lewis B antigen, a fucosylated blood-group antigen present on gastric epithelial cells. Human colostrum secretory immunoglobulin A was observed to inhibit the binding of H. pylori to gastric surface mucus cells, which is caused by digestion with α-L-fucosidase, implying that fucose-containing factors are involved in this inhibition rather than immunoglobulin.61 The fucose-sensitive receptor for H. pylori was identified as Lewis B, a fucosylated carbohydrate structure that is the predominant blood group-related antigen expressed on gastric surface mucus cells in the stomach epithelium of individuals with positive secretor status.62 BabA adhesin is located on the outer membrane at frequency of ca. 500 per cell with Ka value of ≈1×10¹⁰ M^-1.62 There are three BabA genes according to the strain (babA1, babA2, and babB), which differ by a repeating motif in the signal peptide sequence or translation start point. The signal peptide starts at position -20 in BabA and position -18 in BabB. The babA2 gene responsible for the production of the functional BabA adhesin is identical to the babA1 gene with the exception of a 10-bp insert with a repeat motif in the signal peptide sequence resulting in the creation of a translational initiation codon. Most isolates contain babA zero to two copies in the three possible loci of the chromosome.63 Chimeric babB/babA alleles have been detected among clinical isolates, and frame shift mutation and polymorphisms are present throughout the babA alleles.63 European and American strains (generalist strains) bind to blood group A, B, and O type 1 determinants, which are the most prevalent in the population. In contrast, native South American strains (specialist strains) bind only to the blood group O type 1 determinants (Lewis B and H type 1) of the local prevalent blood type.64 In addition, during infection of rhesus macaques, babA is replaced by babB or not expressed at all because of alteration in the dinucleotide CT repeats in the 5' coding region, revealing that babA is regulated to facilitate adherence to the gastric epithelium and to promote chronic infection through both antigenic variation and phase variation.65 These observations indicate that BabA adhesin progressively evolves in the direction for bacterial adaption to their individual hosts in order to sustain the extraordinary chronicity of gastric infection.66 However, many strains do not contain babA bound to Lewis B antigen as well as Lewis B binding activity. Consequently, these strains do not reflect the severity of mucosal damage or the clinical outcome, implying that the expression of BabA adhesin is not critical for persistent colonization with H. pylori in the gastric mucosa of humans.

AlpAB, adherence-associated lipoprotein A and B, was found to be involved in adherence to the gastric cell line Kato III and is annotated as omp20-omp21 in the 26695 genome and as hopC-hopB in strain G27.67 AlpAB was identified as the abundant protein of the sarcosine-insoluble fraction.68 AlpAB genes were identified by gene disruption using transposon mutagenesis and are encoded by two adjacent homologous genes organized in an operon with co-transcription. AlpAB has a porin-like b-barrel in the outer membrane, consisting of 14 transmembrane amphipathic β-strands. H. pylori strains lacking both alpA and alpB or containing only alpB have severely reduced laminin binding, demonstrating laminin to be a target of AlpB, while both are required for H. pylori laminin binding.67 In Guinea pigs, alpA and alpB mutants could not be isolated from animals when infected with a mixture of the wild-type and mutant strains, and IgG antibody responses were lower in animals infected with either the alpA or the alpB mutants, suggesting that the absence of AlpA or AlpB presents a serious disadvantage for colonization of the stomach by H. pylori.69

SabA, a sialic acid-binding adhesion, was first observed as a receptor of gastric mucosa from an H. pylori-infected patient with gastritis, to which mutants of both babA1A2 or only inactivated babA2 could bind to.70 Whereas the gastric surface of patients infected with H. pylori can be bound by BabA mutants, that of healthy individuals without H. pylori infection cannot. A gastric surface target of SabA is sialyl-dimeric-Lewis X glycosphingolipid, which is induced in the gastric epithelium of humans and Rhesus monkeys by H. pylori infection. SabA was identified to be a 66 kDa protein encoded by JHP662 in strain J99 and HP0725 in strain 26695. The binding activity for sLewis X is spontaneously lost in 1% of the colonies expended from a single colony. Instability of sLewis X-binding may come from the regulation of the on/off frameshift modulated by poly T/CT tracts upstream, as well as from within the start of sabA. Although the on/off status of sabA was investigated to find no correlation with gastric disease, severe neutrophil infiltration and atrophy are associated with functional sabA.71

OipA, the outer inflammatory protein adhesin encoded by HP0638, is directly associated with the promotion of interleukin-8 (IL-8) production rather than binding ability to the gastric epithelial surface. OipA knockout mutant strains showed significant reduction in their binding abilities to AGS, and restoration of the adherences was shown in Kato-III by complementation of the oipA gene.72 Although the presence of a functional oipA gene allows significant IL-8 secretion to be maintained above the control values even with disruption of cagPAI, it was supposed that OipA induces IL-8 secretion in conjunction with CagPAI. Live H. pylori induces IL-8 secretion though direct contact with gastric epithelial cells, in which regulation of transcription of the IL-8 gene is mediated by the interferon-stimulated responsive element (ISRE)-like element, activator protein (AP)-1, and NF-κB.73 OipA and the cag PAI are involved in inducing interferon regulatory factor (IRF)-1 to bind and activate the ISRE-like element. Whereas CagPAI is involved in activating AP-1 and NF-κB, OipA is involved in STAT1 phosphorylation of upstream signaling of IRF-1, demonstrating that OipA and the cag PAI are both necessary for full activation of the IL-8 promoter but act via different pathways.74 H. pylori induces focal adhesion kinase (FAK) phosphorylation, which in turn activates Erk, followed by actin stress fiber formation. H. pylori Cag A induces the phosphorylation of FAK Y407, whereas OipA adhesin phosphorylates Y397, Y576, Y577, Y861, and Y925, revealing OipA to be a better effector of stress fiber formation than CagA.75 In addition, H. pylori OipA reduces IL-10 secretion and suppresses DC maturation, promoting the establishment of chronic infection.76 The oipA gene is regulated by the slipped strand repair mechanism of CT dinucleotide repeats in the signal-sequence coding regions of the 5'-region. The number of CT repeats ranges from 6 to 13.77 Isolates with different oipA CT repeat patterns, including inter-niche variation, are frequently found in a single host.77

HopZ, first identified by the preparation of sarkosyl-insoluble outer membrane protein, is a major adhesive, as knocking out the hopZ gene results in a significant decrease in the adhesive ability of bacteria.78 Two alleles of on and off were found to regulate hopZ expression by slipped-strand mispairing within a CT dinucleotide repeat motif located in the signal-peptide coding region.78 Strains harboring the iceA1 allele, sabA functional status, and hopZ "off" status were reported to be associated with the development of MALT lymphoma, demonstrating 10 times higher odds.79 Up-regulation of hopZ was observed in the transcriptional profiling of a chronic atrophic gastritis strain during infection of a mouse gastric stem cell-like line. Genetic inactivation of hopZ produced a fitness defect in the stomachs of gnotobiotic transgenic mice with a chronic atrophic gastritis-like phenotype, but not in wild-type littermates, implying that HopZ is important for bacterial survival in the gastric ecosystem of chronic atrophic gastritis.80 Although the Guinea pig model showed that HopZ-deleted mutants were still able to infect the stomach, HopZ is thought to play an important role in colonization of the gastric mucosa in early infection due to the stable HopZ-ON status of isolates recovered during early infection.81 In addition, during chronic infection, the amino acid residues of HopZ proteins are more frequently changed into ON than OFF status isolates, demonstrating hopZ to be one of the bacterial tools for adaptation to the host environment.

Among OMP adhesin, BabA, SabA, and AlpAB were identified to bind to blood group antigen, sialylated proteins, and laminin, respectively. However, the receptors for OipA and HopZ on gastric cells need to be investigated. In addition to physical adherence activity to the surface of gastric cells, OMP adhesins participate in or partly augment the pathogenic processes of gastric cells, including altered transcription, cytoskeletal rearrangements, disruption of cell junctions, and onset of inflammation. In the case of OipA especially, IL-8 production is articulated more frequently than adherence to the gastric epithelial surface.