Comparative Genomics Revealed Multiple Helicobacter pylori Genes Associated with Biofilm Formation In Vitro

Background Biofilm formation by Helicobacter pylori may be one of the factors influencing eradication outcome. However, genetic differences between good and poor biofilm forming strains have not been studied. Materials and Methods Biofilm yield of 32 Helicobacter pylori strains (standard strain and 31 clinical strains) were determined by crystal-violet assay and grouped into poor, moderate and good biofilm forming groups. Whole genome sequencing of these 32 clinical strains was performed on the Illumina MiSeq platform. Annotation and comparison of the differences between the genomic sequences were carried out using RAST (Rapid Annotation using Subsystem Technology) and SEED viewer. Genes identified were confirmed using PCR. Results Genes identified to be associated with biofilm formation in H. pylori includes alpha (1,3)-fucosyltransferase, flagellar protein, 3 hypothetical proteins, outer membrane protein and a cag pathogenicity island protein. These genes play a role in bacterial motility, lipopolysaccharide (LPS) synthesis, Lewis antigen synthesis, adhesion and/or the type-IV secretion system (T4SS). Deletion of cagA and cagPAI confirmed that CagA and T4SS were involved in H. pylori biofilm formation. Conclusions Results from this study suggest that biofilm formation in H. pylori might be genetically determined and might be influenced by multiple genes. Good, moderate and poor biofilm forming strain might differ during the initiation of biofilm formation.


Introduction
Helicobacter pylori is strongly associated with gastroduodenal diseases, such as chronic gastritis, peptic ulcer, duodenal ulcer and gastric cancer [1]. In fact, this bacteria has been classified as a group I carcinogen by the International Agency for Research on Cancer [2]. Although H. pylori is susceptible to most antibiotics in vitro, only few antibiotics can be used for eradicating H. pylori in vivo, e.g. amoxicillin, clarithromycin, metronidazole and tetracycline [3]. Combination therapy is needed for successful eradication of H. pylori, Increasing prevalence of antibiotic resistance in H. pylori is problematic since it is one of the important causes of therapy failure [4]. The ability of H. pylori to form biofilm in vitro and also in vivo has been demonstrated [5][6][7]. H. pylori formed biofilm within in vivo environment especially gastric mucosa had been well demonstrated by previous study [7]. The formation of biofilm within in vivo environment was believed to be one of the factors which cause H. pylori eradication failure [7].
According to Donlan et al. (2002) [8], biofilms are defined as complex microbial ecosystems adherent to each other and/or to surface or interface. It is a microbially derived sessile community characterized by cells that are embedded in a matrix of extracellular polymeric substances that they have produced, and exhibit an altered phenotype with respect to growth rate and gene transcription [8]. Biofilms are important in bacterial pathogenesis as they plays a central role in helping microbes survive or spread within the host. This is because the biofilm matrix acts as shield, protecting bacteria from host defenses and antibiotics [9,10]. Bacterial biofilm can be 10 to 1000 times less susceptible to antimicrobial substances than the same organism in suspension [11]. Biofilm formation makes it difficult for antibiotics to reach the bacteria, underneath aiding the development of antibiotic resistance and increase the risk of treatment failure [8]. Therefore, biofilms are often associated with chronic infections [12]. Cammarota et al. demonstrated that H. pylori biofilm on the gastric mucosa epithelium was associated with eradication failure and N-acetylcysteine pretreatment before antibiotic therapy was effective in overcoming antibiotic resistance [13,14].
The ability of biofilm formation in bacteria is genetically controlled [15][16][17][18][19]. Based on previous studies, one of the most common genes associated with biofilm formation in bacteria (including H. pylori, S. mutans and L. monocytogenes) is the luxS gene [15][16][17]. Apart from luxS gene, few other genes have also been reported to be associated with biofilm formation in different bacteria [18,19]. According to Yoshida and Kuramitsu (2002) [20], dgk, sgp and com also play a role in S. mutans biofilm formation. Wu and colleagues [21] revealed that treC and sugE affect biofilm formation in K. pneumoniae by modulating capsular polysaccharide synthesis production. Taken together, biofilm formation in bacteria may be regulated by multiple genes working together.
Despite extensive studies conducted on H. pylori biofilm formation, precisely how genetic differences of different H. pylori strains influence biofilm formation remains not wellestablished. The objective of this study is to identify genes associated with H. pylori biofilm formation.

H. pylori strains
H. pylori standard strain J99 (ATCC 700824) was obtained from the American Type Culture Collection (USA). The H. pylori clinical strains used in this study were from the H. pylori collection of the Helicobacter Research Laboratory at the University of Malaya (Malaysia).
H. pylori strain 26695 (ATCC 70392) cagA and cagPAI deletion mutants were constructed using the PCR-based method as described by Chalker et al. [22] and Tan and Berg (2004) [23] in A/Prof. Ho Bow's laboratory at the Department of Microbiology (National University of Singapore, Singapore).

Culturing of the H. pylori and biofilm
H. pylori were cultured on non-selective chocolate blood agar plate and incubated in a humidified incubator with 10% CO2 at 37˚C for three days. The 3-days old H. pylori plate cultures were harvested with a sterile cotton swab and emulsified in brain heart infusion (BHI) broth (Oxoid, Hampshire, UK) supplemented with 1% β-cyclodextrin (Sigma-Aldrich, St. Louis, MO, USA) and 0.4% yeast extract (Oxoid, Hampshire, UK). After 3 days incubation, the total bacterial population of each of the cultures was standardized with the same bacterial turbidity value (bacterial turbidity = 0.042). Following which, 500 μl of bacterial suspension was then inoculated into each well of a 24-well tissue culture plate (BD Falcon, Franklin Lakes, NJ, USA). BHI broth without H. pylori served as a negative control. The growth of the biofilm was observed at the day 3, day 5, day 7 and day 14 with the amount of biofilm formed analyzed based on crystal violet staining.
At each time interval, the bacterial suspensions of the 24-well tissue culture plate were first aspirated out. An aliquot of 650μl of 0.1% crystal violet (Friendemann Schmidt, WA, Australia) was added to each well and the plate was gently swirled for 30 minutes to ensure uniform staining of the biofilm. The excess crystal violet stain was removed and the wells were washed thoroughly using 800μl sterile distilled water to remove any excess crystal violet stain. The crystal-violet stained biofilm was finally destained using 1ml of 95% ethanol (VWR Prolabo, Lutterworth)-5% acetic acid (Friendemann Schmidt, WA, Australia) which serve as destaining solution. The solution collected was measured at OD 600nm on a SmartSpec Plus spectrophotometer (Bio-Rad, Hercules, CA, USA). The amount of crystal violet absorbed by the biofilm (CV 600nm ) was determined by taking the mean absorbance value. All experiments were performed technical triplicates and in difference day of biological triplicates.
At the same time, bacterial turbidity value of each bacteria strains were taken for every time points. The bacterial turbidity value of each strain was taken in order to determine the growth rate of the strains. All experiments were performed in biological and technical triplicates and specific biofilm unit (CV 600nm /OD 600nm ) value was calculated.

Illumina Library Preparation and Sequencing
DNA was extracted from 3-day old H. pylori using the Bacterial DNA mini kit (Stratec Molecular, Berlin, Germany) according to the manufacturer's instructions. Illumina sequencing library was prepared and samples were sequenced on a MiSeq sequencer (Illumina, San Diego, CA, USA) as described in Perkins et al. [24]. Sequences used in this study have been deposited with GenBank under accession numbers listed in Table 1.

Identifying genes associated with biofilm formation
Biofilm forming ability of H. pylori standard strains J99 and 31 sequenced clinical strains were determined. Based on highest biofilm specific unit value, there strains were classified into 3 groups (good, moderate and poor biofilm formers. Cumulative frequency (%) was calculated using the formula: Where, a = previous cumulative frequency; x = number of strains which showed the particular highest biofilm specific unit value The rate of biofilm formation for each of the strains was determined by using the formula below: Rate of biofilm formation ¼ biofilm f ormation day b À biofilm f ormation day a day b À day a H. pylori genomic sequences were annotated and compared using RAST (Rapid Annotation using Subsystem Technology) (http://rast.nmpdr.org/) [28][29][30]. Sequence-based comparison among H. pylori strains was carried out using the SEED viewer of RAST with percentage similarity above 80% was used. Two-tailed Fisher's exact test was carried out using SPSS (IBM statistic 20, Armonk, New York, USA). P-value <0.05 was considered statistically significant.
Next, the presence and absence of genes-of-interest identified by comparative genomic analysis were confirmed by conventional end-point PCR using primers and conditions listed in Table 2.

Optimization of conditions for in vitro biofilm formation
H. pylori biofilm formation was observable at the air-liquid interface of wells after three days of incubation in a humidified 10% CO2 incubator at 37˚C. Optimum conditions for biofilm formation were chosen based on maximum biofilm formation without compromising bacterial growth. Comparing between different concentrations of β-cyclodextrin (0.5% to 2.0%) and fetal bovine serum (5% to 15%) as supplement to BHI broth, 1% β-cyclodextrin was shown to be optimum for biofilm formation. Biofilm formation in BHI with 1% β-cyclodextrin broth showed at least one-fold higher than BHI supplemented with fetal bovine serum (5% to 15%) and β-cyclodextrin (0.5% and 2.0%). Biofilm formation in microaerophilic condition was shown to be 3-fold higher than ambient atmospheric condition. As osmolality of the medium was increased (from original 0.5% to 1% and 2% sodium chloride), biofilm formation and specific biofilm unit were reduced. In addition, the optimum condition for both biofilm formation and growth occurred at pH 7.0. Thus, BHI broth supplemented with 0.4% yeast extract and 1% β-cyclodextrin at pH 7.0 and microaerophilic condition was determined to be optimum for the study of biofilm formation of H. pylori in this study.

Identification of genes associated with biofilm formation
The median biofilm specific unit value (cumulative frequency = 50%) was determined to be between specific biofilm unit of 1.0 and 1.1. Strains producing biofilm with specific biofilm unit <1.0 were classified as poor biofilm formers while those >1.1 were classified as good biofilm formers. Those specific biofilm units between 1.0 and 1.1 were classified as moderate biofilm formers. Twelve (37.5%) of the tested strains were classified as good biofilm formers, while another 12 (37.5%) strains were poor biofilm formers. The remaining 8 (25.0%) strains were moderate biofilm formers ( Table 3). Highest specific biofilm unit was shown to be significantly different between poor, moderate and good biofilm formers (one-way ANNOVA pvalue = 0.001). Based on rate of biofilm formation from day 0 to day 7, there was clear segregation between poor and good biofilm forming strains ( Table 4). Student's t-test revealed that there were significant differences in rate of biofilm formation between poor, moderate and good biofilm forming strains across all time point (p-value <0.05), except between poor and moderate groups at day 3 to day 5 (p-value = 0.260). While good and poor biofilm forming strains displayed relatively consistent rate of biofilm formation across 7 days, moderate biofilm forming strains varied in rate of biofilm formation across time course of study. The moderate biofilm forming strains showed either declined biofilm forming rate by day 3 or delayed optimum biofilm formation starting from day 3 onwards. The rate of biofilm formation profiles was consistent with the highest specific biofilm unit profiles across multiple H. pylori strains.
Through the comparison of annotated genes present in H. pylori genomic sequences of poor, moderate and good biofilm forming strains, with 3 hypothetical genes (Hypothetical protein K747_10375, Hypothetical protein K747_09130 and Hypothetical protein K747_06625) and 4 functional genes (flagellar protein, cag pathogenicity island protein, outer membrane protein Table 3. Classification of H. pylori strains based on highest specific biofilm unit. One-way ANOVA was used to demonstrate statistically significant differences in highest specific biofilm unit between poor, moderate and good biofilm formers (p-value = 0.001).

Strain
Highest specific biofilm unit (CV 600nm /OD 600nm ) Cumulative frequency (%) BF Group (homD) and alpha-(1,3)-fucosyltransferase) were identified to be significantly correlating with capability of forming biofilm (p-value <0.05) ( Table 5). The presences and absences of these interested genes were verified by using PCR method ( Table 5). One of the genes identified is annotated as cag pathogenicity island protein. To confirm that cag pathogenicity island (cag-PAI) genes were involved in H. pylori biofilm formation, biofilm formation by wild-type and deletion mutants of cagA and cagPAI were compared. Bacteria turbidity and biofilm formation were increasing day 3 to day 14 (Fig 2). This was consistent with observation by SEM that spiral form of H. pylori could still found in the biofilm on day 14. Deletion of cagA and cagPAI were demonstrated to reduce biofilm formation (Figs 2 and 3). The microcolonies in the biofilm were denser and bigger in wild type as compared to ΔcagA and ΔcagPAI. cagPAI knockout mutants formed sparsely distributed microcolonies at the air-medium interface (Fig 3). At higher magnification, it was observed that the extracellular matrix formed by the wild-type strain was denser than that of ΔcagA and ΔcagPAI. In addition, extracellular matrix formed by ΔcagA was film-like in structure compared to the filament-like structure form by wild-type. Extracellular matrix was scanty in the biofilm of ΔcagPAI.

Discussion
The identification of a set of genes that partition the good biofilm forming H. pylori strains from their poor biofilm forming counterparts demonstrate that biofilm formation phenotype is influenced by multiple genetic factors of the bacteria. Motility and adhesion capabilities apparently are critical for the initiation of biofilm formation (and probably the subsequent propagation of biofilm). The movement ability by different species of bacteria is linked to their ability to colonize various ecological niches, and is frequently related to pathogenesis and biofilm formation. The gene encoding for hypothetical protein K747_06625 is predicted to contain ParB-like and HNH nuclease domains. In Actinomyces oris K20, genes containing ParB-like nuclease domain has been shown to be involved in the formation of meshwork-like structures, which are found in some biofilm-producing bacteria [32]. In P. aeruginosa, par mutants have been observed to be impaired in motility suggesting a direct or indirect role of Par proteins in regulation of these processes [33]. ParA/ParB family of proteins together with the centromere-like DNA sequence parS is involved in chromosome and plasmid partitioning during bacterial cell division. In the absence of ParB, ParA is most probably in its dimeric state bound with ATP proficient to bind non-specifically to DNA [34]. Furthermore, par mutation also induced genes induced involve in c-di-GMP turnover and signaling [33]. Cyclic-di-GMP (cyclic diguanylate) is an important messenger ubiquitous in bacterial cells controlling various processes, e.g. switch between the motile planktonic and biofilm lifestyles of bacteria, virulence of animal and plant pathogens, antibiotic production, progression through the cell cycle and other cellular functions [34].
Another gene-of-interest encodes for a flagellar protein, which is also involved in bacterial motility. It is believed that the presence of flagella provide motility that aids kinetics of biofilm formation [35]. Motility provides by flagella enhance the recruitment of planktonic cells to the biofilm [36]. Besides, the adhesive properties of flagella proteins might also promote biofilm formation in H. pylori. This was supported by previous studies that demonstrate that Campylobacter jejuni flagella are needed for the initiation of biofilm formation by mediating adhesion on the surface [37]. In one study, microcolonies were formed on a coverslip with flagella forming bridges between organisms [38]. During initial reversible attachment, flagella permit individual planktonic bacteria to swim toward an appropriate biotic or abiotic surface. Subsequently, during irreversible attachment, these flagella are lost and adhesive organelles became elaborated [39].
Biofilm forms a protective layer for Gram-negative bacteria to prevent attack by host immune system against one of its cell wall component, lipopolysaccharide (LPS) [40]. H. pylori alpha-(1,3)-fucosyltransferase, one of the many genes that involved in H. pylori LPS synthesis [41], is also predicted to contain the biofilm formation protein (YliH/bssR) domain, which has been shown to be induced in Escherichia coli biofilms [42]. In addition, H. pylori, alpha-(1,3/ 4)-fucosyltransferase is required for fucosylation of both type I (Le a ) and type II (Le x ) Lewis antigens [43]. The adhesive properties of Lewis X antigen enable first colonist to adhere to the surface of the liquid through weak adhesion force in the formation of biofilm [44].
One of the proteins found to be prevalent with good biofilm forming strains encodes for a cag pathogenicity islands protein. CagA protein, encoded by cag pathogenicity islands, has been identified to be induced in H. pylori biofilms [45]. CagA, which serves as a virulent factor of H. pylori, is injected into the host through the type IV secretion system and affect the individual [46]. Type IV secretion system-mediated transfer is essential in direct cell-cell contact [47]. It is believed that this direct cell-cell contact can control the biofilm behavior in H. pylori [48]. Furthermore, CagE, another protein encoded by the cag pathogenicity islands, has also been reported to be involved in the H. pylori biofilm formation [15,48]. cagE encodes for a cytoplasmic ATPase to allow the translocation of CagA protein that induces the gastric epithelial cells to secrete interleukin-8 (IL-8) and activate many intracellular signaling pathway [25]. Thus, the cag pathogenicity island may play an important role in H. pylori biofilm formation. Besides its role in bacteria-host interaction, CagA and the cag pathogenicity island may also have a role to play in bacteria-bacteria interaction in H. pylori biofilm formation.

Conclusion
Using a comparative genomics approach, data from this study suggest that biofilm formation in H. pylori might be influenced by multiples genes. Good and poor biofilm forming strain may differ genetically in terms of motility, adhesion and bacteria-bacteria interactions, which are important during the initiation of biofilm formation.