Intensive formation of coccoid forms as a feature strongly associated with highly pathogenic Helicobacter pylori strains

The variability of Helicobacter pylori morphology and the heterogeneity of virulence factors expressed by these bacteria play a key role as a driving force for adaptation to the hostile stomach environment. The aim of the study was to determine the relationship between the presence of certain genes encoding virulence factors and H. pylori morphology. One reference and 13 clinical H. pylori strains with a known virulence profile (vacA, cagA, babA2, dupA, and iceA) were used in this study. Bacteria were cultured for 1 h and 24 h in stressogenic culture conditions, i.e., serum-free BHI broths at suboptimal conditions (room temperature and atmosphere, without shaking). H. pylori cell morphology was observed by light and scanning electron microscopy. The vacA polymorphism and the cagA and babA2 presence were positively correlated with the reduction in cell size. Exposure to short-time stressogenic conditions caused more intense transformation to coccoid forms in highly pathogenic H. pylori type I strains (35.83% and 47.5% for type I s1m2 and I s1m1, respectively) than in intermediate-pathogenic type III (8.17%) and low pathogenic type II (9.92%) strains. The inverse relationship was observed for the number of rods, which were more common in type III (46.83%) and II (48.42%) strains than in type I s1m2 (19.25%) or I s1m1 (6.58%) strains. Our results suggest that there is a close relationship between the presence of virulence genes of H. pylori strains and their adaptive morphological features.


Introduction
Helicobacter pylori is a microaerophilic, spiral, Gramnegative rod capable of effectively colonizing the gastric mucosa (Marshall and Warren 1984). Bacteria spread directly from person to person and are responsible for the development of a wide range of gastrointestinal diseases, including chronic (active) gastritis, gastric and/or duodenal ulcers, gastric adenocarcinomas, and mucosa-associated lymphoid tissue lymphomas (Malfertheiner et al. 2017). The presence of gastrointestinal pathologies and their severity is influenced by the type of bacterial strains and individual predispositions of the host, which determine the switching of H. pylori between commensalism and pathogenicity (Larussa et al. 2015). The long-lasting equilibrium between the host and H. pylori, covering at least 60,000 years, has enabled these microbes to develop evolutionary adaptations (Atherton and Blaser 2009;Talebi Bezmin Abadi 2017). The ability to avoid immune clearance and rapid response to changing environmental conditions are key determinants of the persistent host colonization by H. pylori (Larussa et al. 2015;Talebi Bezmin Abadi 2017).
One way of adaptation to environmental changes is the capability of H. pylori to undergo a morphological transformation. This microorganism occurs in two morphological forms, i.e., spiral-live, culturable, and coccoid-viable but nonculturable. The morphological transition from spiral to spherical form is observed under suboptimal environmental conditions, such as aerobiosis, temperature or pH changes, prolonged culture, or exposure to antibiotics or proton pump inhibitors (Loke et al. 2016). In proteomic studies, Loke et al. (2016) found that coccoid forms have high levels of proteins participating in DNA replication and carcinogenesis initiation and should therefore be regarded as highly virulent. The ability of H. pylori to produce elongated/ filamentous forms was also demonstrated, both in vitro (in amiA mutants and also after exposure to aztreonam, elevated NaCl concentrations or during oxygenated CO 2 -depleted culture) and in vivo (Takeuchi et al. 2006;Chaput et al. 2006;Park et al. 2011;Shao et al. 2012;Singh et al. 2015). It is postulated that increased cell filamentation may be a property of commensal bacteria, while the reduction of elongation is a hallmark of pathogenicity (Rossetti et al. 2013).
The heterogeneity of H. pylori strains and the variability of virulence factors expressed by these bacteria play a key role in adaptation to the stomach environment of the host (Odenbreit et al. 2009). The production of CagA (cytotoxin-associated gene A) is a feature of more virulent H. pylori strains, because this oncoprotein is capable of modulating host cell physiology (initiation of increased motility and elongated shape, the so-called hummingbird phenotype), induction of carcinogenesis, and IL-8dependent initiation of inflammation. The vacuolating cytotoxin A (VacA) is a 140-kDa polypeptide responsible for the host tissue damage associated with the formation of anion-selective pores in cell membranes, mitochondria disruption, and eukaryotic cell death (Roesler et al. 2014;Amieva and Peek 2016). The VacA toxicity is determined mainly by two variable regions, called signal (s1 and s2) and middle (m1 and m2). The vacAs1 strains have a higher cytotoxic activity and are responsible for a greater risk of developing gastrointestinal diseases (especially vacAs1m1) than non-vacuolating vacAs2 strains (Roesler et al. 2014). Based on the expression of these two virulence factors, H. pylori strains were classified into three groups, i.e., with the highest virulence potential, cagA + /vacAs1 (type I) and the lowest virulence, cagA − /vacAs2 (type II), and strains with intermediate virulence with genotype, cagA − /vacAs1 or cagA + /vacAs2 (type III) (Xiang et al. 1995). Blood group antigen-binding adhesin A (BabA) is a protein associated with adhesion to the gastric mucosa, by binding to various components of the host (Ansari and Yamaoka 2017). Duodenal ulcer promoting gene A (DupA) is one of many important virulence factors often referred to as important in the promotion of gastrointestinal diseases. The production of this protein is coupled positively with the secretion of proinflammatory cytokine IL-8 (Queiroz et al. 2011). Another important virulence determinant involved in the pathogenicity of H. pylori is induced by contact with epithelium gene A (IceA). Production of this protein is stimulated by contact with the gastric epithelium, which contributes to the increase in IL-8 secretion (Shiota et al. 2012).
The aim of the study was to determine the relationship between the presence of certain genes encoding virulence factors and morphological changes of H. pylori strains.

Bacterial strains and growth conditions
Virulence genes of all 13 clinical H. pylori strains used in this study (Table 2) have been identified and published previously by Biernat et al. (2010Biernat et al. ( , 2014. In these reports, we have shown that H. pylori with specific virulence factors (cagA, vacA, babA2, dupA, and iceA in Table 2) were associated with the induction of gastric mucosa inflammation and tissue destruction. Detection of vacA, iceA1, and babA2 was carried out using multiplex PCR, whereas the cagA and iceA2 genes were detected by single PCR. H. pylori strains used in the study were formerly isolated from antral biopsy specimens of patients with gastritis. In addition, the reference strain J99 (ATCC 700824) was used as a control strain. The genome sequence of H. pylori J99 is available at GenBank with the accession number AE001439.
Culture methods of H. pylori strains were performed according to Macegoniuk et al. (2017) with minor modifications. The frozen bacterial suspensions of clinical and reference H. pylori strains were stored at − 70°C in tryptic soy broth (Oxoid) with 15% glycerol. After reviving from frozen stocks, bacterial strains were pre-cultured onto Columbia agar (Difco) enriched with 7% hemolysed horse blood and H. pylori-selective supplement (Oxoid). Examined strains were incubated under microaerophilic conditions (Genbox microaer kits, BioMerieux) at 37°C for 3 days and then sub-cultured on the same fresh media and incubated for 3 more days under the same conditions. In all of the following experiments, each H. pylori strain was passaged for only two or three times to reduce the risk of phase-variable changing of virulence genes (Odenbreit et al. 2009).

Preparation of cultures for light microscopy
H. pylori strain J99 was used to find the best environment that yielded the lowest average cell length and highest percentage of coccoid forms. Bacteria from solid cultures were inoculated into 2 ml of fresh brain heart infusion (BHI, Sigma) broths to obtain OD 600 of 0.3 (early log phase) or 1 (stationary phase). Each bacterial culture was then incubated for 1 h and 24 h under specified conditions: BHI with 7% fetal bovine serum (Gibco) at optimal conditions (37°C, microaerophilic, 100 rpm), BHI without the serum at optimal conditions, BHI with the serum at suboptimal conditions (room temperature and atmosphere, without shaking), and BHI without the serum at suboptimal conditions. Cell length, shape determination, and image acquisition Light microscopy analysis was performed similarly as described elsewhere (Rossetti et al. 2013) with small modifications. For each time point and strain, 50 μL of bacterial suspension was dropped onto a coverslip, stained by Gram's method and observed under an Olympus BX50 microscope (Olympus Optical) using a × 100 oil immersion objective with numerical aperture of 1.3. The measurement of cell length was performed manually using the CellTool software. The average length of bacterial cells was counted using two independent experiments. From each experiment, three slides were taken and 50 bacterial cells per slide were counted (n = 300 cells/strain for each time point).

Scanning electron microscopy
In order to precisely determine the morphological structure of H. pylori cells and to confirm observations made using light microscopy, scanning electron microscopy (SEM) was used. The morphology of two H. pylori strains with the greatest variation in the virulence profile, namely 6171/T.II (vacAs2m2, cagA − , dupA − , iceA − , babA2 − ) and J99 (vacAs1m1, cagA + , dupA + , iceA2 + , babA2 + ), was analyzed by SEM. These strains from solid cultures were inoculated into 2 ml of fresh BHI (Sigma) broths to obtain OD 600 of 0.3 (early log phase) and were cultured for 1 h in serum-free BHI broths at suboptimal conditions (room temperature and atmosphere, without shaking). After this time point, microbes in liquid media were prefixed in 2.5% glutaraldehyde for 24 h and p e l l e t e d b y c e n t r i f u g a t i o n ( 6 0 0 g , 5 min) in microcentrifuge tubes. Pellets were washed a total of three times in 0.1 M cacodylate buffer (pH 7.4) after each step cetrifugating bacteria again (600g, 5 min). Microbes were treated with a series of ethanol concentrations (50, 70, 90, 99.99%) and pelleted by centrifugation (600g, 5 min). After dehydration in a graded series of ethanol, the specimens were coated with 10 nm of gold. The samples were examined with a scanning electron microscope (Auriga 60, Zeiss) operating at 2 kV.

Statistical analysis
From two independent experiments, six slides per strain were taken and results were expressed as the mean ± standard deviation of the mean (SDM). All data were analyzed by STATISTICA v. 10.0. The distribution normality of continuous variables was calculated using the Shapiro-Wilk test. As data from bacterial cell sizes were not normally distributed, the correlation between variables was analyzed by the nonparametric Spearman's rho. The differences in statistical significance between bacterial cell sizes at 1-h and 24-h incubation periods were analyzed by the Wilcoxon test. Cell morphology type frequencies were estimated by the Pearson χ 2 analysis. The reduction of bacterial cell size after 1-day incubation period for H. pylori strains with specific virulence factors was analyzed by the Mann-Whitney U test and the Kruskal-Wallis test. The differences were found to be statistically significant at a p ≤ 0.05.

Optimization of conditions inducing H. pylori morphological changes
After 1-h and 24-h incubation periods, there were no significant differences in the cell length distribution of H. pylori J99 in broths with different initial densities of bacteria, but at a density of OD 600 = 1 bacterial cells were intensively autoaggregated (data not shown). Therefore, further studies were carried out using a density of OD 600 = 0.3 to accurately determine the size and shape of a single bacterial cell.
It was observed that the reduction of H. pylori J99 in size and the changes in morphology were most prominent in serum-free BHI media at suboptimal culture conditions (Table 1); these counted for 1.02 ± 0.02 μm (56.3% coccoid forms and 42.3% short rods) and 0.9 ± 0.04 μm (71% coccoid forms and 29% short rods) at 1-h and 24-h incubation, respectively. Average bacterial cell length was highest in BHI broths with the serum at optimal culture conditions and equaled 2.26 ± 0.02 μm (70.3% rods and 26.7% short rods) after 1 h and 2.16 ± 0.04 μm (60% rods and 34.7% short rods) after 1 day of incubation periods. Other culture conditions contributed to the formation of intermediate-sized cells, i.e., 1.81 ± 0.04 and 1.45 ± 0.03 μm (suboptimal culture conditions, BHI with the serum) and 1.47 ± 0.03 and 1.06 ± 0.02 μm (optimal culture conditions, serum-free BHI) at 1-h and 24-h incubation, respectively.
For this reason, bacterial inoculations (OD 600 = 0.3) were incubated in serum-free BHI broths at suboptimal culture conditions for 1 h and 24 h in the studies of the remaining H. pylori strains.

Analysis of cell length variability
Light microscopy observation of tested H. pylori strains at 0 h showed no significant differences in average cell length (p > 0.05), which ranged from 2.49 ± 0.03 to 2.1 ± 0.03 μm (Table 2). After 1-h and 24-h incubation, the average cell length of tested H. pylori strains changed profoundly and ranged between 2.44 ± 0.05 and 1.02 ± 0.02 μm and between 1.5 ± 0.05 and 0.9 ± 0.04 μm, respectively. Despite the gradual gradation in average bacterial cell length, coupled with the increasing number of virulence genes, these differences were not statistically significant (p > 0.05) (data not shown).
The statistical analysis of the reduction in bacterial cell size with a specific virulence factor was found to be significant for vacA polymorphism (p = 0.0039) and the presence of cagA (p = 0.0113) and babA2 (p = 0.0179). This significance was not observed for dupA (p = 0.8465) and iceA (p = 0.9485) genes.

Analysis of cell shape variability
For analyzing the differences in the morphological properties, tested strains were classified into four groups, namely type II, type III, type I s1m2 (cagA + , vacAs1m2), and type I s1m1 (cagA + , vacAs1m1). After 1-h incubation, the presence of several times higher numbers of spherical forms in H. pylori type I strains (35.83% and 47.5% for type I s1m2 and I s1m1, respectively) than in type III (8.17%) (p < 0.000) and II (9.92%) (p < 0.000) strains was demonstrated (Fig. 3). The inverse relationship was observed for the number of rods, which were more common in type III (46.83%) and II (48.42%) strains than in type I s1m2 (19.25%) (p < 0.000) or I s1m1 (6.58%) (p < 0.000) strains (Fig. 3). Despite no statistically significant differences in the morphological properties of H. pylori type I s1m1 and I s1m2 strains (p = 0.141), the existence of a feature distinguishing both types was noted. In the case of H. pylori type I s1m1 strains, filamentous/ elongated forms were never observed, while in H. pylori type I s1m2 strains, these forms were present and counted for 3% of all examined cells (Table 3).

Discussion
H. pylori in the natural habitat is mainly present as spiral forms, whereas culture in artificial in vitro conditions or exposure to suboptimal environmental factors is accompanied by morphological transformation into spherical or filamentous forms (Azevedo et al. 2007;Cellini et al. 2008;Park et al. 2011;Shao et al. 2012;Faghri et al. 2014). This pleomorphic nature is not specific to H. pylori only, since the ability to In column Bstressogenic factors^: + indicates the presence of tested factor during culture of microbes; free space indicates the absence of tested factor during culture of microbes, and the presence of 7% fetal bovine serum in BHI media and optimal culture conditions (37°C, microaerophilic, 100 rpm), respectively  undergo morphological transformation has been observed for many Gram-negative rods, i.e. Campylobacter jejuni, Enterobacter aerogenes, Escherichia coli, Klebsiella pneumoniae, Legionella pneumophila, Proteus mirabilis, Salmonella enterica, and Vibrio cholerae (Krogfelt et al. 1993;Justice et al. 2006Justice et al. , 2014Chaiyanan et al. 2007;Krebs and Taylor 2011;Horvath Jr et al. 2012;Ramamurthy et al. 2014;Ghaffar et al. 2015).
The current study has found that adaptational changes in cell morphology were closely related to the virulence profile of H. pylori strains. Rossetti et al. (2013) postulated that increased cell size (including the intensification of the filamentation process) is potentially a feature of commensal microbes, while reduction of elongation intensity can be traced as a pathogenic hallmark. This hypothesis is in agreement with the results obtained in our study. We noted that lower virulence strains had longer cells than highly virulent strains. Moreover, the reduction in cell size of the tested H. pylori strains during short-term exposure to suboptimal conditions was positively correlated with the presence of certain virulence genes, e.g., vacA, cagA, and babA.
It seems that strains with higher adaptability (and higher virulence) are capable of achieving reduced cell size in a shorter time during exposure to stressogenic conditions. The results in this study, showing that H. pylori strains with a higher virulence potential are capable of morphological adaptations in a short period to adverse environmental conditions, are consistent with reports by Vitoriano et al. (2011). The authors undertook co-culture experiments of peptic ulcer disease (PUD) and nonulcer dyspepsia (NUD) H. pylori strains with eukaryotic cell lines. It was demonstrated that after 1 day of co-culture, NUD strains possessed their original spiral shape, whereas PUD strains obtained spherical form as early as after 12-h co-incubation time. It has been suggested that this phenomenon is caused by the more severe destruction of eukaryotic cells and necrosisdependent changes in environmental conditions and thus faster transformation of H. pylori PUD strains into coccoid forms. There are also scientific reports suggesting that minimization of bacterial cell size is a protective feature that contributes to the reduction of the cellular surface exposed to the immune system clearance (complement deposition or opsonophagocytosis) (Dalia and Weiser 2011;Veyrier et al. 2015). Their suggestions are within the scope of our observations showing that under suboptimal conditions (including nutrient depletion or exposure to the oxygenic and thermic stress), the process of reducing cell size and creating short rods or coccoid forms may be a potentially adaptive feature.
The direct molecular mechanism underlying the relationship between the presence of virulence genes and H. pylori morphology is not known. The presence of H. pylori virulence genes, however, does not have to exert   There are many proteins that become active or inactive depending on various pH environments. In a similar fashion, the phenotype of H. pylori may be influenced by other environmental factors. The infection caused by H. pylori strains, especially those producing many virulence factors, contributes to the formation of the inflammatory altered gastric environment, which is characterized by occurrence of numerous immune cells (neutrophils, macrophages, and T cells), proinflammatory cytokines (IL-1β, IL-6, IL-8, and TNF-α), chemokines, and matrix metalloproteinases (Ma et al. 2016). Survival in the inflamed environment requires the presence of multiple genes that protect against the immune clearance, which are most likely co-expressed with the expression of virulence genes, such as cagA, vacA, or babA. Some of these genes, protecting against immune attacks and the presence of unfavorable environmental conditions, could encode information about factors responsible for rapid morphological changes in H. pylori.  Fig. 3 The impact of virulence type on the morphology distribution of H. pylori strains after 1-h incubation period in stressogenic culture conditions. Cell morphology measured from light microscopy images reports as percent of the total bacterial population examined. Data combined from two independent bacterial cultures incubated in serumfree BHI broths at suboptimal culture conditions for 1 h period (n = 300 cells/strain). Results are given as mean ± SDM (n = 6)

Limitations
Our study has some limitations. Firstly, the analysis was conducted with selected virulence genes (cagA, vacA, babA2, dupA, iceA). The selection of these virulence determinants was dictated by literature data indicating the significance of the abovementioned genes in the development of gastrointestinal disorders (Zambon et al. 2003;Torres et al. 2009;Shiota et al. 2010Shiota et al. , 2012. However, it cannot be ruled out that the presence of other genes, not investigated in this study, also influences H. pylori morphology. On the other hand, this method seems to be surprisingly sensitive because it detected differences in cell morphology in relatively low amounts of tested virulence determinants. Secondly, the analysis was performed with s1/2 and m1/2 vacA alleles only. The relevance of the vacA intermediate (i) region in the cytotoxic activity of VacA has been demonstrated (Rhead et al. 2007). In the current study, H. pylori strains were categorized into three groups, namely types I, II, and III. This classification does not include the i-region, but only the s-and m-regions (Xiang et al. 1995). Hence, the impact of this region has not been investigated. The link between the morphology of H. pylori cells and the presence of virulence genes needs to be explored in different populations.

Conclusion
Based on microscopic observations, we demonstrated that the morphology of H. pylori strains might be influenced by the virulence profile. Less virulent H. pylori strains had a lower potential for the formation of coccoid forms than highly pathogenic strains. Furthermore, strains with lower pathogenicity were characterized by the presence of filamentous/ elongated forms, not observed in highly virulent strains.