Understanding the dimorphic lifestyles of human gastric pathogen Helicobacter pylori using the SWATH-based proteomics approach

Helicobacter pylori may reside in the human stomach as two morphological forms: the culturable spiral form and the viable but non-culturable (VBNC) coccoid form. This bacterium transforms from spiral to coccoid under in vitro suboptimal conditions. However, both spiral and coccoid have demonstrated its infectivity in laboratory animals, suggesting that coccoid may potentially be involved in the transmission of H. pylori. To determine the relevance of the coccoid form in viability and infectivity, we compared the protein profiles of H. pylori coccoids obtained from prolonged (3-month-old) culture with that of 3-day-old spirals of two H. pylori standard strains using SWATH (Sequential Window Acquisition of all Theoretical mass spectra)-based approach. The protein profiles reveal that the coccoids retained basal level of metabolic proteins and also high level of proteins that participate in DNA replication, cell division and biosynthesis demonstrating that coccoids are viable. Most interestingly, these data also indicate that the H. pylori coccoids possess higher level of proteins that are involved in virulence and carcinogenesis than their spiral counterparts. Taken together, these findings have important implications in the understanding on the pathogenesis of H. pylori-induced gastroduodenal diseases, as well as the probable transmission mode of this bacterium.

be associated with chronic active gastritis, coccoid H. pylori is more frequently associated with chronic inactive gastritis in symptomatic adult patients 4 . Symptomatic pediatric patients and pediatric patients with recurrent epigastric pain have 4-fold higher seroprevalence for coccoid antigen compared to that for spiral antigen indicating a possible infective role of the coccoid form of H. pylori in these patients 10 .
H. pylori undergoes morphological conversion from spiral to coccoid when cultured under mild sub-optimal growth conditions 11 . These conditions include aerobiosis 12,13 , acidic and alkaline pH [12][13][14] , high temperature 15 , extended incubation 10,13 , treatment with a proton pump inhibitor 13 and treatment with antibiotics 16 . Coccoid form of H. pylori has been shown to preserve its RNA, DNA and structural components for at least 3 months 16 . Extended incubation is gradual and least stressful for the bacterium as compared to other methods of inducing formation of coccoids. Furthermore, this approach is probably more relevant for H. pylori as the organism is known to colonize the human stomach for life unless eradicated 17 . H. pylori coccoid form expresses high-molecular weight antigens that are not expressed by the spiral form 18 indicating that H. pylori coccoids that are formed under prolonged culture may still be viable and immunogenic. However, no coccoid-specific protein has been identified using two-dimensional gel electrophoresis 19 limiting our current understanding of the coccoid lifestyle of H. pylori. The lack of known coccoid-specific proteins also hinders the development of H. pylori coccoid-specific detection assays. The main hurdles of using gel-based to analyze the proteome of H. pylori coccoid sample include low protein content and the accumulation of metabolic or degradation byproducts that can interfere with gel-electrophoresis. Thus, there is a need to apply a more sensitive and gel-free method for proteome comparison between H. pylori spiral and coccoid forms. In this study, liquid chromatography-mass spectrometry (LC-MS) with the capability to efficiently differentiate the two proteomes was used. To address this issue, we compared the protein compositions of 3-day old spiral cells in liquid cultures to that of 3-month old coccoid cells grown in stationary liquid culture of two strains of H. pylori, NCTC 11637 and J99, using the LC-MS-based SWATH (Sequential Window Acquisition of all Theoretical mass spectra) approach. H. pylori NCTC 11637 is the type strain 20 while J99 is one of the earliest strains that were fully sequenced 21 . Both H. pylori strains have been used in laboratory experiments. The SWATH strategy makes use of the fragment ion spectral libraries to mine the complete fragment ion maps generated based on a data-independent acquisition method. The method enables the detection of large number of proteins reproducibly at near quantitative accuracy in a single measurement 22 . Unlike traditional multiple reaction monitoring (MRM)-based quantitation method, where the target protein has to be known before data acquisition during to method setup. SWATH-MS is for unbiased untargeted quantitation of proteins using MS. This provides an added advantage for this study as we are interested to understand the global differential expression using a non-label strategy. Furthermore, this data-acquisition method will be most productive for future retrospective mining of quantitative data when potential target candidates of interest have been identified.

Results and Discussion
In this study, H. pylori coccoids from 3-month-old culture of two H. pylori strains were harvested for proteome analysis in comparison with the spirals from 3-day-old cultures ( Fig. 1). An average of 914 proteins per sample was identified from the spiral and coccoid protein extracts of H. pylori NCTC 11637 and J99 strains with < 1% global protein false discovery rate and 99% confidence threshold (Supplementary data). Proteins showing statistical significant difference by Student's t-test (p-value < 0.05) using MarkerView software were annotated and classified based on KEGG (Kyoto Encyclopedia of Genes and Genomes) Orthology (KO) using STRING version 9.1 (www.string-db.com) 23 . For both H. pylori strains, 81 proteins were significantly different (p-value < 0.05) with coccoid/spiral fold-change < 1.0 (Supplementary Table S1). On the other hand, 66 proteins showed significant difference in expression (p-value < 0.05) for coccoid forms in comparison to their spiral counterparts with coccoid/spiral fold-change > 1.0 (Supplementary Table S2).
Metabolism. Among those proteins that were significantly less abundant in coccoids when compared to the spirals, 35 were enzymes involved in diverse metabolism pathways (Supplementary Table S1). These include enzymes that are involved in core carbon metabolism (Fig. 2), amino acid metabolism, nucleotide metabolism, lipid metabolism and the metabolism of vitamins and co-factors (Supplementary Table S1). Conversely, the abundances of 13 metabolic enzymes were increased in coccoids (Supplementary Table S2). Based on the comparison of the relative abundance of key metabolic proteins expressed in the two morphologically differentiated forms of H. pylori, coccoids were less metabolically active than the spirals but maintained basal levels of these enzymes possibly paving the way for its survival over long period. It is well established that reduced metabolism is a means of survival strategy in living cells 24 . Though low in metabolic activity, these coccoids remain viable 7 , or commonly known as the viable but non-culturable (VBNC). It is therefore poignant to note that the 3-month old cultures did not yield any growth on sub-culturing in fresh liquid or solid media but proteome composition of the coccoid form shows the continuous presence of metabolic proteins, rightfully qualified its VBNC nature.
Endospore forming bacteria maintain longevity by means of forming the metabolically dormant endospores as a survival strategy 25 . Similarly, H. pylori, in the form of coccoid, may actively restrict its metabolic activity to basal level in the face of adversity 26,27 . As we know, the human stomach is a very dynamic environment with continuous peristalsis and shedding of its overlying mucin layer and gastric epithelial cells 28 . In order to chronically persist in the human stomach 3,4 , it is only rational that the bacterium must maintain a basal level of replication and multiplication to sustain its presence in the host stomach. Thus, the coccoid form may be considered as an example of 'longevity as a solution to adversity' in the bacterial world. (The phase "longevity as a solution to adversity" is borrowed from the book "Life" of Martha Holmes and Michael Gunton 29 , in which they used to describe the survival of several thousand years old bristlecone pine in the harsh environment in the White Mountains of western USA).
Scientific RepoRts | 6:26784 | DOI: 10.1038/srep26784 DNA replication. Whether coccoid H. pylori is capable of DNA synthesis and repair remains debatable. Bode et al. 16 had detected newly synthesized DNA among coccoid H. pylori, together with polyphosphates that can serve as energy and phosphorus source permitting basal level of endogenous metabolism to preserve the integrity of nucleic acid 16 . Similarly, Hua and Ho (1998) had also revealed that the conservation of DNA composition in H. pylori of various ages in the course of morphological changes. On the contrary, Kusters et al. 30 claimed that coccoid were merely morphologic manifestation of bacterial cell death with reduced quantity and integrity DNA and RNA 30 . Our proteome analysis of the coccoids revealed that the abundances of DNA gyrase subunit A (GyrA) and B (GyrB), which are involved in DNA replication in bacteria, were significantly increased as compared to the spirals (Supplementary Table S2). These essential enzymes regulate the conformational changes in DNA topology by catalyzing the concerted breakage and rejoining of DNA strands during normal cellular growth.
Homologous recombination plays important role in maintaining integrity of the genome. Recombinase A (RecA) was elevated in H. pylori coccoid (Supplementary Table S2) suggesting that RecA was expressed at high level in order to repair DNA damage or facilitate recombination, which is an important function even in its coccoid form. H. pylori RecA has been reported to be highly expressed in both exponential and stationary phase as part of the constitutive DNA damage adaptation system 31 . In addition to the constitutive DNA damage adaptation system, H. pylori also relies on other recombinational repair proteins (RecN and RecR) for repairing DNA damage induced by oxidative and acid stresses 32,33 . The abundance of RecN and RecR decreased in coccoid compared to spiral (Supplementary Table S1) suggesting that H. pylori coccoid may have limited ability to repair DNA damage induced by oxidative and acid stresses. This augurs well with the fact that coccoid maintains low metabolic activity to avoid undue stresses, thus keeping it viable but not culturable. Alternatively, the coccoid form may be less exposed to stress-induced DNA damage and thus the need for RecN and RecR is reduced as compared to its spiral counterpart.
Environmental adaptation. The abundances of nickel responsive regulator (NikR) and two putative transcriptional regulators were increased in coccoid (Supplementary Table S2). NikR is a nickel-dependent outer membrane protein that regulates the expression of multiple genes by binding to DNA, as well as controlling enzymes that require nickel as co-factor (e.g. urease and hydrogenase) by regulating nickel homeostasis 34,35 . Thus, it is not surprising that the abundance of urease accessory protein (UreG), a specific nickel-dependent GTPase 36 , was also increased in coccoid (Supplementary Table S2). However, urease structural subunit B (UreB) and accessory protein (UreF) were reduced in coccoid (Supplementary Table S1), which explained the observation of reduced urease activity of this morphological form.
On the other hand, the abundance of response regulator (CheY) decreased in coccoid (Supplementary Table S1) while that of another response regulator jhp0403 increased in coccoid (Supplementary Table S2). These findings indicate that the coccoid form may possess an active response or ability of adaptation by the bacterium to nutrient depletion stress in a prolonged culture. These CheY response regulators or outer membrane proteins R (OmpR) are essential for chemotaxis and persistent colonization of the gastric mucosa, as well as a part of the stress-responsive operon 37 . These response regulators may also be involved in regulating the physiology of H. pylori in response to environmental stimuli and stress, and important in facilitating adaptation to the different environments of the gastric mucosa. CheY has been associated with repression of biofilm formation or dispersion in Vibrio species 38 . It is tempting to speculate that CheY may have a significant role in the survival of H. pylori in extra-gastric environment.
Rod shape-determining protein (MreB) was enhanced in coccoids (Supplementary Table S1). Actin-like MreB is not involved in the maintenance of cell shape, but affects the progression of the cell cycle in H. pylori showing that MreB is needed for cell division 39 . The enhancement of MreB suggests it may play a role in the survival of coccoids. In addition, the abundance of cell division inhibitor (MinD) was increased in coccoids (Supplementary  Table S2). In H. pylori, MinD interacts with MinC to play a role in maintaining proper cell morphology and cell division as minC mutant has been demonstrated to form filamentous cells indicating a deficiency in ability to divide 40 . Figure 2. Core carbon metabolism pathway. Glucokinase (Glk), fructose-1,6-biphosphate (Fbp), fructosebiphosphate aldolase (Fba), phosphogluconate dehydratase (Edd), 2-keto-3-deoxy-6-phosphogluconate aldolase (Eda), phosphoenolpyruvate synthase (PpsA), biotin carboxyl carrier protein (AccB), bifunctional aconitate hydratase 2/2-methylisocitrate dehydratase (AcnB), type II citrate synthase (GltA) and fumarate hydratase (FumA) were enzymes of the core carbon metabolism that were significantly reduced (p-value < 0.05) in H. pylori coccoids compared to spirals. On the other hand, phosphoglycerate kinase (Pgk) was significantly elevated (p-value < 0.05) in coccoids. Each bar on the graph represents mean of triplicates and y-axis represents relative abundance (i.e. total sum of the areas under the peaks). Error bar represents standard deviation.
Scientific RepoRts | 6:26784 | DOI: 10.1038/srep26784 Virulence factors. The abundance of tumor necrosis factor-alpha (TNF-alpha)-inducing protein (Hps) was found to be increased in coccoid compared to spiral of H. pylori culture (Supplementary Table S2). Hps binds to nucleolin on cell surface and is incorporated into cytosol and nucleus, where it strongly induces expression of TNF-alpha and chemokine genes mediated through NF-kappaB activation, resulting in tumor development 41,42 . Hps is an essential factor in H. pylori inflammation and cancer microenvironment in the human stomach 43 . Furthermore, Hps could possibly contribute to persistent survival of the bacterium inside the stomach by virtue of its ability to suppress acid production through the induction TNF-alpha and to induce of macrophage apoptosis. The presence of these immune response activators in coccoid further stresses the probable viability of this morphological form of H. pylori.
The Sec dependent pathway and type IV secretion system (T4SS) are mechanisms used by bacteria to translocate proteins across their cytoplasmic membrane. H. pylori possesses a functional Sec machinery 44 . The abundance of putative inner membrane protein translocase component (YidC) of the Sec-dependent pathway increased in H. pylori coccoid (Supplementary Table S2). In addition, the abundances of proteins of the T4SS (CagE and CagV) also increased in coccoid (Supplementary Table S2). CagT, which is an essential structural component of the cag PAI apparatus, acts as a chaperone-like protein to promote the translocation of CagA, a bacterial oncoprotein, into host epithelial cells 45 . Consistent with our finding, antibiotic-induced coccoid forms of two clinical isolates has also been demonstrated to express higher rate of cagE mRNA than their spiral counterparts 46 . These results suggest that H. pylori coccoid is not a passive entity but may actively infect the human by the expression of various virulence genes over a long period of time in the stomach and probably plays a role in chronic as well as severe gastric diseases.
H. pylori coccoid has been observed to be closely associated with severely damaged gastric mucous cells 3 . This suggests that H. pylori coccoid can be as virulent, if not more, than the spiral which has been directly or indirectly responsible for greater host cell destruction. Alternatively, the conversion from spiral to coccoid may be the bacterium's response to host immune system thereby inflicting damage to the gastric mucosa. Our findings are in favor of the former explanation as the abundance of various proteins conferring virulence increased in coccoid when compared to the spiral form, suggesting that H. pylori coccoid may be more virulent than spiral in contributing to severe pathogenesis (gastric cancer).
In summary, we have demonstrated that the protein profile of H. pylori coccoid from prolonged (3-month old) culture suggests that despite its low metabolic activities, might still preserve its ability to replicate and multiply as coccoid without reverting to spiral form. Most interestingly, these data also indicate that H. pylori coccoid may be potentially more virulent than its spiral counterpart. Therefore, it is necessary to further investigate the roles of H. pylori coccoid in pathogenesis, as well as in environmental transmission.

Methods
H. pylori Cultures. H. pylori NCTC 11637 (ATCC 43504; American Type Culture Collection, USA) and J99 strains (ATCC 700824) were grown in 100 ml bottles containing 10 ml of brain heart infusion broth (Oxoid, UK) supplemented with 0.4% yeast extract (Oxoid) and 1% β -cyclodextrin (Sigma-Aldrich, USA) in a humidified 10% CO 2 incubator under stationary conditions. The spiral form was harvested after 3 days while coccoid form was harvested after 3 months. Samples were prepared in independent biological triplicates for protein extraction. Purity of the coccoid culture was confirmed by showing no growth on non-selective chocolate agar supplemented with 5% defibrinated horse blood (HemoStat Laboratories, USA) when incubated for a period of 2 weeks.
Preparation of Proteome Samples. For the preparation of protein extracts from H. pylori, bacteria were harvested by centrifugation (12,000 g, 10 min, 4 °C) from 3-day and 3-month old cultures. The bacterial pellets were key enzymes involve in lipid A biosynthesis that were significantly reduced (p-value < 0.05) in H. pylori coccoids compared to spirals. Each bar on the graph represents mean of triplicates and y-axis represents relative abundance (i.e. total sum of the areas under the peaks). Error bar represents standard deviation.
were washed thrice with sterile Phosphate buffered saline (pH 7.4). The collected cells were lysed and protein extracted using the ProteoSpin detergent-free total protein isolation kit (Norgen Biotek, Canada). Halt protease and phosphatase inhibitors cocktail (Thermo Scientific, USA) were added to the lysate. The lysates were subsequently treated with 10 mM dithiothreitol (DTT; Bio-Rad, USA) at 37 °C for 1 h and alkylated with 55 mM iodoacetamide (IAA; Bio-Rad) for 30 min at room temperature. The proteins in the sample were digested with 1:50 (trypsin: protein) of MS grade porcine trypsin (Calbiochem, Germany) at 37 °C overnight. The samples were desalted using a Pierce C-18 spin column (Thermo Scientific) and dried to completeness in a refrigerated CentriVap centrifugal vacuum concentrator (Labconco, USA) before mass spectrometry analysis. Protein extracts from biological triplicates were pooled for LC-MS/MS analysis.
LC-MS/MS Analysis. TripleTOF ™ 5600 (AB SCIEX, USA) was used for the H. pylori spiral and coccoid proteomes as previously described 47 . The instrument was coupled with an Eksigent NanoLC-ultra 2D+ with Nanoflex cHiPLC nanoflex system in Trap-Elute mode for peptide separation. Solvent A contained 0.1% formic acid (v/v) in 2% acetonitrile while solvent B contained 0.1% formic acid (v/v) in 98% acetonitrile. The samples (2 μ g each) were loaded onto a cHiPLC trap column (200 μ m × 500 μ m ChromXP C18-CL, 3 μ m, 300 Å) with a flow rate of 3 μ L/min for 10 min and separated on a nano cHiPLC analytical column (75 μ m × 15 cm ChromXP C18-CL, 3 μ m, 300 Å) with a linear gradient of 5 to 40% solvent B for 55 min; 40 to 60% for 8 min; and 60 to 90% for 3 min at a flow rate of 300 nL/min on the Nanoflex cHiPLC system. The Chip nanoLC column was rinsed with 90% solvent B for 5 min and equilibrating with 95% solvent A for 13 min. Data Analysis. All spectra generated from information-dependent acquisitions (IDA) were searched against the H. pylori J99 database (1,488 predicted protein entities derived from specific NCBInr database) using ProteinPilot (version 4.5). Peptide library was generated from triplicate runs of IDA analysis for each of the 4 samples (giving a total of 12 runs) and the libraries were combined to create a single common peptide library used for SWATH analysis. The ProteinPilot software generates ion libraries based on FDR of 1%, 5% and 10% automatically during database search 48 . The identified peptides in the search result served as an ion library to apply to the SWATH acquisition runs. For SWATH-MS data, information (including FDR) for the correct peptide peaks extracted (i.e. retention time, amino acid sequence, protein information, etc) compiled from the ProteinPilot ion libraries were processed using the SWATH Acquisition Microapp (version 1.0) in the PeakView software (version 1.2) 22,48 . Following the SWATH workflow, the number of proteins first imported into the SWATH analysis software was restricted to the reported protein number based on false discovery rate of 1% and peak mass accuracy of 50 ppm by user input. Peaks were extracted based on software criteria (non-user input), such as mass accuracy (within 50 ppm), charge state, retention time, peak width and area, etc. The peak extraction mass window was 50 ppm and within 4 min of their expected retention time. Peak area for each protein was extracted using 3 peptides (excluding shared) and 3 transitions for all the samples. The peptide list was also further filtered by peptide confidence level (set as 95% based on ProteinPilot scoring definition) and excluding Shared or Modifications. All quantitative analyses (e.g. normalization, statistical calculations) were processed using the MarkerView (version 1.2.1) software. Samples were grouped according to NCTC 11637 spiral, NCTC 11637 coccoid, J99 spiral and J99 coccoid as triplicate data were imported. Normalization was performed based on total sum of the areas under the peaks in the software. Following that, Student's t-test was performed to compare between these groups, using p-value scoring of 0.05 as cutoff and area fold change as significant expression changes.