Gamma‐glutamyltranspeptidase expression by Helicobacter saguini, an enterohepatic Helicobacter species isolated from cotton top tamarins with chronic colitis

Helicobacter saguini is a novel enterohepatic Helicobacter species isolated from captive cotton top tamarins with chronic colitis and colon cancer. Monoassociated H. saguini infection in gnotobiotic IL‐10−/− mice causes typhlocolitis and dysplasia; however, the virulent mechanisms of this species are unknown. Gamma‐glutamyltranspeptidase (GGT) is an enzymatic virulence factor expressed by pathogenic Helicobacter and Campylobacter species that inhibits host cellular proliferation and promotes inflammatory‐mediated gastrointestinal pathology. The aim of this study was to determine if H. saguini expresses an enzymatically active GGT homologue with virulence properties.


| INTRODUCTION
Enterohepatic Helicobacter species (EHS) are gram-negative, spiralshaped bacteria that colonise the lower intestine, liver, and gall bladder of mammals, birds, and reptiles and are associated with the occurrence of gastrointestinal inflammatory diseases and cancers (Hansen, Thomson, Fox, el-Omar, & Hold, 2011;Mitchell et al., 2014;Thomson et al., 2011;. Infection by EHS is proposed to potentiate the risk of developing inflammatory bowel disease (IBD) in animal models and humans Mitchell et al., 2014;Thomson et al., 2011;, analogous to the causative relationship between Helicobacter pylori infection and the occurrence of gastric peptic ulcers and cancer (Marshall, 1995). Helicobacter hepaticus and Helicobacter bilis are EHS whose infection in mice are well-established models of pathogeninduced IBD and intestinal carcinogenesis Fox et al., 1994;Fox et al., 1995;Fox et al., 1996a;Fox et al., 1996b;Romero et al., 1988;Ward et al., 1994)

. Experimental infection by
Helicobacter cinaedi and Helicobacter fennelliae, both EHS originally isolated from homosexual men with proctitis, elicit diarrhoea and gastrointestinal inflammation in pigtail macaques (Flores et al., 1990;Totten et al., 1985). A current meta-analysis has also found a significant association between EHS infection and IBD status in human patients; however, specific Helicobacter spp. linked to human IBD remain to be elucidated (Castano-Rodriguez, Kaakoush, Lee, & Mitchell, 2015).
Our lab discovered a novel EHS, named Helicobacter saguini, cultured from colonic biopsy and faecal samples of captive cotton top tamarins (CTTs; Saguinus oedipus; Saunders et al., 1999;Shen et al., 2016). CTTs are an endangered new world primate species with a high incidence of idiopathic chronic colitis and colonic adenocarcinomas when maintained in captive colonies (Johnson, Ausman, Sehgal, & King, 1996;Saunders et al., 1999;Wood et al., 1998). The clinical and histopathological manifestations of large bowel inflammation in CTTs strongly resemble those of human ulcerative colitis, making these animals the ideal model to study etiopathogenesis of this disease. Although the protected status of CTTs has prevented establishing a direct etiological relationship between Helicobacter infection and IBD, we recently demonstrated that monoassociated infection of gnotobiotic C57BL/6 IL-10 −/− mice with H. saguini induces IBD and procarcinogenic changes in the large intestine (Shen et al., 2016).
These data suggest that H. saguini infection may influence the onset and progression of chronic colitis and colon cancer in CTTs.
Though a substantial body of evidence supports a causative association between EHS infection and IBD, the pathogenic mechanisms of these organisms remain incompletely defined. The best characterised virulence factor gene for EHS is cytolethal distending toxin (CDT), a DNA nuclease with proinflammatory and genotoxic effects to the gastrointestinal tract (Pratt, Sachen, Wood, Eaton, & Young, 2006;Shen et al., 2009;Young et al., 2004). Nevertheless, not all EHS associated with IBD encode CDT, indicating the need to identify and characterise other virulence factors expressed by these pathogens. Recently, gamma-glutamyltranspeptidase (GGT) expression by H. pylori, Helicobacter suis, H. bilis, and Campylobacter jejuni has been shown to be important for host colonisation persistence and inflammatory disease pathogenesis (Barnes et al., 2007;De Bruyne et al., 2016;Floch et al., 2014;Javed, Mejías-Luque, Kalali, Bolz, & Gerhard, 2013;Rossi et al., 2012;Wustner et al., 2017;. GGT is a constitutively expressed periplasmic enzyme hypothesised to promote bacterial survival by metabolising extracellular glutamine into glutamate, which is then imported by the bacterial cell to fuel energy and metabolic needs (Ling, Yeoh, & Ho, 2013;Ricci et al., 2014;Rossi et al., 2012;Shibayama et al., 2007). Our initial characterisation of H. saguini determined this organism does not express CDT but does display biochemical GGT activity (Shen et al., 2016 was most similar to bgh1 (Table 1), a second ggt gene detected in H. bilis that contains mutations in conserved functional motifs putatively rendering it enzymatically inactive (Rossi et al., 2012).
Multi-sequence alignments showed that HSGGT1 preserves all residues necessary for enzymatic function and putative virulence, including a signal sequence for periplasmic secretion, the threonine dyad for autocatalytic maturation and transpeptidase activity, and the lid loop for substrate binding ( Figure 1). Conversely, HSGGT2 shared identical amino acid mutations with H. bilis bgh1 ( Figure S1), presumably responsible for its inactivity (Rossi et al., 2012). This indicated that HSGGT1 and not HSGGT2 is most likely the enzymatically active paralog. Additionally, the predicted three-dimensional structure of HSGGT1 strongly resembled solved crystal structures of HPGGT (see Supporting Information). Together, these bioinformatics analyses suggested that HSGGT1 has enzymatic and virulence functionality.
As HSGGT1 showed the greatest sequence homology to HBGGT, subsequent experiments used HBGGT as a positive control for GGT activity and virulence.

| Construction of isogenic H. saguini GGT-knockout mutant (HSΔGGT1)
To test the virulence potential of HSGGT1, an isogenic GGTknockout mutant of H. saguini (HSΔGGT1) was created for comparison against the wild-type strain. The HSGGT1 gene was successfully replaced with catNT, a chloramphenicol resistance gene ( Figure S3).
Flanking upstream and downstream gene expression at the site of mutagenesis was detected in the wild-type and mutant strains, indicating a polar effect was not induced ( Figure S3E). Likewise, HSGGT1 gene expression was detected in the wild-type, but not the mutant strain ( Figure S3E). Mutant bacteria grew normally in vitro, consistent with other GGT-knockout Helicobacter spp. mutants (Chevalier, Thiberge, Ferrero, & Labigne, 1999;Rossi et al., 2012). H. bilis (HB) and wild-type H. saguini (HS) sonicate cleaved the glutamate substrate analogue GpNA into pNA at comparable levels ( Figure 2a (Altschul, et al. 1997).
FIGURE 1 Multisequence alignment of bacterial and human ggt genes generated with Clustal Omega (Sievers et al., 2011). Signal sequence predicted with SignalP 4.1 Server using the "Sensitive (reproduce SignalP 3.0's sensitivity)" setting (Petersen, Brunak, von Heijne, & Nielsen, 2011). Green highlighted residues designate the signal sequence exclusion site. Yellow highlighted residues and brackets designate disulfide bonds. Orange highlighted residues designate the autocatalytic cleavage site. Grey highlighted residues designate the conserved GXXGGXXI motif. Purple-coloured font residues designate the threonine catalytic dyad. Red-coloured font residues designate the lid loop. Boxed residues designate the substrate binding and processing sites. Helicobacter and Campylobacter spp. GGT (Flahou et al., 2011;Floch et al., 2014;Rossi et al., 2012). Whereas rHBGGT appeared maturated immediately after purification, rHSGGT1 matured into   GGT activity by HB and HS sonicates at 2000 μM of GpNA was completely inhibited by pretreatment with the GGT-specific inhibitor acivicin (pretreatment, 30-min incubation with 0-or 1-mM acivicin at 37°C). Results shown for two experiments performed in duplicate. (c) Partially purified rHBGGT and rHSGGT1 proteins from fraction #3 exhibited GGT activity. Enzyme activity for both rGGTs conformed to the Michaelis-Menten kinetic profile. Binding affinities to the substrate GpNA for rHBGGT and rHSGGT1 were 6.73 ± 0.31 and 12.3 ± 2.35 μM, respectively. Likely due to a higher concentration of mature enzyme (see Figure S5), rHBGGT had a higher transpeptidase V max of 299.6 mUnits/mg compared with 145.6 mUnits/mg for rHSGGT1. Results shown for representative curve performed in duplicate at each GpNA concentration. (d) GGT activity by rHBGGT and rHSGGT1 at 2,000 μM of GpNA was completely inhibited by pretreatment with the GGT-specific inhibitor acivicin (pretreatment, 30 min incubation with 0-or 1-mM acivicin at 37°C). Results shown for two experiments performed in duplicate Interestingly, HSΔGGT1 sonicate still yielded significant antiproliferative effects compared with the PBS control, suggesting that H. saguini may express other antiproliferative factors besides HSGGT1 ( Figure 3a).

| HSGGT1 inhibits intestinal epithelial and lymphocyte cellular proliferation
Though not readily apparent from the genomic or biochemical characterisation, these other factors will require further investigation to identify. HB and HS sonicates and rGGT proteins also blocked proliferation in a dose-dependent manner by 72 hr of treatment ). This effect may be due to expression of CDT, which is expressed by H. bilis (Chien et al., 2000;Kostia et al., 2003) and has been shown to induce apoptosis in lymphocytes and other cell types (Pratt et al., 2006;Shen et al., 2009 (Floch et al., 2014;Rossi et al., 2012).

| Glutamine supplementation and GGT enzyme inhibition prevent antiproliferative effect
Glutamine is an essential nutrient for intestinal epithelial cell proliferation because it serves as an energy source and precursor for de novo nucleic acid and amino acid biosynthesis (DeBerardinis & Cheng, 2010; Kim & Kim, 2017). Restriction of glutamine has been shown to impair the proliferation of HT-29 and other intestinal epithelial cell lines in vitro (Rhoads et al., 1997;Wiren, Magnusson, & Larsson, 1998). Therefore, we hypothesised that the antiproliferative effect caused by H. saguini is due to GGT-mediated metabolism of glutamine into glutamate and that glutamine supplementation and/or inhibition of GGT enzyme activity could restore proliferation.
To exclude that glutamate produced by HSGGT1 activity ( Figure   S4B) affected cell proliferation, we found that incubating HT-29 cells with up to 10 mM of glutamate for 72 hr had no effect on proliferation (data not shown). Media supplemented with glutamine partially but significantly restored proliferation after sonicate or rGGTs treatment; however, proliferation was still significantly less compared with (c) (d) GGT inhibition also significantly restored cell proliferation (Figure 5c).
Boiling sonicate and rGGT proteins rescued cell proliferation as well ( Figure S9). Together, these results suggest that the antiproliferative effect of HSGGT1 and HBGGT is in part due to the enzymatic metabolism of glutamine.

| rHSGGT1 induces proinflammatory gene expression
GGTs from Helicobacter spp. are capable of inducing proinflammatory changes in vitro and in vivo. HPGGT and HBGGT both increase activation of NF-κB and IL-8 expression (Gong, Ling, Lui, Yeoh, & Ho, 2010;Javed et al., 2013), whereas infection by GGT-knockout H. pylori and H. suis mutants causes significantly less inflammatory-mediated pathology in the rodent stomach compared with wild-type strains (Zhang, Ducatelle, et al., 2015). Therefore, we assessed the proinflammatory potential of HSGGT1 towards HT-29 cells. Interestingly, there was no statistical difference in IL-8 or TNF-α gene expression levels between infections with H. saguini wild-type versus HSΔGGT1, even though both cause significantly higher expression compared with the media control ( Figure 6a). However, treatment with rHSGGT1 or rHBGGT significantly increased expression of IL-8 and TNF-α ( Figure 6b).
Nevertheless, although these data show that HSGGT1 is capable of promoting proinflammatory gene expression in colon epithelial cells, it suggests that factors other than HSGGT1 may be primarily responsible for inflammatory processes in vitro. has enzymatic and virulent properties. The finding of two ggt genes for H. saguini agrees with a previous report that H. bilis also contains two ggt genes paralog, but only one of which (bhg2, HBGGT) was found to have enzymatic and virulence activity (Rossi et al., 2012). The role of these additional ggt genes in H. saguini and H. bilis requires further study (Rossi et al., 2012).

| DISCUSSION
To test our hypothesis, we created a viable isogenic knockout of H. saguini that lacked GGT activity. This finding also supports our bioinformatic predictions that HSGGT2 is enzymatically nonfunctional because the mutant H. saguini strain still harbours this gene, but lacked GGT activity. Likewise, we purified enzymatically active rHSGGT1 protein that metabolised the GGT substrate analogue GpNA and was blocked by the GGT-selective inhibitor acivicin.
Purified rHSGGT1 had a similar binding affinity for GpNA as rHBGGT and rHPGGT but appeared to autocatalytically mature into~40-and 20-kDa subunits less efficiently than rHBGGT. The purification protocol in this study was modelled after that used to purify rHBGGT (Rossi et al., 2012) and as a consequence may have not been optimised for complete rHSGGT1 maturation. Nevertheless, rHSGGT1 was able to undergo the expected maturation after 24-hr incubation at 37°C.
The common in vitro effect shared by virulent GGTs is the ability to impair gastrointestinal epithelial and lymphocyte proliferation. We found that HS sonicate and rHSGGT1 significantly inhibited HT-29, T84, AGS, HeLa, and Jurkat T cell proliferation on par with HBGGT. Like HBGGT and CJGGT, the antiproliferative effect by HSGGT1 occurred without evidence of cell death (Floch et al., 2014;Rossi et al., 2012).
We also showed that rHSGGT1 alone is capable of stimulating significant chemokine and cytokine gene expression by colon epithelial cells, consistent with the proinflammatory nature of HPGGT, HSuGGT, and HBGGT. Together, these data show that HSGGT1 exhibits enzymatic GGT activity with potential virulence properties in vitro.
Unexpectedly, we were unable to detect an antiproliferative or  Asterisk (*) designates statistical difference (*, P value < 0.05; **, P value < 0.01; ***, P value < 0.001) between indicated groups. Results of one representative experiment are shown. (c) Cellular proliferation of HT-29 cells after 72-hr treatment with 0.5 mUnits Helicobacter sonicate or rGGT that pretreated with acivicin (pretreatment, 24-hr incubation with 0.0-or 0.5-nM acivicin at 37°C). Asterisk (*) designates statistical difference (*, P value < 0.05; **, P value < 0.01; ***, P value < 0.001) between indicated groups. Results shown for three experiments performed in triplicate GGT expression has been shown to be pivotal for infection and disease pathogenesis by H. pylori, H. suis, and C. jejuni (Chevalier et al., 1999;Rossi et al., 2012;Zhang et al., 2013). GGT-knockout mutants of H. pylori and C. jejuni have an impaired ability to establish colonisation and cause pathology in their hosts (Barnes et al., 2007;Chevalier et al., 1999;Wustner et al., 2017;Zhang et al., 2013). In contrast, GGT expression by H. suis had no effect on stomach colonisation levels in mice or Mongolian gerbils but causes more severe inflammation and pathology compared with the GGT-knockout mutant (Zhang et al., 2013). In human patients, the occurrence of peptic ulcer disease is also associated with higher GGT activity by H. pylori (Park et al., 2014).
Whether GGT endows H. bilis with similar pathogenic properties in vivo has not been published to date.
The putative virulence mechanism of GGT is proposed to be due to enzymatic degradation of host glutamine into glutamate (Schmees et al., 2007;Shibayama et al., 2007;Wustner et al., 2015). Glutamate In conclusion, we have shown H. saguini expresses an enzymatically active GGT homologue with potential virulence properties. Gibco/Thermo Fisher Scientific, Grand Island, NY) at 37°C with 5% CO 2 .

| Construction of an isogenic GGT-knockout mutant (HSΔGGT1)
A full description of this method is described in the Supporting Information. Briefly, 500-bp fragments upstream and downstream of the H. saguini HSGGT1 gene were amplified and spliced together with an intervening HincII restriction enzyme site by overlap extension PCR (Heckman & Pease, 2007;Lee, Shin, Ryu, Kim, & Ryu, 2010).
After inserting this product into a pCR2.1-TOPO vector, a chloramphenicol resistance gene cassette (catNT) was inserted utilising the HincII restriction enzyme site. The recombinant plasmid was transformed into H. saguini by electroporation. Mutants were selected as previously described on blood agar plates containing 25-μg/ml chloramphenicol under microaerobic conditions (Ge et al., 2005;Ge et al., 2008;Ge et al., 2014). Mutants were confirmed for genetic authenticity by PCR amplification and sequencing. To rule out a polar effect on upstream and downstream gene expression at the mutagenesis site, cDNA from these genes was amplified from the wild-type and mutant strains.

| Trypan blue exclusion assay for cell viability
Cell viability was measured using trypan blue, a dye that is excluded by viable cells due to intact membrane integrity (Strober, n.d.;Strober, 2001). Fifty thousand HT-29 cells were plated in 1 ml of media in 24-well plates for 24 hr and then treated with 5 mUnits of HS and HSΔGGT1 sonicates or rHSGGT1 and rHBGGT protein for 72 hr. Cells were treated with 0.1% saponin (Sigma-Aldrich) for 5 min as a positive control for cell death. Media supernatant was collected and pelleted at 13,523 × g for 5 min to collect any floating or dead cells. Cells were detached from the plate with 0.25% trypsin-EDTA (Gibco) for 5 min and recombined with their supernatant pellets. Five microlitres of cells were mixed with 5 μl of 0.4% trypan blue (Gibco) and then loaded into a haemocytometer to count live and dead cells under a light microscope.

| Proinflammatory cytokine expression
Fifty thousand HT-29 cells were plated in 1 ml of media in 24-well plates. Cells were incubated for 24 hr at 37°C with 5% CO 2 to allow cells to adhere to the plates. Cells were treated with 5.0 mUnits dose of rHSGGT1, rHBGGT, or PBS for 4 hr at 37°C with 5% CO 2 in duplicate. Cells were collected in 1 ml of Trizol reagent (Invitrogen/Thermo Fisher Scientific) for total RNA extraction following the manufacturer's protocol. RNA was extracted using Trizol reagent. Total RNA (5 μg) was converted into cDNA using a high capacity cDNA Archive kit following the manufacturer's protocol (Applied Biosystems, Foster iments, respectively. The plates were centrifuged at 200 × g to facilitate bacterial cell adhesion and then incubated at 37°Cunder 5% CO 2 . Cell proliferation was quantified after 72 hr using the MTT assay as described above. Proinflammatory cytokine expression was assessed after 4 hr of infection. Media supernatant was collected for quantification of glutamate levels using the colorimetric Glutamate Assay Kit (Cell Biolabs, Inc., San Diego, CA) following the manufacture's instruction. HT-29 cells were washed once with phosphatebuffered saline (PBS) and collected in TRIzol reagent for total RNA extraction and qPCR of TNF-α and IL-8, as described above.

| Statistical analysis
Data are presented as mean ± standard deviation. Statistical analysis was performed by one-way analysis of variance with a Tukey post-hoc test using GraphPad Prism 5.0 (GraphPad Software, Inc.).
Results were considered significant at P value < 0.05. All graphs were generated using GraphPad Prism 5.0 (GraphPad Software, Inc.).