Helicobacter monodelphidis sp. nov. and Helicobacter didelphidarum sp. nov., isolated from grey short-tailed opossums (Monodelphis domestica) with endemic cloacal prolapses

Abstract

In a search for potential causes of increased prolapse incidence in grey short-tailed opossum colonies, samples from the gastrointestinal tracts of 94 clinically normal opossums with rectal prolapses were screened for Helicobacter species by culture and PCR. Forty strains of two novel Helicobacter species which differed from the established Helicobacter taxa were isolated from opossums with and without prolapses. One of the Helicobacter species was spiral-shaped and urease-negative whereas the other Helicobacter strain had fusiform morphology with periplasmic fibres and was urease-positive. 16S rRNA gene sequence analysis revealed that all the isolates had over 99 % sequence identity with each other, and were most closely related to Helicobacter canadensis . Strains from the two novel Helicobacter species were subjected to gyrB and hsp60 gene and whole genome sequence analyses. These two novel Helicobacter species formed separate phylogenetic clades, divergent from other known Helicobacter species. The bacteria were confirmed as novel Helicobacter species based on digital DNA–DNA hybridization and average nucleotide identity analysis of their genomes, for which we propose the names Helicobacter monodelphidis sp. nov. with the type strain MIT 15-1451^T (=LMG 29780^T=NCTC 14189^T) and Helicobacter didelphidarum sp. nov with type strain MIT 17-337^T (=LMG 31024^T=NCTC 14188^T)

Introduction

Grey short-tailed opossums (Monodelphis domestica) are small marsupials (family Didelphidae) native to Brazil, Bolivia, Paraguay and Argentina. They have become an important animal model in many basic, comparative and biomedically oriented research applications including immunologic and developmental studies [1]. Opossums successfully breed in captivity after reaching sexual maturity at 5–6 months. Their average litter size is eight, and they are capable of producing three litters per year. They can be easily maintained in standard rodent cages and fed commercial diets. Grey short-tailed opossums have been raised in pedigreed laboratory colonies for over 40 years [2–5]. Because of their physical and reproductive attributes, these opossums are the most widely used laboratory-bred research marsupial mammals in the world.

Cloacal (or ‘rectal’) prolapse has been described as the most prevalent health problem in laboratory opossums [5]. Helicobacter species have been associated with several gastrointestinal diseases in a variety of animals including humans; importantly, enterohepatic Helicobacter species (EHS) have been linked to rectal prolapse in mice [6]. Natural Helicobacter infections in animal models used in biomedical research have the potential to cause clinical disease and may influence the outcome and reliability of experimental studies. In 2014, an increased incidence of cloacal prolapse in postpartum dams maintained at a research colony at Texas A&M University prompted a search for the potential infectious agents that might be responsible. Prolapsed colonic segments had moderate to severe oedema, haemorrhage, inflammatory cells and necrotic mucosa lined by large colonies of bacteria.

The opossum colony at Texas A&M University (TAMU colony), the primary focus of this study, was established from animals representing nine random-bred ‘stocks’ and partially inbred ‘lines’ obtained from the Texas Biomedical Research Institute, San Antonio, TX (TBRI colony) between 2006 and 2015. Six were sublines descended entirely from the TBRI population one founder group, while three were of various admixed descent (Pop1/Pop2, Pop1/Pop4 and Pop1/Pop5). Information is available regarding the geographic origins of the TBRI colony, which comprises multiple genetic stocks descended from five wild-caught founder groups (Pop1–Pop5) from Brazil and Bolivia [7]. Opossums at the TAMU colony were generally housed in standard plastic hanging rat cages with shaved aspen bedding, provided with a longitudinally halved, 4 inch PVC pipe for shelter and shredded paper towels or (more recently) standard rodent crinkle paper for nesting material. Diet consisted exclusively of LabDiet 5ATD (Short-tailed Opossum #2, PMI Nutrition International) fed ad libitum and water prepared by reverse osmosis from lightly chlorinated (1–2 p.p.m.) water obtained from the Texas A&M University water supply. Opossums housed at the TBRI had similar husbandry and housing arrangements. Cloacal prolapses have been observed sporadically throughout the history of the TBRI and TAMU opossum colonies. The incidence of prolapse is generally low and has become less frequent since the development of standardized diets. The TAMU colony has experienced two episodes of increased prolapse frequency since its inception. In 2007/2008, unexplained inflammatory colitis presenting with anal bleeding, bloody stools and frank cloacal prolapses was noted in animals of the LL1 and LL2 stocks (the only two stocks maintained at TAMU at that time). Juveniles and adults of both sexes were affected. LL1 was the more severely affected of the two stocks. The increased morbidity and mortality led to severely reduced reproductive success, especially in LL1. As a result, the LL1 stock became reproductively unsustainable and a decision was made to euthanize the remaining LL1 animals. The condition eventually disappeared from LL2, and LL1 was subsequently replaced with new animals in 2009. Additional random-bred stocks and partially inbred lines were obtained in 2014 and 2015. The second episode, between late 2014 and early 2016, was observed in five of the nine stocks/lines being maintained at TAMU. Cloacal prolapse was generally the first clinical sign and was restricted almost exclusively to lactating females, 4–6 weeks postpartum. Few males and no juveniles were affected. In this episode, the LL2 stock was more severely affected than LL1. A few cases of prolapse were also observed in one additional random-bred stock and three partially inbred lines. A low incidence of the condition in this random-bred stock and its absence in three of the six partially inbred lines might be attributable to the fact that there were few litters produced in those groups during that time period rather than to any innate resistance to the clinical condition affecting other stocks. Elevated incidence of prolapse was not observed in animals at the TBRI or in its current location at the University of Texas, Rio Grande Valley vivarium.

In light of the fact that Helicobacter species have previously been isolated from the intestines of three species of Australian marsupials [8, 9], we investigated whether grey, short-tailed opossums maintained in captivity might be colonized with Helicobacter species and, if so, whether intestinal colonization with Helicobacter species could be associated with the development of cloacal prolapses in this species. Additional samples from the TBRI opossum colony (recently relocated to University of Texas, Rio Grande Valley), which has a history of infrequent, sporadic cloacal prolapses, was also surveyed for the presence of Helicobacter species.

Helicobacter isolation, host and ecological niche

A total of 94 opossums of both sexes from 11 stocks (LL1, LL2, LSD, AH11L, ATHHN, AH11H, ATHL, FD8X, PD2M, AC-P+, PBP), obtained from two different colonies, were screened for Helicobacter species. Ages ranged from 4 to 15 months. Faecal samples and select tissue samples (liver, gallbladder, stomach, colon and cecum) from 74 animals maintained at TAMU and faecal samples from 20 opossums from the TBRI colony were collected in freeze media (20 % glycerol in Brucella broth) and maintained at −80 °C. Samples from about 30 mg of tissues were homogenized in tissue grinders. The tissue homogenates and about 200 mg of faeces in freeze media were directly placed on CVA (cefoperazone, vancomycin and amphotericin) plates (Remel Laboratories) or filtered through 0.65 µm filters on to blood agar plates and incubated under microaerobic conditions with a gas mixture of N₂, CO₂ and H₂ (80 : 10 : 10) at 25, 37 and 42 °C. Suspect colonies were identified on the basis of colony morphology, Gram staining and biochemical reactions. Detailed biochemical characterization analysis was performed on ten individual isolates using the RapID NH System (Remel Laboratories) and the API Campy kit (bioMérieux). Biochemical characterization of urease, catalase, oxidase production, and sensitivity to nalidixic acid and cephalothin as well as growth in the presence of 1 % glycine were conducted as previously described by our laboratory [10]. A disc assay was used for indoxyl acetate hydrolysis.

Helicobacter species colonized most of the opossums surveyed from the two research colonies. A total of 40 helicobacter strains were isolated from the faecal, ceca and colon samples of 47 animals after 4–7 days of incubation under microaerobic conditions. The Gram-negative bacteria grew as thin, spreading films on the agar surfaces on CVA and blood agar plates. All the isolates were motile under phase contrast microscopy. Two different biochemical profiles were observed with two different morphologies. The Helicobacter isolates, named Helicobacter monodelphidis and Helicobacter didelphidarum , were isolated from opossums housed at both institutes. Twenty-five strains had phenotypes of H. monodelphidis and 15 strains had phenotypes of H. didelphidarum. H. monodelphidis was isolated from 8/11 stocks, H. didelphidarum was isolated from 6/11 stocks and three stocks had both organisms isolated (data not show). The biochemical characteristics of five isolates in each morphotype were compared with those of other Helicobacter species (Table 1) [11–13]. All of the isolates were oxidase- and catalase-positive, had γ-glutamyl transpeptidase activity, did not hydrolyse indoxyl acetate and did not grow in 1 % glycine nor at 25 or 42 °C. H. monodelphidis was positive for alkaline phosphatase hydrolysis and did not have urease activity. One of five strains reduced nitrate to nitrite. H. didelphidarum had urease activity and did not reduce nitrate to nitrite. Two of five strains were positive for alkaline phosphatase hydrolysis. H. monodelphidis was resistant to nalidixic acid and cephalothin, whereas H. didelphidarum was sensitive to nalidixic acid and resistant to cephalothin.

Table 1.

Phenotypic characteristics that differentiate H. monodelphidis and H. didelphidarum from other species of the genus Helicobacter

Species: 1, H. monodelphidis ; 2, H. didelphidarum ; 3, H. acinonychis ; 4, H. ailurogastricus ; 5, H. anseris ; 6, H. apri ; 7, H. aurati ; 8, H. baculiformis ; 9, H. bilis ; 10, H. bizzozeronii ; 11, H. brantae ; 12, H. canadensis ; 13, H. canis ; 14, H. cetorum ; 15, H. cholecystus ; 16, H. cinaedi ; 17, H. cynogastricus ; 18, H. equorum ; 19, H. felis ; 20, H. fennelliae ; 21, H. ganmani ; 22, H. heilmannii ; 23, H. hepaticus ; 24, H. himalayensis ; 25, H. jaachi ; 26, H. japonicus ; 27, H. macacae ; 28, H. marmotae ; 29, H. mesocricetorum ; 30, H. muridarum ; 31, H. mustelae ; 32, H. pametensis ; 33, H. pullorum ; 34, H. pylori ; 35, H. rodentium ; 36, H. saguini ; 37, H. salomonis ; 38, H. suis ; 39, H. trogontum ; 40, H. typhlonius ; 41, H. valdiviensis. +, All strains examined give a positive result; −, all strains examined give a negative result; (+), 80–94 % strains positive; ±, 33–66 % strains positive; (−), 7–33 % strains positive; NA, nalidixic acid; CE, cephalotin; I, intermediate resistance; B, bipolar; M, monopolar; St, subterminal; Pt, peritrichous; U, unknown. Data are from references [11–13].

Characteristic	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16	17	18	19	20
Oxidase	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+	+
Catalase	+	+	+	+	+	+	+	+	+	+	+	+	−	+	+	(+)	+	+	+	(+)
Nitrate reduction	(−)	−	−	+	−	+	−	+	+	+	−	±	−	−	+	+	+	+	+	−
Indoxyl acetate hydrolysis	−	−	−	−	+	−	+	−	−	+	+	+	+	−	−	(−)	−	−	+	+
Urease	−	+	+	+	+	−	+	+	+	+	−	−	−	+	−	−	−	−	+	−
Alkaline phosphatase	+	±	+	+	−	+	−	+	U	+	−	−	+	−	+	(−)	+	+	+	±
γ-Glutamyl transpeptidase	+	+	U	U	−	−	+	+	U	+	−	−	U	+	−	U	−	−	U	U
Growth at 42 °C	−	−	(−)	−	+	+	+	−	±	±	+	+	+	+	+	±	−	−	±	(−)
1 % glycine	−	−	−	U	+	−	−	−	+	−	+	−	−	U	+	−	+	−	−	−
Resistance to: NA (30 mg)	+	−	+	U	−	−	−	I	+	+	−	±	−	±	I	−	I	+	−	−
Resistance to CE (30 mg)	+	+	−	U	+	+	+	+	+	−	+	+	(−)	+	+	+	+	+	−	−
Periplasmic fibres	−	+	−	−	−	−	+	+	+	−	−	−	−	−	−	−	+	−	+	−
Distribution of flagella	B	B	M	B	St	B	B	B	B	B	St	B	B	B	M	B	B	M	B	B
Number of flagella	7–14	6–12	2–5	6–8	2	2	7–10	11–22	3–4	10–20	2	2	2	2	2	1–2	6–12	1	14–20	2
DNA G+C content (mol%)	35	32	30	37	30	40	36	45	35	46	39	34	48	36	35	37–38	44	38	45	35

Characteristic	21	22	23	24	25	26	27	28	29	30	31	32	33	34	35	36	37	38	39	40	41
Oxidase	+	+	+	+	+	+	+	+	U	+	+	+	+	+	+	+	+	+	+	+	+
Catalase	(−)	+	+	+	+	+	+	+	+	+	+	+	(+)	+	+	+	+	+	+	+	+
Nitrate reduction	+	+	+	+	−	−	−	−	+	−	+	+	+	−	+	−	−	−	+	+	−
Indoxyl acetate hydrolysis	−	−	+	−	+	−	+	−	U	−	+	−	−	(−)	−	−	(−)	−	U	−	+
Urease	−	+	+	−	+	−	−	+	−	+	+	−	−	(+)	−	−	+	+	+	−	(±)
Alkaline phosphatase	−	−	U	+	−	−	−	+	+	−	+	+	−	+	−	−	±	+	(−)	−	−
γ-Glutamyl transpeptidase	U	+	U	+	−	−	−	−	−	U	U	U	U	+	−	+	+	+	+	−	−
Growth at 42 °C	−	−	−	+	+	+	+	+	+	−	±	+	+	(−)	+	+	−	−	+	+	+
1 % Glycine	−	−	+	−	+	−	+	+	−	−	−	+	−	−	+	+	−	−	−	+	+
Resistance to: NA (30 mg)	−	U	+	−	−	+	+	+	−	−	−	−	+	(+)	+	+	−	U	+	−	(±)
Resistance to CE (30 mg)	+	U	+	+	+	+	+	+	+	−	+	+	−	(−)	+	+	±	U	+	+	+
Periplasmic fibres	−	−	−	−	+	−	−	−	−	+	−	−	−	−	−	+	−	−	+	−	−
Distribution of flagella	B	B	B	B	B	M	B	B	B	B	Pt	B	M	M	B	B	B	B	B	B	M
Number of flagella	2	4–10	2	1–2	7–14	1	2	2	2	10–14	4–8	2	1	4–8	2	6–12	10–23	4–10	5–7	1–2	1
DNA G+C content (mol%)	37	47	36	40	41	38	41	40	34	34	43	38	34–35	35–37	37	35	46	40	33	39	32

One isolate from each novel species was negatively stained with 1 % (w/v) phosphotungstic acid (pH 6.5) and examined under a jeol 2100F transmission electron microscope. Cells of H. monodelphidis were 3–4 µm long and 0.5–0.7 µm wide (Fig. 1). The organism was curved with 7–14 bipolar-unsheathed flagella. Cells of H. didelphidarum were 2–5 µm long, 0.5–0.7 µm wide and had a fusiform to slightly spiral appearance, with 6–12 bipolar sheathed flagella and periplasmic fibres that were generally tightly coiled around the bacterial cell (Fig. 1).

Fig. 1.

Transmission electron microscopy image of the opossum Helicobacter isolates.

The faecal and tissue samples from opossums were also tested for Helicobacter species by PCR. A High Pure PCR template preparation kit (Roche Molecular Biochemicals) was used for tissue and bacteria DNA extraction. The QIAamp DNA Stool Mini Kit (Qiagen) was used for faecal DNA extraction. Helicobacter genus-specific primers which amplify 1.2 kb PCR products were used for PCR [14]. By combining Helicobacter culture or Helicobacter genus-specific PCR methods, Helicobacter species were identified in 98 % of opossums in TAMU and in 100 % of opossums in the TBRI. Helicobacter species were not isolated from the liver/gallbladder and stomach samples, although one of the seven stomach samples was positive for helicobacter by PCR. (Table 2).

Table 2.

Helicobacter species prevalence in opossums analysed

Helicobacter species were identified by culture or PCR analysis in 100 % of opossums at the Texas Biomedical Research Institute (TBRI) and 98 % of opossums at Texas A&M University (TAMU). Helicobacter species were detected or isolated from faeces, colon and cecum, but not liver/gallbladder. Only one stomach sample was PCR-positive for Helicobacter species. nd, not determined.

Institution	Assay	Faeces	Colon	Cecum	Liver/gallbladder	Stomach
TBRI	Culture	20/20 (100 %)	nd	nd	nd	nd
TAMU	Culture	16/23 (70 %)	8/13 (61.5 %)	4/4 (100 %)	nd	0/4 (0 %)
TAMU	PCR	60/61 (98 %)	7/7 (100 %)	7/7 (100 %)	0/7 (0 %)	1/7 (14 %)

16S rRNA, hsp60 and gyrB gene phylogeny

All 40 opossum isolates were subjected to 16S rRNA PCR and sequencing with the conserved primers 9F 5′-GAGTTTGATYCTGGCTCAG-3′ and 1541R 5′-AAGGAGGTGWTCCARCC-3′ to amplify the 1.5 kb products [15]. The sequences of three strains of each species are represented in Fig. 2a. The sequences were almost identical within the two novel species and shared 99% sequence similarity between the two novel species despite their differences in morphology and biochemical tests. Consistent with other studies, 16S rRNA analysis alone does not always provide conclusive evidence for species level identification [13, 16] The most closely related previously recognized species was Helicobacter canadensis (97 %) (Fig. 2a). The 16S rRNA gene sequence similarity with the novel Helicobacter species isolated from brushtail and ringtail possums was 92–95% [8, 9].

Fig. 2.

Phylogenetic analysis of 16S rRNA (a), hsp60 (b) and gyrB (c) gene sequences. Neighbour-joining trees were based on the comparison of genes from different Helicobacter species. C. jejuni was used as an outgroup. Bootstrap values (>50 %) based on 1000 replications are indicated. GenBank accession numbers (in parentheses) are provided for each strain. Bars indicate the number of nucleotide substitutions.

The housekeeping gene sequences for heat shock protein 60 (hsp60) and DNA gyrase subunit B (gyrB) have been used to identify and infer phylogenetic relationships among Helicobacter species [17, 18]. Three strains from each novel species were subjected to sequencing for hsp60 with primers HSP60AF 5′-GCTAATCCTATTGAAGTGAAAAGAGGNATGGAYAA-3′ and HSP60DR 5′-CACTAAGGTAGTTAAAGCTTCCCCTTCDATRTCYT-3′ [19] as well as gyrB with primers specific for Helicobacter opossum isolates gyrB-opo-F 5′-GAYACTTATAAAGTWTCTGG-3′ and gyrB-opo-R 5′-AAATTCATCWCCAATMCCACA-3′. Their phylogenetic relationships with other Helicobacter species were compared by the neighbour-joining method using dnastar Lasergene software (Fig. 2b, c). These two novel Helicobacter species formed a distinct cluster that separated them from other Helicobacter species. H. monodelphidis , which was urease-negative, clustered with other urease-negative Helicobacter species, H. ganmani and H. pullorum. H. didelphidarum, which was urease-positive and had fusiform morphology, clustered with other urease-positive, fusiform Helicobacter species, H. aurati and H. bilis . The sequences of gyrB and hsp60 genes shared 98–100 % identity within H. monodelphidis or H. didelphidarum , while between H. monodelphidis and H. didelphidarum , the sequence identity was only 65–77 % (Fig. 2b. c).

Genome analysis

The draft genomes for H. monodelphidis MIT 15-1451^T and H. didelphidarum MIT 17-337^T were sequenced using Illumina MiSeq as described previously [20]. A pan-genomic phylogenetic tree was econstructed from a binary matrix of orthogroups determined with OrthoFinder (version 2.3.1) followed by iq-tree (version 1.6.12) using 1000 bootstrap replicates [21, 22]. pyani (version 0.2.10) with blastn was used to calculate average nucleotide identity (ANIb) between genomes in order to differentiate species at a 95 % similarity threshold [23]. Digital DNA–DNA hybridization (dDDH) was performed with the Genome-to-Genome Distance Calculator (GGDC) using a 70 % similarity threshold to differentiate species [24]. OrthoFinder was used to determine homologous protein coding genes (i.e. orthogroups) in order to identify the number of shared and unique genes between genomes [25]. Statistics for the draft genomes of H. monodelphidis MIT 15-1451^T and H. didelphidarum MIT 17-337^T are summarized in Table 3. Whole genome-based phylogenic analysis indicated that both novel Helicobacter species were associated with EHS rather than gastric Helicobacter species. However, each genome was aligned with different clades despite being isolated from the same host (Fig. 3). H. monodelphidis MIT 15-1451^T has a genome size of 1.99 Mb comparable with other spiral-shaped EHS, like H. hepaticus and H. typhlonius [26, 27]. H. didelphidarum MIT 17-337^T has a genome size of 2.59 Mb, which is comparable with fusiform-shaped EHS, like H. bilis [28]. ANIb and dDDH comparisons indicated that H. monodelphidis and H. didelphidarum are different from each other as well as all available Helicobacter species genomes and Campylobacter jejuni subsp. jejuni NCTC 11168^T. The ANIb and dDDH values for H. monodelphidis and H. didelphidarum versus other closely related Helicobacter species are shown in Table 4. Along with a substantially larger genome, H. didelphidarum also had about 500 more protein-coding genes than H. monodelphidis . The total 4775 protein-coding genes from both genomes contained 3359 orthogroups; 1186 of these orthogroups were shared between genomes, while 862 and 1311 were unique to H. monodelphidis and H. didelphidarum , respectively.

Table 3.

Genome summary statistics for novel Helicobacter strains isolated from opossums

The draft genomes for H. monodelphidis MIT 15-1451^T and H. didelphidarum MIT 17-337^T were sequenced by Illumina MiSeq.

Genome	Contigs	Coverage	Genome length (bp)	G+C content (mol%)	Protein-coding sequences	tRNA	rRNA	NCBI GenBank accession
H. monodelphidis MIT 15-1451T	131	288.4×	1 985 479	35.33	2125	36	2	NHYN00000000
H. didelphidarum MIT 17-337T	249	232.8×	2 589 753	31.60	2650	36	2	NXLQ00000000

Fig. 3.

Pan-genomic phylogenetic tree. Genomes for Helicobacter species and C. jejuni were accessed from public databases. Type strains were used when available, and GenBank accession numbers for each genome are shown in brackets. The pan-genomic phylogenetic tree was created from a binary matrix of orthogroups using 1000 bootstrap replicates with iq-tree and support values are shown on the tree.

Table 4.

Average nucleotide identity (ANIb) and digital DNA–DNA hybridization (dDDH) comparisons of novel Helicobacter isolates with other closely related enterohepatic Helicobacter strains

Strains: 1, H. monodelphidis MIT 15-1451^T; 2, H. didelphidarum MIT 17-337^T; 3, H. aurati MIT 97-5075^T;4, H. bilis ATCC 51630^T; 5, H. canadensis MIT 98-5491^T;6, H. ganmani MIT 99–5101; 7, H. mesocricetorum ATCC 700932^T; 8, H. muridarum CCUG 29262^T; 9, H. pullorum NCTC 12824^T; 10, H. rodentium ATCC 700285^T; 11, H. saguini MIT 97-6194-5^T; 12, H. trogontum ATCC 700114^T. All ANIb and dDDH comparisons were below the 95 and 70% thresholds, respectively, indicating H. monodelphidis MIT 15-1451^T and H. didelphidarum MIT 17-337^T are each novel species.

	Genome ID	1	2	3	4	5	6	7	8	9	10	11	12
H. monodelphidis MIT 15–1451T	ANIb	100	78.69	71.11	71.17	71.43	70.96	71.31	70.95	71.1	71.59	70.57	71.39
	dDDH	100	49.6	19.3	19.8	19	19.2	19.4	22.3	18.9	23.5	30	25.6
H. didelphidarum MIT 17–337T	ANIb	79.54	100	73.59	74.29	71.96	71.35	71.33	73.24	72.24	71.21	72.98	74.63
	dDDH	49.6	100	21.7	21.1	25.2	19.3	29.2	20.1	37.5	23.7	23.7	23.6

kaas (kegg Automatic Annotation Server) was used to assign KO (KEGG Orthology) annotations for metabolic reconstruction and functional predictions [29]. kaas analysis indicated that several metabolic/functional pathways were differentially enriched in H. monodelphidis versus H. didelphidarum. H. monodelphidis contained substantially more genes associated with the two-component system, flagellar assembly, ABC transport, pyruvate metabolism, and quorum sensing. In contrast, H. didelphidarum harboured more genes associated with bacterial secretion systems, propanoate metabolism, phenylalanine/tyrosine/tryptophan biosynthesis, fatty acid metabolism, and biotin metabolism. This suggests these species may have different metabolic potentials regarding chemotaxis/environmental sensing and the processing of carbohydrates, fatty acids and amino acids.

Virulence factors genes were identified using blastp against the VFDB and victors databases as well as against known Helicobacter species virulence gene sequences from the literature (threshold parameters: identity ≥25 %, coverage ≥75 % and E-value ≤10e-10) [30]. blast analysis indicated that H. monodelphidis and H. didelphidarum contained 157 and 168 homologous gene sequences to virulence factors expressed by Helicobacter and Campylobacter species, respectively. Several notable virulence factors identified in both genomes were flagella, γ-glutamyl transpeptidase (ggt), high-temperature requirement A protein-secreted serine protease (htrA) and neutrophil activating protein (napA). H. monodelphidis also encodes campylobacter invasion antigen B (ciaB) and fibronectin/fibrinogen-binding protein (fbps), while H. didelphidarum contains urease and catalase genes. Interestingly, H. didelphidarum also encodes homologous sequences for type VI secretion system components (HH_0242–HH_0252) and its secreted effectors (Hcp and VgrG proteins) harboured within the HHGl1 pathogenicity island of H. hepaticus . The presence of HHGl1, in particular the VI secretion system loci, enhances the pathogenicity of H. hepaticus in mice [31]. Neither vacuolating cytotoxin A (VacA) or cytolethal distending toxins (CDT) were identified in either genome. Overall, whole genomic analysis indicated that H. monodelphidis and H. didelphidarum are each novel species and have pathogenic properties.

Histopathologic findings

Distal colon tissues from six clinically normal opossums which were positive for helicobacter by culture or PCR were subjected to pathological examination. Four archived paraffin embedded colon samples from opossums with prolapses sent from TAMU were also evaluated. In the colon sections which had Helicobacter species isolated, the lamina propria and submucosa had low to moderate numbers of macrophages, lymphocytes, and plasma cells. The mucosa was lined by mild hyperplastic colonic epithelium characterized by crowded, overlapping epithelial cells with prominent basophilic nuclei, frequent mitotic figures, and loss of goblet cells. The submucosa was expanded by multiple lymphoid follicles with germinal centres. The mucosa and submucosa were disrupted by haemorrhage, oedema, and few inflammatory infiltrates. In four animals with cloacal prolapses, the prolapsed colonic segment was affected by moderate to severe oedema, haemorrhage, inflammatory cells and necrotic mucosa lined by large colonies of bacteria (Fig. 4a–c). The colonic tissue sections from four opossums were subjected to fluorescent in situ hybridization (FISH) staining for Helicobacter species with Helicobacter genus-specific probes [15]. Fluorescently labelled bacteria were identified in all the samples tested. Large colonies of helicobacter were present in the lumen and the crypts of the colon (Fig. 4d)

Fig. 4.

Pathological evaluations of the opossum colon tissues. (a) The lamina propria and submucosa had low to moderate numbers of macrophages, lymphocytes and plasma cells. The mucosa was lined by mild hyperplastic colonic epithelium characterized by crowded overlapping epithelial cells with prominent basophilic nuclei, frequent mitotic figures and loss of goblet cells. The bar corresponds to 100 µm. (b) The submucosa was expanded by multiple lymphoid follicles with germinal centres. The bar corresponds to 500 µm. (c) The prolapsed colonic segment was affected by moderate to severe oedema, haemorrhage, inflammatory cells and necrotic mucosa lined by large colonies of bacteria. The bar corresponds to 1 mm. (d) fish staining was used to identify Helicobacters species (red) present in the colonic lumens and crypts of opossums with prolapse. Cell nuclei were stained with DAPI (blue). The image was taken at ×630 magnification.

Discussion

Rectal prolapse is commonly associated with severe diarrhoea and tenesmus. Predisposing factors include intestinal inflammation from bacteria, parasites, rectal trauma, neoplasia of the rectum or distal colon, urolithiasis, urethral obstruction, cystitis and dystocia [32]. Prolapse of the rectum has been reported in many domestic, wild and laboratory animals, including dogs, cats, ferrets, cattle, horses, ewes, pigs, rodents, marsupials and macaque monkeys with recurrent diarrhoea [5, 6, 8, 9, 33]. Rectal prolapse is a common clinical condition in laboratory animals especially in mice under different settings that are highly influenced by the genotype, extent of intestinal inflammation, underlying neoplastic process or exposure to intestinal toxic insults such as DSS/AOM treatments [34]. Severe inflammation in the distal colon can result in thickened edematous tissue and lead to ano-rectal mucosal eversion and prolapse. We have established that two novel Helicobacter species have been isolated in a high percentage of clinically normal animals from two colonies of grey short-tailed opossums. Although the organisms have been identified in prolapsed rectal tissues in select animals, it is not known if antibiotic eradication of Helicobacter species would eliminate the occurrence of rectal prolapse in these colonies. Rectal prolapse associated with microbial pathogens can occur in immune competent as well as in immunocompromised mice [35, 36]. Certain strains of genetically engineered mice, for example IL-10^-/- and Rag-deficient mice, are highly susceptible to typhlocolitis with Helicobacter infection and are used to model IBD and colitis-associated carcinoma. Foltz et al. reported that H. hepaticus was isolated in 85 % of mutant mice affected with rectal prolapses [35]. Also, lamellipodin-deficient (Lpd^-/-) mice infected with Helicobacter species were highly susceptible to rectal prolapse and development of rectal carcinoma [6, 36–38]. Helicobacter -free Lpd^-/- mice achieved through embryo transfer rederivation abrogated the incidence of rectal prolapse in the colony [38]. In humans, H. cinaedi and H. fennelliae were first isolated from homosexual men suffering from proctitis, or proctocolitis and subsequently from diarrheic children residing in Africa [39, 40]. Experimentally, IL10^-/- and Rag2^−/− mice colonized with H. cinaedi resulted in chronic typhlocolitis [41, 42].

The Australian common brushtail possum (Trichosurus vulpecula) is a large tree-dwelling marsupial herbivore that feeds on a variety of leaves, particularly eucalyptus. Physiologically the brushtail possum is a cecum fermenter with a well-developed cecum, proximal colon, and a simple stomach [8]. Three morphologically different Helicobacter species: common-shaped, S-shaped and fusiform-shaped organisms were identified in the brushtail possums [8]. Subsequently, the same authors isolated another novel Helicobacter species (S-shaped) from the lower bowel of a second species of Australian marsupial, the ringtail possum (Pseudocheirus peregrinus) [9]. Similar to our study, Helicobacter species were only isolated from the lower gastrointestinal tracts (colon, cecum and rectum), but not from the liver or stomach. In neither of these two marsupial species was there evidence of intestinal or liver pathology [8, 9]. Although identified by Helicobacter genus PCR, no Helicobacter species were isolated from another marsupial, the koala [9]. The H. didelphidarum isolated from the grey short-tailed opossum had similar morphological characteristics with the fusiform-shaped Helicobacter species isolated from brushtail possums [8]. They were both urease-positive, γ-glutamyl transpeptidase-positive and nitrate reduction-negative; however, the H. didelphidarum from grey short-tailed opossum was alkaline phosphatase-positive, did not grow on 1 % glycine and was sensitive to nalidixic acid. By 16S rRNA analysis, none of these Helicobacter species isolated from Australian possums were categorized taxonomically with the opossum Helicobacter species identified in this study, sharing only 95 % sequence identity with each other [8, 9].

In summary, almost all opossums examined from the TAMU and TBRI colonies were colonized by novel Helicobacter species. This observation suggests that one or both of these novel Helicobacter species could be normal constituents of the opossum gut community. However, both colonies have historically experienced sporadic cases of cloacal prolapse, and one (TAMU) has had two episodes of substantially elevated cloacal prolapses, with one restricted primarily to early postpartum mothers. Inasmuch as Helicobacter infections have been implicated in colitis and rectal prolapse in various other mammals, we hypothesize that physiological and immunological changes occurring during mammalian pregnancy or lactation may predispose opossums to increased risk of bacterial infection and lower bowel inflammation. The ‘adaptive Th2 phenomena’ which provides a protective mechanism against rejection of feti and miscarriage, can also influence production of specific immunoglobulins produced as a reservoir of Th1 mediated immunoglobulins rendering the pregnant host more susceptible to infections by opportunistic bacterial organisms [43–45]. The disruption of normal opossum microbial flora during pregnancy and parturition can be followed by Helicobacter -induced colonic inflammation and subsequent prolapses. Further, during parturition, female opossums may have breaches in the epithelial surface in their cloacal tissue with resulting mucosal barrier dysfunction and dysbiosis. Notably, the incidence of cloacal prolapses at TAMU has been substantially mitigated by feeding the lactating females a gut lubricant (Laxatone, Vetoquinol, Princeville, Quebec, Canada) and fresh fruit. This has reduced the presence of large, dry, hard faeces, which may contribute to tenesmus and subsequent prolapses. Although Helicobacter infection in these opossums raises the possibility that these novel Helicobacter species could play a role in the pathogenesis of cloacal prolapses. The virtually universal colonization of these novel Helicobacter species in both the clinically normal and animals with cloacal prolapses surveyed, makes it impossible to assess the degree to which Helicobacter infection contributes to prolapse susceptibility in the opossum colonies. Mechanisms whereby these Helicobacter species might induce disease in this animal model require further investigation.

We isolated two novel species from two opossum research colonies, 25 strains had phenotypes of H. monodelphidis , and 15 strains had phenotypes of H. didelphidarum . The two novel Helicobacter species from our study shared 99 % sequence similarity in their 16S rRNA gene, but only had 60–70 % sequence identity with other Helicobacter species in the housekeeping genes gyrB and hsp60. We also sequenced the whole genomes of selected strains of the two species. ANIb and dDDH values were calculated and compared with other closely related Helicobacter species. All the ANIb scores were below the threshold of 95 % and all dDDH scores were below the threshold of 70 %, indicating they are two novel Helicobacter species [13]. We propose the name Helicobacter monodelphidis sp. nov. with the type strain MIT 15-1451^T; and Helicobacter didelphidarum sp. nov with type strain MIT 17-337^T.

Description of Helicobacter monodelphidis sp. nov.

Helicobacter monodelphidis (mo.no.del′phi.dis. N.L. gen. n. monodelphidis of the opossum species Monodelphis).

Cells are slender and slightly curved to rod-shaped (0.5–0.7 µm wide and 3–4 µm long). The bacterium is Gram-negative and non-sporulating. Coccoid forms were observed in old cultures. The organism is motile, having multiple, bipolar, unsheathed flagella and does not have periplasmic fibres. The organism grows slowly and appears on solid agar as a spreading film on the surface. The bacterium grows under microaerobic conditions, but not aerobically or anaerobically, at 37° but not at 42 °C, or 25 °C. The bacterium is oxidase, catalase, alkaline phosphatase and γ-glutamyl transpeptidase-positive but urease-negative. The organism does not hydrolyse indoxyl acetate. It does not grow in 1 % glycine and is resistant to nalidixic acid and cephalothin. The type strain, MIT 15-1451^T, isolated from the colon of an opossum in Texas, USA, has been deposited in the Belgian Coordinated Collections of Microorganisms (BCCM) as LMG 29780 and in The National Collection of Type Cultures as NCTC 14189. It has a DNA G+C content of 35.33 mol%, and its genome size is ~2.0 Mb. The 16S rRNA and the whole genome sequences of the type strain have been deposited in GenBank under accession numbers MH726195 and NHYN00000000.

Description of Helicobacter didelphidarum sp. nov.

Helicobacter didelphidarum (di.del.phi.da′rum. N.L. gen. pl. n. didelphidarum of the family Didelphidae).

Cells are fusiform (0.5–0.7 µm wide and 2–5 µm long) with periplasmic fibres. The bacterium is Gram-negative, non-sporulating and motile, having multiple, bipolar, sheathed flagella. The organism grows slowly and appears on solid agar as a spreading film on the surface. Coccoid forms were observed in old cultures. The bacterium grows at 37 °C but not at 42 or 25 °C, under microaerobic conditions, but not aerobically or anaerobically. The bacterium is oxidase-, catalase-, urease-, and γ-glutamyl transpeptidase-positive, but indoxyl acetate hydrolysis-negative. It does not grow on 1 % glycine and is resistant to cephalothin and sensitive to nalidixic acid. The type strain, MIT 17-337^T, isolated from the faeces of an opossum in Texas USA, has been deposited in the Belgian Coordinated Collections of Microorganisms (BCCM) as LMG 31024 and in The National Collection of Type Cultures as NCTC 14188. It has a DNA G+C content of 31.60 mol% and its genome size is ~2.6 Mb. The 16S rRNA and the whole genome sequences of the type strain haveposic been deposited in GenBank under accession numbers MH726196 and NXLQ00000000, respectively.