Characteristics of Unique Imported Vibrio cholerae Strains Which Caused Cases of Acute Intestinal Infection in Moscow in 2023

Full Text

BACKGROUND

Over the past 15 years, rare cases of cholera caused by strains of epidemic significance have been sporadic in Russia and always related to importations from India. Their causative agents belonged to the Haitian (2010) and post-Haitian (2012, 2014, 2023) groups, which hold a dominant position in the global etiology of cholera [1, 2]. In 2023, Moscow recorded its first two cases of acute intestinal infections caused by non-toxigenic (lacking cholera toxin genes ctxAB) V. cholerae strains, which carried the preCTX prophage and the classical-type tcpA gene encoding the structural subunit of toxin-coregulated pili. The tcpA gene differed from the prototype by three single nucleotide polymorphisms (SNPs). The patients, as part of a group of Russian citizens, visited Indonesia, Bangladesh, and India during a business trip, returning to Moscow on a special flight from Delhi shortly before admission. Notably, a member of this group, a Russian citizen, developed cholera due to a toxigenic post-Haitian strain, suggesting an independent source of infection. The second toxigenic strain was isolated in the same year in the Tambov Region as a result of independent importation by Indian citizens.

A phylogenetic analysis of whole genome sequences (WGSs) showed that both of the toxigenic isolates were grouped on the dendrogram with a number of strains isolated in 2021–2023 in foreign countries (Africa, Australia, Pakistan, US), whereas the non-toxigenic strains carrying the preCTX prophage formed a separate cluster, which was distanced from all the others, i.e. they had a unique genotype [2]. These strains were etiological agents of acute intestinal infections, so the initially published brief data clearly required more detailed characterization. This was needed to perform a bioinformatic analysis of individual virulence factor genes and their genomic arrangements (clusters and genomic islands) to assess their potential for expression, which is a critical determinant of virulence potential and etiological hazards.

AIM

The work aimed to provide phenotypic and genotypic characteristics of Vibrio cholerae O1 strains imported to Moscow in 2023 and isolated from patients with acute intestinal infections.

MATERIALS AND METHODS

We studied strains of V. cholerae O1 isolated in 2023 in Moscow from hospitalized patients with acute intestinal infections who had returned from a business trip to Southeast Asian countries. Strain 132 was obtained from one patient, and three almost identical subcultures of another strain (226, 228, and 264) were obtained from another patient. The strains were stored lyophilized in the collection of live cultures of the Reference Center for Cholera Monitoring at the Rostov-on-Don Anti-Plague Institute. After subculturing from vials, the cultures were tested for all previously characterized properties. This assessment included phenotypic and genotypic markers, including antibiotic susceptibility/resistance, with PCR testing performed according to Guidelines MUK 4.2.3745-22¹.

We used WGSs of these strains obtained using the MiSeq Illumina platform (Illumina Inc., US). We identified genetic determinants of pathogenicity and persistence in WGSs using the BLASTN 2.2.29 (http://blast.ncbi.nlm.nih.gov) and BioEdit 7.2.5 (http://www.mbio.ncsu.edu/bioedit) software. Antibiotic resistance was determined using the CARD database (https://card.mcmaster.ca) and a set of relevant genes found in cholera vibrios described earlier [3]. Genes and their translation products were analyzed using components of the Vector NTI Advance 11 software package (Invitrogen). We searched for homologous amino acid sequences of proteins and identified potential active domains using the BLASTP program (http://blast.ncbi.nlm.nih.gov). The nucleotide sequences of V. cholerae strains N16061 of the El Tor biovar (AE003852, AE003853) and O395 of the classical biovar (CP045718, CP045719) were used as prototypes. Manipulations with cluster sequences and data visualization were performed using Python scripts and the pygenomeviz and biopython packages (https://github.com/moshi4/pyGenomeViz).

Ethics Approval

Not applicable.

RESULTS

The strains under study had all the phenotypic characteristics of the V. cholerae species and belonged to the El Tor biovar and Ogawa serovar. All of them had the same antibiotic sensitivity/resistance profile: they were sensitive to ciprofloxacin, gentamicin, ceftriaxone, chloramphenicol, doxycycline, tetracycline, clarithromycin, azithromycin, cefotaxime, moxifloxacin, and cefaclor and resistant to ampicillin, rifampicin, tobramycin, streptomycin, amikacin, and nalidixic acid.

Previously, all genes of the preCTX prophage, the RS2 element (rstRAB) and additional pathogenicity factors such as cep (core encoded pilin), ace (accessory cholera enterotoxin), and zot (zonula occludens toxin), were identified in WGSs of the study strains, and the C-terminal sequence of the latter differed significantly from that of the prototype gene included in the CTX prophage in the genomes of toxigenic strains. This is typical for preCTX, which was considered to be a precursor of CTX, which, together with the ctxAB cholera toxin genes, acquired a new (canonical) zot end [4]. The prophage also contained the psh and orfU (pIII^CTX) genes; however, these determinants were located across several short contigs with overlapping terminal sequences. In this study, we assembled the complete sequence of preCTX and RS1 prophage (Fig. 1).

 

Fig. 1. Structure of preCTX and RS1 prophages of imported strains isolated from patients in Moscow (2023).

 

In its RS2 element, the preCTX prophage had the classical type rstR gene and an additional gene, which we designated as orfX. Most Vibrio cholerae isolates lack this gene; however, its homologs have been identified in specific strains from the US National Center for Biotechnology Information (NCBI) gene database, in particular in isolates from Taiwan (2002) and Uganda (2015). Their translation products (HAS3381368, HAS3410378, PYC77001) did not have potential active domains but were annotated as a “chemotaxis protein” with an unspecified function.

Interestingly, the 180-bp rstR gene of the RS1 prophage belonged to the Calcutta type, which is typically associated with O139 serogroup vibrios [5, 6] but has also been occasionally reported in V. cholerae O1 strains (WP_000201563, EGQ7979904, EGQ8141952). In the study strains, this prophage was located on the small chromosome, whereas hybrid genome assembly is required to determine the localization of preCTX.

Both pathogens have a complete Vibrio Pathogenicity Island 1 (VPI-1), which has AldA clusters with aldA, tagA genes and open reading frames with an unspecified function; tcp, which is responsible for the production of adhesion pili; and Acf, which includes genes for additional colonization factors [7]. The core tcp cluster showed overall similarity to those of El Tor and classical one; however, its tcpF gene diverged so dramatically from prototypes that it was undetectable by BLAST search and was only identified in WGSs. Despite having a different amino acid profile, the product of its translation had the same active domain, and its homologs, 99% identical, were found in the NCBI database (EGR4403235, NOE84825, NOE95797, NOF00903, NOF18360). The toxT gene showed significant divergence from the prototypes, particularly in its proximal region (1–436 bp). The distal region (437–831 bp) contained 17 SNPs, most of which were silent mutations. The C-terminal domain (144–277) contained only two substitutions, whereas the N-terminal domain (1–143) had numerous changes. However, the protein retained all its active domains and had full homologs in the NCBI database (6P7R_A, AAG31795, EHE0024350, EGR4117644, EGR1124498, EGR4402848), which was not surprising because its active domains were concentrated mainly at its distal end.

Fig. 2 shows products of the tcp-cluster genes of the study strains compared with those of the typical strains of the classical biovar and El Tor. Differences in the structure of these and some other proteins can be seen.

 

Fig. 2. Comparison of amino acid sequences of tcp cluster gene products from V. cholerae strains N16961, O395, and 226 (Moscow, 2023). Connecting lines in blue, light blue, pink and gray indicate 100% identity, and identity in the range of 95.0%–99.9%, 80.0%–84.9%, and 70.0%–79.0%, respectively. The coverage parameter for pairwise comparison of proteins (70%) was applied. We did not perform comparison for sequences with coverage of less than 70%; no connective lines are shown.

 

As noted above, the tcpA gene belonged to the classical type but differed from the prototype by the presence of 3 SNPs (G117T, A530C, A625G), two of which were missense mutations that led to substitutions in the protein (i.e. N177T and K209E). Its homologs were found in the NCBI database, differing only by one (i.e. the first) substitution (i.e. ACK75631, AHY02152, EJL6303812, KEH05472, KNH49515). These homologs belonged to strains of the non-O1/non-O139 serogroups, including the toxigenic V52, which was the causative agent of a large epidemic outbreak in Sudan in 1968.

The VPI-2 pathogenicity island was present in its complete form and showed near-identity to the prototype, except for a few SNPs. Its nan-nag region demonstrated 100% homology to the reference sequence.

The RTX cluster responsible for synthesizing a high-molecular-weight self-processing actin modulator called the MARTX cytotoxin was 99% homologous to that of the El Tor reference strain. The rtxA toxin gene had the same length as the El Tor prototype and contained 13 SNPs. None of these SNPs resulted in a premature stop codon; six were silent (synonymous); the remaining seven were missense mutations leading to non-synonymous amino acid substitutions in the protein. These substitutions did not disrupt the protein structure or its active domains.

There was no gene cluster for the Type 3 Secretion System (T3SS); however, T6SS clusters were present, namely the main cluster and auxiliary clusters (AUX1, AUX2, and AUX3). They contained determinants for structural components and effectors of this secretion system and were 99%–100% homologous to their reference prototypes.

The msh cluster, which is responsible for producing mannose-sensitive pili (Mannose-Sensitive Hemagglutinin, MSHA), was 99% identical to the prototype. The clusters responsible for synthesizing polysaccharides (Vibrio Polysaccharide, VPS) and the vps-I-rbm-vps-II proteins showed more differences and matched only 97.5%.

The genes encoding additional pathogenicity factors, including hemolysin/cytolysin HlyA, cytotoxic factor CefE1, hemagglutinin/protease HA/P, collagenase VchC, metalloprotease PrtV, serine proteases VesA, VesB, VecC, IvaP, and RssP; outer membrane proteins OmpU, OmpT, and OmpW; and transcriptional regulators ToxR, HapR, LuxO, CdgA, AphA, AphB, Hns, and CytR, were either identical or highly similar to their prototypes.

There were no pathogenicity islands VSP-I and VSP-II; only their flanking genes (VC0141 and VC0153 for VSP-I; VC0459 and VC0517 for VSP-II, respectively) were present. At the VSP-II insertion site, a sequence corresponding to tRNA-Met8 was identified.

No genes for the chxA cholix toxin and the stn/sto heat-stable toxin were identified.

As for drug resistance determinants, the CARD program detected only the alm operon (resistance for polymyxin), the β-lactamase gene varG (for carbapenems), the qnrVC4 gene (for fluoroquinolones), as well as CRP efflux pumps. Using the BioEdit software, we found determinants of VcmA, B, D, H, N, VcrM, VC1634 efflux pumps.

DISCUSSION

A comprehensive genomic analysis of imported clinical strains confirmed previous findings regarding their uniqueness, demonstrating no genetic relatedness to other isolates and the formation of a distinct, phylogenetically distant cluster [2]. The uniqueness concerned primarily the structure of the preCTX and RS1 prophages, because the NCBI database did not contain any identical complete sequences for them, although some matches were found for individual genes. This is not surprising, because this reflects the extreme heterogeneity of these prophages and their mosaic structure. Different types of preCTX prophages are known to exist not only as individual copies but also to form tandems within a single genome, integrating into both the large and small chromosomes. Tandems may include preCTX-preCTX, preCTX-CTX, CTX-RS1-preCTX, and other variants, whereas prophages may be represented by both identical and different types [8, 9]. Prophage types are classified depending on the structure of the RS2 element genes, primarily the rstR gene, which has the largest number of alleles. The most famous of them include rst^ET (El Tor), rstR^class (classical), rstR^calc (Calcutta), rstR^env (environmental) [10, 11], but in recent years many additional alleles have been described [6, 9]. The coexistence of prophages with different alleles likely results not from simultaneous infection by two distinct CTXφ/preCTXφ phages [12], but rather because each prophage provides immunity only against superinfection by homologous phages but not by other phage types [5, 6].

As noted above, the rstR^class gene was detected in the RS2 element of the study strains, and rstR^calc was detected in the RS1 prophage. The origin of the “extra” orfX gene, which is localized downstream of rstB in preCTX of the study strains, remains unknown. Only a single case of an additional gene, designated rstU, in the RS1 prophage has been reported in the published data. This strain was isolated from the environment in China (KP768424) [8]. It was completely homologous to the gene encoding fumarate reductase subunit D and had nothing in common with the orfX we found, the product of which was described as a “chemotaxis protein” in the NCBI database.

Another distinctive feature of the study strains was an unusual architecture of their tcpA^class gene, which diverged from the prototype by three SNPs, and significant alterations in tcpF and toxT genes of the tcp cluster, though their translated products still had all functional domains. In the NCBI database, we found five 99% homologs for the first variant and five 99% homologs for the second one; the third variant had six 100% matches. Only one strain contained both the second and third variants simultaneously. All others represented distinct strains, none of which carried the aforementioned TcpA variant. This is quite consistent with the mosaic structure of the tcp cluster, where tcpA has the greatest variability [7, 13]. Different authors have described multiple alleles of this gene and some others in this cluster [14, 15]. It has been repeatedly suggested that genetic variability within both VPI-1 and CTX/preCTX elements plays a crucial role in V. cholerae evolution by fostering new clonal variants with enhanced persistence, pathogenicity, and epidemic potential through interstrain genetic exchange [14–16].

Acute intestinal infections caused by V. cholerae O1 strains harboring the preCTX prophage were identified in Russia for the first time. Previously, they were occasionally found only in environmental waters (Moscow, 1977; Crimea, 1991; Rostov-on-Don, 1982, 1987, 2007) or isolated from patients in Uzbekistan (1988, 1990). The only clinical isolate (Rostov-on-Don, 1981) belonged to the nonO1/nonO139 (O8) serogroup [17]. Detected sporadically in Russia, preCTX⁺VPI⁺ strains represent imported cases originating from other countries outside Russia, likely Asian, where they are more widely distributed and cause human infections. Specifically, such strains have been detected in China among isolates belonging to genetic lineages L3b and L9. In their characterization, the authors designated these strains as epidemic with potential for global spread [18, 19]. Indeed, two preCTX⁺VPI⁺ strains of the L3b lineage were isolated from patients in Switzerland [20]. Regardless of their isolation source and serogroup, all previously studied strains of our collection with this genotype caused a significant enteropathogenic effect in the model of infant rabbits, which sometimes resembled the cholera effect [17]. Their increased pathogenetic potential is directly related to the presence of preCTX, which contains genes of the Zot and Ace toxins and additional colonization Cep factor, and to the VPI-1 island, which includes determinants of not only the key colonization factor TCP but also additional Acf. Furthermore, they usually have a virtually complete “standard” set of intact determinants of pathogenicity typical for El Tor vibrios such as VPI-2 island, T6SS, RTX, msh clusters, genes for hemolysin hlyA, cytotoxic toxin cef, hapA, prtV, vchC metalloproteases, vesA, vesB, vecC, ivaP, rssP serine proteases, ompU, ompT, ompW outer membrane proteins; each of them can contribute to pathogenesis [3].

As for drug resistance determinants, very few were found in WGSs, which did not reflect phenotypic resistance in vitro. Resistance to ampicillin could be associated with the varG β-lactamase gene, whereas qnrVC4 did not result in resistance to fluoroquinolones (ciprofloxacin, moxifloxacin). No genes associated with resistance to rifampicin, tobramycin, streptomycin, or amikacin were found. These strains likely possess other, as-yet-undescribed resistance determinants, or their resistance may be mediated by efflux pumps [21]. Such inconsistencies between phenotype and genotype were also found for other strains [3, 22, 23]. This is why, when selecting the most effective etiotropic therapy for acute intestinal infections, health professionals should focus mainly on the phenotype of the isolate.

The pathogens were likely to originate in India, though infection could potentially have occurred in either Indonesia or Bangladesh, where the patients shared the same hotels and traveled on the same special flight, suggesting possible transmission through foodborne exposure, person-to-person contact/fomites, or waterborne routes (including iced beverages). Reports on Vibrio cholerae detection (including toxigenic and multidrug-resistant strains) in food-grade ice and iced beverages [24, 25], beverage production line equipment [26], and street-vended fruits and vegetables [27] have originated predominantly from Indonesia. Meanwhile, cholera vibrios have also been detected in street-vended foods in Bangladesh, although their characteristics were not reported;² in vegetables and greens, only unspecified Vibrio spp. were identified without species-level determination [28]. In India, similar studies of salad vegetables found many pathogenic bacteria, but vibrios were not mentioned [29]. However, these studies were limited and irregular, whereas the risk of foodborne or waterborne cholera transmission is unequivocal. For example, an outbreak of cholera in Vietnam in 2010 was associated with consuming iced tea [30].

CONCLUSION

The bioinformatic analysis of whole-genome sequences from V. cholerae O1 strains isolated in 2023 from two patients with acute intestinal infections, who had returned to Moscow after a short-term visit to Indonesia, Bangladesh, and India, revealed the composition and structural features of their genetic determinants, including gene clusters, genomic islands, and prophages. The genetic identity of the pathogens suggested a single source of infection and a common transmission factor. We called these strains unique strictly in the sense that their genomic elements, including the preCTX and RS1 prophages and tcp cluster, lacked complete homologs among both our collection strains and those available in the NCBI database, despite some partial matches being found. The uniqueness lies rather in the specific allele combinations within these genetic elements. From this perspective, many other isolates with different combinations could also be considered unique; however, these particular variants represent the first such strains ever imported into Russia. With their extensive repertoire of intact pathogenicity determinants enabling full pathogenic expression, these strains constitute a potential threat to public health.

The risk of further importation of such or similar strains persists, and the characterized isolates can be used as reference strains for cholera surveillance studies in Russia.

ADDITIONAL INFORMATION

Authors’ contribution. E.V. Monakhova―analysis of literature and whole-genome sequencing data, writing the text and preparing the article for publication; V.D. Kruglikov―data analysis, discussion of results, editing the article; O.A. Podoinitsyna―comparative analysis of protein structure, preparation of figures; A.S. Vodopyanov―whole-genome sequencing, identification of genetic determinants; N.B. Nepomnyashchaya―phenotypic and PCR identification of the studied strains; A.V. Evteev―determination of antibiotic resistance; N.E. Gaevskaya―collecting and analysis of information on the importation and isolation of strains, their characteristics. Thereby, all authors provided approval of the version to be published and agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Ethics approval. Not applicable.

Funding sources. None.

Disclosure of interests. The authors have no relationships, activities or interests for the last three years related with for-profit or not-for-profit third parties whose interests may be affected by the content of the article.

¹ Guidelines MUK 4.2.3745-22. 4.2. Control methods. Biological and microbiological factors. Methods for laboratory diagnostics of cholera. Available at: https://docs.cntd.ru/document/350413501. Accessed on: November 15, 2024.

² Ahamad R. Harmful street-vended food in Dhaka. NEWAGE Bangladesh; 2020. Available at: https://www.newagebd.net/article/100735/harmful-street-vended-food-in-dhak

About the authors

Elena V. Monakhova

Rostov-on-Don Plague Control Research Institute

Author for correspondence.
Email: monakhova_ev@antiplague.ru
ORCID iD: 0000-0002-9216-7777
SPIN-code: 3091-5680

Dr. Sci. (Biology)

Russian Federation, Rostov-on-Don

Vladimir D. Kruglikov

Rostov-on-Don Plague Control Research Institute

Email: kruglikov_vd@antiplague.ru
ORCID iD: 0000-0002-6540-2778
SPIN-code: 9767-2936

MD, Dr. Sci. (Medicine)

Russian Federation, Rostov-on-Don

Oksana A. Podoinitsyna

Rostov-on-Don Plague Control Research Institute

Email: podoynitcina_oa@antiplague.ru
ORCID iD: 0000-0002-9996-4189
SPIN-code: 4424-8963

Cand. Sci. (Biology)

Russian Federation, Rostov-on-Don

Alexey S. Vodopyanov

Rostov-on-Don Plague Control Research Institute

Email: vodopyanov_as@antiplague.ru
ORCID iD: 0000-0002-9056-3231
SPIN-code: 7319-3037

MD, Cand. Sci. (Medicine)

Russian Federation, Rostov-on-Don

Natalia B. Nepomnyashchaya

Rostov-on-Don Plague Control Research Institute

Email: nepomniashchaia_nb@antiplague.ru
ORCID iD: 0000-0003-0868-6791
SPIN-code: 3314-8774
Russian Federation, Rostov-on-Don

Artem V. Evteev

Rostov-on-Don Plague Control Research Institute

Email: evteev_av@antiplague.ru
ORCID iD: 0000-0002-0087-9153
SPIN-code: 6050-0361
Russian Federation, Rostov-on-Don

Natalia E. Gaevskaya

Rostov-on-Don Plague Control Research Institute

Email: gaevskaya_ne@antiplague.ru
ORCID iD: 0000-0002-0762-3628
SPIN-code: 3143-4788

MD, Cand. Sci. (Medicine)

Russian Federation, Rostov-on-Don

References

Supplementary files