Background
Members of the genus Vibrio are abundant in marine environments [1] and in inland rivers where seawater intrusion occurs [2]. The levels of Vibrio spp. in various seafood commodities have been reported to be significantly higher in the summer than in other seasons [3]. Under this genus, V. furnissii is a motile, oxidase-positive, gram-negative, halophilic bacterium with phenotypic characteristics highly similar to those of V. fluvialis [4]. Although V. furnissii is closely related to V. fluvialis, it differs in its ability to produce gas through carbohydrate fermentation [4]. V. furnissii is a potential pathogen of European eel (Anguilla anguilla) [5] and also one of the non-cholera Vibrio species pathogenic in humans that can spread through the consumption of contaminated seafood products or exposure to coastal waters [6, 7]. V. furnissii has been associated with outbreaks or sporadic cases of gastroenteritis with cholera-like symptoms including diarrhoea, abdominal cramps, nausea and vomiting [7, 8]. Cases of V. furnissii bacteraemia associated with skin lesions or cellulitis have also been reported [9, 10]. V. furnissii infection has been reported to be rarer than V. fluvialis infection. The US Centers for Disease Control and Prevention (CDC) reported only 10 isolates of V. furnissii from 1997 to 2008: three from blood, two from a wound and five from stool [9]. However, since 2006, 24 sequences of V. furnissii have been submitted to GenBank in succession, which means that V. furnissii infection is increasingly reported.
V. fluvialis infection has been reported worldwide [11‐13] and shows features of multi-drug resistance [12, 14] including resistance to fluoroquinolones and β-lactam antimicrobials, blaNDM−1-mediated carbapenem resistance and azithromycin resistance [15‐17]. V. fluvialis has the ability to cause epidemics, so the rapid increase in and spread of antibiotic resistance in this pathogen during the past 20 years has become a major cause of concern [18]. V. furnissii is phylogenetically close to V. fluvialis, and recently, a mph(A)- and blaOXA−1-bearing conjugative plasmid, which mediates resistance to cephalosporins and azithromycin, was also discovered in a V. furnissii strain isolated from hospital sewage in Zhuhai, a coastal city in China [19].
Anzeige
Genome analysis of V. furnissii NCTC11218 isolated from estuaries [20] has shown that this species has a dynamic and fluid genome that can quickly adapt to environmental perturbation and has a series of virulence-related genes, such as quorum sensing-related genes cqsA, cqsS, luxS, luxU/O and luxP/Q; biofilm formation-related genes hapR and vpsT; major pilin subunit-encoding gene tcpA; and haemolysis-related genes rtx and hlyA. These genes are also widely distributed in other pathogenic Vibrio spp.
In this study, we isolated seven V. furnissii strains from the stool samples of patients with diarrhoea; surveyed the clinical characteristics of the sampled patients; and analysed the antibiotic resistance and virulence phenotypes of, and related genes in, the isolated strains by using whole-genome sequencing.
Materials and methods
V. furnissiiisolates
We collected 1,985 stool samples from patients aged over 14 years who presented with acute diarrhoea at a general hospital in Beijing, China, between April and October 2018. All of the patients completed an epidemiological questionnaire on clinical history and physical fitness. We first enriched the stool samples in an alkaline peptone water solution (Beijing Land Bridge, China) [21], then cultured 20 µL of the resulting mixture on Columbia agar (Oxoid, UK) containing 5% sheep erythrocytes and 20 µg/mL ampicillin [22] (Sigma, USA), which was a selective agar for Aeromonas spp., and on gentamicin selective medium (Beijing Land Bridge, China), which was widely used as a selective Vibrios agar in China. Next, we performed oxidase tests (BioMerieuX, France) to screen for gram-negative rod colonies on the agar plate. The oxidase test results of Vibrio spp., Aeromonas spp. and Plesiomonas shigelloides colonies were positive, while those of Enterobacteriaceae colonies were negative. Finally, microorganisms were identified taxonomically using a VITEK matrix-assisted laser desorption/ionisation time-of-flight mass spectrometry (MALDI-TOF MS) system (BioMerieuX, France) and the microbial identification system VITEK II (BioMerieuX, France). Per protocol, the stool samples were also simultaneously tested for Salmonella spp., Shigella spp., Aeromonas spp., Plesiomonas shigelloides and other Vibrio spp. All of the strains were stored in a Luria broth (LB)–glycerol mixture (80:20) at -80 °C until identification.
Genome sequencing
The genomic DNA of the isolates was extracted using TIANamp Bacteria DNA Kit (Beijing Tiangen, China) and then sequenced using Illumina NovaSeq PE150 at Beijing Novogene Bioinformatics Technology Co., Ltd. Sequencing libraries of ∼ 350 bp were prepared using NEBNext® Ultra™ DNA Library Prep Kits and analysed for size distribution using an Agilent 2100 Bioanalyzer and quantified using real-time polymerase chain reaction. SOAP denovo [23] was used to assemble paired reads. All of the genomic sequences are available at the National Center for Biotechnology Information (NCBI; accession nos. SAMN21988700, SAMN22062944-49). Further, as VFBJ05 and VFBJ07 were found to own transposon islands, the extracted DNA of them were further subjected to 250-bp paired-end whole-genome sequencing with 150× coverage using the Nanopore sequencer. The filtered subreads were assembled using Canu v1.5 [24], and then Circlator v1.5.5 (https://github.com/sanger-pathogens/circlator) was used to cyclise the assembled genomes. Coding gene prediction was performed using Prodigal v2.6.3 [25]. Genome annotation was performed using the RAST server (https://rast.nmpdr.org/rast.cgi). The whole-genome sequences of VFBJ05 and VFBJ07 are available at the NCBI (accession nos. SAMN35555507 and SAMN35555508).
Anzeige
Average nucleotide identity (ANI) analysis
ANI analysis was used to evaluate the evolutionary distance of bacteria at the genomic level based on a Perl script previously described [26]. ANI between 19 genome assemblies was calculated using pyani (https://pypi.org/project/pyani/), and strains with ANI values > 95% were considered as the same species [27, 28]. These strains included seven newly sequenced clinical isolates of V. furnissii and 11 reference strains of Vibrio from GenBank.
Genome analysis
Prokka [29] was used to annotate genomes. The Virulence Factor Database (VFDB) [30] was used to predict virulence-related genes by BLAST+, as described in a previous study [31]. Potential antimicrobial resistance genes were predicted using the Antibiotic Resistance Genes Database [32]. Plasmids were searched through the NCBI database (https://ftp.ncbi.nlm.nih.gov/refseq/release/plasmid/) and confirmed using Platon. Insertion sequences were found using BLAST in ISfinder (https://www-is.biotoul.fr/index.php) to obtain the transposon islands.
Genome comparison
To obtain the phylogenetic tree of V. furnissii strains, including seven sequences from our study and 24 available sequences from GenBank, MUMmer (Version 3.23) was first used to filter out gaps and single-nucleotide polymorphisms that were less than 5 bp long and FastTree was used to construct the tree. The phylogenetic tree and gene presence/absence profiles of 31 V. furnissii strains were integrated and rendered using iTOL 3 [33].
Antibiotic susceptibility tests
Antibiotic susceptibility tests were performed using antimicrobial susceptibility testing panel for Vibrios (Shanghai Biofosun, China). After V. furnissii strains were refreshed on blood agar plates at 35 °C for 16–18 h, 0.5 McFarland standard of direct colony suspension were prepared and diluted into panel according to the manufacturer’s instructions, the panel were then incubated at 35 °C for 18–20 h. The minimum inhibitory concentrations of the following 19 antibiotics were determined according to the guidelines of Clinical and Laboratory Standards Institute [34, 35]: ampicillin, ampicillin/sulbactam, cefazolin, ceftazidime, cefepime, aztreonam, amikacin, gentamicin, ciprofloxacin, levofloxacin, imipenem, meropenem, tetracycline, doxycycline, chloramphenicol, sulphonamides, trimethoprim–sulfamethoxazole, azithromycin and streptomycin. Escherichia coli ATCC 25,922 was used as the quality-control strain for susceptibility testing.
Bacterial killing assay
The assay was performed as described previously [36]. Overnight cultures of V. furnissii (predator) were mixed with E. coli MG1655 (prey) at a 5:1 (predator: prey) ratio. The mixture (5 µL) was spotted onto LB agar with a 0.22-µm filter membrane. After incubating at 30 °C for 4 h, the surviving cells of the prey MG1655 were determined by 10-fold serial dilutions on rifampin-containing LB agar plates.
Analyses of type VI secretion system (T6SS) expression
Protein samples were prepared as previously described [37]. Overnight cultures of V. furnissii were inoculated 1% in fresh LB medium and grown at 30 °C until an OD600 of ∼ 1.0 was reached. The cultures (1 mL) were then centrifuged for 2 min at 8000 rpm, and the obtained pellets were resuspended in sodium dodecyl sulphate-loading dye and boiled for 10 min. Proteins were loaded on an SDS-PAGE gel and separated by electrophoresis, after which proteins were transferred to a polyvinylidene difluoride (PVDF) membrane (Bio-Rad). The protein-bound membrane was blocked with 5% (wt/vol) non-fat milk in PBS with Tween 20 (PBST) buffer for 2 h at room temperature. The resulting membrane was cut according to the location of interest proteins, then incubated with primary antibodies (anti-Hcp and anti-RpoB) and secondary antibodies (goat anti-rabbit IRDye and goat anti-mouse IRDye), and finally imaged by two-color infrared laser imaging system (LI-COR odyssey CLx). These antibodies were prepared as previously described [38].
Haemolysin assay
For haemolysin assays, V. furnissii strains were grown to the mid-log phase at 37 °C. Portions (5 µL) of concentrated cultures were then spotted onto Columbia blood agar plates and incubated for 48 h at 30 °C [39].
Results
Clinical features
There were ultimately 349 diarrhoea cases related to bacterial infection, which included 145 cases caused by Vibrio spp. Apart from V. parahaemolyticus, V. cholerae and V. fluvialis, seven V. furnissii strains were recovered from the stool samples of the patients. Among these V. furnissii, three of them (42.9%, 3/7) was only isolated from AMP blood agar, and four of them (57.1%, 4/7) could be isolated from both AMP blood agar and gentamicin selective medium.
Anzeige
The clinical and epidemiological characteristics of the seven patients with diarrhoea related to V. furnissii infection are shown in Table 1. The sex ratio (male: female) was 2.5. Among these patients, one (14.3%) also had vomiting, two (28.6%) had abdominal pain and three (42.9%) had watery stool, but none had fever. Leukocytes were found under high magnification (×40) in the stool samples of two (28.6%) of the seven patients. These seven patients reported having had no exposure to seawater or seafood.
Table 1
Clinical characteristics of the seven patients with diarrhoea caused by Vibrio furnissii infection
Strain | Sexa | Age | Month | Feverb | Diarrhoea No./day | Abdominal pain | Stool consistency | Vomiting | Erythrocytes No./Hpc | Leukocytes No./Hpc |
|---|---|---|---|---|---|---|---|---|---|---|
VFBJ01 | M | 45 | May | - | 3 | - | watery | - | 0 | 0 |
VFBJ02 | F | 65 | June | - | 10 | - | loose | - | 0 | 0 |
VFBJ03 | M | 15 | July | - | 10 | - | watery | + | 0 | 1 |
VFBJ04 | M | 62 | July | - | 4 | - | loose | - | 0 | 0 |
VFBJ05 | F | 36 | July | - | 3 | + | loose | - | 3 | 30 |
VFBJ06 | M | 64 | July | - | 6 | - | watery | - | 0 | 0 |
VFBJ07 | M | 72 | August | - | 7 | + | loose | - | 0 | 0 |
General bioinformatic features of the seven V. furnissiistrains
V. furnissii was first identified using VITEK MS and had 99.0% identified rate, then identified biochemically using VITEK II. We found that the seven V. furnissii strains were misidentified as V. fluvialis, at a rate of 85.0–99.0% by VITEK II, so we further identified these strains based on ANI [20]. Dendrograms constructed based on the ANI values (Fig. 1) revealed that these seven strains indeed belonged to the V. furnissii species. The ANI values of the seven strains ranged from 98.1 to 99.9%, providing accurate identification at the species level. Further genomic analysis showed that these strains contained chromosomes of 4.80 Mb to 5.11 Mb in length, with GC contents ranging from 50.46 to 50.78%. These strains had 8 to 15 genomic islands and one to five prophages in their genomes (Table 2).
Fig. 1
Average nucleotide identity (ANI) analysis of the genomes of 19 Vibrio strains. The strains sequenced in this study are marked with *
×
![]()
Anzeige
Table 2
Genome characteristics of the seven Vibrio furnissii strains
Sample ID | Genome size (bp) | Fold-coverage | No. of contigs (> 500 bp) | Sequence GC% | No. of genes | No. of genomic islands | No. of Prophages | No. of tRNA | No. of rRNA | No. of sRNA |
|---|---|---|---|---|---|---|---|---|---|---|
VFBJ01 | 5,109,544 | 652 | 7 | 50.46 | 4,727 | 15 | 2 | 74 | 1 | 16 |
VFBJ02 | 4,906,640 | 679 | 24 | 50.75 | 4,596 | 8 | 4 | 86 | 7 | 15 |
VFBJ03 | 5,062,567 | 658 | 8 | 50.78 | 4,725 | 12 | 3 | 61 | 4 | 14 |
VFBJ04 | 4,904,255 | 679 | 25 | 50.58 | 4,594 | 13 | 2 | 95 | 6 | 21 |
VFBJ05 | 4,965,951 | 671 | 33 | 50.56 | 4,656 | 15 | 2 | 99 | 8 | 17 |
VFBJ06 | 4,802,519 | 694 | 28 | 50.74 | 4,489 | 11 | 1 | 71 | 6 | 15 |
VFBJ07 | 5,016,530 | 664 | 44 | 50.56 | 4,686 | 12 | 5 | 87 | 6 | 20 |
Population structure analysis
Next, we conducted a comparative genome analysis of the seven isolated strains and 24 V. furnissii reference sequences from GenBank (Fig. 2). These reference strains were mainly isolated in China, Columbia, the USA, Japan, the UK and Bangladesh during 2006–2022. Phylogenetic trees showed three major clades among the 31 V. furnissii strains. Our isolated strains were scattered in these three clades: clade 1 contained VFBJ02, VFBJ04 and VFBJ07; clade 2 contained VFBJ01 and VFBJ06; and clade 3 contained VFBJ03 and VFBJ05. The isolates in the same clade had shorter pairwise evolutionary distances but also occurred in different smaller monophyletic clades, meaning that they did not originate from the same colony.
Prediction of antimicrobial resistance genes
The antimicrobial resistance gene presence/absence profiles of V. furnissii were integrated with phylogenetic trees (Fig. 2). Among our seven isolated strains, VFBJ07 had the largest number of antibiotic resistance genes, including those for aminoglycoside resistance (strA, strB, aph(3’’)-Ib and aph(6)-Id), sulphonamide resistance (sul1 and sul2), tetracycline resistance (tetA, tetB and tetR), florfenicol/chloramphenicol resistance (floR) and quinolone resistance (aac(6’)-IIa). The second largest number of antibiotic resistance genes was found in VFBJ05, which had the same genes mentioned above except for aph(3’’)-Ib, sul1, tetB and aac(6’)-IIa. The strain VFBJ01 had the antibiotic resistance genes strA, strB, aph(3’’)-Ib, aph(6)-Id and sul2. In all of the 31 V. furnissii sequences, antimicrobial resistance genes were mainly found in strains isolated after 2017, including VFBJ07, VFBJ05, VFBJ01, GCA_007050385 (isolated from river sediments, Bangladesh, 2017), GCA_022289035 (isolated from a water sample, Columbia, 2019), GCF_024220035 (strain 104,486,766 [40], isolated from a stool sample, China, 2022) and GCA_021249365 [19] (isolated from hospital sewage, China, 2021). Among these strains, GCA_021249365 (strain VFN3), submitted by Zhuhai People’s Hospital, had the largest number of antimicrobial resistance genes.
Susceptibility to antibiotics
The antibiotic resistance profiles of the seven V. furnissii isolates to 19 antibiotic agents and the related antibiotic resistance genes are shown in Table 3. High resistance to cefazolin, tetracycline and streptomycin was found in 100.0%, 57.1% and 42.9% of the isolates, respectively, and intermediate resistance to ampicillin/sulbactam and imipenem was found in 85.7% and 85.7% of the isolates, respectively. Four strains – VFBJ01, VFBJ02, VFBJ05 and VFBJ07 – exhibited multi-drug resistance patterns: VFBJ02 was resistant to ampicillin, cefazolin, imipenem, meropenem and tetracycline, and VFBJ01, VFBJ05 and VFBJ07 were resistant to cefazolin, tetracycline and streptomycin.
Anzeige
Table 3
Antibiotic susceptibilitya patterns and the related antibiotic resistance genes in the seven Vibrio furnissii strains
Antibiotic1 (µg/mL) | VFBJ01 | VFBJ02 | VFBJ03 | VFBJ04 | VFBJ05 | VFBJ06 | VFBJ07 |
|---|---|---|---|---|---|---|---|
β-lactam/β-lactamase inhibitor combinations | |||||||
AMP | 16 (I) | 64 (R) | 16 (I) | 16 (I) | 16 (I) | 32 (R) | 16 (I) |
AMS | 16/8 (I) | 16/8 (I) | 16/8 (I) | 16/8 (I) | 8/4 (S) | 16/8 (I) | 16/8 (I) |
Cephems | |||||||
CFZ | 16 (R) | 32 (R) | 16 (R) | 16 (R) | 8 (R) | 32 (R) | 32 (R) |
CAZ | 1 (S) | 2 (S) | 0.5 (S) | 1 (S) | 0.5 (S) | 1 (S) | 0.5 (S) |
FEP | 1 (S) | 8 (I) | 1 (S) | 1 (S) | 1 (S) | 1 (S) | 1 (S) |
AZM | 2 (S) | 8 (I) | < 1 (S) | 2 (S) | 2 (S) | 1< (S) | < 1 (S) |
Aminoglycosides | |||||||
AMI | < 4 (S) | < 4 (S) | < 4 (S) | < 4 (S) | < 4 (S) | < 4 (S) | < 4 (S) |
GEN | < 1 (S) | < 1 (S) | < 1 (S) | < 1 (S) | < 1 (S) | < 1 (S) | < 1 (S) |
STR | > 32 (R) | 8 (S) | 16 (I) | 16 (I) | > 32 (R) | 8 (S) | > 32 (R) |
Aminoglycoside resistance genes | |||||||
StrA | + | - | - | - | + | - | + |
StrB | + | - | - | - | + | - | + |
aph (3’’)-Ib | + | - | - | - | - | - | + |
aph (6)-Id | + | - | - | - | + | - | + |
Fluoroquinolones | |||||||
CIP | 0.125 (S) | 0.25 (S) | 0.06 (S) | < 0.03 (S) | 0.125 (S) | 0.125 (S) | 0.125 (S) |
LEV | 0.5 (S) | 0.5 (S) | 0.25 (S) | < 0.125 (S) | 0.25 (S) | 0.25 (S) | 0.25 (S) |
Fluoroquinolone resistance gene | |||||||
aac (6’)-IIa | - | - | - | - | - | - | + |
Carbapenems | |||||||
IMI | 2 (I) | > 8 (R) | 2 (I) | 2 (I) | 2 (I) | 2 (I) | 2 (I) |
MEM | 0.25 (S) | 4 (R) | 0.25 (S) | 0.125 (S) | 0.125 (S) | 0.25 (S) | 0.125 (S) |
Tetracyclines | |||||||
TET | 32 (R) | 16 (R) | 2 (S) | 4 (S) | 16 (R) | 8 (I) | 32 (R) |
DOX | 8 (I) | 2 (S) | 2 (S) | 2 (S) | 4 (S) | 2 (S) | 8 (I) |
Tetracycline resistance genes | |||||||
tetA | - | - | - | - | + | - | + |
tetB | - | - | - | - | - | - | + |
tetD | - | - | - | - | - | - | - |
tetR | - | - | - | - | + | - | + |
Phenicols | |||||||
CHL | 8 (S) | < 2 (S) | < 2 (S) | < 2 (S) | 16 (I) | < 2 (S) | 8 (S) |
Phenicol resistance genes | |||||||
floR | - | - | - | - | + | - | + |
Folate pathway inhibitors | |||||||
SUL | 128 (S) | 64 (S) | < 32 (S) | 128 (S) | 128 (S) | < 32 (S) | 128 (S) |
SXT | > 8/152 (R) | < 0.25/4.75 (S) | < 0.25/4.75 (S) | < 0.25/4.75 (S) | < 0.25/4.75 (S) | < 0.25/4.75 (S) | < 0.25/4.75 (S) |
SUL resistance genes | |||||||
Sul1 | - | - | - | - | - | - | + |
Sul2 | + | - | - | - | + | - | + |
Macrolides | |||||||
AZI | < 2 (S) | < 2 (S) | < 2 (S) | < 2 (S) | < 2 (S) | < 2 (S) | < 2 (S) |
Plasmid and transposon island analyses
NCBI and Platon searches revealed no plasmid in our seven isolates, with only VFBJ05 and VFBJ07 showing a low coverage, ranging from 20 to 53%, with the sequence of the known plasmid isolated from V. furnissii (Fig. 2). Two of the 31 V. furnissii sequences, GCF_024220035 [40] and GCA_021249365 [19], contained the same plasmid NZ_CP089604.1 (pVFN3-blaOXA-193 K), an antimicrobial resistance gene-bearing conjugative plasmid [19]. However, whole-genome sequences revealed two plasmids present in VFBJ05 and VFBJ07. The plasmid in VFBJ05 was 85,626 bp long and showed 99.88% identity and 65% coverage with the plasmids CP040989.1 and CP064380.1 reported in V. furnissii strain FDAARGOS_777 and V. furnissii strain PartQ-Vfurnissii-RM8376, respectively. The plasmid of VFBJ07 was 71,247 bp long and showed 99.90% identity and 79% coverage with the plasmids CP040989.1 and CP064380.1.
Fig. 2
Phylogenetic tree, antimicrobial resistance and virulence gene presence/absence profiles, and plasmid analysis of Vibrio furnissii strains, including seven sequences from our study and 24 V. furnissii sequences from GenBank, were integrated and rendered using iTOL 3. The isolation year, regions and sources are also shown in the figure. The robustness of tree topologies was evaluated using 1,000 bootstrap replications
×
![]()
Transposon islands were found in VFBJ05 and VFBJ07 and contained antibiotic resistance genes. The transposon islands in VFBJ05 were located on the plasmid, while those in VFBJ07 were located on the chromosome. As shown in Figs. 3 and 4, there were two transposon islands in VFBJ05, one carrying strB, strA, tetA and sul2 and the other carrying floR. There were also two transposon islands in VFBJ07, one carrying strB, strA, tetA, sul1, sul2, floR and aac(6’)-IIa and the other carrying tetB.
Fig. 3
Transposon islands containing the antibiotic resistance genes in VFBJ05. One transposon island contained strB, strA, tetA and sul2, and the other contained floR
×
![]()
Fig. 4
Transposon islands containing the antibiotic resistance genes in VFBJ07. One transposon island contained strB, strA, tetA sul1, sul2, floR and aac(6’)-IIa, and the other contained tetB
×
![]()
Prediction of virulence-associated genes
The virulence gene presence/absence profiles of V. furnissii were also integrated with phylogenetic trees (Fig. 2). According to the VFDB, T6SS-related genes (vipA, vipB, vasB, vasD, vasE and hcp-2), ilpA (immunogenic lipoprotein A) [41] and quorum sensing-related gene luxS are widespread in V. furnissii. As V. furnissii was found to be phylogenetically close to V. fluvialis, the virulence factor-encoding genes vfh (V. fluvialis haemolysin) [42], hupO (hemin-binding outer membrane protein) [43] and vfp (V. fluvialis metalloprotease) [44], which widely occur in V. fluvialis, were also screened in the genome of V. furnissii.
Virulence phenotypes of V. furnissii
First, to confirm whether V. furnissii encodes the complete T6SS, we analysed the whole genomes of the seven V. furnissii strains and found that they all carry genes encoding T6SS elements, including several structural gene clusters and auxiliary gene clusters (Table 4). For example, the representative strain VFBJ01 encodes three core gene clusters and two auxiliary gene clusters (Fig. 5A). To assess whether the T6SS of these strains is active, a western blot assay was used to detect the expression of Hcp, a hallmark of functional T6SS [45]. The results showed that the seven strains constitutively expressed T6SS (Fig. 5B). We further evaluated the effect of their T6SS on inter-bacterial competition. In a competitive killing assay, E. coli MG1655 as prey was co-cultured without and with V. furnissii strains as predator. The result of this assay showed that the survival of MG1655 co-cultured with V. furnissii strains was evidently lower than that of MG1655 co-cultured without V. furnissii strains (Fig. 5C). Collectively, these data suggest that V. furnissii encodes an activated T6SS and uses it to compete with neighbouring bacterial cells. Further, the haemolysin-related gene vfh of V. furnissii showed 93–95% similarity to that in V. fluvialis, and haemolysin assays showed that V. furnissii strains grew in large colonies with strong beta-haemolysis on blood agar, similar to V. fluvialis [39] (Fig. 5D).
Table 4
Genetic characterisation of the T6SS in Vibrio furnissii
Strains | T6SS | Structural cluster | Auxiliary cluster | |
|---|---|---|---|---|
VFBJ01 VFBI02 | + + | 3 2 | 2 2 | |
VFBJ03 VFBJ04 VFBI05 VFBI06 VFBI07 | + + + + + | 2 2 2 2 2 | 3 2 2 2 2 | |
Fig. 5
(A) The genomic organisation of the structural genes of the T6SS gene clusters from the representative strain VFBJ01. Components encoded by the core genes are shown in various colours. (B) Western blot analysis of Hcp expression in seven Vibrio furnissii strains. RNA polymerase beta subunit (EF-Tu) was used as the loading control. Original images were included in Supplemental Fig. 1. (C) Quantification of Escherichia coli MG1655 survival after a T6SS attack by the indicated V. furnissii strains. Surviving prey cells (MG1655) were determined by serial dilution and plating on rifampin-containing LB agar plates. Representative images from three independent biological replicates are shown in B and C. (D) Quantification of haemolysin expression. Haemolysin assays showed these V. furnissii strains grew in large colonies with strong beta-haemolysis on blood agar. Haemolysin-positive control: V. cholerae N16961; haemolysin-negative control: E. coli MG1655
×
![]()
Discussion
Although V. furnissii and V. fluvialis are two closely related species, V. furnissii differs from V. fluvialis in its ability to produce gas through glucose fermentation. However, gas production from glucose fermentation has been reported be variable and thus not reliable for differentiating between the two species [46]. In our study, VITEK MS identified V. furnissii strains as accurately as did whole-genome sequencing, and other studies also used MS to confirm V. furnissii colonies [7, 10, 46]. These results indicate that MS could be used in clinical laboratories to identify V. furnissii quickly and easily.
Many selective-differential media have been developed for isolation of Vibrio spp., however, none of the media developed to date combines the sensitivity to low numbers with the specificity necessary to inhibit growth of other organisms [47]. AMP is frequently added to isolation medium as a selective agent when culturing Aeromonas [48]. In this study, as the isolated V. furnissii strains caused strong beta haemolysis on blood agar and had intermediate resistance or high resistance to AMP, blood agar plates with 20 µg/mL AMP also helped to isolate more V. furnissii compared with only using gentamicin selective medium.
V. furnissii has previously been isolated in inland cities such as Beijing, and its total isolation rate in our study was 0.4% (7/1985), which is close to the rate (0.5%) reported in Recife, Brazil [49], an Atlantic seaport city. The reason for such similar rates might be that seafood products are widely consumed worldwide and bacteria of the Vibrio genus can contaminate seafood [22]. Although the patients in our study reported no exposure to seawater or seafood, we inferred that seafood might have polluted the food they ate.
Phylogenetic analysis showed that the V. furnissii strains were clustered into three main clades, which were not directly related to the isolation location or host source. Our seven V. furnissii isolates were in different monophyletic clades in the phylogenetic tree, suggesting that these strains were not associated with outbreaks but were seven independent cases of gastroenteritis.
V. furnissii strains express many putative virulence factors, such as a series of virulence factors described in the genome analysis of V. furnissii NCTC11218 [20]. In addition, the virulence genes vfh, vfp and hupO, which occur widely in V. fluvialis, were also found in V. furnissii. VFH expression is associated with strong beta-haemolysis on blood agar [39] and can induce interleukin-1β secretion through the activation of the NLRP3 inflammasome [42]. Given the high similarity of vfh between V. fluvialis and V. furnissii, the haemolysin characterisation of V. furnissii was also likely to be related to vfh. VFP is a metalloprotease that exhibits haemagglutinating, permeability-enhancing, haemorrhagic and proteolytic activities [43]. HupO, an iron-regulated hemin-binding outer membrane protein, stimulates haemolysin production and resistance to oxidative stress [44]. V. furnissii also showed the presence of IlpA, a potent immunogenic lipoprotein that triggers cytokine production in human monocytes by activating Toll-like receptor 2 in V. vulnificus [41]. T6SS was first identified in V. cholerae [50] and Pseudomonas aeruginosa [51]. To date, it has been shown to be widely distributed in approximately 25% of all gram-negative bacteria [51]. T6SS can mediate V. cholerae infection in humans [50], and the T6SS of V. fluvialis is important for interbacterial competition [52]. In our study, V. furnissii was found to express functional T6SS.
V. furnissii strains were resistant to ampicillin (85.7%) and cephalothin (21.4%), a first-generation cephalosporin, in a Peru survey in 1995 [8]. Some case reports have also reported that their V. furnissii isolates were sensitive to most antibiotics except ampicillin [7, 9, 10]. In comparison, the V. furnissii strains in our study showed higher resistance to streptomycin, tetracycline and cefazolin (a first-generation cephalosporin), as well as high rates of intermediate resistance to ampicillin/sulbactam and imipenem. VFBJ02 was also resistant to imipenem and meropenem, both of which are carbapenems, although carbapenem resistance-related genes were not detected. This suggests the presence of an undiscovered carbapenem-resistant mechanism in V. furnissii. Imipenem is thus not recommended to treat V. furnissii infections. The CDC recommends combination therapy with doxycycline and ceftazidime to treat V. vulnificus infection [53]. Although antimicrobial therapy for V. furnissii infection has not yet been established, our results suggest that fluoroquinolones and third-generation cephalosporins, such as ceftazidime and doxycycline, are effective at treating V. furnissii infection.
Except for the streptomycin resistance shown by VFBJ01, VFBJ05 and VFBJ07 being completely consistent with the presence of strA, strB, aph(3’’)-Ib and aph(6)-Id genes, the antibiotic susceptibility patterns of the seven V. furnissii strains were partly consistent with the presence of other antibiotic resistance genes. For example, although VFBJ01, VFBJ05 and VFBJ07 carried sul genes, their sensitivity to sulphonamides only slightly declined. Similarly, prior studies have shown that resistance genes do not completely correlate with phenotypic resistance [54, 56]. In the present study, the tetracycline resistance of VFBJ05 and VFBJ07 was consistent with the positive detection of tetA, tetB and tetR; however, VFBJ01 and VFBJ02 were also resistant to tetracycline despite carrying no tet genes. This finding suggests that other unknown genes influence the relationship between the genotype and phenotype of tetracycline resistance. In addition, although VFBJ07 carried aac(6’)-IIa, its fluoroquinolone resistance level was only slightly raised and it remained sensitive to fluoroquinolones. The reason may be that several resistance mechanisms together contribute to fluoroquinolone resistance, including mutations in the quinolone resistance-determining genetic regions and plasmid-mediated quinolone resistance [57]. Ultimately, although the detection and bioinformatic analysis of antibiotic resistance genes cannot completely replace antibiotic susceptibility tests, the existence of antibiotic resistance genes might mean that the strains’ sensitivity to the corresponding antibiotics is declining, thus providing useful guidance for infectious disease treatment.
Several mobile genetic elements, such as insertion sequences, transposons and gene cassettes/integrons, can move within or between DNA molecules and transfer between bacterial cells [58]. GCA_021249365 was isolated from hospital sewage in Zhuhai, Guangdong province, China [19], and GCF_024220035 [40] was isolated from the stool sample of a 63-year-old man in Zhongshan, Guangdong province. These two strains were quite close on the phylogenetic tree, indicating that they might have originated from the same colony. The two strains also possessed the same antimicrobial resistance gene-bearing conjugative plasmid. In our samples, antibiotic resistance genes occurred on transposon islands in VFBJ05 and VFBJ07. The existence of transposon islands carrying antimicrobial resistance indicates that these strains may greatly increase the spread of drug resistance among clinical isolates through the process of infection, especially VFBJ05, in which transposon islands are located on a plasmid.
In conclusion, we found that diarrhoea associated with V. furnissii infection occurred sporadically and was more common than expected in the summer in Beijing, China. Fluoroquinolones and third-generation cephalosporins, such as ceftazidime and doxycycline, may be effective at treating V. furnissii infection. Two strains – VFBJ05 and VFBJ07 – were found to carry transposon islands containing antibiotic resistance genes. V. furnissii had unique virulence characteristics indicated mainly by the presence of T6SS and haemolysis. Overall, the results contribute to our understanding of the bioinformatic and clinical features of V. furnissii infections. Future academic and clinical efforts should focus on continual and improved laboratory-based surveillance to prevent and control V. furnissii infections and antibiotic resistance gene dissemination.
Acknowledgements
We thank the medical staff of the intestinal clinic of Beijing Friendship Hospital.
for recording the data of patients with diarrhoea. Professional English language editing support provided by AsiaEdit (asiaedit.com).
Declarations
Ethical approval and consent to participate
The study obtained ethical approval (2017-P2-095-01), in which the requirement for informed consent was waived by the Medical Ethics Committee of Beijing Friendship Hospital, Capital Medical University, Beijing, PR China.
Consent for publication
Not applicable.
Competing interests
The authors declare no conflict of interest.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.