Epidemiologic potentials and correlational analysis of Vibrio species and virulence toxins from water sources in greater Bushenyi districts, Uganda

Adequate water supply is one of the public health issues among the population living in low-income settings. Vibriosis remain a significant health challenge drawing the attention of both healthcare planners and researchers in South West districts of Uganda. Intending to clamp down the disease cases in the safest water deprive locality, we investigated the virulent toxins as contaminants and epidemiologic potentials of Vibrio species recovered from surface waters in greater Bushenyi districts, Uganda. Surface water sources within 46 villages located in the study districts were obtained between June and October 2018. Standard microbiological and molecular methods were used to analyse samples. Our results showed that 981 presumptive isolates retrieved cell counts of 10–100 CFU/g, with, with (640) 65% confirmed as Vibrio genus using polymerase chain reaction, which is distributed as follows; V. vulnificus 46/640 (7.2%), V. fluvialis 30/594 (5.1), V. parahaemolyticus 21/564 (3.7), V. cholera 5/543 (0.9), V. alginolyticus 62/538 (11.5) and V. mimicus 20/476 (4.2). The virulence toxins observed were heat-stable enterotoxin (stn) 46 (82.10%), V. vulnificus virulence gene (vcgCPI) 40 (87.00%), extracellular haemolysin gene {vfh 21 (70.00)} and Heme utilization protein gene {hupO 5 (16.70)}. The cluster analysis depicts hupO (4.46% n = 112); vfh (18.75%, n = 112); vcgCPI and stn (35.71%, & 41.07%, n = 112). The principal component analysis revealed the toxins (hupO, vfh) were correlated with the isolate recovered from Bohole water (BW) source, while (vcgCPI, stn) toxins are correlated with natural raw water (NRW) and open springs (OS) water sources isolates. Such observation indicates that surface waters sources are highly contaminated with an odds ratio of 1.00, 95% CI (70.48–90.5), attributed risk of (aR = 64.29) and relative risk of (RR = 73.91). In addition, it also implies that the surface waters sources have > 1 risk of contamination with vfh and > six times of contamination with hupO (aR = 40, − 66). This is a call of utmost importance to the population, which depends on these water sources to undertake appropriate sanitation, personal hygienic practices and potential measures that ensure water quality.

www.nature.com/scientificreports/ be wanting in their expected capacity to clamp down these diseases conditions that have continued to ravage the ordinary citizens in remote hard to reach areas where the significant population still fall sick, get worst and die without access to medical services and interventions 34,35 . Nevertheless, new pathogenic strains and virulence continue to emerge in endemic, pandemic and spreading to another region. It is worrisome that the precise role of most Vibrio spp., pathogenicity determinants in producing the clinical manifestations remains unclear.
These clearly calls for coordinated local and national response to make society a better place to be. Based on these premises, this study was designed to investigate the epidemiologic potentials of Vibrio species virulence toxins recovered from surface waters in greater Bushenyi districts, Uganda. The correlations association between the level of virulence toxins contamination and water sources were analysed.

Materials and methods
Study locations. The surface waters sources used in the four districts of the Western region of Uganda, including Bushenyi, Mitooma, Rubirizi and Sheema, were sampled for epidemiologic potentials of six pathogenic Vibrio species and virulence toxins. According to WHO, standard classification for drinking-water quality 36 . The surface waters used such as tap water, groundwater (borehole, open spring, ground running water, raw water, www.nature.com/scientificreports/ well water), Lakes, and fish pond were sampled from each of the 19 points in Bushenyi, 8 points in Mitooma, 11 points in Riburizi, and 8 points in Sheema districts as shown in Fig. 1A-D. Focus group discussion 37 was organised with stockholders that assisted in getting to the study's sampling identification. The group comprising (investigators, interpreters, microbiologists, district health officers, village local chairperson (LC1s)) were selected base on their relevance to this study. To be more specific, the investigators outlined the purpose and objectives of the study; the district health officer assisted in suggesting the sampling points that fit the purpose of the study. The microbiologist provided the advice that ensured that sampling was aseptic. The LC1s confirmed that the water for sampling is present in the 46 villages and nominated the village health teams (VHTs). After the focus group discussion, the VHTs nominated by the LC1s took the investigators to the sampling points as shown in Fig. 1A-D.
The maps of the study areas in (Bushenyi; Mitooma; Rubirizi; and Sheema districts) was created using the open-source software QGIS desktop version 3.0.3 38 . The United States Geographical Surveys (USGS) provided the Sentinel-2 image ID: L1C T35MRV A025815 20200601T082835 dated 2020/9/7, which was overlay on a shapefile for Uganda and highways, +e satellite image file was adjusted to show land vegetations.
Collection of samples, processing and enumeration of Vibrio spp.. A total of 46 villages was visited to obtain samples between June 2018 and October 2018.
Using sterilised Nalgene glass bottles (1000 ml) water samples were salvaged for four months in different sampling points each of the four Districts (Bushenyi, Mitooma, Rubirizi and Sheema) of South West of Uganda and conveyed on an ice-cool box to the department of medical microbiology laboratory, Kampala International University, Western-campus, Uganda for analysis within 6 h. Ten-folded dilution was carried out on the water samples as described by 39 , three series (10-1, 10-2 and 10-3) of which 1 ml of each dilution was spread plated onto thiosulphate citrate bile salts sucrose (TCBS) agar (Neogen, Lansing, MI 48912 USA) in triplicates for 24 h at 37 °C. The presumptive Vibrio spp., counted and expressed in colony-forming units per millilitres (CFU/ml) of water samples for the yellow and green colonies identified by colonial morphology and cultural characteristic of the colony 40,41 . A single colony of the presumptive Isolates was subsequently subcultured onto nutrient agar to ascertain purity; each sample's pure culture was picked and stored in glycerol stock for further analysis at the Applied and Environmental Microbiology Research Group (AEMREG) laboratory, Department of Biochemistry and Microbiology, University of Fort Hare, South Africa. Molecular confirmation of presumptive Vibrio genus and delineation into six pathogenic Vibrio species. The glycerol stocks were resuscitated using nutrient broth (Merck, Modderfontein, South Africa) and incubation for 24 h at 37 °C, while the genomic DNA of the 981 presumptive Vibrio spp., isolates were extracted following the boiling procedure described by Refs. 42,43 with modifications. The fresh overnight cultured isolates were subcultured into a sterile 1.5 ml microfuge tube and centrifuged (HERMLE, Siemensstr-25, D-78564 Wehingen, Germany) at a speed of 13,000 rpm for 10 min. The cells pellet was washed twice with phosphate-buffered saline, suspended on 500 µl sterile distilled water, and then lysed to release the DNA by boiling at 100 °C for 10 min pre-heated heating block (Techne heating block Dri-Block, DB-3D; Gauteng, Pretoria, South Africa). Afterwards, the suspension was centrifuged for 5 min at 15,000 rpm, and the supernatant was carefully pipetted into sterile Cryon tubes and stored at − 20 °C pending for use as a genomic DNA template PCR assays.
The primer pair F-5′CGG TGA AAT GCG TAG AGA T-3′ and R-5′TTA CTA GCG ATT CCG AGT TC-3′ previously described by Ref. 3 , was purchased from Inqaba Biotechnical Industries (Pty) Ltd., Pretoria, South Africa and used to amplify 16s RNA genes of Vibrio spp. with the amplicon size of 663 which was further delineated into the six pathogenic Vibrio species using the primers and condition in Table 1. The PCR reaction mixture of 25 µl (12 µl PCR master mix (New England BIOLABS), 1 µl of each forward and reverse primers, 6 µl of PCR grade water and 5 µl of genomic DNA template were amplified using BioRad T100 thermal Cycler Lasec. (621BR44012, Singapore). Afterwards, 4 µl of the amplicons were electrophoresed in 1.5% agarose gel using the thermal tank (Labnet, Enduro Gel XL, USA) on staining with ethidium bromide (0.5 µl) and 0.5× Tris-borate EDTA (TBE) buffer with a controlled base size of 100-bp DNA ladder (New England BIOLABS), Madison, WI, USA). A 100 Volt and 60 min electrophoresis process was done, and the gels were visualised under the UV trans-illuminator (Alliance 4.7, UVItec), Merton, London, UK.
Evaluation of virulence genes signature of Vibrio spp. recovered isolates. The virulence genes signature distribution in the confirmed Vibrio spp. isolates using PCR technique as described by 49,50 with modifications. The sets of primers indicating the targeted genes, sequence and conditions are presented in Table 2. The genomic DNA templates of confirmed isolates of Vibrio spp., including (62-Vibrio alginolyticus, 30-Vibrio fluvialis, 46-Vibrio vulnificus, 20-Vibrio mimicus, 21-Vibrio parahaemolyticus, 5-Vibrio cholerae) recovered from surface waters by adopting the earlier reported protocol described by Refs. 43,51,52 and the PCR reaction mixture was made up to a final volume of 25 μl while the amplified amplicons were electrophoresed and visualised as stated earlier.
Safety for research staff and environment. After the entire experiment, all specimens and isolates were decontaminated using autoclave at 121 °C, 15 PSI for 15 min. The decontaminated specimen and isolate were incinerated, and the ash was buried at the designated spot.
Statistical analysis. The result was entered into Microsoft excel. The distribution of toxins in the water sources was analysed using the violin box plot by considering Dunn's post hoct test to compare the occurrence of virulence toxin across the surface waters sources. Using multi-cluster analysis and Spearman's correlation   Table 3. Surface waters sources sampled in the districts. Author compilation as found in the districts.

Surface waters sources Uses
Bohole water Irrigation and domestic use www.nature.com/scientificreports/ coupled with Principal Component Analysis (PCA) was used to understand the correlations between the Vibrio spp., virulence toxins and surface water sources. All in RStudio version 3.5.1 software 55 . Furthermore, the significant epidemiological prevalence and risk estimate of toxins contaminations at 95% confidence interval evaluated in WINPEPI software version 11.65 56 . All statistical significant differences were recognised at p < 0.05.
Ethical consideration. The protocol for this study was reviewed by the research ethics committee of the Kampala International University, Western-Campus, Uganda, and obtained a clearance number of Nr.UG-REC-023/201919.

Results
The Surface waters sources studied in the region. South-Western Uganda is endowed with water resources and forest game reserved. The temperature ranges relatively between 19 and 24 °C through the year, even with the seasonal variations. We identified and studied nine primary types of surface waters sources used by the populaces in 46 villages of the greater Bushenyi districts. These surface waters were surveyed for the distribution and prevalence of Vibrio spp. and virulence toxins. The surface waters used by the general population across numerous purposes of life from domestic to agricultural uses and recreational/medical tourist attractions in the districts are highlighted in Table 3. The map of sample collection points in the four districts are shown in . Nevertheless, all other virulence toxins investigated in this study showed a 0.00% prevalence and no contamination (aR = − 100). The burden of risk of vcgCPI virulence toxin contaminations of the surface waters sources is of significant concern to the public (RR = 6.67) identified as shown in Table 5.
Using the multi-way cluster analysis, the pathogenic species virulence toxins was grouped according to their frequency of occurrence. Four clusters of virulence determinants were distinguished on the base of their frequency of occurrence (percentage number) and the surface water they were identified in. The calculations showed that the virulence toxins (vcgEP2, vfpA, trd, trh, flaE, ompU, zot, hylA) clustered into no. 1, (0.00%, n = 112); hupO clustered into 2, (4.46% n = 112); vfh clustered into no. 3, (18.75%, n = 112); and vcgCPI and stn clustered into no. www.nature.com/scientificreports/ Correlations patterns between Vibrio spp. virulence toxins and surface water sources. We use principal component analysis PCA to study the multivariate association between the distribution of Virulence toxins and Surface waters sources. The results obtained from the PCA in Fig. 7 showed correlations between virulence toxins level contamination and Surface water sources of the isolated Vibrio spp. Interestingly, the Vibriosis virulence toxins (trd, vfpA, trh, vcgEP, zot, flaE, ompU) show no correlation with the isolates recovered from the water sources. On the other hand, the toxins (hupO, vfh) are positively correlated with the isolate recovered from the bohole water (BW) source. Similarly, (vcgCPI, stn) toxins are positively correlated with natural raw water (NRW) and Open Springs (OS) water sources isolates.
To better understand the result, details of variables contained in PCA1 and PCA2, which explain the total correlation of 91.7%, were analysed further as recommended by scree plot of eigenvalues evaluation (p < 0.05). The www.nature.com/scientificreports/ PCA1 accounts for 82.3% variability of which vcgCPI is the most strongly corrected, with sufficient correlation significant (r = 0.9531264 and p. value of 0.00007), stn (r = 0.9059934 and p. value of 0.00076), vfh (r = 0.8876029 and p. value of 0.00140) and hupO (r = 0.8807374 and p. value of 0.00171) associated with surface waters isolates. While the PCA2 account for only 10.4%, with insufficient significant correlation in Fig. 7.

Discussion
For the past decade, an array of several virulence determinants implicated in the pathogenesis of vibriosis have been reported both from clinical and environmental strains across the world. This study screened the prevalence and epidemiological significance of the virulence toxins stn, vcgCPI, vcgEP2, vfh, hupO, vfpA, trd, trh, flaE, ompU, and zot by PCR. The density of the Vibrio bacteria enumerated varies between months and water sources, with the highest densities record in September compared to other months ( Fig. 2A-D). This may be due to changes in moisture and poor hygienic practice in some districts where poor probability orientation-also, the majority of the population living in the hard-to-reach region practice open defecation 58 . The geographical location of the sampling sites has a significant contribution to the changing pattern of Vibrio densities observed in the respective months. However, the Vibrio densities obtained are sufficient to initiate an infection in humans, which is in harmony with the studies of various investigators 54 . In addition, Vibrio species quantity (of about 2 × 10 5 to 3 × 10 7 CFU/ml) with a cultivation time spanning 4-96 h (approximately 15 h) is sufficient to cause acute gastroenteritis 59,60 as previously reported by various investigators. Table 4 and Fig. 3 show that the pattern or profile of Vibrio strains recovered from the environmental samples depicts the magnitude of pathogenicity seen among Vibrio strains. Interestingly, the two major classes of Vibrio spp. and its associated infections were adequately represented among the isolates recovered. Thus, cholerae strain (ompW) and non-cholera strains (V. vulnificus (toxR), V. fluvialis (toxR), V. parahaemolyticus (toxR), V. alginolyticus (Vg gyrB), V. mimicus (VM) were the observed prevalence strains. Such Vibrio spp., infections could be worse in the immune-compromised individual 61 ; all the same, the infections are often self-limited 62 . This is also similar to 63 reports that the worst infection is observed in immunesuppressed patients. Major enterotoxins are expressed by Vibrio strains, including V. cholerae, V. paraheamolyticus, and V. mimicus. Interestingly, our result shows a very much high prevalence of 46/56 (82.10%) in the studied water sources as depicted in Table 4 (Figs. 4, 5), compared to the study of 64 , which shows a relatively low frequency of 28.2% amongst non-O1/non-O139 strains. Their study revealed 10.5% of toxigenic V. cholerae O1 and 14.3% among O139 serogroups belonging to Vibrio cholerae, which are recovered from environmental samples in Europe. In Thailand, 10/21 of clinical isolates were recovered 65 , and 26/193 (13.5%) in V. fluvialis of an environmental specimen of South Africa 11 .
All the Vibrio strains studied show 35.44% positive to the virulence genes, and specifically, its result is as follows: Lake (27.68%), natural raw water sources (18.75%) and open Springs (13.39%), as shown in Table 5, indicating a zero-tolerance limit for vibriosis infections.
The absence of thermostable direct haemolysin (tdh) and the tdh related haemolysin (trh) virulence toxins responsible for the pathogenicity of V. parahaemolyticus is in agreement with the recent report of genes for cholerae toxin (ctx), thermostable direct hemolysin (tdh), or zonula occludens toxin (zot) as there were not detected in any of 116 isolates of seawater in Norway 31 . Similar to the study of 68 Lake isolates in Ohio US, and 69 ctxA, tcpA, and zot were not detected in the V. cholerae strains, while hlyA, rtxA, and rtxC were positive for water sample isolate in China 69 . However, these toxins have been reported in a relatively low occurrence in environmental samples in Malaysian 70 , in Turkey 9 in Europe and Atlantic coast in Spain 71 , in Italy 72 . Also, it has been reported that the highly cytotoxic and human gastrointestinal infecting Vibrio parahaemolyticus strains of environmental origin with no detection of the tdh or trh genes 10 were observed. We are not surprised about the difference and absence of some virulence toxins in the Vibrio strains. Most virulence toxins predominate in clinical isolate sources of toxigenic vibriosis, e.g. ctxAB or tdh and trh are predominant in a clinical strain of V. cholera and V. parahaemolyticus 17 . Although Virulence toxins/genes hupO, vfh and vfpA are often detected in V. fluvialis of both patient isolates, and seafood isolates strains, some Vibrio strains virulence can be prevalent irrespective of the origin.
The Vibrio strains may have acquired the virulence toxins stn by horizontal gene transfer or natural genetic exchange by Organism interactions in the ecosystems or human host 73,74 . The significant role of stn is unclear in vibriosis pathogenesis (Table 5). However, the high frequency of concern is a threat to the population using the water sources.
The result of the virulence-correlated gene (vcg) of Vulnificus, vcgCPI for clinical (C-) genotypes and vcgEP2 environmental (E-) genotypes were observed to vcgEP2 40/46 (87%), and vcgCPI 0/46 (0%) is similar to the findings of 75 76 . It also implies a difference to the report where almost an equal per cent of vcgE (46.9%) and vcgC (53.1%) were detected from oyster isolates 77 and water areas surrounding oyster harvest 75 . Consequently, it is of necessity to continuously monitor surface water source uses, although Vulnificus infection is frequent in aquaculture and rare in humans but can be fatal in immunocompromised persons, causing wound ulceration infections, gastroenteritis or septicemia.   www.nature.com/scientificreports/ The result also showed a prevalence of 21/30 (70%) of vfh genes in the Vibrio flavilis strains, depicting health as significant as the virulence phenotypes were predominant in this species. Vibrio cholera and Vibrio vulnificus 78,79 utilised 70% of vfp toxin acting as homologous precursor proteins of metalloproteases during pathogenicity in humans. Specifically, Vfh expressed by Vibrio vulnificus protease is implicated in proteolytic activity as well in haemagglutinating enhancing permeability and haemorrhagic activities 17,80 . Also, the report showed that vfp virulence is predominant in Vibrio strains of clinical origin of and the expression of these toxins could be more virulent in their pathogenesis 17 .
Applying multi-way statistical computations provides a novel technique to interpret data on the prevalence of virulent toxins and contaminations levels associated with the water sources. Food and water research rarely adopt such calculations, except in few studies 57,81,82 . However, this approach has not been untaken in literature; it gave us a more in-depth characteristic of Vibrio virulence toxin prevalence in the water samples. The multi-way computations enabled us to identify stn and vcgCPI as the most frequent toxins occurrence in sources terms of percentages. The PCA results indicate the prevalence of toxins in correlations to their sources; the analysed shown hupO and vfh are associated with bohole water (BW) source while vcgCPI, and stn, are positively correlated with  Table 5. Prevalence distribution of virulence toxins of human pathogenic Vibrio spp. by PCR (relative risk and attributable risk), odds ratios in surface waters of each district. CI confidence interval, LL lower limit, UL upper limit, RR relative risk, aR attributed risk. www.nature.com/scientificreports/ natural raw water (NRW) and open springs (OS) water (Fig. 7). This study has expanded the baseline databases for Vibrio spp., and associated infections in this region. Therefore, the findings of this study have provided the basis for future studies lasting up to 3 years design to establish a trend.

Conclusion
In this study, the analysed Vibrio spp., recovered from the water sources used in the region of Uganda, were found to harbour virulence toxins of significant potential health concern. This is concerned with causing diseases associated explicitly with diarrhoeagenic infections, septicemia, and outbreaks of vibriosis in the region where there is an inadequate water supply or water treatment. The heat-stable enterotoxin (stn) and V. vulnificus virulence genes (vcgCPI) were the most frequently occurring toxins in Lakes and natural raw water in the region. The use