Introduction

Antimicrobial resistance in bacteria has become a serious public health problem worldwide (WHO 2015). Initially, concerns about the problem of resistance focused on hospitals, but recently, the presence of clinically relevant antimicrobial resistance genes in the natural environment has become a source of worry to the scientific community (Pruden et al. 2013; Berendonk et al. 2015). Routes of exposure to antimicrobial resistance outside the hospital setting include contact with human and animal carriers, and contaminated food and water (Valverde et al. 2008; Kennedy and Collignon 2010; Coleman et al. 2012). Exposure through the water route can occur by drinking contaminated water or by contact with recreational water at freshwater and marine beaches.

Beach tourism has evolved to be an important and rapidly growing activity in the world due to its suitability for recreation and sporting activities (Orams 2003) leading to a concomitant increase in the incidence of infections associated with recreational waters over the past decade (Halliday and Gast 2011). Thus, the current challenge posed by the presence of antimicrobial-resistant bacteria in the environment is an important public health issue for beach goers and those living in coastal areas. This problem is more acute in the developing world where the aquatic ecosystem serves as a sink for untreated wastewater from domestic, industrial, hospital, agricultural, and aquaculture sources. Through this practice, residual antimicrobial agents, antimicrobial-resistant bacteria, and antimicrobial resistance genes enter the aquatic environment (Taylor et al. 2011). In addition, the usual flow of human traffic to coastal areas in search of jobs and recreation with the attendant anthropogenic impact, and the corresponding lack of surface water quality monitoring in many developing countries, make recreational beaches in these countries an area of special interest for the development and spread of antimicrobial resistance (Overbey et al. 2015). Unfortunately, very little is known about the health status of beaches in Nigeria, especially as it relates to the presence of bacteria carrying clinically relevant antimicrobial resistance genes.

Lagos State, located in the south-western part of Nigeria on the narrow coastal floodplain of the Bight of Benin, consists of a long coastline of about 255 km long stretching from Epe in the east to Cotonu in the Republic of Benin to the west. This long coastline is home to a number of marine recreational beaches who by virtue of their location in the tropical ecosystem makes them potential breeding ground for the proliferation of pathogenic or potentially pathogenic bacteria such as Vibrio species whose presence in the aquatic ecosystem has been known to be positively correlated with mean water temperature (Blackwell and Oliver 2008; Rosec et al. 2009; Johnson et al. 2010; Urquhart et al. 2016).

Potentially, human pathogenic species of Vibrio such as V. parahaemolyticus, V. vulnificus, and non-O1/non-O39 V. cholera are common inhabitants of marine and brackish water environments worldwide (Fabbro et al. 2010; Johnson et al. 2012; Oliver et al. 2013; Ceccarelli et al. 2013; Wu et al. 2014). These organisms are important causal agent of septicemia, and gastrointestinal and wound infections, especially in immunocompromised patients (Daniels and Shafaie 2000; Drake et al. 2007). V. parahaemolyticus and V. vulnificus in particular are leading causes of human intestinal and wound infections after consumption of raw and undercooked seafood or contact with seawater during beach activities (Letchumanan et al. 2014). However, despite the growing importance of environmental Vibrio species as agents of opportunistic infections and reservoir of antimicrobial resistance genes (García-Aljaro et al. 2014), most studies of antimicrobial resistance in Vibrio species have focused on clinical and aquaculture isolates (Okoh and Igbinosa 2010). Only few studies have investigated the resistance of environmental isolates of Vibrio to antimicrobial agents (Baker-Austin et al. 2008; Okoh and Igbinosa 2010; Mudryk et al. 2013; Shaw et al. 2014; Baron et al. 2016; De Menezes et al. 2017; Ghenem and Elhadi 2018); fewer still among these studies focused on recreational beaches and none of these studies emanated from Nigeria, the most populous nation in Africa where little is known of antimicrobial resistance in environmental reservoirs.

In this study, we investigated the prevalence of resistance to selected antimicrobials, and the presence of virulence and antimicrobial resistance genes in V. parahaemolyticus and V. vulnificus isolated from public beaches along the shoreline of Lagos in southwestern Nigeria. We also used pulse field gel electrophoresis (PFGE) to detect clonal relationship among V. parahaemolyticus isolates. Our aim is to generate important baseline information on the health status of Nigerian recreational beaches and their role as reservoir of antimicrobial resistance that can pose a threat to the health of beach goers.

Material and methods

Strains of V. parahaemolyticus and V. vulnificus used in this study

Eighty-five strains of V. parahaemolyticus and V. vulnificus were isolated from Eleko (ELK), Lekki (LEK), Takwa Bay (TAK), and Badagry (BDG) beaches, Lagos, Nigeria, as previously described (Oyelade et al. 2018). Typical V. parahaemolyticus and V. vulnificus colonies were selected from CHROMagar™ Vibrio plates (CHROMagar, Paris, France) after enrichment in alkaline peptone water. Presumptive V. vulnificus colonies were further subcultured onto mCPC agar and TCBS agar before the identities of both V. parahaemolyticus and V. vulnificus were confirmed using API20E (API system, France) (Dumontet et al. 2000) and qPCR detection of tlh genes (V. prahaemolyticus) and vvhA genes (V. vulnificus) as previously described (Nordstrom et al. 2007; Campbell and Wright 2003). The isolated organisms were screened for resistance to ceftazidime (16 μg/ml) and meropenem (4 μg/ml) as indicators of extended spectrum beta-lactamase (ESBL) and carbapenemase production based on CLSI breakpoints listed in Table 20 of CLSI document M45-A3 (CLSI 2016). Forty isolates (26 V. parahaemolyticus and 14 V. vulnificus) that showed resistance to ceftazidime and meropenem were selected for further analysis. The origins, sources, and dates of isolation of the strains are listed in Table 1.

Table 1 Origin and source of V. parahaemolyticus and V. vulnificus used in this study
Full size table

DNA extraction

Genomic DNA was extracted from the isolates by boiling lysis as described by Blackstone et al. (2003) and used as template in the qPCR detection of virulence and antimicrobial resistance genes as described below.

Real-Time qPCR amplification of tlh, tdh, trh, and vvha

Real-time qPCR was used for the multiplex detection and quantification of the tlh, tdh, trh (V. parahaemolyticus), and vvha (V. vulnificus) genes with primers (Table 2) and conditions previously described (Campbell and Wright 2003; Nordstrom et al. 2007) using an ABI 7500 instrument (Applied Biosystems Carlsbad, CA, USA). The reaction mixture (25 μl) contains the following: 1× PCR buffer (Invitrogen, Carlsbad, CA, USA), 5.0 mM MgCl2 (Invitrogen), 300 nM of each of the dNTPs (Roche, Indianapolis, USA), 300 nM each of the trh and vvhA primers, 200 nM tlh primers, 100 nM of tdh primers, 75 nM of tlh and vvha IAC primers, 25 nM of tdh and trh IAC primers, 150 nM tlh probe and IAC, 75 nM tdh and trh probes, 200 nM vvha probe, and 1.50 U, 1.12 U, 2.25 U platinum Taq polymerase (Invitrogen) for tlh, vvha, and tdh/trh respectively. All primers, IACs, tlh, and vvha probes were obtained from Integrated DNA Technologies (Coralville, IA, USA) while the tdh and trh probes were obtained from Life Technologies (Carlsbad, CA, USA). The rest of the reaction mixture volume consisted of a previously quantified IAC DNA (2 μl), nuclease-free water and DNA templates (2 μl of boiled cell lysates). V. parahaemolyticus ATCC 17802 carrying tlh, tdh, and trh, and V. vulnificus ATCC 27562 possessing vvha gene were used as positive controls while nuclease-free water served as negative control.

Table 2 Serotype and detection of tlh and pathogenic determinants tdh and trh among the V. parahaemolyticus isolates
Full size table

Serotyping of V. parahaemolyticus

Serotyping of the V. parahaemolyticus isolates was performed using a commercially available V. parahaemolyticus antiserum test kit (Denka Seiken, Tokyo, Japan) according to the manufacturer’s instructions. For O antigen typing, a heavy cell suspension was prepared in a normal saline solution and was then autoclaved at 121 °C for 2 h. A 1-μl loopful of the autoclaved suspension was added to one drop of each O antiserum which was observed for agglutination. For K antigen typing, a 1-μl loopful of isolate colony was mixed with one drop of each pool of K antisera and was observed for agglutination. If an isolate agglutinated in any K pool, the individual antisera of that pool were tested.

Antimicrobial resistance analysis

Antimicrobial resistance analysis was carried out using MicroScan Dried Gram Negative antibiotic panel (Beckman Coulter, Inc., Brea, CA, USA) containing 30 antimicrobial agents. The list of antimicrobial agents used, breakpoints, and references are listed in Table S3. The production of extended spectrum beta-lactamase (ESBL) in isolates showing phenotypic resistance to third-generation cephalosporins was confirmed using MicroScan ESBL plus panel. Inoculum preparation was performed in 2% NaCl solution according to the manufacturer’s instruction based on CLSI document M07-A10 (CLSI 2015a). Test panels were incubated for 16–20 h at 35 °C and results interpreted using the CLSI breakpoints described for Enterobacteriaceae in Table 2A of document M100-S25 for the following antimicrobial agents: aztreonam, cefotetan, ceftriaxone, cephalothin, ertapenem, nitrofurantoin, ticarcillin/K clavulanate, and tobramycin (CLSI 2015b). Breakpoints specific for Vibrio spp. in Table 20 of document M45-A3 was used for interpretation of the following antimicrobial agents: amikacin, ampicillin, ampicillin/sulbactam, cefazolin, cefepime, cefotaxime, cefoxitin, ceftazidime, cefuroxime, chloramphenicol, ciprofloxacin, gentamicin, imipenem, levofloxacin, meropenem, piperacillin, piperacillin/tazobactam, tetracycline, and trimethoprim/sulfamethoxazole (CLSI 2016). Cefotaxime/K clavulanate and ceftazidime/K clavulanate do not have interpretive breakpoint criteria, a ≥ 3-fold MIC decrease between CAZ and CAZ/CA; CTX and CTX/CA are considered positive for ESBL production according to the manufacturer’s manual. For tigecycline, the manufacturer’s breakpoint was used.

Detection of beta-lactams resistance genes

qPCR was carried out to detect β-lactams resistance genes from ESBL-producing V. parahaemolyticus) and V. vulnificus isolates with primers targeting the following beta-lactam resistance genes: blaKPC, blaNDM-1, blaTEM, blaSHV, blaCTX-M-1 and -M-9 (Cluster A), and blaCMY (Table S1). Total genomic DNA for antimicrobial resistance gene detection was extracted from each isolate according to the Center for Disease Control and Prevention (CDC) protocol for Multiplex Real-Time PCR Detection of blaKPC and blaNDM-1 (CDC 2011). The reaction mixture (25 μl) contains 1× PCR buffer (Invitrogen, Carlsbad, CA, USA), 5.0 mM MgCl2 (Invitrogen), 300 nM of each of the dNTPs (Roche, Indianapolis, USA), 200 nM forward and reverse primers (Integrated DNA Technologies), 75 nM for IAC forward and reverse primers (Integrated DNA Technologies), 150 nM probes (Integrated DNA Technologies), and 1.50 U, platinum Taq polymerase (Invitrogen). The rest of the reaction mixture volume consisted of a previously quantified IAC DNA (2 μl), nuclease-free water, and a DNA template (2 μl cell lysates). The cycling parameters consisted of a 95 °C hold for 3 min for the initial denaturation and activation of the hot-start Taq polymerase, followed by 35 cycles of amplification, with each cycle consisting of denaturation at 95 °C for 3 s and a combined annealing/extension step at 60 °C for 30 s. Klebsiella pneumoniae ATCC BAA-1705 strain possessing blaKPC (as described by ATCC) and K. pneumoniae ATCC BAA-2146 carrying the blaNDM-1 (as described by ATCC), blaTEM, blaSHV, blaCTX-M-15, and blaCMY (Leski et al. 2012) were used as positive controls while the negative control was nuclease-free water added to the PCR mix as a template.

Pulse field gel electrophoresis analysis of V. parahaemolyticus

Pulse field gel electrophoresis (PFGE) was performed according to the CDC PulseNet protocol for V. parahaemolyticus (CDC 2013) using Sfi I (40 U/sample) (Roche Diagnostics, Indianapolis, IN, USA) for restriction digestion. The restricted fragments were resolved in 1% seakem gold agarose gel in 0.5% Tris-boric acid-EDTA buffer using CHEF Mapper system (Bio-Rad Laboratories, Richmond, CA, USA). The following electrophoresis conditions were used: Auto Algorithm of 78 kb and 396 kb for low and high molecular weight, initial and final switch time of 10 s and 35.03 s, and run time of 16.5 h at 14 °C. Voltage density and included angle were 6.0 V/cm and 120° respectively, while linear ramping and pump speed was 1 l/min. Xba I digested DNA from Salmonella enterica serovar Braenderup strain H9812 was used as a molecular size marker. Stained gels were visualized using Gel Doc XR+ (Bio-Rad Laboratories, Richmond, CA, USA). The PFGE patterns were analyzed using BioNumerics software version 7.6 (Applied Maths, Austin, TX, USA). Clustering was performed using the unweighted pair group method and the Dice correlation coefficient with position tolerance of 1.5%. Clusters were defined on the basis of an 80% similarity cutoff (CDC 2013).

Results and discussion

In Nigeria, studies of Vibrio species are mostly focused on the detection of these organisms in seafoods (Oramadike and Ogunbanwo 2015; Adebayo-Tayo et al. 2011) with few studies focusing on environmental isolates (Eyisi et al. 2013). In this study, we isolated V. parahaemolyticus and V. vulnificus from recreational beaches along the shoreline of Lagos, Nigeria. V. parahaemolyticus was isolated from all the beaches sampled and occurred more frequently in beach sand compared to seawater samples (Table 1). V. vulnificus was, however, only detected from water samples of ELK, LEK, and TAK, and sand from ELK and LEK.

Detection of tlh and tdh/trh genes and serotyping of V. parahaemolyticus

The results of real-time PCR assay showed that all the tested V. parahaemolyticus were positive for thermolabile hemolysin (tlh) gene (Table 2) confirming the identity of these isolates as V. parahaemolyticus (Brasher et al. 1998; Bej et al. 1999; Rojas et al. 2011). Similarly, all V. vulnificus isolates were positive for the cytotoxin hemolysin protein encoded by vvha. Further, results of real-time PCR assay for detection of virulence genes tdh and trh showed that five isolates were positive for tdh and two isolates for trh while one isolate (LEK 108) carried both tdh and trh (Table 2). tdh and trh are the most predictive indicators of potential virulence in V. parahaemolyticus (Baker-Austin et al. 2008) and are responsible for hemolytic activity in many clinical strains (Ceccarelli et al. 2013). Similar to our results, previous studies (Cook et al. 2002; Hara-Kudo et al. 2003; Robert-Pillot et al. 2004; Baker-Austin et al. 2008; Li et al. 2015; He et al. 2016; Ghenem and Elhadi 2018) have reported a low prevalence of these virulence factors in food and environmental V. parahaemolyticus strains, confirming that non-clinical isolates of this bacteria rarely possess virulence factors. However, the prevalence of virulence factors reported in our study is higher than those reported in the studies cited above. Additionally, we also detected the co-occurrence of tdh and trh genes in an isolate from one of our study sites.

Serotyping revealed 10 different serogroups among the 26 isolates based on O and K antigens (Table 2). The most prevalent serogroup is the O5:Kut (K untypeable) with six isolates, followed by O1:Kut (4 isolates), O3:K31 and O4:Kut (3 isolates each), O2:K3, O3:K6, O8:K70, and O11:Kut with two isolates each, and O5:K30 and O10:Kut with one isolate each. Two serotypes of public health importance were isolated in three of the beaches included in this study. The first of such serotype, O3:K6, was detected in V. parahaemolyticus strain isolated from ELK. Serotype O3:K6 has been reported in large outbreaks of V. parahaemolyticus infections worldwide (Raszl et al. 2016) but are rarely detected in food and environmental isolates (Raghunath 2014). Secondly, serotype O4:Kut, known to be more virulent than other pathogenic strains of V. parahaemolyticus and has previously caused large outbreaks in the USA, Europe, and South America (Martinez-Urtaza et al. 2013; Newton et al. 2014; Haendiges et al. 2015), was isolated from ELK, TAK, and BDG. The detection of these serotypes in the Nigerian coastal environment, especially from recreational beaches, is of great public health concern for beach users.

Antimicrobial resistance analysis of V. parahaemolyticus and V. vulnificus

Results of susceptibility testing using 30 antibiotics revealed that all the V. parahaemolyticus and V. vulnificus isolates showed resistance to multiple antibiotics with the least being resistance to seven antibiotics in combination. There was 100% resistance to AMP, CFZ, CEF, and PIP among the isolates; and all V. parahaemolyticus isolates were in addition resistant to CAZ, TIM, AMK, and TOB (Fig. 1, Table S2). In contrast, all isolates were sensitive to CHL, LVX, and TGC while in addition to these three antibiotics, all V. parahaemolyticus were sensitive to CTX, CTX/CA, CRO, CTT, SXT, CIP, NIT, and TET (Fig. 1, Table S2). One V. parahaemolyticus isolate from Takwa Bay (TAK 145) and ten V. vulnificus isolates from ELK (ELK 125, 173, 175), LEK (95B, 164), and TAK (TAK 141, 178, 181B, 184, 196) showed resistance to carbapenems (IMP, MER, and ETP) (Table S2).

Fig. 1
figure 1

Frequency of resistance to antimicrobial agents among V. parahaemolyticus and V. vulnificus isolates from the Nigerian Beaches. AMK amikacin, AMP ampicillin, ATM aztreonam, CAZ Ceftazidime, CAZ/CA ceftazidime/clavulanate, CEF cephalothin, CFZ cefazolin, CHL chloramphenicol, CIP ciprofloxacin, CRO ceftriaxone, CTT cefotetan, CTX cefotaxime, CTX/CA cefotaxime/clavulanate, CXM cefuroxime, ETP ertapenem, FEP cefepime, FOX cefoxitin, GEN gentamicin, IMP imipenem, LVX levofloxacin, MER meropenem, NIT nitrofuratoin, PIP piperacillin, SAM ampicillin/sulbactam, SXT trimethoprim/sulfamethoxazole, TET tetracycline, TGC tigecycline, TIM ticarcillin/clavulanate, TOB tobramycin, TZP piperacillin/taxobactam,

Full size image

The widespread resistance shown by the isolates of the present study to third (CAZ, CRO, CTX) and fourth-generation (FEP) cephalosporins, which remains rare among Vibrio species (Han et al. 2007; Baker-Austin et al. 2008; Wong et al. 2012; Li et al. 2015), is worrisome. Further, in contrast to previous observations by Baker-Austin et al. (2008), we found no significant difference in the resistance (12.6 antimicrobials per isolate) of V. parahaemolyticus isolates possessing the tdh and trh virulence determinants their counterparts lacking these genes (13.5 antimicrobials per isolate).

Detection of β-lactams resistance genes in ESBL-producing Vibrio isolates

The New Delhi metallo-β-lactamase gene blaNDM-1 was found in five strains of V. vulnificus from ELK (ELK125, ELK175), LEK (LEK164), and TAK (TAK141, TAK196) and one V. parahaemolyticus isolated from TAK (TAK145) (Table 3). All the blaNDM-1-positive isolates except one which carried blaTEM (ELK175) were also positive for blaTEM and blaCMY. None of the tested genes was found in one isolate of V. vulnificus (TAK 178). All the blaNDM-1 producers showed resistance to multiple antibiotics with the least being resistant to 17 antibiotics. One blaNDM-1 producer (ELK175) was resistant to 24 of the 30 antibiotics tested. blaNDM-1 producers are always resistant to several classes of antibiotics (Rolain et al. 2010; Nordmann and Poirel 2014).

Table 3 Profile of antibiotic resistance genes detected among ESBL-producing isolates
Full size table

There are increasing reports of the carriage of cephalosporinase-encoding genes among Vibrio species in other parts of the world. These include blaCMY-2 located on IncA/C plasmids in V. parahaemolyticus (Li et al. 2015) and V. alginolyticus (Ye et al. 2016), blaVEB-2 in V. parahaemolyticus (Li et al. 2016), and blaPER-1 located on an IncN plasmid in V. parahaemolyticus, all isolated from seafood in China. Further complicating this disturbing observation is the increasing report of carbapenemases, which in some cases are novel (Bier et al. 2015; Mangat et al. 2016) in members of the genus Vibrionaceae. Previous studies have detected blaNDM-1 in V. parahaemolyticus isolated from shrimps (Briet et al. 2018) and other Vibrio species such as environmental V. cholera in India (Walsh et al. 2011), clinical V. cholera O1 E1 Tor in india (Mandal et al. 2012), environmental non O1/non O39 V. cholera in Vietnam (Diep et al. 2015); clinical non O1/non O39 V. cholera in the UK (Darley et al. 2012), and V. fluvialis of clinical origin in India (Chowdhury et al. 2016). However, reports of blaNDM-1 in V. parahaemolyticus and V. vulnificus are not very common. The presence of these genes in bacteria of public health importance such as V. parahaemolyticus and V. vulnificus portend great danger for beach goers as they not only limit treatment options in case of infection, but they can also transmit their resistance elements to other pathogenic and potentially pathogenic bacteria in the beach ecosystem.

Pulse field gel electrophoresis analysis of V. parahaemolyticus strains

The PFGE profiles of the SfiI-digested DNA of V. parahaemolyticus are presented in Fig. 2. A total of 23 pulsotypes were recognized among 26 V. parahaemolyticus strains with 19 isolates showing distinct banding patterns; the remaining four banding patterns included at least two isolates. This suggested a large degree of genomic variation among the isolates even within the same sampling sites. It also suggested that the isolates included in this study are not just clonal replicates. The isolates sharing the same banding patterns are TAK 145 and BDG 151B both belonging to serotype O3:K31, ELK 169 and ELK 170 both with serotype O11:Kut, TAK 139 and TAK 142 both with serotype O8:K70, and ELK 124 and ELK 127 with serotype O3:K6 respectively. It is interesting to note that the only clade found among this organism is between isolate TAK 145 from wet sand sample collected from Takwa Bay beach and BDG 151B from dry sand sample collected from Badagry beach.

Fig. 2
figure 2

PFGE profile of 26 strains of V. parahaemolyticus isolates after SfiI enzyme restriction. V. parahaemolyticus ATCC 17802 was used as a control

Full size image

In conclusion, we isolated V. parahaemolyticus and V. vulnificus showing resistance to multiple antimicrobial agents from four recreational beaches in Lagos, southwestern Nigeria. The virulence gene tdh and trh was detected in six isolates of V. parahaemolyticus with one isolate positive for both genes, confirming the presence of potentially pathogenic strains of V. parahaemolyticus in Nigerian recreational waters. Five V. vulnificus and one V. parahaemolyticus tested positive for the metallo-β-lactamase gene blaNDM-1which occurred together with blaTEM and blaCMY in five isolates. The detection of these genes in public beaches visited by beach goers is of great public health concern and highlights the need for urgent mitigation efforts to protect beach goers and curbs further dissemination of these resistance genes to other bacterial population through horizontal gene transfer.