Ampli ﬁ cation of tlh gene in other Vibrionaceae specie by specie-speci ﬁ c multiplex PCR of Vibrio parahaemolyticus

Background: The surveillance of Vibrio parahaemolyticus in the Chilean coast has been mainly performed by multiplex PCR ampli ﬁ cation of three different hemolysin genes, which are specie-speci ﬁ c virulence factors. These genes are also employed in the determination of V. parahaemolyticus pathogenic load in seafood and for characterization of pathogenic strains associated to diarrhea cases in human. During environmental surveillance that we performed every summer, we occasionally observed a thermolabile hemolysin ( tlh ) PCR product of a slightly smaller size than expected, which was coincident with low loads of V. parahaemolyticus in the environment. In order to understand this observation, we probed the speci ﬁ city of tlh primers for the detection of V. parahaemolyticus at different bacterial loads and DNA concentrations. Results: Primers used for the detection of V. parahaemolyticus speci ﬁ c tlh ampli ﬁ ed a slightly smaller tlh gene, which is found in Vibrio alginolyticus and other related strains. These amplicons were observed when V. parahaemolyticus was absent or in undetectable loads in the environment. Conclusions: Surveillance of V. parahaemolyticus using tlh primers can be imprecise because ampli ﬁ cation of a V. parahaemolyticus speci ﬁ c marker in V. alginolyticus and other related strains occurs. This situation complicates potentially the estimation of bacterial load in seafood, because do not ensure the correct identi ﬁ cation of V. parahaemolyticus when his load is low. Additionally, it could complicate the tracking of outbreaks of V. parahaemolyticus infections, considering the genetic markers used would not be specie-speci ﬁ c.


Introduction
Vibrio parahaemolyticus is a halophilic gram negative bacterium widely distributed in marine environments and is recognized as the world's leading cause of gastroenteritis in humans due to raw or undercooked seafood [1]. Some strains are capable of causing disease and most of them are related to the pandemic serotype O3:K6 [2]. This serotype comprises a clonal complex originally observed in Southeast Asia that reached Antofagasta, northern Chile, in 1998 [3], and reached and remained in Region de Los Lagos, southern Chile since 2004. During this period, seafood consumption-related diarrhea increased drastically in Chile, because approximately 69% of the country's shellfish are produced in this region [4]. Filter-feeding bivalve, such as mussels, can concentrate pathogenic strains of V. parahaemolyticus resulting in bacterial loads capable of producing infection to the people that ingest them [5]. Consequently, virulent V. parahaemolyticus strains are clearly a concern for seafood safety, and hence, their reliable detection in mussels is crucial.
In Chile, surveillance of V. parahaemolyticus in seafood has been performed mainly by multiplex PCR (mPCR) [6] including amplification of thermolabile hemolysin (tlh), thermostable direct hemolysin (tdh) and tdh-related hemolysin (trh) genes [7]. The tlh gene encodes a phospholipase A2 [8], while its contribution to V. parahaemolyticus pathogenicity is not clear, the expression of tlh is upregulated under conditions mimicking the human intestine [9]. This gene is considered a specie-specific marker for V. parahaemolyticus and is frequently employed to identify this specie [10]. On the other hand, most of pathogenic strains additionally possess tdh and/or trh hemolysin genes, which are considered important markers of clinical strains [1,11].
Using mPCR, Fuenzalida et al. [2] observed tlh amplicons slightly smaller than expected (450 bp) derived from shellfish extracted in Antofagasta. Authors sequenced tlh amplicon and observed 91% similarity with tlh gene found in Vibrio alginolyticus [12] and 82% Electronic Journal of Biotechnology 18 (2015) 459-463 similarity with V. parahaemolyticus tlh gene. Smaller tlh amplicons were later observed in Puerto Montt, coinciding with lower loads of V. parahaemolyticus (geometric average b 0.3 to 2 g -1 ) than observed in previous years (geometric average b 2 to 110 g -1 ). To understand these observations and the potentially critical consequences for V. parahaemolyticus surveillance, we performed sampling to quantify the bacterial load in mussels using mPCR, including tlh primers. Additionally, we studied the specie-specificity of these primers for isolates obtained from Chile and we performed PCR at different loads of V. parahaemolyticus and V. alginolyticus, simultaneously.

Sampling and load estimation
Mussel samples obtained in Quillaipe, Puerto Montt, were analyzed by mPCR as described previously [13]. Briefly, samples of shellfish soft tissue were enriched for V. parahaemolyticus in three-tube serial dilutions in alkaline peptone water for the assessment of bacterial load by the most probable number (MPN) method; tubes with bacterial growth were tested for tlh, tdh and trh by multiplex PCR [6]. PCR was performed using approximately 10 ng of total bacterial DNA per reaction tube. The tlh amplicons were visualized in 8% polyacrylamide gels and stained with silver nitrate. Total and pandemic V. parahaemolyticus loads were calculated according to the number of tubes positive for tlh and for tdh and trh, respectively.

Phenotypic characterization
According to the mPCR results, tlh positive enrichment tubes were plated on TCBS agar (Difco) and chromogenic agar CHROMagar Vibrio (CHROMagar Microbiology, Paris, France). Furthermore, purified colonies were characterized with Api20E kit for enteric bacteria (Biomerieux, Halzelwood, MO) and referred as A1 to A7. V. alginolyticus strains R4 and VA1 were isolated from Crete, Greece and were identified by BIOLOG GENIII and molecular methods (16s RNA sequence and PCR using collagenase gene primers) [14].

tlh restriction assay
To determine the sequence similarity of tlh PCR products, in silico restriction analysis was performed with Webcutter 2.0 (http://rna. lundberg.gu.se/cutter2/) using tlh sequence of V. alginolyticus (JF718655.1) and V. parahaemolyticus (JX262976.1). Of the tested enzymes, HhaI and MboI were selected because they showed clear differences between species. Experimental restriction analysis of tlh was performed using HhaI (Fermentas) and MboI (Fermentas) for V. alginolyticus and V. parahaemolyticus amplicons according to the manufacturer's instructions, and PCR products were visualized in 1% agarose gels and stained with ethidium bromide.

Sequencing and phylogenetic analysis
PCR product of tlh gene from colonies obtained from sampling (A1 to A7) along with V. alginolyticus R4 and ANw1 (Antofagasta's strain referred as X in [2]) were sequenced in Macrogen (Korea) and results were compared with tlh sequences obtained from Genbank [15] using BLAST [16]. Phylogenetic tree was constructed using MEGA5 software v5.2 [17] based on UPGMA method.

tlh PCR assay specificity
The strains V. alginolyticus R4 and V. parahaemolyticus Vp222 were selected to determine tlh primer specificity at different DNA concentrations and bacterial load of both species. For PCR with different concentrations of DNA, the genome was extracted with Fig. 1. tlh amplification. 8% polyacrylamide gel shows that in samples A1, A2 and A4 PCR amplification rendered a tlh amplicon slightly smaller than obtained with bona fide V. parahaemolyticus (A6 and A7). VpKX and ANw1 were used as positive control. Molecular weight, in the first line, corresponds to ladder 100 bp.
Wizard Genomic DNA purification kit (Promega). The PCR was performed using DNA concentrations from 0.0001 ng to 30 ng of each bacterial strain and templates were mixed using proportional inverse concentration of genomic DNA. For PCR with different bacterial loads, cells were counted under microscope and serial dilutions were prepared from 10 0 to 10 5 total cells for each species. Six mixtures of R4 and Vp222 were performed using proportional inverse bacterial load. PCR was performed according amplification conditions mentioned in [6]. The tlh amplicons were visualized in 8% polyacrylamide gels and stained with silver nitrate or gel red as appropriate.

Vibrio parahaemolyticus load estimation and isolates characterization
Of 28 samples analyzed, tlh was detected in 89% and all samples were negatives for tdh and trh genes. Vibrio parahaemolyticus load was estimated between 0.7 g -1 and 110 g -1 , with a geometric mean of 2.1 g -1 . tlh amplicons obtained from some of these samples (A1 to A5; size 440 bp) were slightly smaller than that obtained with bona fide V. parahaemolyticus (A6, A7; size 450 bp), as was reported previously in Antofagasta [2]. To confirm this observation, PCR with tlh primers was repeated for isolates A1, A2, A4, A6 and A7, using VpKX and ANw1 as positive control (Fig. 1). All the isolates with a smaller tlh amplicon showed phenotypic characteristics of V. alginolyticus: they were yellow and white colonies in TCBS and CHROMagar Vibrio, respectively, and Voges Proskauer positive and ara negative according Api20E. While isolates with expected tlh amplicon showed phenotypic characteristic from V. parahaemolyticus.

Analysis of tlh in V. alginolyticus strains
To corroborate that tlh amplicons were effectively amplified from V. alginolyticus or related isolates, tlh primers were used to perform PCR using strains R4 and VA1 isolated in Crete, Greece and V. parahaemolyticus Vp7, Vp222, and VpKX. Slightly size differences of approximately 10 bp were observed between V. alginolyticus and V. parahaemolyticus amplicons (Fig. 2), as was observed in isolates obtained during sampling (Fig. 1).
Additionally, restriction pattern analysis of tlh amplicon showed difference between V. alginolyticus R4 and VA1 strains and V. parahaemolyticus strains. This distinctiveness was observed using HhaI (Fig. 3A) or MboI (Fig. 3B) restriction enzymes, suggesting that tlh sequence were not equal in both species.

Sequencing and phylogenetic analysis of tlh gene from isolates obtained during sampling
According to the suggested differences between tlh sequences of the reference strains (Fig. 3), tlh of isolates from sampling (A1 to A7) was sequenced, including R4 and ANw1 strains as control. Additionally, other tlh sequences reported for other Vibrio were obtained from GenBank and considered for analysis. Phylogenetic tree showed that isolates with smaller tlh amplicons (A1 to A5) do not clustered with V. parahaemolyticus (Fig. 4). Of this smaller tlh amplicons, isolates A1, A2, and A3 were clustered with V. alginolyticus strains, including reference strain R4. Isolates A4 and A5 grouped with Vibrio diabolicus strains and formed a cluster with ANw1 strain, previously isolated in Antofagasta [2] (Fig. 4).

Specificity of tlh primers
As we mentioned before, our previous observations suggest that tlh primers could amplify a tlh gene from other Vibrio when  The results show that at lower concentration of V. parahaemolyticus DNA (30 ng R4/0.0001 ng Vp222), the tlh from V. alginolyticus is amplified. A decrease in the DNA of V. alginolyticus and increased in the DNA of V. parahaemolyticus will lead to the amplification of both amplicons (Fig. 5, upper panel, lane 2) until the tlh from V. alginolyticus is no longer amplified (Fig. 5, upper panel, lanes 3-8). Similar results were observed when the template DNA was determined by different loads of cells. Polyacrylamide electrophoresis show that a lower V. parahaemolyticus load determine tlh amplification corresponding to V. alginolyticus (10 5 R4/10 0 Vp222) which is characterized by smaller size. Again, it is observed that when V. parahaemolyticus load is low, smaller tlh amplicon take place. Inversely, when V. parahaemolyticus load increased is observed bona fide tlh amplification (Fig. 5, bottom panel).

Discussion
Since V. parahaemolyticus reached Chile in 1998, the mPCR proposed by Bej and collaborators [6] has provided a high-throughput and convenient means to determine total and pathogenic V. parahaemolyticus numbers in shellfish. However, these last years, we observed that during surveillance when V. parahaemolyticus loads were lower in Chilean coasts, mPCR was clearly a subject to false positives because thermolabile hemolysin gene was amplified from other Vibrio species (V. alginolyticus and V. diabolicus).
The presence of tlh in other species of Vibrio is not novel and had already been previously reported. Wang et al. [18], showed that tlh is widespread in vibrios [18], including V. alginolyticus, Vibrio harveyi, Vibrio fischeri, Vibrio mimicus, Vibrio natriegens, V. parahaemolyticus, Vibrio proteolyticus, Vibrio anguillarum, and Vibrio vulnificus among others. In other study, Xie et al. [12] showed that tlh sequence data was recovered from non-V. parahaemolyticus strains, indicating that tlh gene sequences in V. diabolicus, Vibrio tubiashii-like, and Photobacterium damselae are highly similar to those in V. parahaemolyticus. Worryingly,  Xie also reported that V. alginolyticus strains often possess homologues of the V. parahaemolyticus and V. cholerae virulence genes such as toxR, and VPI [12], which suggests that V. alginolyticus maybe contain other known virulence genes of V. parahaemolyticus, besides tlh, potentially complicates the analysis employed during surveillance of V. parahaemolyticus, including pathogenic strains. Additionally, Klein and their coworkers [19] reported that virulence-related genes occurred at high frequencies in non-V. parahaemolyticus-Vibrionaceae species. In this study, V. diabolicus was of particular interest, as several strains were recovered, and the large majority (N83%) contained virulence-related genes [19].
Similarity sequence between different Vibrio tlh genes makes the specie distinction imprecise. Consequently, the use of tlh as marker for V. parahaemolyticus loads may lead to overestimation by including other Vibrio species carrying tlh. The level of overestimation may be variable among different systems and sample types. Considering that tlh detection is formally recommended to the detection of V. parahaemolyticus [20,21], the results of this study are critical for surveillance in coasts. It is essential to find additional specie-specific probes and perform supporting analysis, to find reliable markers for V. parahaemolyticus quantitative detection, and to avoid overestimation of loads and erroneous identification of this specie in the environment.