The Vibriolysin-Like Protease VnpA and the Collagenase ColA Are Required for Full Virulence of the Bivalve Mollusks Pathogen Vibrio neptunius

Vibrio neptunius is an important pathogen of bivalve mollusks worldwide. Several metalloproteases have been described as virulence factors in species of Vibrio that are pathogenic to bivalves, but little is known about the contribution of these potential virulence factors to Vibrio neptunius pathogenesis. In silico analysis of the genome of V. neptunius strain PP-145.98 led to the identification of two hitherto uncharacterized chromosomal loci encoding a probable vibriolysin-like metalloprotease and a putative collagenase, which were designated VnpA and ColA, respectively. Single defective mutants of each gene were obtained in V. neptunius PP-145.98, and the phospholipase, esterase and collagenase activities were studied and compared with those of the wild-type strain. The results showed that the single inactivation of vnpA resulted in a 3-fold reduction in phospholipase/esterase activity. Inactivation of colA reduced the collagenase activity by 50%. Finally, infection challenges performed in oyster larvae showed that ΔvnpA and ΔcolA—single mutant strains of V. neptunius—are between 2–3-fold less virulent than the wild-type strain. Thus, the present work demonstrates that the production of both VnpA and ColA is required for the full virulence of the bivalve pathogen V. neptunius.


Introduction
Bivalve mollusks are an important source of animal protein for human nutrition, representing about 15% of the average animal protein consumed per capita [1]. The decrease in natural beds has required the establishment of hatcheries to provide farms with juveniles. Unfortunately, the optimum conditions for the growth and development of bivalve larvae in hatcheries favor the occurrence of bacterial infections [2,3].
Several species of the bacterial genus Vibrio can cause disease in aquatic animals and/or humans [4][5][6]. Specifically, some species are pathogenic to mollusks and cause vibriosis, which leads to high mortality ratios in larvae and spats of farmed bivalves worldwide [7,8]. Vibrio species belonging to the Corallilyticus and Orientalis clades, such as Vibrio coralliilyticus, V. tubiashii and V. neptunius, are among the best-known species of bivalve pathogens [7]. In particular, V. neptunius has been isolated from samples of sea water and marine animals, including turbot larvae (Scophthalmus maximus) and rotifers (Brachiomus plicatilis) [9,10]. This bacterium has also been reported to be responsible for larval and spat mortality episodes in clam and oyster cultures, leading to significant economic losses in shellfish hatcheries [3]. V. neptunius is able to rapidly invade the tissues of bivalve larvae after entering by filtration, and it produces thermolabile extracellular products demonstrating enzymatic activities, such as amylase, gelatinase and lipase, that are cytotoxic to fish and homeothermic animal cell lines [11]. However, the genetic basis of the pathogenicity of this bacterium remains unclear, as there are no known studies on the characterization of genes that encode proteases or other hydrolytic enzymes that might play an active role in the pathogenic character of V. neptunius.
Virulence is a multifactorial trait that results from the sequential action of several microbial factors that enable colonization, proliferation in the host and, finally, the infection process [12]. Some metalloproteases, including vibriolysin-like proteases and collagenases with a role in virulence, have been described in vibrios isolated from seawater, fish and shellfish [13,14]. The metalloproteases produced by human pathogenic vibrios have been extensively studied, and the evidence shows a broad range of pathological effects [15,16]. In V. cholerae, vibriolysin is a direct toxic factor that causes host tissue damage, forms edematous lesions and facilitates heme utilization [17]. Furthermore, it has indirect toxic effects by converting the precursor of an enterotoxic hemolysin into the mature toxin through propeptide release [18,19], increasing bacterial attachment to the intestinal surface through digestion of the mucosal substances [20] and accelerating bacterial detachment via the digestion of V. cholerae adhesins [21]. In V. coralliilyticus, one of the most studied pathogens of corals and mollusks, vibriolysin-like metalloproteases encoded by vcpA and its orthologs vcpB and vchA facilitate bacterial invasion and the infection process [22,23]. Recently, an immunoassay to detect VcpA was proposed as a diagnostic tool capable of tracking pathogens involved in coral disease outbreaks [24]. Metalloproteases with roles in virulence have also been described in several bivalve pathogens, such as V. coralliilyticus [22,23], V. tubiashii [25], V. aestuarianus [26] and V. splendidus [27]. However, the influence of these proteases on the virulence of different pathogenic bacteria is diverse. On the other hand, collagen is a major component of the extracellular matrix of animal tissues, and thus, collagenase production can also accelerate bacterial dissemination and facilitate nutrient acquisition [28]. Unfortunately, the role of collagenases in the virulence of bivalve pathogens is poorly studied, as most works have focused on the pathogenic effect of collagenase in humans, considering bivalves as reservoirs of bacteria that are pathogenic to humans [29,30].
In this work, an in silico analysis of the genome of V. neptunius strain PP-145.98 led to the identification of two hitherto uncharacterized chromosomal loci encoding a probable vibriolysin-like metalloprotease and a putative collagenase, which were designated VnpA and ColA, respectively. The construction of mutants defective in each of these genes resulted in a reduction in the respective phospholipase/esterase and collagenase activities. In addition, infection challenges in oyster larvae showed that the production of both proteases is required for the full virulence of V. neptunius in bivalve mollusks. Our results show for the first time that collagenases may be key virulence factors in Vibrio species that are pathogenic to bivalve mollusks.

Bacterial Strains, Plasmids and Growth Conditions
The strains and plasmids used, as well as those derived from this study, are included in Table 1. V. neptunius strains were grown at 25 • C in Tryptic Soy Agar and Broth (Pronadisa, Madrid, Spain) supplemented with 1% NaCl (TSA-1 and TSB-1, respectively), as well as in M9 minimal medium supplemented with 0.2% Casamino Acids (Difco, Leeuwarden, The Netherlands) (CM9) [31]. The V. neptunius strains used to evaluate enzymatic activities belong to the laboratory strain collection. They were isolated from vibriosis outbreaks affecting bivalve hatcheries in Galicia (Spain). Escherichia coli strains were grown at 37 • C in Luria-Bertani (LB) medium (Pronadisa) or LB supplemented with the appropriate antibiotics. Ampicillin sodium salt and kanamycin were used at 100 µg mL −1 and 50 µg mL −1 , respectively (final concentrations).

General DNA Manipulation Techniques and PCR Assays
Total genomic DNA from V. neptunius was purified with the InstaGene™ Matrix (BioRad, Hercules, California, CA, USA) and stored at −20 • C. The extraction of DNA from agarose gels and purification of plasmid DNA were carried out using NucleoSpin ® Gel and a PCR clean-up kit (Macherey-Nagel, Düren, Germany) and a GeneJET Plasmid Miniprep Kit (Thermo-Fisher, Waltham, MA, USA). PCR reactions were all carried out in a T-Gradient Thermal Cycler (Biometra, Göttingen, Germany) with Taq polymerase NZYTaq (Nzytech, Lisboa, Portugal). Thermal cycling conditions were as follow: 95 • C for 5 min, 30 cycles at 95 • C for 45 s, 58 • C for 45 s, and an elongation step at 72 • C for 1 min per kb.

The Search for Candidate Genes and Construction of vnpA and colA Mutants by Allelic Exchange
Candidate genes to encode metalloproteases related to virulence in the genome of V. neptunius PP-145.98 (Accession No. JAFHLB000000000) were identified using NCBI services (http://ncbi.nlm.nih.gov, accessed on 1 January 2021). For this purpose, the protein sequences of metalloproteases HapA (Accession No. WP_000782181.1), VcpA (Accession No. AFK08684.1) and ColP (Accession No. WP_069531114.1) were used as queries in the BLAST program to identify homologs in V. neptunius PP-145.98. The prediction of protein domains was carried out using the UniProt database facilities (https://www.uniprot.org, accessed on 1 February 2021).
In-frame deletions of vnpA and colA in V. neptunius PP-145.98 were constructed using PCR amplification of two fragments of each gene and the flanking regions, which, when ligated together, should result in an in-frame (nonpolar) deletion. The oligonucleotides used to amplify the upstream and downstream ends of each gene are shown in Table 2. Suitable restriction sites were added to the oligonucleotides. The PCR-amplified 3 -end gene fragments were ligated into pWKS30, and the resulting plasmids were cut with suitable enzymes and ligated to the 5 -ends of the PCR fragments of each gene. Each deleted allele cloned in pWKS30 was digested with NotI and ApaI and ligated into the suicide vector pCAR109 [34]. As a pCVD442 derivative, pCAR109 contains R6K ori, requiring the pir gene product for replication, and the sacB gene, conferring sucrose sensitivity. The resulting plasmids were mated from E. coli S17-1-Lpir into V. neptunius PP-145.98, and transformants, in which the plasmid was integrated into the chromosome by homologous recombination, were selected on agar medium containing 50 mg mL −1 kanamycin (resistance conferred by pCAR109) and 100 mg mL −1 ampicillin (specific antibiotic to select for V. neptunius PP-145.98). A second recombination event was obtained by selecting for sucrose (10%) resistance. This process led to the generation of ∆colA and ∆vnpA, which were named FG56 and FGE5, respectively, as defective mutant strains of V. neptunius PP-145.98. Sanger sequencing using primers 1 and 4 ( Table 2) was used to confirm mutant construction.

Enzymatic Activity
Enzymatic activities were evaluated out by placing 5 µL of a TSB-1 culture of each strain at OD 600 ≈ 0.5 on the surface of TSA-1 plates supplemented with the appropriate substrate. All assays were repeated three times to ensure reproducibility of results. TSA-1 plates supplemented with 1% egg yolk were used to determine phospholipase activity [35]. TSA-1 plates with 1% Tween 20 were used to confirm lipolytic activity, as Tween 20 is easily hydrolyzed by esterase because it contains esters of lower chain fatty acids [36]. Both lipolytic activities (phospholipase and esterase) were evaluated after 24 h incubation at 25 • C. The gelatinase activity test was carried out on TSA-1 plates supplemented with 0.4% gelatin, and the results were recorded after 48 h incubation at 25 • C by covering the surface of the agar plate with a 12.5% (w/v) HgCl 2 solution. For the hemolysis test, 5 µL of an inoculum in TSB-1 at OD 600 ≈ 0.5 was placed on sheep blood agar plates (Oxoid, Hampshire, UK) and incubated at 25 • C for 48 h.

Virulence Assays
Pathogenicity assays were carried out in 6-well sterile plastic cell-culture microplates. Oyster larvae of Ostrea edulis were placed in each well with 9 mL of autoclaved seawater at an average density of 11 larvae mL -1 . Microplates were incubated on an orbital shaker (100 rpm) at room temperature (22 • C). No food was added during the experimental period. The wild-type strain and the mutants ∆colA and ∆vnpA of V. neptunius were cultured in CM9 medium for 12 h at 25 • C to achieve an OD 600 = 0.5. Bacteria were added to the wells to achieve final concentrations ranging from 10 4 to 10 5 cfu mL -1 . Wells without added bacteria served as negative controls for the assay. Bioassays were performed in triplicate. The larvae were observed under an optical microscope (4× and 10×) to report the effect of the bacteria: live larvae were identified as those that were swimming or had closed valves but showed internal movement, and larvae without internal movement were considered dead. Pathogenicity results were evaluated at 96 h and expressed as a percentage of survival. Mortalities were recorded daily, and the statistical significance of differences in percentage survival for V. neptunius strains was determined using the Kaplan-Meier method with the Mantel-Cox log-rank test using SPSS (version 20; IBM SPSS Inc., Chicago, IL, USA). P-values were considered significant when p was <0.05, <0.01 and <0.001.

V. neptunius Strains Show Proteolytic and Lipolytic Activities
To assess the extracellular enzymatic activity of V. neptunius, a collection of 19 strains isolated from disease episodes that occurred in diverse bivalve larvae was screened in three replicates for proteolytic, lipolytic and hemolytic activities. Most V. neptunius isolates were positive for gelatinase, esterase and phospholipase activities ( Table 3). The unique exception was strain PP-256, which was considered negative for esterase activity since it failed to hydrolyze Tween 20. In addition, most strains (15 out of 19) showed β-hemolysis, producing a translucent hemolytic halo between 2 and 5 mm wide on sheep blood agar plates. However, two strains (PP-266 and PP-273) were non-hemolytic, and the PP-256 and PP-258 strains showed α-hemolysis since they did not produce translucent halos on sheep blood agar. The results clearly show that all V. neptunius strains possess both proteolytic and lipolytic activities. Since PP-145.98 is a V. neptunius strain that shows one of the highest proteolytic/lipolytic activities, and its genome was readily available, we chose this strain for an in silico analysis to identify genes that are suspected to encode proteases with a putative role in virulence.

The V. neptunius Genome Contains Genes Encoding Two Metalloproteases Putatively Related to Virulence
Many of the toxic proteases produced by vibrios fall into the vibriolysin and collagenase groups, which belong to the metalloprotease family [28]. The in silico analysis of the genome of V. neptunius PP-145.98 (Accession No. JAFHLB000000000) showed that it harbors a number of loci encoding putative metalloproteases. According to homology searches, two of them, loci JYA62_21125 of 1821 bp and JYA62_12825 of 2415 bp, are likely to encode virulence-related genes ( Table 4). The predicted product of locus JYA62_21125 (protein ID MBN3580162.1) shares close homology (between 74% and 90% similarity) with several vibriolysin-like proteases belonging to the thermolysin family (M4) of metallopeptidases described in other marine Vibrio species, including VcpA of V. coralliilyticus and VtpA of V. tubiashii, proteins that were shown to be major virulence factors of their respective bacterial species [22,37]. Thus, we designated locus JYA62_21125 as vnpA (Vibrio neptunius metalloprotease A). Notably, VnpA is a homolog of HapA (82% similarity), which is an extracellular Zn-dependent metalloprotease, hemagglutinin/protease produced by V. cholerae and also known as vibriolysin [19]. Thus, VnpA is predicted to be a protein of 607 amino acids, containing the characteristic domain architecture conserved between vibriolysins [15,38]. It contains a signal peptide signature in the N-terminal region (be-  coralliilyticus and VtpA of V. tubiashii, proteins that were shown to be major virulence factors of their respective bacterial species [22,37]. Thus, we designated locus JYA62_21125 as vnpA (Vibrio neptunius metalloprotease A). Notably, VnpA is a homolog of HapA (82% similarity), which is an extracellular Zn-dependent metalloprotease, hemagglutinin/protease produced by V. cholerae and also known as vibriolysin [19]. Thus, VnpA is predicted to be a protein of 607 amino acids, containing the characteristic domain architecture conserved between vibriolysins [15,38]. It contains a signal peptide signature in the N-terminal region (between positions 20 and 24 (VTA-AE)), two conserved propeptide motifs between 51 and 189 (FTP domain between 51 and 90 and PepSy domain 114 and 189), mature termolysin matallopeptidase domains (207 and 496) and a C-terminal pre-peptidase (PPC) domain (530 and 594). Conserved domains and motifs found in VnpA metalloproteases are shown in Figure 1.  On the other hand, the predicted product of locus JYA62_12825 (protein ID MBN3578544.1) is a protein of 805 amino acids of which the closest homolog (93% similarity) is a putative collagenase of V. coralliilyticus belonging to class II metalloproteinases (Table 4). This protein is predicted to degrade collagen since it contains the conserved zinc-binding motif comprising the sequence HEXXH-E at positions 436 and 466 and a double glycine motif upstream of the HEXXH motif ( Figure 1): both domains are required for collagenase activity [39]. These motifs are also present in collagenases from other pathogenic bacteria, such as Bacillus anthracis, B. cereus, Clostridium tetani and C. botulinum [40]. Notably, ColA showed significant homology with ColP (39% identity and 60% similarity), a collagenase that was characterized as a major virulence factor of the fish pathogen Photobacterium damselae subsp. damselae [41]. Thus, locus JYA62_12825 was termed colA (collagenase A). On the other hand, the predicted product of locus JYA62_12825 (protein ID MBN3578544.1) is a protein of 805 amino acids of which the closest homolog (93% similarity) is a putative collagenase of V. coralliilyticus belonging to class II metalloproteinases (Table 4). This protein is predicted to degrade collagen since it contains the conserved zinc-binding motif comprising the sequence HEXXH-E at positions 436 and 466 and a double glycine motif upstream of the HEXXH motif ( Figure 1): both domains are required for collagenase activity [39]. These motifs are also present in collagenases from other pathogenic bacteria, such as Bacillus anthracis, B. cereus, Clostridium tetani and C. botulinum [40]. Notably, ColA showed significant homology with ColP (39% identity and 60% similarity), a collagenase that was characterized as a major virulence factor of the fish pathogen Photobacterium damselae subsp. damselae [41]. Thus, locus JYA62_12825 was termed colA (collagenase A).

Inactivation of VnpA and ColA Reduced Proteolytic Activity
To study the role of vnpA and colA in the extracellular enzymatic activities and virulence of V. neptunius, in-frame deletion mutants of each gene were produced in the V. neptunius PP-145.95 background. Then, the ability to hydrolyze egg yolk (phospholipase Antibiotics 2021, 10, 391 7 of 13 activity), Tween 20 (esterase activity) and gelatin (gelatinase activity) was evaluated in both mutant strains and compared with those of the wild type ( Figure 2). The wild-type V. neptunius PP-145.98 strain yielded an opaque precipitation halo in the egg yolk and Tween 20 plates, which is indicative of phospholipase and esterase activities. Interestingly, the ∆vnpA mutant was virtually unable to hydrolyze egg yolk and showed a 3-fold reduction in esterase activity. The gelatinase activity of this mutant was almost identical to that of the wild type. Thus, inactivation of vnpA in V. neptunius disables the phospholipase activity and greatly reduces esterase activity.

Inactivation of VnpA and ColA Reduced Proteolytic Activity
To study the role of vnpA and colA in the extracellular enzymatic activities and virulence of V. neptunius, in-frame deletion mutants of each gene were produced in the V. neptunius PP-145.95 background. Then, the ability to hydrolyze egg yolk (phospholipase activity), Tween 20 (esterase activity) and gelatin (gelatinase activity) was evaluated in both mutant strains and compared with those of the wild type ( Figure 2). The wild-type V. neptunius PP-145.98 strain yielded an opaque precipitation halo in the egg yolk and Tween 20 plates, which is indicative of phospholipase and esterase activities. Interestingly, the ΔvnpA mutant was virtually unable to hydrolyze egg yolk and showed a 3-fold reduction in esterase activity. The gelatinase activity of this mutant was almost identical to that of the wild type. Thus, inactivation of vnpA in V. neptunius disables the phospholipase activity and greatly reduces esterase activity. When the phenotype of the V. neptunius ΔcolA mutant was evaluated, a major effect on gelatin hydrolysis was detected, as the halo was 2-fold narrower (Figure 2). By contrast, a slight reduction in Tween 20 hydrolysis was observed, and the halo formed in the egg yolk was indistinguishable from that produced by the wild-type strain. In addition, since hemolysis may result from phospholipase enzyme activity [42], the hemolytic phenotype of ΔvnpA and ΔcolA mutants was also evaluated. Interestingly, although the When the phenotype of the V. neptunius ∆colA mutant was evaluated, a major effect on gelatin hydrolysis was detected, as the halo was 2-fold narrower (Figure 2). By contrast, a slight reduction in Tween 20 hydrolysis was observed, and the halo formed in the egg yolk was indistinguishable from that produced by the wild-type strain. In addition, since hemolysis may result from phospholipase enzyme activity [42], the hemolytic phenotype of ∆vnpA and ∆colA mutants was also evaluated. Interestingly, although the hemolysis produced by the ∆colA mutant was the same as that shown by the wild-type strain, the hemolytic halo of the ∆vnpA mutant was halved. This result shows that VnpA activity is involved in the lysis of erythrocytes. However, since the ∆vnpA mutant retains some degree of hemolytic activity, it is not likely a major hemolytic factor of V. neptunius. Thus, Antibiotics 2021, 10, 391 8 of 13 inactivation of vnpA in V. neptunius greatly reduces phospholipase and esterase activities and slightly reduces hemolytic activity. Taken together, the results suggest that VnpA is responsible for the phospholipase and esterase activities in V. neptunius and that ColA is a protease that hydrolyzes gelatin.
Screening for the presence of vnpA and colA by PCR showed that these genes are widespread in V. neptunius strains. However, some strains failed in the amplification of the PCR target for either vnpA (two strains of 19) or colA (one of 19) but had an enzymatic activity profile that was almost identical to that of the PP-145.98 strain (Table 3). Thus, additional proteases must be present in the genome of this strain.

VnpA and ColA Have a Major Role in V. neptunius Virulence
Proteases can play major roles in colonization and persistence within a host or the degradation of host biomass for nutrient acquisition [28,43]. To study the contribution of VnpA and ColA to V. neptunius pathogenesis, the virulence of the PP-145.98 wild-type strain, as well as that of both single ∆vnpA and ∆colA mutants, was evaluated by performing experimental infections in oyster larvae of O. edulis. Infection challenges were performed in triplicate; similar results were obtained in each replicate, and the data were pooled together. The survival curves observed after infection challenges are shown in Figure 3. The mortality was zero in the control group (non-challenged clam larvae). By contrast, inoculation of the PP-145.98 wild-type strain caused high mortality ratios, killing 88% of the group after 96 h. This result was comparable to the mortalities previously reported by Prado et al. for the same strain [3]. Interestingly, the mortality observed after a challenge with either ∆colA or ∆vnpA mutant strains decreased to 21% and 30%, respectively. In summary, the inactivation of either colA or vnpA produces a significant reduction in V. neptunius virulence, and thus, both genes (colA and vnpA) are required for the full virulence of V. neptunius in oyster larvae. hemolysis produced by the ΔcolA mutant was the same as that shown by the wild-type strain, the hemolytic halo of the ΔvnpA mutant was halved. This result shows that VnpA activity is involved in the lysis of erythrocytes. However, since the ΔvnpA mutant retains some degree of hemolytic activity, it is not likely a major hemolytic factor of V. neptunius. Thus, inactivation of vnpA in V. neptunius greatly reduces phospholipase and esterase activities and slightly reduces hemolytic activity. Taken together, the results suggest that VnpA is responsible for the phospholipase and esterase activities in V. neptunius and that ColA is a protease that hydrolyzes gelatin. Screening for the presence of vnpA and colA by PCR showed that these genes are widespread in V. neptunius strains. However, some strains failed in the amplification of the PCR target for either vnpA (two strains of 19) or colA (one of 19) but had an enzymatic activity profile that was almost identical to that of the PP-145.98 strain (Table 3). Thus, additional proteases must be present in the genome of this strain.

VnpA and ColA Have a Major Role in V. neptunius Virulence
Proteases can play major roles in colonization and persistence within a host or the degradation of host biomass for nutrient acquisition [28,43]. To study the contribution of VnpA and ColA to V. neptunius pathogenesis, the virulence of the PP-145.98 wild-type strain, as well as that of both single ΔvnpA and ΔcolA mutants, was evaluated by performing experimental infections in oyster larvae of O. edulis. Infection challenges were performed in triplicate; similar results were obtained in each replicate, and the data were pooled together. The survival curves observed after infection challenges are shown in Figure 3. The mortality was zero in the control group (non-challenged clam larvae). By contrast, inoculation of the PP-145.98 wild-type strain caused high mortality ratios, killing 88% of the group after 96 h. This result was comparable to the mortalities previously reported by Prado et al. for the same strain [3]. Interestingly, the mortality observed after a challenge with either ΔcolA or ΔvnpA mutant strains decreased to 21% and 30%, respectively. In summary, the inactivation of either colA or vnpA produces a significant reduction in V. neptunius virulence, and thus, both genes (colA and vnpA) are required for the full virulence of V. neptunius in oyster larvae.

Discussion
Vibrionaceae accounts for a dominant proportion of bivalve mollusk microbiota, providing them with a wide range of beneficial services, including defense against ex- Figure 3. Survival curves after challenge in oyster larvae (Ostrea edulis) with the wild-type V. neptunius strain or ∆colAor ∆vnpA-defective mutants. Asterisks denote statistically significant differences between strains (*** p < 0.001).

Discussion
Vibrionaceae accounts for a dominant proportion of bivalve mollusk microbiota, providing them with a wide range of beneficial services, including defense against external pathogens and facilitation of nutrient acquisition [44][45][46][47][48][49]. However, bacterial inhabitants themselves can act as opportunistic pathogens [50,51]. Notably, the production of certain potential virulence factors, such as siderophores and proteases, is a widespread property of healthy bivalve microbiota [52]. Although siderophores are considered indirect virulence factors that play roles in the establishment of the infection, proteases are usually toxic factors that cause disease symptoms [15].
V. neptunius is an opportunistic pathogen in bivalve mollusk microbiota and is responsible for larval and spat mortality episodes in clam and oyster cultures [3,53,54]. In this work, enzymatic activities such as phospholipase, esterase and gelatinase, as well as hemolysis, were evaluated using a collection of 19 V. neptunius strains isolated from outbreak episodes affecting oysters and clams. The results show that V. neptunius displays lipolytic and proteolytic activities, with some degree of variability among strains. The same strain collection was used in a recent work to study siderophore production, and major differences among isolates were also detected [55]. The absence of diversity usually indicates a rapid spread of pandemic strains [56]. In contrast, the results obtained in this work suggest a long-term association between V. neptunius and bivalves, which supports the idea that V. neptunius is an endemic opportunistic pathogen present in mollusk ecosystems [3]. The optimum conditions for the growth and development of bivalve larvae in hatcheries (densities, temperature, load organic matter, etc.) favor the multiplication of bacteria and the occurrence of bacterial infections [2,3].
Extracellular metalloproteases, including vibriolysins and collagenases, are widely distributed among commensal and pathogenic species of Vibrio [13,14,28]. In this work, two hitherto uncharacterized chromosomal loci encoding a vibriolysin-like protease and a probable collagenase were identified in the genome of V. neptunius strain PP-145.98 and were designated vnpA and colA, respectively. The phenotype of the V. neptunius ∆vnpA mutant strongly suggests that VnpA is the main enzyme responsible for the lipolytic activity of this bacterium. VnpA shares high similarity with vibriolysin-like proteases with phospholipase activity from a variety of bivalve pathogens, such as V. aestuarianus [26], V. tubiashii [25], V coralliilyticus [22] and V. splendidus [27]. Interestingly, the mutation of vnpA also induced a slight reduction in the gelatinase and hemolytic activity of V. neptunius. These results are congruent with those found in V. coralliiilyticus, where inactivation of vcpA, a close homolog of vnpA, produced a significant reduction in phospholipase and collagenase activities [22]. V. neptunius ColA contains the catalytic domain structure that is conserved among collagenases, and its defective mutant strain showed reduced gelatinase activity. Gelatin is frequently used as a substrate to detect proteases, including collagenases [57]. Thus, collectively, all results strongly suggest that colA encodes a collagenase.
Virulence is a context-dependent multifactorial trait, where the contribution of a particular virulence factor depends on the genetic repertoire of a specific pathogen and on host-pathogen interactions [58]. The results obtained from infection challenges strongly suggest that the genes analyzed, namely, vnpA and colA, are required for the full virulence of V. neptunius. While vibriolysin-like proteases can play multiple roles in pathogenesis [15,16], collagenases can be involved in colonization and proliferation in the host tissues [59]. Mutagenesis of vnpA homologs in V. tubiashii and V. splendidus was also associated with a significant reduction in lethality in oysters [27,60,61]. By contrast, in V. coralliilyticus, the single inactivation of vcpA did not appear to reduce virulence [22]. Notably, the genome of V. coralliilyticus harbors three vcpA paralogs named vcpA, vcpB and vchA [23]. Thus, it is tempting to hypothesize that differing contributions to virulence are based on the existence of gene duplications (paralogs). However, since V. tubiashii also harbors three paralogs of these genes (vtpA, vtpB and vthA) [23], this hypothesis is not fully sustainable. Bacterial pathogens are highly specific in their infection behavior against different host organisms [62]. Thus, a plausible explanation relies on the fact that virulence challenges reported for the V. coralliilyticus vcpA-defective mutant were performed using Drosophila and Artemia as animal models, and the results might not be comparable [23]. Notably, our results also showed that inactivation of vnpA led to a decrease in hemolytic activity, an effect that was also reported in V. tubiashii after inactivation of vtpA [61]. By contrast, vcpA inactivation in V. coralliilyticus resulted in a significant increase in hemolytic activity [22]. In this regard, the contribution to virulence of hemolysins from the fish pathogen Photobacterium damselae subsp. damselae is dependent on their combined expression with damselysin, a metalloprotease with phospholipase activity that has a synergistic effect on virulence, depending on the animal infected [63].
A significant number of collagenolytic organisms associated with pathogenic processes that involve the destruction of collagen-containing tissues have been described [64]. Notably, to date, no known studies have evaluated the contribution of collagenases to the virulence of bivalve pathogens. Inactivation of colA in V. neptunius significantly reduced the mortality of oyster larvae with respect to the wild-type strain. This result strongly suggests that ColA is required for the full virulence of this species. Several collagenases isolated from various species of Vibrio (V. alginolyticus, V. anguillarum, V. cholerae, V. fluvialis, V. parahaemolyticus, V. proteolyticus and V. vulnificus) have been postulated as probable virulence factors [64]. However, their role in host colonization is still unclear. In V. alginolyticus, the colA mutant showed no collagenolytic activity, but no virulence assays were performed [29]. Collagenases of Clostridium perfringens, V. parahaemolyticus and Fusobacterium nucleatum act as virulence factors for tissue invasion and damage [65][66][67]. In addition, collagenase inactivation in Leptospira sp. results in a decrease in virulence for mice [59]. Recently, deletion of the colP gene reduced the virulence of P. damselae subsp. damselae in sea bass [41].

Conclusions
This work shows that vnpA encodes a gene that is largely responsible for the phospholipase and esterase activities in V. neptunius and that colA encodes a collagenase. The results show that both vnpA and colA are required for the full virulence of V. neptunius in larvae of Ostrea edulis. Although some vibriolysin-like metalloproteases with a role in virulence were previously characterized in some pathogens of bivalves, our work provides the first evidence of the role of a collagenase in the virulence of this type of pathogen. In addition, some degree of variability in enzymatic activity phenotypes and in vnpA and colA distribution was observed among V. neptunius strains. However, it is currently unknown whether this phenotypic variability correlates with virulence. Further work must be performed to study the geographical distribution of V. neptunius and its implications for virulence. Funding: This work was supported by grants AGL2017-86183-R (AEI/FEDER, EU), co-funded by the FEDER Programme from the European Union, and PID2019-103891RJ-100 (AEI) from the State Agency for Research (AEI) of Spain.

Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.

Data Availability Statement:
The data for this manuscript is available from correspondence author.