Biocontrol of Acute Hepatopacongreatic Necrosis Disease (AHPND)

Acute hepatopancreatic necrosis disease (AHPND) causes mass mortalities in farmed penaeid shrimp and has proven difficult to control using typical disease control measures. The causative agent of AHPND has been identified as Vibrio parahaemolyticus strains possessing the 69 kbp plasmid pVPA3-1 containing genes homologous with Photorhabdus insect-related (Pir) toxin-like genes (pirAand pirBlike). Probiotics have been used successfully in shrimp aquaculture to control disease outbreaks caused by pathogenic Vibrio, but there are currently no probiotics available that have been proven to control AHPND. The goal of this study was to screen and characterize marine bacterial isolates as potential agents to prevent Artemia nauplii and Litopenaeus vannamei post-larvae (PL) mortality by the pathogen Vibrio parahaemolyticus. Twelve candidate probiotic organisms were tested in an Artemia sp. model. Phaeobacter inhibens was the only candidate probiont that significantly increased the survival of Artemia nauplii challenged with AHPND V. parahaemolyticus (p<0.001). Candidate probionts Pseudoalteromonas piscicida, Pseudoalteromonas flavipulchra, and Pseudoalteromonas arabiensis were lethal to Artemia nauplii (p<0.001). Six species of candidate probiotic organisms were tested in L. vannamei. P. inhibens was the only candidate probiont tested which was not harmful to L. vannamei PLs and significantly increased the survival of PLs challenged with AHPND V. parahaemolyticus (p<0.001). Genome analysis of V. parahaemolyticus PSU5579 revealed the presence of the multiple putative virulence genes including nine hemolysins, six secreted proteases, and six secretion systems including one T3SS and two T6SS. The genome also contains the 69 kbp pVPA3-1 plasmid encoding the pirAand pirB-like toxin genes. Genome analysis of Bowmanella denitrificans JL63 revealed several gene clusters potentially involved in the production of the following antibacterial compounds: colicin V (or bacteriocin), lanthionine, the broad-spectrum antibacterial protein marinocine encoded by the lodAB operon, a secreted hemolysin-type calcium-binding bacteriocin, lantipeptide, bacteriocin, and a nonribosomal peptide.

v I would also like to thank Paul Johnson and Janet Atoyan at the URI Genomics and Sequencing Center for their tremendous help with sequencing.
I would like to thank the entire faculty as well as all the graduate students of the Cell and Molecular Biology department. It was an honor to get to know every one of you, to learn from you, and to work with you over these last couple of years.
vi DEDICATION I dedicate this thesis to my parents Carol and James Ionata. Thank you for always believing in me and encouraging me to further my education and professional career. I am unendingly grateful for your support and continuous confidence in me.
I would also like to dedicate this thesis to Linda LaPorte. Thank you for your continuous love and support. I am incredibly grateful that we share a passion for microbiology and that I was able to undertake this journey through graduate school together with you. vii

PREFACE
The following thesis has been prepared in manuscript format according to the guidelines of the Graduate School of the University of Rhode Island. This thesis contains a literature review and three manuscripts.
The first manuscript "Biocontrol of acute hepatopancreatic necrosis disease (AHPND)" will be submitted to BMC Microbiology.
The second manuscript "Draft genome sequence of Vibrio parahaemolyticus PSU5579, isolated during an outbreak of acute hepatopancreatic necrosis disease (AHPND) in Thailand." will be submitted to Genome Announcements.
The third manuscript "Draft genome sequence of Bowmanella denitrificans JL63, a bacterium isolated from whiteleg shrimp (Litopenaeus vannamei) that can inhibit the growth of Vibrio parahaemolyticus" will be submitted to Genome Announcements. viii

Introduction
In 2009 an emerging disease now known as acute hepatopancreatic necrosis disease (AHPND) began to affect penaeid shrimp farms in southern China [1,2]. The disease has spread to Vietnam, Malaysia, Thailand, and Mexico and global losses from AHPND are estimated to amount to more than one billion US dollars annually [2][3][4][5].
AHPND causes serious production losses in affected areas which negatively impacts local employment, social welfare, and international markets [6]. The causative agent of AHPND has been identified as Vibrio parahaemolyticus strains possessing the 69 kbp plasmid pVPA3-1 containing genes homologous with Photorhabdus insect-related (Pir) toxin-like genes (pirA-and pirB-like) [2,7,8]. AHPND has proven difficult to control using typical disease control measures such as water disinfection and antibiotic treatment [9,10].
Beneficial microbes known as probiotics have been used to improve the health and disease tolerance of terrestrial farm animals since the 1940s, and research on probiotics in aquaculture has continued to increase since the late 1980s [11][12][13][14].
Studies have shown that probiotics can be used in aquaculture to prevent diseases in a variety of farmed species while also improving harvest yields [14][15][16][17][18][19][20][21]. Probiotics can provide various benefits in aquaculture including improvement of water quality, enhancement of nutrition of host species, reduced incidence of diseases, higher survival rates, and improved host immune response [15,16,22]. Probiotics have been used successfully in shrimp aquaculture to control disease outbreaks caused by pathogenic Vibrio spp. [14][15][16][17][18] and may have the potential to control AHPND.
Probiotics provide an alternative to the use of antibiotics in aquaculture, which have become increasingly controversial and ineffective due to the emergence of antibiotic resistance in bacteria [15,[23][24][25]. Members of the genus Phaeobacter have been shown to be effective probiotic organisms by protecting cod and turbot larvae from the pathogen Vibrio anguillarum [26,27], as well as eastern oyster (Crassostrea virginica) larvae from the pathogens Aliiroseovarius crassostreae CV919-312 T and Vibrio coralliilyticus RE22 [19,28]. The marine bacterium Phaeobacter inhibens S4Sm is an excellent biofilm former [28], produces the broad-spectrum antibiotic tropodithietic acid (TDA) [28], can quench/inhibit the quorum sensing-dependent production of the virulence factor protease in V. coralliilyticus RE22 [29], and is nontoxic to eukaryotic organisms [30], which makes it an ideal candidate for the control of bacterial diseases in aquaculture such as AHPND.

Acute Hepatopancreatic Necrosis Disease (AHPND)
Aquaculture is the world's fastest growing food production sector with cultured shrimp increasing at an annual rate of 16.8 % [31]. In 2007, shrimp harvested from aquaculture surpassed wild-caught shrimp, and in 2013, aquaculture produced 4.45 million metric tons of shrimp [32]. As of 2012, the shrimp farming industry was worth an estimated $19.4 billion [32]. Southeast Asia and China have the largest and most productive shrimp farming regions in the world with 77% of globally produced shrimp coming from Asia [32]. In 2009 an emerging disease first called early mortality syndrome (EMS) began to affect shrimp farms in southern China [1]. The disease has recently been given a more descriptive name, acute hepatopancreatic necrosis disease (AHPND) [2]. Since its emergence, the disease has spread to Vietnam, Malaysia, Thailand, and Mexico [2][3][4]. AHPND affects both whiteleg shrimp (Litopenaeus vannamei) and black tiger shrimp (Penaeus monodon) and can lead to 100% mortality in affected populations [3]. The causative agent of AHPND has been identified as V. parahaemolyticus strains possessing the 69 kbp plasmid pVPA3-1 containing genes homologous with Photorhabdus insect-related (Pir) toxinlike genes (pirA-and pirB-like) [2,7].
Initial studies determined that the pathology of AHPND is limited to the hepatopancreas (HP) which suggests that the disease may have a toxin-mediated etiology [19,33]. It has also been shown that cell-free supernatant from V.
parahaemolyticus strains possessing pVPA3-1 can cause AHPND, supporting the conclusion that a toxin is associated with the disease [19]. AHPND develops approximately eight days after ponds are stocked with shrimp post-larvae (PLs) and severe mortalities occur within the first 20-30 days [19,33]. Early signs of AHPND include a pale to white HP, reduced HP size, empty stomach, and empty midgut ( Figure 1) [19]. Histological analysis of the HP has revealed three stages of AHPND: initial, acute, and terminal. In the initial stage, the epithelial cells of the HP are elongated into tubular lumen and there is a reduction of the vacuole size in R (resorptive) and B (blister like) cells [33]. In the acute stage, the tubular epithelium is necrotic with severe desquamation of the cells showing hemocytic infiltration as a response to the necrotic epithelium [33]. In the terminal stage of the disease, the HP tubules show a severe inflammatory response and the tubular epithelium becomes entirely necrotic with massive sloughing of epithelial cells ( Figure 2) [33][34][35]. At this stage, low levels of Vibrio can be found in the necrotic tissue in the HP and higher loads of Vibrio can be found in the stomach [33]. Additionally, there is increased hemocyte infiltration and black streaks or spots develop in the HP due to melanin deposition from hemocyte activity [19][20][21][22][23][24][25][26][27][28][29][30][31][32][33]. The absence of an inflammatory response that is usually elicited by a pathogen during the early stages of AHPND strongly supports the conclusion that this disease has a toxin-mediated etiology [2].  hepatopancreas were observed when challenged at 10 5 CFU/ml (c), and these signs were more severe at 10 6 CFU/ml (d); however, non-AHPND pathology was found at 10 3 CFU/ml (a) and 10 4 CFU/ml (b) [35]. Sloughing can be observed as cells round up and detach into the tubule lumens.
AHPND causes serious production losses in affected areas which negatively impacts local employment, social welfare, and international markets [6]. Global losses from shrimp disease are estimated to amount to around three billion US dollars annually [31] with losses from AHPND amounting to more than one billion US dollars annually [5]. Disease prevention can be challenging for shrimp farmers because most farmers do not have the resources to treat seawater before it is used to fill their ponds [36] and by the time shrimp are showing signs of AHPND, it is difficult to treat as antibiotic treatment has proven unsuccessful in most cases [10].
Additionally, treating water sources with chlorine, ozone, or UV before stocking does not provide total sterility [9]. Further, disinfection of water perturbs the natural microbial balance and leaves the environment open to opportunistic bacteria which survived disinfection. This can actually favor the growth of Vibrio as Vibrio grow rapidly after their competitors are removed [37]. V. parahaemolyticus has been reported to have a generation time as short as 12 minutes [38].
Current recommendations to prevent AHPND outbreaks in shrimp farms include the use of greenwater systems [39] or the application of biocontrol strategies such as probiotics [5], phage [40], or Bdellovibrio-and-like organisms (BALOs) [41].
It has been observed that AHPND is less prevalent in ponds colonized by copepods [39]. Copepods require a constant supply of phytoplankton and bacteria as feed, so their presence in an indicator of a mature ecosystem [42]. The use of greenwater systems has also been observed to reduce the incidence of AHPND [39]. Greenwater systems are characterized by a mature micro-algal and bacterial community. These systems have been shown to maintain decreased Vibrio levels and decreased animal mortality [43,44]. The beneficial effect of greenwater systems can be attributed to the algal and bacterial production of antibacterial substances [45,46] and compounds which quench/inhibit quorum sensing-dependent production of virulence factors in pathogens [47]. Additionally, the bacteria in greenwater systems compete with pathogens for available nutrients and occupy niches which would otherwise be left open for invading pathogens [46]. Occurrences of overgrowth of pathogenic bacteria such as Vibrio spp. in shrimp grow-out ponds can be reduced by minimizing disturbances such as water disinfection which lead to sudden variations in nutrient levels, and by colonizing pond water with nonpathogenic bacteria and/or algae [48].
Probiotic bacteria have been used successfully in shrimp aquaculture to control disease outbreaks caused by pathogenic Vibrio spp. [14][15][16]. A recent study determined that the probiotics which are currently commercially available to shrimp farmers in Malaysia are not effective at controlling AHPND [49]. More research needs to be conducted to develop and test new probiotic formulations which may have the potential to control AHPND. The use of phage has also been proposed as a potential strategy to control AHPND, and a virulent Siphoviridae phage, pVp-1, has been shown to have effective bacteriolytic activity against 74% of AHPND strains of V. parahaemolyticus tested, but has yet to be tested in an aquaculture setting [40].
Another promising biocontrol strategy to prevent AHPND involves the use of BALOs.
A recent study isolated a BALO, identified as Bacteriovorax sp. BV-A, from a sediment samples in a shrimp farm in Thailand, which could kill all AHPND strains of V. parahaemolyticus tested as well as Vibrio vulnificus, Vibrio cholerae, and Vibrio alginolyticus [41]. Bacteriovorax sp. BV-A was also shown to increase the survival of L. vannamei PLs challenged with AHPND V. parahaemolyticus by 50% [41]. In field studies, BALOs in combination with photosynthetic bacteria have been shown to provide increased survival rates of Chinese mitten crab (Eriocheir sinensis) and decreased Vibrio concentrations in cultured pufferfish (Fugu obscurus) [50]. The use of BALOs is a promising prospect for the control of diseases caused by bacterial pathogens in aquaculture.

Vibrio parahaemolyticus
Vibrio is a genus of Gram-negative motile marine bacteria of the family Vibrionaceae within the Gammaproteobacteria [51]. Members of this genus are facultative anaerobes with a curved-rod shape [52]. Vibrio species can be found in a wide range of aquatic environments, including the water column, in association with hosts (both pathogenic and symbiotic), and even in extreme habitats (hydrothermal vents) [53]. Pathogenicity in Vibrio is not species dependent, but rather strain specific as different strains of the same species can cause diseases in different hosts, or can be nonpathogenic [54,55] campbellii, V. alginolyticus) [53]. The aquaculture industry suffers multibillion-dollar losses due to these pathogens annually [5,53,61].  [53,[62][63][64]. Virulence gene expression in Vibrio is regulated by quorum sensing and has been studied extensively in V. harveyi [53,64].
V. harveyi uses a three-channel quorum-sensing system (Figure 3), secreting chemical signal molecules that include HAI-1 (Harveyi autoinducer 1), AI-2 (Autoinducer 2), and CAI-1 (Cholera autoinducer 1) [53]. The concentration of these molecules in the extracellular environment is proportional to cell density. These autoinducers are detected at the cell surface by membrane bound histidine sensor kinase proteins that feed a phosphorylation ⁄ dephosphorylation signal transduction pathway which controls the production of the quorum-sensing master regulator protein LuxR (V. harveyi)/OpaR (V. parahaemolyticus) [53]. LuxR/OpaR directly activates the Lux operon, whereas most of the other genes regulated by quorum sensing are controlled indirectly [53]. Several species of Vibrio, including V. cholerae and V.
parahaemolyticus, have virulence factors which are controlled by the ToxR regulon [53]. In V. cholerae, the ToxR regulon controls the expression of the ctx gene encoding the cholera toxin [53]. In V. parahaemolyticus the toxR operon controls the expression of the thermostable direct hemolysin gene (tdh) as well as the T3SS [65,66]. The toxR operon is found in both clinical and environmental isolates of V.
Vibrio parahaemolyticus is commonly found in marine coastal waters and estuarine environments including water, sediment, suspended particles, plankton, fish and shellfish [53]. Strains of this species are a leading cause of seafood-associated bacterial gastroenteritis globally and can also cause eye, ear, and wound infections [68]. While most environmental strains of V. parahaemolyticus are not pathogenic to humans, strains possessing the tdh and trh genes and the T3SS2 gene cluster are pathogenic [69,70]. Strains of V. parahaemolyticus are also an important shrimp pathogen and have been identified as the causative agent of AHPND. One of the challenges of preventing and treating AHPND is the high frequency of antibiotic resistance found in V. parahaemolyticus isolates. Jiang et al. [71]  Photorhabdus insect-related (Pir) toxin-like genes (pirA-and pirB-like) [2,7,8].
These genes are located within a 3.5 kbp fragment flanked by inverted repeats of a transposase-coding sequence (1 kbp) which is a mobile genetic element that can induce horizontal gene transfer [2]. The GC content of the pirA-and pirB-like genes is only 38.2%, which is considerably lower than that of the rest of the plasmid (45.9%), suggesting that these genes were recently acquired [2]. Similar to the Pir toxins that affect insects, the Pir-like toxins act as binary proteins, which form a heterodimer and both pirA-and pirB-like genes are required for pathogenesis [2,8,75]. The crystal structure of the PirAB-like heterodimer has similar structural topology to that of the Bacillus Cry insecticidal toxin-like proteins, despite the low sequence identity (<10%), which suggests that the putative PirAB-like toxin might emulate the functional domains of the Cry protein and its pore-forming activity [75].
While the PirA-and PirB-like toxins affect the hepatopancreas in shrimp, the Pir toxins primarily affect the midgut of insects, which may suggest different mechanisms of action [2].
While AHPND strains of V. parahaemolyticus have been shown to possess between 1 and 121 copies of the pVPA3-1 plasmid per cell [2,8], the copy number of this plasmid does not correlate with virulence [8]. Instead, the amount of secreted PirA-and PirB-like proteins determines virulence to shrimp [8]. AHPND strains of V.
parahaemolyticus have been shown to possess other virulence factors as well, including T3SS1, T6SS1, and T6SS2 genes [74]. Additionally, a unique sequence encoding a type IV pilus has been found in the genomes of AHPND strains of V.
parahaemolyticus isolated in Thailand and Mexico [7,76], but was not detected in strains isolated in India [34]. It has also been shown that AHPND strains of V.
parahaemolyticus lack the tdh and trh genes [7,8,19,33,34] as well as the T3SS2 gene [7,34,74] required for pathogenesis in humans, indicating that these strains are not human pathogens. The role of virulence factors other than the PirA-and PirB-like toxins in the pathogenesis of AHPND strains of V. parahaemolyticus to shrimp has yet to be determined.

Probiotics in aquaculture
For more than 70 years, beneficial microbes known as probiotics have been used to improve the health and disease tolerance of terrestrial farm animals such as swine and chickens [11][12][13]. Probiotics are defined as "live microorganisms, conferring a healthy benefit to the host when being consumed in adequate amounts" [77]. Probiotics are now widely used for enhancing production of land animals due to the fact that they are better, cheaper, and more effective in promoting animal health than antibiotics or chemical substances [21]. Research on the use of probiotics in aquaculture dates back to the late 1980s and has continued to increase since then [14].
Studies have shown that probiotics can be used in aquaculture to prevent diseases in bivalves (oysters, scallops), fish (salmon, cod, trout, halibut, turbot, catfish), and crustaceans (shrimp, Artemia spp.) [14-18, 20, 21]. Although probiotics can prevent disease when applied prophylactically, they are not meant to be used therapeutically and are unlikely to cure animals which are already infected with a pathogen [78,79].
Currently the main rate limiting factor in the shrimp aquaculture industry is disease control. Intensive (high-density) shrimp culture systems have become common practice because they produce substantially higher shrimp yields than do any other invertebrates due to the lack on an adaptive immune system [20]. In aquaculture, bacterial disease is generally controlled through water disinfection and the application of antibiotics both prophylactically and therapeutically. The use of antibiotics in aquaculture has become increasingly controversial and ineffective due to the emergence of antibiotic resistance in bacteria [15,[23][24][25]. Water disinfection also has limited success and in some cases may actually increase the likelihood of an outbreak, most notably in controlling diseases caused by Vibrios spp. such as AHPND [9,10,39]. Additionally, disinfecting water with chlorine has been shown to increase the proportion of multiple antibiotic resistance bacteria [83].
The overuse of antibiotics in aquaculture has become a major concern due to the emergence of antibiotic resistance in bacteria and the potential for residual contamination in harvested fish and shellfish. The aquaculture industry uses massive quantities of antibiotics which are released into the environment [24,37]. For example, antibiotic usage in shrimp farms in Thailand in 1994 was estimated to be as much as 500 -600 tonnes [37]. The leaching of these antibiotics into the environment contributes to the development of antibiotic resistance determinants in bacteria which can be spread to other species by horizontal gene transfer [24,84]. These determinants can spread by horizontal gene transfer to bacteria of the terrestrial environment as well, including human and animal pathogens [24]. There is also public health concern over potential exposure of human consumers to antibiotic residues or other chemical contaminants in shrimp harvested from aquaculture [86]. Undetected consumption of antibiotics in food can cause allergy and toxicity problems, alter normal flora and increases susceptibility to infections, and select for antibiotic-resistant bacteria [24]. In 2006, there was an antibiotic residue crisis in the flatfish industry in China where 25,000 tonnes of turbot could not be sold, costing the industry an estimated 200 million Euro [50]. For these reasons, the use of probiotics in aquaculture is becoming an increasingly popular alternative to the use of antibiotics [15].
Probiotics provide several benefits in aquaculture including improvement of water quality, enhancement of nutrition of host species, reduced incidence of diseases, higher survival rates, and improved host immune response [15,16,22]. Although there have been many studies on probiotics in aquaculture, they are still not widely used. Greenwater systems are a new water management strategy which use mature microalgal and bacterial communities and have been shown to have reduced Vibrio levels and increased animal survival rates [43,44]. Probiotics can be used not only as a biocontrol strategy, but can also be used in conjunction with algae treatment to make greenwater systems. In aquaculture, probiotics are typically added to the feed or directly into the culture water [15]. Probiotics have been used successfully in shrimp aquaculture to control disease outbreaks caused by pathogenic Vibrio spp. while also improving harvest yields [14][15][16][17][18]. The most common probiotics used in aquaculture are photosynthetic bacteria (purple non-sulfur bacteria), antagonistic bacteria (Pseudoalteromonas spp., Flavobacterium spp., Alteromonas spp., Phaeobacter spp., Bacillus spp.), microorganisms for improving digestion (lactic acid bacteria and yeast), bacteria for improving water quality (nitrifying bacteria, denitrifiers), and predatory bacteria that kill other bacteria (e.g. BALOs) [50].
Candidate probiotic organisms are typically selected based on their ability to produce antibacterial and/or antivirulence compounds. These compounds give an ecological advantage to the producing bacteria against other microorganisms, and may also provide an advantage against bacteriovorous eukaryotic predators [30]. Some organisms that have been shown to inhibit the growth of bacterial pathogens also produce compounds that are toxic to eukaryotic organisms [30,87,88]. These organisms should tested thoroughly before being applied in aquaculture as they might cause adverse effects on the farmed animals, their prey species (algae, rotifers, or culture water with non-antagonistic probiotic bacteria protected Artemia sp. from the pathogenic effects of a Vibrio proteolyticus and hypothesized that this protection may be due to competition with the pathogen for available nutrients. 5) Enhancement of host digestion through the production of enzymes which can break down chitin, starch, protein, cellulose, and lipids [16]. 6) Stimulation of host immune response [15].
Although shrimp lack an adaptive immunity, they still possess an innate immune system that effectively protects them from harmful microorganisms and probiotic treatment has been shown to modulate the cellular and humoral immune responses in shrimp [95]. 7) Production of siderophores which compete with pathogens for ferric iron in the iron-limited environment of the host [15]. 8) Production of acylhomoserine lactones (AHLs) which quench/inhibit the quorum sensing-dependent production of virulence factors in pathogens [29,95]. Quorum sensing has been shown to be one of the virulence mechanisms of many pathogenic bacteria, including V. harveyi and V. parahaemolyticus [53,98]. Organisms that are not harmful to host species and possess some, if not all, of these properties make ideal candidates for use as probiotics in aquaculture.

Phaeobacter inhibens S4Sm
The Roseobacter clade consists of organisms that occupy diverse marine  The probiotic activity of P. inhibens S4Sm can be attributed to at least three factors: 1) excellent biofilm forming ability [28]; 2) production of TDA [28]; and 3) ability to quench/inhibit the quorum sensing-dependent production of the virulence factor protease in V. coralliilyticus RE22 [29]. Toxicology studies using Artemia sp.
and Caenorhabditis elegans have shown that both P. inhibens as well as purified TDA are innocuous for these organisms [30]. P. inhibens S4Sm produces a more robust biofilm than the fish pathogen V. anguillarum or the oyster pathogens A. crassostreae or V. coralliilyticus (Table 1) [28]. Knockout mutants of P. inhibens S4Sm which were deficient in biofilm formation (exoP) or antibiotic production (clpX) were shown to provide significantly less protection to oyster larvae after challenge with V.
coralliilyticus RE22 compared to wild-type P. inhibens S4Sm, demonstrating the importance of these activities for probiotic function ( Figure 5) [28]. Additionally, TDA knockout mutants of P. gallaeciensis do not protect cod larvae challenged with V. anguillarum as well as wild type P. gallaeciensis [26] and do not reduce cell densities of V. anguillarum as well as wild type P. gallaeciensis [113]. P. inhibens S4Sm has also been shown to produce acyl-homoserine lactones (AHLs) which downregulate the virulence factor protease activity in V. coralliilyticus by disrupting the quorum-sensing pathway that activates protease transcription of V. coralliilyticus [29]. , TdaR3 facilitates the glutamate-dependent acid-response system by converting glutathione to 5-oxo-proline, which is then hydrolyzed to glutamate via a 5-oxoprolinase. This glutamate is then decarboxylated to form γ-aminobutyric acid (GABA), which is exchanged by an antiporter for glutamate, resulting in the export of 1 H + per glutamate [115]. The strong biofilm forming ability combined with production of antivirulence compounds (AHLs) and a broad-spectrum antibiotic (TDA) which pathogens are unlikely to become resistant to, make P. inhibens S4Sm a promising candidate for use as a probiotic to control bacterial diseases in aquaculture such as AHPND.

Goals of this study
The overall goal of this study was to isolate and characterize bacteria inhibitory towards the growth of V. parahaemolyticus and determine if they can be used to prevent or reduce losses due to the AHPND strains of V. parahaemolyticus in aquaculture systems. AHPND causes significant losses in the shrimp aquaculture industry and current strategies to control the disease are not effective [9,10]. The use of probiotics has the potential to control AHPND, but new formulations are needed as it has been shown that probiotics which are currently available to shrimp farmers in Malaysia are not effective at controlling AHPND [49].
The first aim of this investigation was to isolate potential probiotic bacteria from the environment which can inhibit the growth of V. parahaemolyticus, quantify their biofilm formation, and identify their species. More than 300 bacterial isolates were cultured from a variety of sources and used in a zone of inhibition assay to determine if they could inhibit the growth of V. parahaemolyticus on an agar surface.
Biofilm formation was then quantified using the crystal violet method. The 16S rRNA gene for each isolate was sequenced for species-level identification.
The second aim of this study was to determine if any of the candidate probiotic organisms can increase the survival of Artemia nauplii challenged with AHPND V. parahaemolyticus. A model system using Artemia nauplii challenged with AHPND V.
parahaemolyticus PSU5579 was developed and used to test candidate organisms for probiotic activity.
The third aim of this study was to determine if any of the candidate probiotic organisms can increase the survival of L. vannamei PLs challenged with AHPND V.

Introduction
In 2009 an emerging disease now known as acute hepatopancreatic necrosis disease (AHPND) began to affect penaeid shrimp farms in southern China [1,2]. The disease has spread to Vietnam, Malaysia, Thailand, and Mexico with global losses from AHPND estimated to be more than one billion US dollars annually [2][3][4][5].
AHPND affects both whiteleg shrimp (Litopenaeus vannamei) and black tiger shrimp (Penaeus monodon) and can lead to 100 % mortality in affected populations [3]. The disease causes serious production losses in affected areas, which negatively impacts local employment, social welfare, and international markets [6]. The causative agent of AHPND has been identified as Vibrio parahaemolyticus strains possessing the 69 kbp plasmid pVPA3-1 containing genes homologous with Photorhabdus insect-related (Pir) toxin-like genes (pirA-and pirB-like) [2,7]. AHPND has proven difficult to control using typical disease control measures such as water disinfection and antibiotic treatment [8,9].
Disease prevention can be challenging for shrimp farmers because most farmers do not have resources necessary to treat seawater before it is used to fill their ponds [10]. Further, by the time shrimp show signs of AHPND, it is difficult to treat as antibiotic treatment has proven unsuccessful in most cases [9] and antibiotic treatment will select for antibiotic resistant bacteria. Additionally, treating water sources with chlorine, ozone, or UV before stocking does not provide total sterility [8]. Further, disinfection of water perturbs the natural microbial balance, leaving the environment open to opportunistic bacteria that survive disinfection, and can actually favor the growth of Vibrio species, which grow rapidly after their competitors are removed [11]. Occurrences of overgrowth of pathogenic bacteria such as Vibrio in shrimp grow-out ponds can be reduced by minimizing disturbances such as water disinfection that can lead to sudden variations in nutrient levels, and by colonizing pond water with nonpathogenic bacteria and/or algae [12].
Studies have shown that probiotics can be used in aquaculture to prevent diseases in a variety of farmed species while also improving harvest yields [16][17][18][19][20][21][22][23]. Probiotics provide several benefits in aquaculture including improvement of water quality, enhancement of nutrition of host species, reduced incidence of diseases, higher survival rates, and improved host immune response [16,18,24]. Probiotics have been used successfully in shrimp aquaculture to control disease outbreaks caused by pathogenic Vibrio spp. [16-18, 20, 21] and may have the potential to control AHPND.
Probiotics provide an alternative to the use of antibiotics in aquaculture, which have become increasingly controversial and ineffective due to the emergence of antibiotic resistance in bacteria [16,[25][26][27]. Currently there are no probiotics commercially available to shrimp farmers that have proven to be effective at preventing AHPND. A recent study determined that the probiotics which are available to shrimp farmers in Malaysia are not effective at controlling AHPND [28].
Before potential probiotic organisms can be used in aquaculture, they must be tested to confirm that no pathogenic effects can occur in the host. Artemia spp. have been used as a model organism not only for toxicology studies but also to test the effectiveness of probiotics and the role of quorum sensing in pathogenesis [29][30][31][32][33][34][35][36].
Artemia spp. are useful model organisms because they adapt easily to changes in nutrients, salinity, temperature, and oxygen, are easy to culture, are resistant to manipulation, have a short life cycle, and are inexpensive [35].
In this study, ten newly isolated potential probionts, as well as two oyster probionts, Phaeobacter inhibens S4Sm and Bacillus pumilus RI06-95, were identified as having in vitro antibiotic activity against an AHPND strain of V. parahaemolyticus.
These 12 candidate probionts were tested in vivo for their ability to protect Artemia nauplii or L. vannamei post-larvae (PL) from AHPND V. parahaemolyticus challenge.
It was found that P. inhibens S4Sm was the only candidate probiont tested which significantly increased the survival of Artemia nauplii challenged with AHPND V.
parahaemolyticus. All species of Pseudoalteromonas tested were found to be pathogenic to Artemia sp. P. inhibens S4Sm was also the only candidate probiont which was not harmful to L. vannamei PLs and significantly increased the survival of PLs challenged with AHPND V. parahaemolyticus.
parahaemolyticus strains were grown in LB20 (10 g/L tryptone, 5 g/L yeast extract, 20 g/L NaCl, pH 8). Spontaneous streptomycin-resistant mutants were selected by passing on increasing concentrations of streptomycin, up to 200 µg/ml. These strains are indicated by "Sm" at the end of their strain name. All bacterial strains were maintained and stored in 25% glycerol stocks at -80 °C.

Isolation of candidate probiotic bacteria
Environmental samples, such as seawater or small marine invertebrates, such as shrimp, were collected for the isolation of bacteria (Table 1

Zone of inhibition assay
Zones of inhibition were quantified using a modification of a method described previously [38]. V. parahaemolyticus PSU5429 was grown for 24 h in LB20, diluted 10 3 -fold, and 100 µl of this diluted culture was spread on YP30IOS agar. The candidate probionts (10 µl of a 24 h culture) were then spotted on the same plate.
Plates were incubated at 27 °C for 24-48 h. Inhibition zones were measured between growth of the candidate probiont (edge of spot) and the V. parahaemolyticus lawn (edge of lawn). Each candidate probiont was tested three times.

Biofilm assay
Biofilm formation was quantified using a modification of the crystal violet staining method [39]. Bacteria were grown for 24 h before being diluted 10 3

Statistical analysis
Statistical data analysis was performed using the Student's t-test. Data with p<0.05 was considered to be statistically significant.

Isolation of candidate probiotic bacteria and 16S sequencing
More than 300 bacterial isolates were screened for antibiotic activity against V.
parahaemolyticus PSU5429 by zone of inhibition (ZOI) assay. A total of 30 isolates were found to inhibit the growth of V. parahaemolyticus PSU5429. The 16S rRNA genes of these isolates were sequenced to identify their species, and these data combined with the ZOI and biofilm data were analyzed to rule out strains that were isolated more than once. This analysis revealed that of the 30 original isolates, 10 were unique strains (

Zone of inhibition assay
Zones of inhibition (ZOI) produced by candidate probionts against V.
parahaemolyticus PSU5429 were quantified to evaluate each organism's ability to inhibit V. parahaemolyticus growth on an agar surface. ZOIs are areas around the candidate probiont spot where V. parahaemolyticus was plated, but was not able to grow due to the presence of growth-inhibiting compound(s) secreted by the candidate probiont. Of the 12 candidate probionts, Pseudoalteromonas flavipulchra JL1 and parahaemolyticus culture was spread on YP30 agar and 10 µl of a stationary phase culture of each candidate probiont was spotted over the V. parahaemolyticus lawn.
Plates were incubated at 27 °C for 24-48 h. Inhibition zones were measured between growth of the candidate probiont and the V. parahaemolyticus lawn. Representative of three independent experiments. Error bars equal one standard deviation.

Biofilm assay
The biofilm forming ability of each candidate probiont, as well as V.  parahaemolyticus PSU5579 (t-test, p<0.05).

Duplex PCR for the detection of pirA-and pirB-like genes
In order to confirm that the putative AHPND strains of V. parahaemolyticus contained the genes necessary to cause this shrimp disease, ten strains of V.
parahaemolyticus isolated from shrimp farms located in Pattani and Songkla provinces, southern Thailand during an AHPND disease outbreak were screened for

Artemia challenge studies
To determine if any of the candidate probiotic organisms have the potential to prevent AHPND, an assay was developed to test if candidate probionts could protect Artemia nauplii from AHPND V. parahaemolyticus challenge. This assay also served as a test to determine if any of the candidate probionts can be harmful to crustaceans, such as Artemia sp., under certain conditions. For this assay, the addition of 1.6 ml YP30IOS allowed for V. parahaemolyticus PSU5579 to consistently induce a 53% -71% mortality rate in Artemia nauplii when applied at 10 5 CFU/ml 24 h after hatching ( Figure S2). When V. parahaemolyticus PSU5579 was applied at 10 4 CFU/ml 24 h after hatching, the mortality rate was lower (39% -49%) ( Figure S2). The higher mortality rate induced by V. parahaemolyticus PSU5579 when applied at 10 5 CFU/ml provides a range of survival between the challenged and unchallenged controls where a level of protection provided by probiotic organisms can be detected, which is why this concentration was used for these experiments. Without the addition of YP30IOS, challenging Artemia nauplii 24 h after hatching with V. parahaemolyticus PSU5579 at 10 6 or 10 7 CFU/ml only induced 0% or 28% mortality, respectively ( Figure S3).
Twelve candidate probiotic organisms were tested for their potential ability to kill Artemia nauplii, as well as their ability to protect nauplii from challenge with V.
parahaemolyticus PSU5579 than those pretreated with P. inhibens S4Sm at 10 5 CFU/ml ( Figure S4). None of the other eleven probiont candidate isolates were able to protect Artemia nauplii from infection and death when challenged with V.

Whiteleg shrimp (Litopenaeus vannamei) post-larvae challenge
An assay was developed to test if any of the candidate probionts could protect L. vannamei PLs from AHPND V. parahaemolyticus challenge. An initial experiment determined that V. parahaemolyticus PSU5579Sm induces a high mortality rate (67%) in L. vannamei PLs when applied at 10 6 CFU/ml at the start of the experiment; however, when PLs were incubated for 24 h prior to the addition of V.
parahaemolyticus, the mortality rate was reduced to 33% ( Figure S5). This may be due to the growth of commensal bacteria from the shrimp during the 24 h nutrients. This 24 h preincubation period is important because probiotics usually require a pretreatment period to effectively protect animals from pathogen challenge [42,43]. This issue was resolved through the addition of 200 µg/ml streptomycin to the PL water at the start of the experiment which allowed for a 24 h preincubation period without a reduced V. parahaemolyticus-induced mortality rate when applied at 10 6 CFU/ml (67%) ( Figure S5). Challenging L. vannamei PLs with 10 5 CFU/ml V.
parahaemolyticus PSU5579Sm after a 24 h preincubation period with streptomycin only induced a 5% mortality rate ( Figure S5).
In order to determine if any of the candidate probionts are harmful to shrimp, flavipulchra JL1Sm with 83% survival compared to 97% in the untreated control (p=0.008) (Figure 8).
In an effort to determine if any of the candidate probionts can protect L.
vannamei PLs from V. parahaemolyticus challenge, PLs were pretreated with candidate probionts at 10 6 CFU/ml for 24 h before V. parahaemolyticus PSU5579Sm challenge and, as described previously, the candidate probionts were also added every  characteristic is an important property for probiotics used for disease control [38,44], and is commonly used as a primary test in selecting candidate probionts, but does not necessarily guarantee that candidate probionts will be effective at protecting live host organisms such as shrimp [45]. Candidate probionts were characterized and tested for their ability to protect both L. vannamei and Artemia sp. from AHPND V.
parahaemolyticus challenge. This study identified P. inhibens S4Sm as a bacterial candidate, which has the potential to be used as a probiotic for control of AHPND in penaeid shrimp aquaculture. This study also showed that Artemia sp. can be used to identify probionts that protect L. vannamei from AHPND V. parahaemolyticus challenge. Under the conditions used in this study, Artemia sp. were also shown to have a higher sensitivity than L. vannamei to organisms which are harmful to crustaceans and, therefore, can be used to identify organisms that should not be used in shrimp aquaculture.
Twelve bacterial strains were selected as candidate probionts for control of AHPND because of their antagonistic properties against V. parahaemolyticus. Two of these strains, P. inhibens S4Sm and B. pumilus RI06-95, were oyster probionts previously identified by Karim et al. [41]. The other ten strains were isolated during this study based on their ability to inhibit the growth of V. parahaemolyticus and were identified by 16S rRNA genes sequencing. The production of antimicrobial compounds by these organisms, as determined by ZOI assay, suppresses the growth of V. parahaemolyticus, allowing them to outcompete V. parahaemolyticus for nutrients and energy sources. Probiotics with known antagonistic activity have been shown to decrease the concentration of Vibrio spp. in black tiger shrimp (P. monodon) rearing water [46,47]. Antagonistic probionts have also been shown to inhibit the colonization of P. monodon by V. harveyi through competitive exclusion [48].
The biofilm forming ability of the twelve candidate probionts as well as V.
parahaemolyticus PSU5579 was quantified by the crystal violet method using polystyrene 96-well plates. Biofilm formation is an important characteristic for probiotic activity because competition for attachment sites within the host is likely to serve as the first barrier of defense against invading pathogenic bacteria [16,38,44].
Eleven of the twelve candidate probionts produced significantly stronger biofilms than V. parahaemolyticus PSU5579 under the conditions tested (p<0.006). The only candidate probiont that did not produce a significantly stronger biofilm than V.
parahaemolyticus PSU5579 was B. pumilus HR1 (p=0.565). The biofilm assay used in this study provides insight into the biofilm forming ability of the organisms tested, but is not comprehensive and may not be predictive of how well organisms will be able to colonize a host. Some of the organisms tested may form stronger biofilms on a biotic surface than on an abiotic surface, and the biofilm forming ability of these organisms may be underestimated using this assay. However, other organisms may form strong biofilms on a variety of surfaces. For example, P. inhibens S4Sm, which formed the strongest biofilm on polystyrene of any of the organisms tested in this study, has also been shown to form a strong biofilm on borosilicate glass [38]. Zhao et al. [38] made an exoP-knockout mutant of P. inhibens S4Sm to study the contribution of biofilm forming ability to the probiotic activity of this organism. The P. inhibens S4Sm exoP mutant had 60% reduced biofilm forming ability and oyster larvae pretreated with this mutant before Vibrio coralliilyticus challenge had 30% lower survival than larvae pretreated with wild-type P. inhibens S4Sm, indicating that biofilm formation is important for the probiotic activity of P. inhibens S4Sm [38].
All twelve candidate probionts demonstrated antagonistic activity against V.
parahaemolyticus on an agar surface and eleven of the twelve candidate probionts also form stronger biofilms than V. parahaemolyticus on a polystyrene surface. Organisms with both of these characteristics may be able to competitively exclude the pathogen from colonizing the host and the surrounding environment, thereby limiting the proliferation of the pathogen and reducing the likelihood of disease. Attachment to the host and production of antimicrobial compounds are critical factors for the ability of lactic acid bacteria to exclude pathogens in both humans [49,50] and fish [51].
Verschuere et al. [31] quantified the colonization of Artemia nauplii by nine candidate probionts as well as the ability of these organisms to protect Artemia nauplii from Vibrio proteolyticus challenge and observed a correlation between colonization potential and the protective ability of the candidate probionts [31]. All twelve candidate probionts used in this study showed promising results in vitro; however, these results were not predictive of their effectiveness in vivo, possibly due to toxicity to the host or other undetermined factors.
Artemia spp. are an advantageous model organism to test the effectiveness of probiotics at reducing pathogen-induced mortality [31,34,36]. Verschuere et al. [31] found several probionts that provide total protection to Artemia nauplii from V.
proteolyticus. Pretreatment of Artemia nauplii with yeast (Saccharomyces boulardii) also provides total protection from Vibrio harveyi challenge [36]. Bacillus licheniformis and Pseudomonas aeruginosa have also been shown to provide nearly maximum survival (78%) to Artemia nauplii from non-AHPND V. parahaemolyticus [34]. This study is the first to test candidate probionts in an AHPND V. inhibens S4Sm is allowed to precolonize the coverslip for 24 h prior to the introduction of the pathogen [38].
For the L. vannamei challenge assay an initial experiment determined that a 24 h preincubation period prior to 10 6 CFU/ml V. parahaemolyticus challenge reduced V.
parahaemolyticus-induced L. vannamei mortality from 67%, when L. vannamei were challenged at 0 h, to 33% even without the addition of candidate probionts. This assay uses a much higher PL density than that of even super-intensive shrimp farming practices which use a maximum density of 7 PLs per 10 L [55]. Due to this high PL density, commensal bacteria from the shrimp likely grew to a high density during the 24 h preincubation period. These commensal bacteria then compete with V.
parahaemolyticus for available nutrients, which may explain why V.
parahaemolyticus-induced L. vannamei mortality was reduced. To inhibit the growth of commensal bacteria from the PLs, streptomycin (200 µg/ml) was added to the shrimp culture water. The addition of streptomycin restored V. parahaemolyticusinduced L. vannamei mortality to 67% when applied at 10 6 CFU/ml ( Figure S5). A similar approach using streptomycin treatment is used in mouse models to allow both pathogenic and nonpathogenic bacteria to colonize the gastrointestinal tract [56][57][58][59].
Streptomycin treatment renders mice highly susceptible to enteric pathogens due to the elimination of commensal facultative intestinal bacteria [59]. This study showed that streptomycin treatment has the same effect on L. vannamei by increasing their susceptibility to AHPND V. parahaemolyticus. These results indicate that the Artemia sp. model used in this study makes a good substitute for L. vannamei to study the effects of probiotics on AHPND. The Artemia sp. model successfully identified P. inhibens S4Sm as being the only candidate probiont tested that is not harmful to crustaceans and can prevent AHPND.

Representatives
The Artemia sp. model also identified B. pumilus as having no effect on AHPND.
Artemia were also more sensitive to the harmful effects of Pseudoalteromonas spp. and B. denitrificans than L. vannamei, demonstrating that these organisms may be harmful to shrimp if used under different conditions and/or long-term.
Phaeobacter inhibens S4Sm is the only candidate probiont tested that is not harmful to L. vannamei PLs and can significantly increase the survival of both L.
Although twelve strains of candidate probionts produced promising results in vitro, our study showed that biofilm formation and growth-inhibiting activity toward a particular pathogen in vitro are not necessarily predictive of how a candidate probiont would perform in vivo. This study found that under the conditions used, Ps. flavipulchra, Ps. piscicida, and Ps. arabiensis were lethal to Artemia sp. and Ps.
flavipulchra was also harmful to L. vannamei. It has been shown that some organisms, such as Pseudoalteromonas spp., that produce compounds inhibitory toward the growth of bacterial pathogens are also toxic to eukaryotic organisms [29,60,61]. Neu et al. [29] determined that Pseudoalteromonas luteoviolacea strains S2607 and S4060 produce the antibacterial compound pentabromopseudilin which is lethal to Artemia nauplii. Ps. piscicida S2049 has also been shown to produce several bromoalterochromides [62] which are inhibitory toward Bacillus subtilis [63] and toxic to sea urchins [62]. Ps. rubra produces prodigiosin [64] which is antagonistic toward bacteria [65] and toxic to algae [66] and eukaryotic parasites [67].
Bowmanella denitrificans did not significantly decrease the survival of Artemia demonstrated potentiation of the fish pathogen Aeromonas salmonicida by lactic acid bacteria. The lactic acid bacteria were able to colonize the intestine of Atlantic salmon (Salmo salar) fry, but surprisingly increased the mortality of fry challenged with A.
Bacteria belonging to the genus Bacillus are some of the most common organisms used as probiotics in aquaculture, have been shown to be effective probiotics for penaeid shrimp, and can reduce incidence of disease caused by Vibrio spp. [16,22,71,72]. However, this study found that in the conditions used in these experiments, B. pumilus was not able to reduce the mortality rate of Artemia sp. or L.
vannamei challenged with AHPND V. parahaemolyticus. This lack of in vivo protection by a candidate probiont with promising in vitro activity has been shown before. For example, Pseudomonas fluorescens can protect rainbow trout (Oncorynchus mykiss) from Vibrio anguillarum [73] but does not protect salmon (S. salar) from A. salmonicida, even though P. fluorescens can inhibit the growth of A. salmonicida in vitro [45]. This emphasizes the need to test candidate probionts for each unique host-pathogen combination in vivo before application in aquaculture.
Members of the genus Phaeobacter have been shown to be effective probiotic organisms for the protection of cod and turbot larvae from the pathogen V.
anguillarum [74,75], as well as eastern oyster (Crassostrea virginica) larvae from the pathogens Aliiroseovarius crassostreae CV919-312 T and V. coralliilyticus RE22 [19,38]. This study demonstrated that P. inhibens S4Sm can also protect Artemia sp. and L. vannamei from AHPND V. parahaemolyticus. The probiotic activity of P. inhibens S4Sm has been studied and can be attributed to at least three factors: 1) excellent biofilm forming ability [38]; 2) production of the broad-spectrum antibiotic tropodithietic acid (TDA) [38]; and 3) ability to inhibit the quorum sensing-dependent production of the virulence factor protease in V. coralliilyticus [76]. It has also been shown that resistance to TDA is hard to select [77], making it unlikely that pathogens with develop resistance to this probiotic over time.
In conclusion, P. inhibens S4Sm has great potential for application in whiteleg shrimp (L. vannamei) aquaculture for prevention of AHPND. P. inhibens S4Sm is a strong biofilm former, showed antibiotic activity against V. parahaemolyticus in vitro, and provided protection to both Artemia sp. and L. vannamei in vivo. Application of Supplemental Data Figure S1. Survival of Artemia nauplii challenged immediately after hatching with suspected AHPND V. parahaemolyticus strains at 10 6 CFU/ml. Representative of one independent experiment with three technical replicates.    Figure S4. Effect of preincubation of Artemia nauplii with P. inhibens S4Sm at 10 5 or 10 6 CFU/ml for 24 h on survival 48 h after challenge with V. parahaemolyticus PSU5579 at 10 5 CFU/ml. P. inhibens S4 was added at the start of the experiment and every 24 h. Representative of one independent experiment with three technical replicates. The genome of V. parahaemolyticus PSU5579 encodes a number of lytic enzymes including two secreted collagenases, one chitinase, one extracellular lipase, phospholipases A and C, nine hemolysins including cytolysin, leukocidin, and delta-VPH, and six secreted proteases including an extracellular serine protease, three secreted trypsin-like serine proteases, and two extracellular zinc proteases including Vibriolysin. Three adherence systems were identified: a type IV pilus, a mannosesensitive hemagglutinin type IV pilus system, and a symbiotic colonization and sigmadependent biofilm formation gene cluster. Several iron acquisition systems were annotated including hemin, enterobactin, vibrioferrin, ferrichrome, and TonB, including the full complement of proteins responsible for the formation of the TonB-ExbB-ExbD complex. Three quorum-sensing systems are present: LuxMN, LuxSPQ, and CqsAS. Six secretion systems were identified: one type I secretion system