Targeting PirAvp and PirBvp Toxins of Vibrio parahaemolyticus with Oilseed Peptides: An In Silico Approach

Acute hepatopancreatic necrosis disease (AHPND), caused by PirAvp- and PirBvp-releasing Vibrio parahaemolyticus strains, has resulted in massive mortality in shrimp aquaculture. Excessive use of antibiotics for AHPND management has led to antibiotic resistance, highlighting the urgency to search for alternatives. Using an in silico approach, we aimed to discover PirAvp/PirBvp-binding peptides from oilseed meals as alternatives to antibiotics. To search for peptides that remain intact in the shrimp digestive tract, and therefore would be available for toxin binding, we focused on peptides released from tryptic hydrolysis of 37 major proteins from seeds of hemp, pumpkin, rape, sesame, and sunflower. This yielded 809 peptides. Further screening led to 24 peptides predicted as being non-toxic to shrimp, fish, and humans, with thermal stability and low water solubility. Molecular docking on the 24 peptides revealed six dual-target peptides capable of binding to key regions responsible for complex formation on both PirAvp and PirBvp. The peptides (ISYVVQGMGISGR, LTFVVHGHALMGK, QSLGVPPQLGNACNLDNLDVLQPTETIK, ISTINSQTLPILSQLR, PQFLVGASSILR, and VQVVNHMGQK) are 1139–2977 Da in mass and 10–28 residues in length. Such peptides are potential candidates for the future development of peptide-based anti-AHPND agents which potentially mitigate V. parahaemolyticus pathogenesis by intercepting PirAvp/PirBvp complex formation.


Introduction
The shrimp is one of the most valuable aquaculture species consumed by people from many parts of the world. However, intensive shrimp culture and improper management have led to the outbreak of several diseases, such as acute hepatopancreatic necrosis disease (AHPND), hepatopancreatic microsporidiosis, infectious hypodermal and hematopoietic necrosis, and white spot disease [1][2][3]. AHPND, commonly known as early mortality syndrome, has caused huge loss in the whiteleg shrimp (Litopenaeus vannamei) industry. This bacterial shrimp disease has resulted in mass shrimp mortality and low shrimp production [3,4]. Administration of antibiotics in shrimp culture has been most widely used in the past in controlling the AHPND infection caused by Vibrio species [4]. Excessive use of antibiotics in shrimp culture has created antibiotic resistance in the pathogenic Vibrio. For instance, AHPND-causing Vibrio campbellii from China was found to carry multiple antibiotic resistance genes [5]. In Mexico, several Vibrio parahaemolyticus AHPND strains were reported to have the tetB gene coding tetracycline resistance [6]. The spread of antibiotic resistance genes among the pathogens may lead to the failure of AHPND control if the management of AHPND mainly depends on antibiotics. Moreover, the

In Silico Tryptic Hydrolysis of Oilseed Meal Proteins
To discover peptides that can potentially block the formation of the PirA vp /PirB vp toxin complex, we focused on peptides resulting from the in silico trypsin digestion of oilseed proteins. The in vitro trypsin digestion of feedstuffs is comparable to that of in vivo whiteleg shrimp digestion in terms of protein digestibility [25]. Thus, in silico tryptic digestion can be used to simulate the in vivo shrimp digestion of oilseed proteins and predict the peptide fragments that are released in the shrimp digestive tract. Peptides resulting from tryptic digestion also represent peptides that can be fed to the shrimp directly and remain undegraded in the shrimp digestive tract. Either way, such peptides could remain intact in the shrimp digestive tract, and would hence be available for neutralizing the toxicity of the pathogenic V. parahaemolyticus strains. In silico trypsin hydrolysis released 809 peptide fragments from 37 oilseed meal proteins (Table 1). Rape and hemp seed proteins released 272 and 247 peptide fragments, respectively, accounting for about a 4.5-fold greater than those from the sunflower proteins (Table 1). Regardless of oilseed type, in silico trypsin hydrolysis generated more peptide fragments from the globulin-type proteins than from the albumin-type. For instance, 93% of the hemp peptides resulting from tryptic hydrolysis was generated from the globulin-type proteins (edestin 1-3 and 7S vicilin-like protein) ( Table 1). A similar trend was observed in other studies of the in silico hydrolysis of oilseed proteins [26,27]. Overall, our results suggest that the globulin-type proteins of oilseeds are prominent sources of peptides that could remain intact in the shrimp digestive tract.  G  190  19  Cruciferin CRU1 alpha chain  P33525  G  296  22  Cruciferin CRU1 beta chain  P33525  G  190  17  Cruciferin CRU4 alpha chain  P33522  G  254  20  Cruciferin CRU4 beta chain  P33522  G  189  17  Cruciferin subunit alpha  P11090  G  275  24  Cruciferin subunit beta  P11090  G  190  16  Napin-1A small chain  P24565  A  31  5  Napin-1A large chain  P24565  A  79  8  Napin-2 small chain  P01090  A  36  6  Napin-2 large chain  P01090  A  81  9  Napin-3 small chain  P80208  A  37  6  Napin-3 large chain  P80208  A  88

Toxicity of Peptides
A bottleneck in the success of drugs used in aquaculture is their toxicity [7]. Peptides to be adopted as a potential anti-AHPND strategy should ideally be non-toxic to shrimp, other aquatic organisms, or human consumers. Therefore, by using the online admetSAR and ToxinPred tools, we screened the 809 peptides liberated from oilseed proteins by in silico hydrolysis for potential toxicity to crustaceans, fish, and humans. Remarkably, all the peptides were predicted as non-toxic to crustaceans (Table 2). Our results suggest that the five oilseed meals appear safe as protein ingredients in shrimp feed. This is in agreement with the report of good survival rate (>87%) in shrimp fed 25-100 g/kg rapeseed meal [28]. In other growth trials, the inclusion of 20-40% sesame meal and 1-10% sunflower cake to shrimp diets also led to good survival rates and health conditions [29,30]. Our results show that 25-35% of the peptides released from the five oilseeds were potentially toxic to fish ( Table 2). Our results concur with previous observations of adverse effects in fish fed oilseed-containing diets at high inclusion levels. For instance, fish fed a diet containing 25% sesame meal as a replacement for soybean meal displayed comparable growth performance and survival rate to the control group; but a diet containing 75-100% sesame meal reduced growth performance and survival [31]. Besides a decline in weight gain, the specific growth rate and survival rate of fish was observed when 25-100% of soybean meal was substituted with pumpkin seed meal in fish feed [32]. Our results indicate that although all of the oilseed-derived peptides we investigated in this study were likely non-toxic to shrimp, some of them may be toxic to fish. The drugs used in aquaculture may enter environment, where they contaminate wild seafood [7]. Hence, in this study, only the 580 peptides that were predicted as being non-toxic to fish were considered in our subsequent analysis when searching for potential anti-AHPND peptides. Meanwhile, our ToxinPred analysis revealed that 96-100% of tryptic peptides generated from the oilseed proteins were non-toxic to human. This consideration is important as residual drugs in treated aquatic organisms, including shrimp, can pose health issues to people who consume aquatic food products [7]. Altogether, our toxicity screening narrowed down the initial pool of 809 peptides to a set of 572 oilseed-derived peptides which are potentially non-toxic to shrimp, fish and humans. Such peptides, 463 derived from globulin-and 109 from albumin-type proteins of the five oilseeds, were promising peptides for subsequent search of anti-AHPND agents in this study.

Water Solubility and Thermal Stability of Peptides
In shrimp farming, feed ingredients with high water solubility may be rapidly leached and lost on exposure to water, thus reducing their availability for shrimp ingestion [33]. Thus, in our search for potential anti-AHPND peptides as in-feed medication for shrimp, we further screened the 572 non-toxic oilseed-derived peptides for water solubility. This revealed 59 low-water-solubility peptides (data not shown), each made of 25% to 100% hydrophobic amino acids (e.g., Ala, Val, and Phe). This is in agreement with the finding that the nonpolar portion of peptides accounts for their low solubility in aqueous solutions [34]. In both feed processing and aquafarming, highly water-soluble compounds are prone to leaching. For instance, water-soluble crystalline amino acids were reported to leach out rapidly from shrimp feed, compromising the nutritional quality of shrimp diets [33].
Another study also pointed out that the leaching of water-soluble protein hydrolysates or peptides from feed may weaken the responses of fish to diets supplemented with such protein hydrolysates or peptides [35]. To minimize the potential issue of leaching, only the 59 low-water-solubility oilseed peptides were selected for further analysis in our search for in-feed anti-AHPND peptides. The 59 peptides included some repetitive sequences. For example, two identical sequences (i.e., GVLYK) could be found among the non-toxic hemp-derived peptides. By excluding the repetitive sequences, the 59 low-water-solubility peptides were consolidated into a set of 49 peptides with unique sequences, as shown in Table 3.
Thermal treatment is an important part of aquaculture feed production [36]. Hence, anti-AHPND peptides to be incorporated into shrimp feed should be heat-stable. In this study, the thermal stability of the 49 low-water-solubility oilseed peptides were predicted based on their aliphatic index and compared with lantibiotic nisin-A ( Table 3). The aliphatic index, computed based on the percentage of aliphatic amino acids (Ala, Ile, Leu and Val), is correlated with thermal stability to proteins [37]. Based on their aliphatic indices, 24 of the 49 peptides were likely thermally more stable than nisin-A (Table 3). This is notable as nisin-A is a heat-stable peptide that retains its antimicrobial activity after thermal treatment at 121 • C for 15 min; it has also been used as food preservative for decades [38]. The thermal stability predicted for the 24 oilseed-derived peptides could be attributed, at least in part, to a relatively high percentage of aliphatic amino acids in their sequences. For instance, PI and PVV, the sesame-and hemp-derived peptides predicted as having the highest aliphatic indices, are composed of 50% and 67% of aliphatic amino acids, respectively (Table 3). Taken together, our analysis led to 24 oilseed-derived peptides with favorable properties namely non-toxicity, low water solubility, and high thermal stability. Table 3. Aliphatic indices of 49 oilseed-derived peptides predicted to be non-toxic and having low water solubility, in comparison with reference peptide lantibiotic nisin-A.

Molecular Docking Analysis
Interception of the complex formation between PirA vp and PirB vp was recommended as a promising anti-AHPND strategy [24]. To discover peptides that could preclude the formation of PirA vp /PirB vp complex, the 24 aforementioned peptides were each docked onto PirA vp and PirB vp . When docking onto PirA vp , 18 peptides produced similar or more negative docking scores (−144.315 to −194.881) than the six PirB vp regions (−142.536 to −174.899) which we used for comparison (Table 4). Hence, when compared with PirB vp , the 18 peptides potentially bound more stably to PirA vp . This implies the potential of the 18 peptides to intercept the binding between PirA vp -PirB vp . Intermolecular interactions between the 18 peptides and PirA vp were analyzed by using LigPlot+. The 18 peptides could each form interactions with 4-10 residues within the two PirB vp -binding regions of PirA vp (Tables 4 and S1). For example, sunflowerderived VIQNLPNQCDLEVQQCTTCTG could bind to one residue (Val23) in the 15-WTVEPNGGVTEVDSKHTPIIPEVGRS-40 region, and to five residues (Thr52, Gln54, Ser71, Gln75, and Arg76) in the 52-TIQYQWGAPFMAGGWKVAKSHVVQRDET-79 region of PirA vp . Overall, most of the predicted peptide-PirA vp binding involved the 52-TIQYQWGA PFMAGGWKVAKSHVVQRDET-79 region of PirA vp . This region was also responsible for most of the predicted binding between the six PirB vp regions and PirA vp (Supplementary Materials Table S1). Overall, hydrophobic interactions formed the majority of the peptide-PirA vp interactions. A similar trend was observed when the six PirB vp regions were docked onto PirA vp (Table 4). In particular, sesame-derived ISTINSQTLPIL-SQLR stood out among the 18 peptides due to its largest number of interactions with residues in 52-TIQYQWGAPFMAGGWKVAKSHVVQRDET-79, including Lys67 and Lys70 (Tables 4 and S1). The two Lys residues were found to be localized in the dimeric interface between PirA vp and PirB vp . The binding of peptides to the two Lys residues may block the cross-linking between PirA vp with PirB vp [24]. A graphical representation of the intermolecular interactions between ISTINSQTLPILSQLR and PirA vp is shown in Figure 1.   When the 24 non-toxic, low-water-solubility, and thermally-stable oilseed peptides were docked onto PirB vp , only six produced more negative docking scores, suggesting more stable binding, than the PirB vp -binding region 15-WTVEPNGGVTEVDSKHTPIIPEVGRS-40 of PirA vp . By contrast, all of the 24 peptides could not form a more stable binding to PirB vp when compared with the PirA vp region 52-TIQYQWGAPFMAGGWKVAKSHVVQRDET-79 (Table 5). A LigPlot+ analysis also revealed that the six best-scored peptides could bind to 2-4 residues within only two of the six PirA vp -binding regions of PirB vp , namely 386-FVVGENSGKPSVRLQL-401 and 426-YELFHPDEF-434. For instance, rape-derived QSLGVPPQLGNACNLDNLDVLQPTETIK, predicted to form the largest number of interactions with PirB vp , could bind to Val397, Arg398, and Gln400 in the 386-FVVGENSGKPSVRLQL-401 region, and could bind to Phe429 in the 426-YELFHPDEF-434 region of PirB vp . Figure 2 shows the intermolecular interactions contributing to QSLGVPPQLGNACNLDNLD VLQPTETIK-PirB vp binding. Similarly, limited interactions between the two PirA vp regions 15-WTVEPNGGVTEVDSKHTPIIPEVGRS-40 and 52-TIQYQWGAPFMAGGWKVAKSHV VQRDET-79 with PirB vp were also observed (Tables 5 and S2). Interestingly, the six oilseed peptides having predicted interactions with PirB vp (Table 5) were also among the 18 bestscoring PirA vp -binding peptides presented in Table 4. The six peptides ranged between 1139 Da and 2977 Da in mass, and 10-28 residues in length, mostly having a net charge of +1. Modeling with PEP-FOLD 3.5 revealed the 3D structures of all the peptides except the rape-derived ISYVVQGMGISGR to be comprised of helical elements (Figure 3). The six peptides were derived from globulin-type proteins from rape, sesame, and hemp. Our findings suggest that the globulin-type proteins of these three seeds are potential sources of peptides that can bind to both PirA vp and PirB vp . As observed in oilseed peptide-PirA vp interactions (Table 4), hydrophobic interactions contributed to the majority of the interactions between the six peptides and PirB vp (Table 5). Taken together, our results suggest that hydrophobic interactions are critical in facilitating stable binding between the six oilseed peptides and the proteins of both PirA vp and PirB vp . When the 24 non-toxic, low-water-solubility, and thermally-stable oilseed peptides were docked onto PirB vp , only six produced more negative docking scores, suggesting more stable binding, than the PirB vp -binding region 15-WTVEPNGGVTEVDSKHTPIIPEVGRS-40 of PirA vp . By contrast, all of the 24 peptides could not form a more stable binding to PirB vp when compared with the PirA vp region 52-TIQYQWGAPFMAGGWKVAKSHVVQRDET-79 (Table 5). A LigPlot+ analysis also revealed that the six best-scored peptides could bind to 2-4 residues within only two of the six PirA vp -binding regions of PirB vp , namely 386-FVVGENSGKPSVRLQL-401 and 426-YELFHPDEF-434. For instance, rape-derived QSLGVPPQLGNACNLDNLDVLQPTETIK, predicted to form the largest number of interactions with PirB vp , could bind to Val397, Arg398, and Gln400 in the 386-FVVGENSGKPSVRLQL-401 region, and could bind to Phe429 in the 426-YELFHPDEF-434 region of PirB vp . Figure 2 (Tables 5 and S2). Interestingly, the six   WTVEPNGGVTEVDSKHTPIIPEVGRS-40) that were previously reported to bind to PirB vp to form a toxic complex were analyzed for comparison.   The virulence factors of AHPND have been shown to be both PirA vp and PirB vp [9,39]. Hence, targeting only one of the proteins may not result in full protection against AHPND. For example, the binding of the chicken egg yolk immunoglobulin (IgY) specific for a recombinant PirB vp -like protein (anti-PirB vp -IgY) to the PirB vp toxin still allowed AHPND infection to progress, resulting in a 14% survival rate of AHPND-challenged shrimp [40]. On the other hand, the binding of the IgY specific for the recombinant PirA vp -like protein (anti-PirA vp -IgY) to PirA vp gave rise to only an 86% survival rate in AHPND-challenged shrimp [40]. Therefore, we speculate that dual-target oilseed peptides that can bind to The virulence factors of AHPND have been shown to be both PirA vp and PirB vp [9,39]. Hence, targeting only one of the proteins may not result in full protection against AHPND. For example, the binding of the chicken egg yolk immunoglobulin (IgY) specific for a recombinant PirB vp -like protein (anti-PirB vp -IgY) to the PirB vp toxin still allowed AHPND infection to progress, resulting in a 14% survival rate of AHPND-challenged shrimp [40]. On the other hand, the binding of the IgY specific for the recombinant PirA vp -like protein (anti-PirA vp -IgY) to PirA vp gave rise to only an 86% survival rate in AHPND-challenged shrimp [40]. Therefore, we speculate that dual-target oilseed peptides that can bind to both PirA vp and PirB vp may be a relatively promising approach for intercepting PirA vp /PirB vp complex formation, thus mitigating AHPND occurrence in shrimp. Based on our results, the six oilseed peptides that can bind to both PirA vp and PirB vp (ISYVVQGMGISGR, LTFVVHGHALMGK, QSLGVPPQLGNACNLDNLDVLQPTETIK, ISTINSQTLPILSQLR, PQFLVGASSILR, and VQVVNHMGQK) deserve attention in future research. In this theo-retical investigation, we considered peptides undegraded by trypsin to be bioavailable in the shrimp digestive tract. Considering the complexity of factors governing bioavailability, investigation into the in vivo availability of the six aforementioned peptides should be the first priority in future research. Next, testing of the in vivo effectiveness of these peptides individually and in combination in live shrimp, delivered as ingredients incorporated into shrimp feed, would be of great interest. Besides validating the properties predicted in this study (absence of toxicity, low water solubility, and thermal stability) and confirming the anti-AHPND potency of the peptides, such studies may also reveal whether the peptides could function synergistically. An interesting recent development is the novel discovery by Almanza-Martínez et al. [41] of an α-amylase-like protein serving as a binding receptor to PirB vp in the digestive tract tissue of the whiteleg shrimp. The specific role of the α-amylase-like protein in the mechanism of action of PirB vp remains to be established, as are the residues participating in the α-amylase-PirB vp binding interface. Once such information has been elucidated, it would be interesting to investigate, through molecular docking and/or in vitro experiments, whether the oilseed peptides reported in this study can also intercept the α-amylase-PirB vp interaction. Meanwhile, the potential off-target interactions of the six selected oilseed peptides should also be addressed to ensure that they do not disrupt other normal biological processes, causing adverse side-effects in the shrimp.

Water Solubility and Thermal Stability of Peptides
Water solubility of peptides were analyzed using an online peptide property calculator (https://pepcalc.com/). Aliphatic index, an indicator of thermal stability, was predicted by using the Peptides Package in R (https://rdrr.io/snippets/) [37]. The two online tools were accessed between 26 to 31 August 2021.

Conclusions
Our in silico study found the globulin-type proteins of the oilseeds investigated to be promising sources of PirA vp -and PirB vpbinding peptides. Upon selection for non-toxicity against crustaceans, fish, and humans, as well as thermal stability and low water solubility, six peptides that could bind to both PirA vp and PirB vp with stability comparable or superior to the binding between the key regions in the native PirA vp /PirB vp complex, were identified. These six peptides could potentially play a role in intercepting the formation of the toxic PirA vp /PirB vp complex, thus serving as promising anti-AHPND agents. The findings of this study can be used to guide the future discovery and development of anti-AHPND peptides from oilseed proteins.

Supplementary Materials:
The following are available online at https://www.mdpi.com/article/ 10.3390/antibiotics10101211/s1, Table S1: Number of residues in the two PirB vp -binding regions of PirA vp protein that are involved in interactions with 18 chosen oilseed peptides, Table S2

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.