Massive Gene Expansion and Sequence Diversification Is Associated with Diverse Tissue Distribution, Regulation and Antimicrobial Properties of Anti-Lipopolysaccharide Factors in Shrimp

Anti-lipopolysaccharide factors (ALFs) are antimicrobial peptides with a central β-hairpin structure able to bind to microbial components. Mining sequence databases for ALFs allowed us to show the remarkable diversity of ALF sequences in shrimp. We found at least seven members of the ALF family (Groups A to G), including two novel Groups (F and G), all of which are encoded by different loci with conserved gene organization. Phylogenetic analyses revealed that gene expansion and subsequent diversification of the ALF family occurred in crustaceans before shrimp speciation occurred. The transcriptional profile of ALFs was compared in terms of tissue distribution, response to two pathogens and during shrimp development in Litopenaeus vannamei, the most cultivated species. ALFs were found to be constitutively expressed in hemocytes and to respond differently to tissue damage. While synthetic β-hairpins of Groups E and G displayed both antibacterial and antifungal activities, no activity was recorded for Group F β-hairpins. Altogether, our results showed that ALFs form a family of shrimp AMPs that has been the subject of intense diversification. The different genes differ in terms of tissue expression, regulation and function. These data strongly suggest that multiple selection pressures have led to functional diversification of ALFs in shrimp.


Introduction
Anti-lipopolysaccharide factors (ALFs) are multifunctional antimicrobial host defense peptides (AMPs) with the ability to bind to microbial surface molecules. They were initially characterized as ALFs clustered into seven distinct groups with specific amino acid sequence signatures ( Figure 1A). In addition to already documented ALFs from Groups A to E, we identified here two novel groups that were conveniently named Group F and G ( Figure 1A).
Mar. Drugs 2018, 16, x 3 of 17 described ALF groups (Groups A to E). Surprisingly, from our sequence analysis, shrimp ALFs clustered into seven distinct groups with specific amino acid sequence signatures ( Figure 1A). In addition to already documented ALFs from Groups A to E, we identified here two novel groups that were conveniently named Group F and G ( Figure 1A).  Across shrimp species, ALFs corresponded to full-length transcripts that encode for precursors composed of a signal peptide (22 to 28 residues), followed by a mature peptide (10.74 to 12.23 kDa) containing two conserved cysteine residues ( Figure 1A). Besides their differences in size and molecular weight, shrimp ALFs also displayed contrasting electrostatic characteristics. Notably, while ALFs from Groups B, C and F showed cationic properties, Groups A, D, E and G were composed of anionic ALFs ( Figure 1B). However, independently of their differences in primary structure and biochemical properties, the seven ALFs shared a similar three-dimensional architecture: three α-helices packed against a four-stranded β-sheet ( Figure 1C). Members of the seven groups were identified in at least four different shrimp species (Farfantepenaeus aztecus, L. vannamei, M. japonicus and P. monodon) (Table S1). Remarkably, Group G ALFs were the only members that were not identified in the genus Fenneropenaeus (F chinensis, F. indicus and F. penicillatus). On the other hand, two different members from Group C were identified in F. chinensis (FcALF2: JX853775 and FcALF3: JX853776), M. japonicus (MjALF-C1: AB210110 and MjALF-C2: KU160498) and P. monodon (ALFPm6: JN562340 and ALFPm7: KX431031).
Our knowledge of ALF intraspecific sequence diversity was also enriched by the discovery of novel sequences in F. aztecus (Groups A to G), F. penicillatus (Groups A to F), L. vannamei (Groups E to G) and P. monodon (Groups D to G). Besides, according to our in silico analyses, some ALFs from M. japonicus were classified in a different Group to that previously categorized by Jiang and colleagues [11]. For instance, the sequence MjALF-A2 [11] is actually a member of Group G and not an ALF from Group A, whereas the cationic MjALF-D1 (GenBank: KU160499) belongs to Group F and not to Group D (which gathers anionic sequences only). More surprisingly, the sequence MjALF-E1 (GenBank: KY627760), previously classified as a cationic member of Group E, did not fit in any ALF Group. Indeed, its mature sequence contains an additional cysteine residue (apart of the two cysteines holding the central β-hairpin structure) that is not found in ALFs from either marine chelicerates or crustaceans. Interestingly, coding sequences related to MjALF-E1 were also found in P. monodon and L. vannamei. Unlike the three-cysteine-containing sequences from M. japonicus and P. monodon, the sequences identified in L. vannamei contain four cysteines ( Figure S1).

ALF Sequence Diversity Is Gene-Encoded
To gain insights into the origin of the molecular diversity of the shrimp ALF family, we searched for ALF gene sequences in both annotated (GenBank Nucleotide) and non-annotated (Whole-Genome Shotgun Contigs) databases. From our in silico mining analysis, seven unique genomic sequences were identified in P. monodon: ALFPm2 from Group A (GenBank: EF523561), ALFPm3 from Group B (GenBank: EF523562), ALFPm6 from Group C (GenBank: JN562340), ALFPm8 from Group D (GenBank: NIUS010164210), ALFPm9 from Group E (GenBank: NIUS010076396), ALFPm10 from Group F (GenBank: NIUS011801312) and ALFPm11 from Group G (GenBank: NIUS010749450). Each genomic sequence corresponded to a specific ALF member and we found no evidences that ALFs from different Groups could be encoded by a same genomic sequence.
Despite their differences in terms of sequence signatures, all genes shared a similar structural organization: three exons interrupted by two introns ( Figure 1D). Every sequence presents a second exon that encodes the four stranded β-sheets, with the two cysteines delimiting the central β-hairpin. This structure holds the seven charged residues involved in LPS binding and is considered as the functional domain of ALFs. As shown in Figure 1D, not all P. monodon ALFs contain those conserved residues found in ALFPm3 from Group B [6]. Regarding the other gene regions, the first exon covers the 5 -untranslated region (UTR), the leader sequence and the hydrophobic N-terminal portion of the mature peptide (the first α-helice), and the third exon encodes the two C-terminal α-helices of the mature peptide and the 3 -UTR.

ALFs Evolved from Gene Duplication Events before Shrimp Speciation
In order to unravel the phylogenetic relationships of the shrimp ALFs, phylogenetic reconstructions were performed with ALF sequences from 38 species of decapod crustaceans (suborders Dendrobranchiata and Pleocyemata) and three species of marine chelicerates (the horseshoe crabs Carcinoscorpius rotundicauda, L. polyphemus and T. tridentatus) (Table S1). Additionally, we analyzed the ALF-related sequences containing three and four cysteine residues and scygonadins (anionic AMPs from crabs that contain two cysteines flanking 17 amino acid residues [12]). Our Bayesian phylogenetic analysis revealed that ALFs comprise a large and diverse gene family in decapod crustaceans. The first striking piece of information is that the three/four-cysteine-containing peptides (including MjALF-E1), as well as scygonadins, are not authentic members of the ALF family since they form a separate and distant clade from all other sequences ( Figure 2A). Indeed, the ALF clade gathered sequences from both crustaceans and marine chelicerates. Regarding the crustacean group, ALFs were split into two main clades ( Figure 2A). The first clade included ALFs from Group A, while the second clade gathered ALFs from six additional groups (B to G). Interestingly, sequences from non-penaeid species (Pleocyemata) were found in all shrimp ALF groups, but they also formed exclusive groups distinct from those found in penaeids (Figure 2A). Cationic and anionic ALF groups from penaeid shrimp are displayed in blue and red, respectively. Posterior probabilities (Bayesian) and bootstrap values (neighbor-joining) higher than 50% are shown in the nodes. The list of the ALF sequences included in analyses (annotations, sequences and GenBank accession numbers) is provided in Table S1.
Then, an additional phylogenetic tree was constructed to determine the phylogenetic relationships among the seven shrimp ALF groups (A to G) ( Figure 2B). In this tree, shrimp ALFs clustered into two main clades: a first clade containing ALFs from Group A and a second clade divided into three branches. Within the second clade, all cationic shrimp ALFs (Groups B, C and F) clustered into one branch, and anionic ALFs from Groups E and G clustered into a second branch ( Figure 2B). Group D ALFs clustered in a third branch ( Figure 2B). Altogether, our results suggest that the sequence diversity found in the ALF family was likely driven by gene duplication events before the divergence of decapod crustaceans (Dendrobranchiata and Pleocyemata). Notably, the gene expansion and subsequent diversification of the ALF family seems to have occurred in crustaceans and not in marine chelicerates. Actually, in marine chelicerates, only one ALF type was identified (Table S1).

ALFs Are All Expressed in Individual Shrimps and Differentially Modulated in Response to Tissue Damage
We further focused on the transcriptional profiles of shrimp ALFs in terms of tissue distribution. For gene expression analyses, we considered the seven ALFs from the Pacific white shrimp L. vannamei (Litvan ALF-A to -G). First, the gene expression distribution of Litvan ALFs was assessed in eight different tissues of healthy juveniles by semiquantitative RT-PCR analysis. Overall, Litvan ALFs were mainly detected in circulating hemocytes and gills ( Figure 3A). Transcripts of Litvan ALF-A and Litvan ALF-B were detected in foregut, midgut, hemocytes, gills and nerve cord, while the expression of Litvan ALF-C was observed in midgut, hemocytes, and gills ( Figure 3A). Besides, while the expression of Litvan ALF-E and Litvan ALF-G was exclusively detected in hemocytes and gills, Litvan ALF-F was mainly expressed in the foregut ( Figure 3A). Unlike the other ALF groups, the expression of Litvan ALF-D was only detected in hemocytes ( Figure 3A). For all genes, no signals were observed in hepatopancreas, hindgut and muscle ( Figure 3A).
ALF gene expression was then studied in response to infections and wounding. We first asked whether ALF genes were all transcribed in a single animal or whether their diversity reflected inter-individual sequence variability. Transcripts of the seven ALF genes were detected in the circulating hemocytes of every individual shrimp, as determined by RT-qPCR ( Figure 3B). However, important variation was observed in the basal transcription of each gene among individuals ( Figure 3B). While the basal gene expression of Litvan ALF-A to Litvan ALF-F varied from 2-to 6-fold among individuals, variations up to 11.3-fold were found for Litvan ALF-G gene expression ( Figure 3B).
Next, we analyzed the gene expression profile of Litvan ALFs in response to microbial challenge and injury. Two unrelated shrimp pathogens were chosen: the Gram-negative V. harveyi and the White spot syndrome virus (WSSV). The transcriptional response of Litvan ALFs was quantified by RT-qPCR in shrimp hemocytes 48 h after infections. This time point was chosen on the basis of previous studies from our group [13][14][15]. Anionic ALFs from Groups A, D and E did not respond to pathogens nor to injury ( Figure 3C). Conversely, cationic ALFs (Groups B, C and F) and the anionic Group G ALF showed significant changes in expression only in response to tissue injury. Indeed, the expression of Litvan ALF-B (2.6-fold), Litvan ALF-C (18.7-fold), Litvan ALF-F (3.6-fold) and Litvan ALF-G (8.3-fold) was significantly induced in circulating hemocytes after the injection of a tissue homogenate prepared from shrimp muscle ( Figure 3C). Similarly, the expression of Litvan ALF-B (2.7-fold) and Litvan ALF-F (4.1-fold) also increased after the injection of sterile seawater. The pathogens (V. harveyi and WSSV) did not modulate further ALF expression. Notably, independently of the experimental condition, a high variability in gene expression was observed for all ALFs ( Figure 3C). Results are presented as mean ± standard deviation of relative expressions (three biological replicates) and statistical differences are indicated by asterisks (*) (one-way ANOVA/Tukey, p < 0.05). N: naïve (non-stimulated) shrimp (white bars), S: sterile seawater injury control, V: V. harveyi ATCC 14126 (6 × 10 7 CFU/animal), W−: tissue homogenate inoculum prepared from WSSV-free shrimp, W+: WSSV (3 × 10 2 viral particles/animal).

Some ALF Genes Are Transcribed Early in Shrimp Development, while Others Are Mainly Expressed in Juveniles
Finally, we studied the expression of the three new L. vannamei ALFs (Groups E, F and G) at different stages of shrimp development: fertilized eggs, nauplii, protozoeae, mysis, postlarvae and juveniles. The expression profile of Litvan ALFs from Groups A to D was previously reported [16]. Three distinct patterns of expression were observed for ALF groups E to G over L. vannamei development.
ALFs from Groups E and F were detected at all developmental stages, but Group F expression could only be quantified from nauplius stages ( Figure 4). Group E expression did not vary significantly over the entire shrimp development, from larvae to juveniles. In contrast, Group F expression was maximum in protozoea III (ZIII) and then decreased significantly in juveniles (PL17) (Figure 4). Finally, Group G ALF was only expressed from protozoea III (ZIII), and its expression increased significantly up to juvenile stages (PL17) (Figure 4). Results are present as mean ± standard deviation. The red dotted line indicates the expression in hemocytes from juveniles while the solid blue underline highlights the stages at which the expression was detected (valid dissociation curve) but not quantified (Cq values higher than the limit of quantification). Different letters indicate significant differences among the developmental stages and asterisks (*) shows significant differences between each developmental stage and hemocytes from juveniles (one-way ANOVA/Tukey, p < 0.05).

Sequence Diversity of Shrimp ALFs Results in Distinct Antimicrobial Properties
The functional domain (central β-hairpin) of the three novel ALF members identified in L. vannamei (Groups E, F and G) was generated by chemical synthesis to evaluate their antimicrobial properties. Indeed, this functional domain is considered a good proxy of the full-length ALF antimicrobial properties [9,10,17]. Minimal inhibitory concentration assays were performed against Gram-positive and Gram-negative bacteria and fungi (yeast and filamentous) ( Figure 5). From the three synthetic peptides, Litvan ALF-G 34-55 displayed the broadest range of antimicrobial activity, being effective against all tested Gram-positive bacteria, the Gram-negative V. nigripulchritudo and the filamentous fungus F. oxysporum ( Figure 5). This peptide could also affect the growth of the Gram-negative bacteria E. coli and V. harveyi at 40 µM (data not shown), but total inhibition was only observed against V. nigripulchritudo. Additionally, Litvan ALF-G 34-55 exhibited bactericidal activity against the Gram-positive bacteria B. cereus, B. stationis and M. maritypicum. On the other hand, synthetic β-hairpins of Litvan ALF-E 32-53 could inhibit only the growth of marine Gram-positive bacteria (B. stationis and M. maritypicum) and F. oxysporum ( Figure 5). Notably, no antimicrobial activity was observed for Litvan ALF-F 30-51 β-hairpin even at 40 µM. None of the synthetic peptides was able to inhibit the growth of the Gram-negative bacteria A. salmonicida, P. aeruginosa, V. alginolyticus and V. anguillarum, and of the yeast C. albicans. Thus, according to their synthetic β-hairpin, Group G and to a lower extent Group E ALFs show a broad spectrum of antimicrobial activities, whereas Group F is devoid of antifungal and antibacterial activity. However, in agreement with a very poor conservation of residues involved in LPS binding ( Figure 1D), ALFs from Groups E-G were almost inactive against Gram-negative bacteria ( Figure 5C).  Figure 5). Notably, no antimicrobial activity was observed for Litvan ALF-F30-51 β-hairpin even at 40 µM. None of the synthetic peptides was able to inhibit the growth of the Gram-negative bacteria A. salmonicida, P. aeruginosa, V. alginolyticus and V. anguillarum, and of the yeast C. albicans. Thus, according to their synthetic βhairpin, Group G and to a lower extent Group E ALFs show a broad spectrum of antimicrobial activities, whereas Group F is devoid of antifungal and antibacterial activity. However, in agreement with a very poor conservation of residues involved in LPS binding ( Figure 1D), ALFs from Groups E-G were almost inactive against Gram-negative bacteria ( Figure 5C).

Discussion
We showed here that shrimp ALFs are composed of seven distinct members (Groups A to G) with contrasting biochemical properties, activities and expression patterns. Particularly, ALF sequences were found to vary from cationic to anionic with important consequences on their antimicrobial activities. Overall, ALFs comprise the most diverse AMPs found in penaeid shrimp. Indeed, such diversity has not been observed in any other gene-encoded AMP families from shrimp, which are exclusively composed of cationic (penaeidins and crustins) or anionic (stylicins) members [8,12]. ALF diversity is encoded by at least seven genes that arose from successive duplications and subsequent mutations (nucleotide substitutions and insertion/deletion events) before decapod crustacean speciation occurred. This indicates that strong evolutionary pressures have driven the functional diversification of ALF genes, giving rise to neo-or sub-functionalization and retention in the shrimp genome.
We found that shrimp ALFs are paralogous genes that evolved before the speciation of the suborder Dendrobranchiata (penaeid shrimp). Indeed, ALF diversity, which is the subject of the present study, goes beyond penaeid shrimp and extends to other decapod species from the suborder Pleocyemata (including crayfish, crabs, lobsters, freshwater prawns, etc.). Some ALF members from non-penaeid decapods fall into the seven groups characterized here for penaeid shrimp. However, the ALF diversity found in the suborder Pleocyemata is different from penaeid shrimp (Dendrobranchiata). It is likely that the remarkable gene expansion and diversification of ALF sequences through gene duplication and subsequent mutation have fueled adaptation to different lifestyles and environments (and their associated pathogens) among crustaceans. Here, we have focused our study on shrimp ALFs, as a good sub-representative of ALF diversity. In order to support our hypothesis, we showed that the biological activities and expression patterns of ALF genes have diverged. In particular, we found that some ALFs are antimicrobial, whereas others are not. Some are expressed early during shrimp development, whereas others are expressed in late developmental stages. Finally, ALFs differ in their tissue distribution and responses to tissue damage. However, much more biological data are still needed on the expression and functions of the different ALF members to understand ALF evolution. This ambitious objective will require the development and use of emerging gene-silencing technologies (such as CRISPR-Cas9 and RNA interference) to achieve specific invalidation of closely-related genes in crustaceans and further phenotyping. Similarly, molecular tools such as in situ hybridization could reveal the tissue specificity of ALFs and thus, they would help in uncovering other biological functions. Finally, a classification of all crustacean ALFs (from both decapod and non-decapod species) based on robust phylogenetic reconstructions may avoid misleading classifications and lead to consensus among researchers.
With the identification of two novel ALF groups, we found that ALFs from Groups E and G share a common ancestor gene. Interestingly, Group G is lacking in species of the genus Fenneropenaeus whereas it is found in species of the genera Farfantepenaeus, Litopenaeus, Marsupenaeus and Penaeus. Although it cannot be ruled out that data are missing from publicly accessible databases, the absence of Group G in Fenneropenaeus could result from a gene loss event within this genus. Alternatively, the duplication event that originated these two genes may not have occurred in the genus Fenneropenaeus. Indeed, the evolutionary history of each group traced a particular trajectory in each shrimp species. For instance, while Group C ALFs from F. chinensis (FcALF2 and FcALF3 [17]), M. japonicus (MjALF-C1 and MjALF-C2 [11]) and P. monodon (ALFPm6 and ALFPm7 [18,19]) are composed of two members, in other penaeids this group appears to be composed by a single gene. However, we do not favor this last hypothesis as Group G is found in a diversity of penaeid species. Interestingly, in L. vannamei, Group G ALF was shown here to (i) have broader and more potent antimicrobial activity than Group E ALF, according to their β-hairpin activity, and (ii) to be expressed at late developmental stages whereas Group E expression tends to decrease over ontogenesis. Therefore, it is tempting to speculate that Group E confers antimicrobial protection at larval stages when Group G is still not expressed, whereas Group G provides a selective advantage to the Litopenaeus genus in facing infections at juvenile and adult stages when they are more exposed to different environmental challenges.
We showed that the expression of the seven ALF genes is simultaneous in the circulating hemocytes of a single shrimp. This result is particularly interesting because it suggests that the different ALF members may act synergistically to improve their antimicrobial properties. However, it is still unknown whether they are produced by the same hemocyte populations. Comparatively, the different penaeidin members of L. vannamei (Litvan PEN1/2, -3 and -4) are constitutively expressed by the granular cell populations [20]. Although the expression of ALFs has been detected in hemocytes from juveniles, some members (Groups C, E and F) appeared to be transcribed in larval stages of shrimp development that precede the emergence of these immune cells [21]. Instead, the expression of ALFs from Groups A, B, D and G was quite similar to that observed for other shrimp AMPs that are exclusively produced by hemocytes [16]. On the one hand, the expression of ALFs early in development could be the result of maternal transmission [16,22] but, on the other hand, those transcripts might originate from other shrimp tissues. Interestingly, in different species, including L. vannamei, the expression of ALFs from Groups C and F was mainly detected in other tissues (digestive system, gills, eyestalk) than in the circulating hemocytes [17][18][19]. However, only the expression of ALFs from Group B (ALFPm3 from P. monodon) was studied by immune staining [23]. More knowledge about the precise sites of ALF production will contribute to understand the involvement of these AMPs in shrimp epithelial defenses, especially those occurring in gills and intestines [15].
One important finding from this study concerns the differential gene expression pattern of shrimp ALFs in response to various challenges. Indeed, the different ALF genes found across penaeids showed to be responsive to various shrimp pathogens, from viruses to bacteria and filamentous fungi [11,14,15,18]. Moreover, RNA interference (RNAi)-mediated gene silencing assays have confirmed that ALFs are directly involved in shrimp survival to infectious diseases [18,24,25]. Additionally, our results provided new evidences for the role of ALFs in other biological processes. Interestingly, while ALFs from Groups A, D and E were not regulated, the expression of the other ALF genes was induced in response to tissue damage. Particularly, ALFs from Groups C and G were shown to be responsive to a tissue homogenate prepared from shrimp muscle (injury control for the WSSV infection), suggesting that they can be modulated by danger/damage-associated molecular patterns (DAMPs). This nonspecific transcriptional response could be associated with additional biological roles involving the promotion of wound healing and the rapid regeneration of tissues [20]. Additionally, we showed that some ALFs are modulated in the shrimp gut in response to infections, suggesting that ALFs can act as a first line of defense in tissues continuously exposed to microbe-rich environments [15,19]. Therefore, it is possible that those ALF variants have evolved novel functions associated with the control of the intestinal microbiota. The shrimp intestinal microbiota is a complex and dynamic community that is directly influenced by both biotic and abiotic factors [13], but probably it is also by the constitutive expression of immune-related genes. In fact, RNAi experiments revealed that ALFs from Groups B [18] and C [25] play an essential role in the control of the bacterial communities residing in the hemolymph. However, more functional genomic studies are required to understand the role of ALF in shrimp intestinal defenses.
Another relevant conclusion taken is that the antimicrobial activity of the functional domain of ALFs (central β-hairpin) is associated with its primary sequence rather than to its charge. Despite their differences in primary structure and biochemical features, the seven ALF groups shared a similar tertiary structure. However, the residues involved in LPS binding are not conserved among the seven groups, confirming the neo-functionalization hypothesis proposed by Rosa and colleagues [10]. Indeed, LPS binding has been demonstrated for the limulus ALF sequence, which shares a common ancestor with all shrimp ALFs. Taking into account previous studies [9][10][11]17] and the present results, shrimp ALFs have proved to display a diverse spectrum of antimicrobial activity. While some members exhibited a broad range of antimicrobial activity (Groups B and G), some others displayed limited (Groups A, C and E) or very weak action (Groups D and F). One possible explanation is that the effectiveness of the antimicrobial activity of each ALF group is directly proportional to the amount of positively charged amino acids in its central β-hairpin structure [26]. However, we showed that the highly cationic central β-hairpin structure of Litvan ALF-F 19-54 (pI = 9.24) was not active against the microorganisms tested in this study. Likewise, synthetic β-hairpins of the FcALF1 (Group F) from F. chinensis was also poorly active against both Gram-positive and Gram-negative bacteria [17]. Thus, besides their overall net charge, other features may interfere directly on their biological activities. Given these results, the determination of the amino acid residues involved in the interaction with other microbial surface molecules (peptidoglycan, lipoteichoic acid, β-glucans, etc.) may provide valuable information of the mechanism of action of ALFs against other microorganisms beyond Gram-negative bacteria [7].

Database Searches and Phylogenetic Reconstructions
ALF sequences were methodically collected from publicly accessible databases and used for the search of homologous sequences in both annotated and non-annotated databases. Only fulllength coding sequences were considered. Homology searches were performed using tBLASTx at NCBI. Exon-intron boundaries were defined by alignment of the cDNA and genomic sequences. All nucleotide sequences were manually inspected and analyzed using open-access bioinformatics tools. Three-dimensional models for L. vannamei ALFs were built with SWISS-MODEL (https: //swissmodel.expasy.org/) using ALFPm3 NMR resolution (PDB: 2JOB1) as a template. Deduced amino acid sequences were aligned using MAFFT multiple alignment program (https://mafft.cbrc. jp/alignment/server/). Bayesian phylogenetic analysis was conducted in MrBayes 3.1.2 (http: //mrbayes.sourceforge.net/), using WAG + G as substitution model, with two runs of 10 7 generations, sample rate of 1000 and burn-in of 25%. Neighbor-joining analysis was conducted in MEGA X [27]. Bootstrap sampling was reiterated 1000 times using a 50% bootstrap cutoff. Trees were drawn using FigTree v1.4.2 (http://tree.bio.ed.ac.uk/software/figtree/).

Animals and Tissue Collection
Litopenaeus vannamei juveniles (10 ± 2 g) and at different development stages were obtained from the Laboratory of Marine Shrimp (Federal University of Santa Catarina, Florianópolis, Brazil). Each developmental stage was identified microscopically and collected as previously described [16] while juveniles were acclimated in controlled conditions for at least one week before any experimentation. Hemolymph was collected from the ventral sinus into a precooled modified Alsever solution (27 mM sodium citrate, 336 mM NaCl, 115 mM glucose, 9 mM EDTA, pH 7.0) and hemocytes were isolated by centrifugation. After hemolymph collection, the following tissues were harvested by dissection: foregut, hepatopancreas, midgut, hindgut, muscle, gills and nerve cord. Tissues were rinsed in Tris-saline solution (10 mM Tris, 330 mM NaCl, pH 7.4), homogenized in TRIzol reagent (Thermo Scientific, Asheville, NC, USA) and processed for semiquantitative RT-PCR analysis.

Experimental Infections
Two unrelated shrimp pathogens were chosen for experimental infections, the Gram-negative Vibrio harveyi and the White spot syndrome virus (WSSV). For the bacterial infection, 6 × 10 7 CFU/animal of V. harveyi ATCC 14126 under 100 µL sterile seawater (SSW) or 100 µL SSW (injury control) were injected. For the viral infection, shrimp were injected with 100 µL of a WSSV inoculum containing 3 × 10 2 viral particles. The WSSV inoculum was prepared from muscle tissues of WSSV-infected shrimp as previously described [14]. Animals injected with 100 µL of a tissue homogenate prepared from WSSV-free shrimp were used as injury control for the viral infection. At 48 h post-infections, hemocytes were collected, pooled (three pools of five animals per condition) and processed for gene expression analysis. Naïve (non-stimulated) animals were used as a control for all experimental conditions.

Semiquantitative RT-PCR Analysis for Tissue Distribution of Gene Expression
Total RNA was extracted using TRIzol reagent (Thermo Scientific, Asheville, NC, USA) according to the manufacturer's protocol. RNA samples were treated with DNase I (Thermo Scientific) at 37 • C for 15 min and precipitated with 0.3 M sodium acetate (pH 5.2) and isopropanol (1:1; v:v). RNA amount and quality were assessed by spectrophotometric analysis and the integrity of total RNA was analyzed by 0.8% agarose gel electrophoresis. First strand cDNA was synthesized from 1 µg of total RNA using the RevertAid Reverse Transcription kit (Thermo Scientific, Asheville, NC, USA) and oligo(dT) [12][13][14][15][16][17][18] primers. PCR reactions were carried out in a 15-µL reaction volume containing 1 µL cDNA, 2 mM MgCl 2 , 0.4 mM dNTP Mix, 0.4 µM of each primer (Table 1) and 1 U Taq DNA Polymerase (Sinapse, São Paulo, SP, Brazil). PCR conditions were as follows: 1 cycle of denaturation at 95 • C for 10 min followed by 30-35 cycles of 95 • C for 30 s, 60 • C for 30 s and 72 • C for 30 s. PCR products were analyzed by electrophoresis (1.5% agarose gel) and stained by ethidium bromide. The expression of the LvActin gene was used as endogenous control.

Fluorescence-Based Reverse Transcription Real-Time Quantitative PCR (RT-qPCR)
RT-qPCR amplifications were performed in a final volume of 15 µL containing 0.3 µM of each primer (Table 1), 7.5 µL of reaction mix (Maxima SYBR Green/ROX qPCR Master Mix 2×; Thermo Scientific, Asheville, NC, USA) and 1 µL of cDNA. The RT-qPCR program was 95 • C for 10 min, followed by 40 cycles of 95 • C for 15 s and 60 • C for 1 min. Melt curve analysis was performed to evaluate primer specificity. The eukaryotic translation elongation factor 1-alpha (LvEF1α) and the ribosomal protein LvL40 were used as reference genes of expression data in hemocytes. Relative transcript levels were determined by the comparative standard curve method using a standard curve derived from 2-fold dilution series of a cDNA pool of all samples. Differences we considered significant at p < 0.05 (one-way ANOVA and Tukey's multiple comparison test). Gene expression of ALFs during shrimp development was assessed in twelve developmental stages as previously described [16].
Then, peptides were oxidized as previously reported [28]. In brief, 5 mg of the crude peptide were first reduced with 10% β-mercaptoethanol (95 • C for 5 min) then dissolved in 50% (v/v) AcOH/H 2 O and later diluted in 32 mL of oxidation buffer (2 mM guanidinium chloride, 10% isopropyl alcohol and 10% dimethyl sulfoxide). The pH was adjusted to 5.8 with ammonium hydroxide. The peptide solution was subjected to air oxidation at room temperature for 18 h. The peptide solution was then acidified to pH 2.5 and purified using a SPE C18 (Waters Corp., Milford, MA, USA). The peptides were eluted with 5%, 20%, 40%, 60% and 80% acetonitrile in 0.05% TFA ultrapure water at a flow rate of 1 mL/min. The fractions were collected, and the acetonitrile was evaporated on a Savant SPD 1010 SpeedVac Concentrator (Thermo Scientific, Asheville, NC, USA). The fractions were analyzed by MALDI-TOF mass spectrometry.
Minimum inhibitory concentrations (MICs) were determined in duplicate by the liquid growth inhibition assay, as previously described [29]. MIC values are expressed as the lowest concentration tested that causes 100% growth inhibition. Poor Broth (PB: 1% peptone, 1% NaCl, pH 7.2) was used for standard bacteria, while PB supplemented with 0.5 M NaCl (PB-NaCl) was used as a culture medium for Vibrio strains. For B. stationis and M. maritypicum cultures, PB-NaCl medium was supplemented with 20 mM KCl, 5 mM MgSO 4 and 1.5 mM CaCl 2 . Potato dextrose broth (Kasvi, São José dos Pinhais, PR, Brazil) at half strength was used for cultures of F. oxysporum while Sabouraud medium (1% peptone, 4% glucose, pH 5.6) was used for yeast cultures. The growth of bacteria and yeast was monitored spectrophotometrically (λ = 595 nm), while F. oxysporum hyphae formation was observed in an inverted microscope. After MIC determination, bacterial cultures were plated in nutrient agar for 24-48 h for the determination of the bactericidal activity of the synthetic peptides.

Conclusions
In conclusion, the combination of our molecular, transcriptional and functional data revealed that ALFs comprise the most diverse AMP family found in penaeid shrimp. We showed that they are composed of seven members encoded by different genes that follow a diverse pattern of expression. Our results also strongly suggest that the expansion and diversification of shrimp ALFs have shaped novel functions for this AMP family beyond their primary antibacterial properties. Thus, ALFs represent an attractive model to explore the impacts of the molecular diversity of immune-related genes on host-microbe interactions. Finally, ALFs possess the broadest spectrum of antimicrobial activity when compared to other shrimp AMPs. These bioactive peptides undoubtedly show biotechnological potential for the development of novel antibiotics derived from AMPs, as well as for the development of selective breeding programs.

Supplementary Materials:
The following are available online at http://www.mdpi.com/1660-3397/16/10/381/ s1. Table S1: Sequences and biochemical properties of ALFs from decapod crustaceans and marine chelicerates. Figure S1: Amino acid sequence alignments of ALF-related sequences containing three and four cysteine residues. The predicted signal peptides are in bold and underlined. Asterisks (*) mark the identical amino acid residues while the cysteines are highlighted with a black background. GenBank accession numbers are indicated in brackets.