Tissue-Biased and Species-Specific Regulation of Glutathione Peroxidase (GPx) Genes in Scallops Exposed to Toxic Dinoflagellates

Marine bivalves could accumulate paralytic shellfish toxins (PSTs) produced by toxic microalgae, which might induce oxidative stress. Glutathione peroxidases (GPxs) are key enzymes functioning in the antioxidant defense, whereas our understanding of their roles in PST challenge in bivalves is limited. Herein, through genome-wide screening, we identified nine (CfGPx) and eight (PyGPx) GPx genes in Zhikong scallop (Chlamys farreri) and Yesso scallop (Patinopecten yessoensis), respectively, and revealed the expansion of GPx3 sub-family in both species. RNA-Seq analysis revealed high expression of scallop GPx3s after D stage larva during early development, and in adult hepatopancreas. However, in scallops exposed to PST-producing dinoflagellates, no GPx was significantly induced in the hepatopancreas. In scallop kidneys where PSTs were transformed to higher toxic analogs, most CfGPxs were up-regulated, with CfGPx3s being acutely and chronically induced by Alexandrium minutum and A. catenella exposure, respectively, but only one PyGPx from GPx3 subfamily was up-regulated by A. catenella exposure. Our results suggest the function of scallop GPxs in protecting kidneys against the oxidative stresses by PST accumulation or transformation. The tissue-, species-, and toxin-dependent expression pattern of scallop GPxs also implied their functional diversity in response to toxin exposure.


Introduction
Reactive oxygen species (ROS) are produced in all living organisms as by-products of aerobic metabolism [1]. The excessive production of ROS can cause lipid peroxidation, protein oxidation, DNA damage, membrane disruption, cellular damage, immune dysfunction, as well as metabolic malfunction [2,3]. Low levels of ROS production are required to maintain the physiological functions, including proliferation, host defense, signal transduction, and gene expression that are involved in the antioxidant defense [4,5]. Aerobic organisms have developed various non-enzymatic and enzymatic antioxidant data expanded our understanding of GPx functions in the detoxification and antioxidant defense of shellfish.

Results and Discussion
2.1. Genome-Wide Identification of GPx Genes in C. farreri and P. yessoensis A total of nine and eight GPx genes were identified in C. farreri (CfGPx) and P. yessoensis (PyGPx) genomes, respectively. General information regarding their genome position, intron number, encoding protein length, GPx domain position, isoelectric point (pI) value, and molecular weight was summarized in Table 1, and the coding nucleotide and deduced amino acid sequences of CfGPxs and PyGPxs were shown in Figure S1. In C. farreri genome, eight SeGPx genes were identified, including two GPx1s (CfGPx1-1 and CfGPx1-2), five GPx3s (CfGPx3-1, CfGPx3-2, CfGPx3-3a, CfGPx3-3b, and CfGPx3-4), and one GPx4 (CfGPx4), whereas only one non-SeGPx gene, CfGPx7, was found. Similarly, in P. yessoensis genome, one PyGPx1, five GPx3s (PyGPx3-1, PyGPx3-2, PyGPx3-3, PyGPx3-4, and PyGPx3-5), one PyGPx4, and one PyGPx7 were obtained. Moreover, the coding sequence of CfGPxs ranged from 366 to 843 bp in length and encoded proteins from 121 to 280 amino acids (aa), whereas the PyGPxs varied from 327 to 834 bp in length and encoded proteins between 108 and 277 aa. The CfGPxs had an average theoretical pI from 4.35 to 8.90 and an average molecular weight from 14.22 kDa to 24.68 kDa, while the theoretical pI of PyGPxs ranged from 4.43 to 9.41 and the average molecular masses varied from 12.73 kDa to 31.41 kDa.
Notably, the GPx3 sub-family has more duplicated copies in scallop genomes than in mammal, teleost, and amphibian, indicating the GPx3 expansion in scallop genome, but members of four GPx sub-families, including GPx2, GPx5, GPx6, and GPx8, were all absent from the scallop genomes (Table 2). Moreover, it was found that the amino acid sequences of GPx genes in C. farreri and P. yessoensis shared great sequence identity of more than 40% with vertebrates and other mollusks (Table S1), suggesting the conservation of the GPx family between vertebrates and invertebrates.  The GPx sequence alignment among selected species revealed the presence of conserved amino acid residues, Cys (C), and the active site residues, Gln (Q) and Trp (W) (Figure 1), which were responsible for the fixation of selenium (Sec) [18,19,40]. In addition, the GPx family members were characterized by the presence of a conserved motif (GPx sig- conserved Cys or Sec residue [41]. Two other conserved GPx domains: GPx signature motif 2, LAFPCNQF and active site motif WNF(S/T)KF, are critical sites for the catalytic activity of this enzyme, which are totally dependent on GSH for its regeneration [41]. It was clear from the sequence alignment that CfGPxs and PyGPxs shared the conserved features found in well-characterized GPxs from other vertebrates. For example, the typical GPx signature motif 2, L(G/A) (F/V)PC(N/D)QF and the active site motif, WNF(E/T/W)K(F/Y) were also highly conserved in scallop (Figures 1 and S1), although the GPx signature motif 2 was missing in some scallop GPx3 sub-family members. Moreover, the GPx signature motif 1 (GQVSLVVNVASECGFT) was only predicted in the non-SeGPx, GPx7, of both scallops (Figures 1 and S1). Qu et al. have also identified the signature sequence motif, LGFPCNQF, an extra active site motif, WNFEKF, and the GPx-1 active site motif, GKVILVENVASLUGTT, in the N-terminal region of thick shell mussel, M. coruscus [25].
In addition, two arginine (Arg, R) residues which direct the GSH donor substrate towards the catalytic center [42] were also well conserved in both scallop GPxs (Figures 1 and S1). Two arginine residues (R, R) involved in binding GSH, and two amino acid residues glutamine and tryptophan (Q and W) responded for fixation of Sec in GPx were also observed in SeGPx of some mollusks, M. coruscus [25], P. yessoensis [18], C. farreri [19], and H. discus discus [21]. Moreover, the structural stability of GPx was reported to depend on three-loop structures where the first loop is Asn (N) to Tyr (Y), the second is Leu (L) to Gln (Q), and the third is Trp (W) to Phe (F) [43]. These three loop structures were also conserved in both scallop GPx sequences and were essential for maintaining the biochemical properties and enzymatic activities of GPxs [17,18]. In addition, three other conserved domains and four highly potential catalytic sites (Cys, Gln, Trp, and Asn) were also present in most amino acid sequences of both scallop GPxs (Figures 1 and S1). These conserved amino acids were likely to stabilize GPx structure and function in evolutionary terms. GenBank. The GPx signature motif 1 (blue), GPx signature motif 2 (red) and active site motif (black) were marked by colored frames. The catalytically important residues were indicated by colored down-arrows, as active sites with red, arginine residues with green, and the loop structures with black. Gaps are indicated by dashes to improve the alignments.

Phylogenetic Relationship of GPxs between Bivalves and Other Organisms
Phylogenetic analysis of GPx amino acid sequences from 19 selected species was conducted ( Figure 2). As the result, the scallop GPx members can be classified into four major clades. One clade contained vertebrate GPx1 and GPx2 members, as well as mollusk GPx1s clustered with vertebrate GPx2. CfGPx1s and PyGPx1 formed a sub-cluster with GPx1s from other bivalve species: pearl oyster (P. fucata), Pacific oyster (C. gigas), and Mediterranean mussel (M. galloprovincialis). In another clade, the GPx3s of both scallops initially formed a marine bivalve GPx3 group and clustered with other mollusk GPx3s, such as Pacific abalone (H. discus hannai), zebra mussel (D. polymorpha), freshwater mussel (U. tumidus), Pacific oyster (C. gigas), and Manila clams (R. philippinarum). All these invertebrate GPx3s formed a sub-clade and clustered with vertebrate GPx3, GPx5, and GPx6 members, which indicated that vertebrate GPx3, 5, and 6 shared the same ancestor with mollusk GPx3 sub-family. Moreover, the GPx4 of both scallops positioned as an independent cluster with a direct linking to the vertebrate GPx4 clade. Besides, the scallop The GPx signature motif 1 (blue), GPx signature motif 2 (red) and active site motif (black) were marked by colored frames. The catalytically important residues were indicated by colored down-arrows, as active sites with red, arginine residues with green, and the loop structures with black. Gaps are indicated by dashes to improve the alignments.

Phylogenetic Relationship of GPxs between Bivalves and Other Organisms
Phylogenetic analysis of GPx amino acid sequences from 19 selected species was conducted ( Figure 2). As the result, the scallop GPx members can be classified into four major clades. One clade contained vertebrate GPx1 and GPx2 members, as well as mollusk GPx1s clustered with vertebrate GPx2. CfGPx1s and PyGPx1 formed a sub-cluster with GPx1s from other bivalve species: pearl oyster (P. fucata), Pacific oyster (C. gigas), and Mediterranean mussel (M. galloprovincialis). In another clade, the GPx3s of both scallops initially formed a marine bivalve GPx3 group and clustered with other mollusk GPx3s, such as Pacific abalone (H. discus hannai), zebra mussel (D. polymorpha), freshwater mussel (U. tumidus), Pacific oyster (C. gigas), and Manila clams (R. philippinarum). All these invertebrate GPx3s formed a sub-clade and clustered with vertebrate GPx3, GPx5, and GPx6 members, which indicated that vertebrate GPx3, 5, and 6 shared the same ancestor with mollusk GPx3 sub-family. Moreover, the GPx4 of both scallops positioned as an independent cluster with a direct linking to the vertebrate GPx4 clade. Besides, the scallop GPx7s were in a clade composed of GPx7s from Pacific oyster, C. gigas and then clustered with vertebrate GPx7s and GPx8s as the sisters of the well-supported GPx7/8 sub-group, which indicated that the mollusk GPx8 was probably lost in their common ancestor during evolution. The phylogenetic analysis provided evidence that CfGPxs and PyGPxs were derived from a common ancestor with other GPx family proteins as reported by De Zoysa et al. [21]. Zhang et al. also stated that the GPx phylogenetic relationships were in good agreement with traditional taxonomy, suggesting that this protein family might have a primarily similar functional role [20].

021, 12, x FOR PEER REVIEW 6 of 14
GPx7s were in a clade composed of GPx7s from Pacific oyster, C. gigas and then clustered with vertebrate GPx7s and GPx8s as the sisters of the well-supported GPx7/8 sub-group, which indicated that the mollusk GPx8 was probably lost in their common ancestor during evolution. The phylogenetic analysis provided evidence that CfGPxs and PyGPxs were derived from a common ancestor with other GPx family proteins as reported by De Zoysa et al. [21]. Zhang et al. also stated that the GPx phylogenetic relationships were in good agreement with traditional taxonomy, suggesting that this protein family might have a primarily similar functional role [20].

Spatio-Temporal Expression of Scallop GPxs During Development and in Adult Tissues
The spatio-temporal expression profiles of scallop GPxs were analyzed using transcriptome data [38,44]. In scallop embryos and larvae, CfGPx1-1, CfGPx3-1, CfGPx3-3b, and CfGPx4 in C. farreri, and PyGPx1, PyGPx3-1, PyGPx3-5, and PyGPx4 in P. yessoensis exhibited relatively higher expression level than other GPxs, suggesting the importance of these GPxs in scallop development (Figure 3a). During the entire developmental process, high expression of GPx genes (CfGPx3-1, CfGPx3-3b, PyGPx1, PyGPx3-1) was mainly present at the larval stages, especially after the D stage veliger, and in the following umbo larvae and juvenile stages (Figure 3a), implying the involvement of these GPxs in antioxidation or detoxification in scallop metamorphosis and post-larval development, which

Spatio-Temporal Expression of Scallop GPxs During Development and in Adult Tissues
The spatio-temporal expression profiles of scallop GPxs were analyzed using transcriptome data [38,44]. In scallop embryos and larvae, CfGPx1-1, CfGPx3-1, CfGPx3-3b, and CfGPx4 in C. farreri, and PyGPx1, PyGPx3-1, PyGPx3-5, and PyGPx4 in P. yessoensis exhibited relatively higher expression level than other GPxs, suggesting the importance of these GPxs in scallop development (Figure 3a). During the entire developmental process, high expression of GPx genes (CfGPx3-1, CfGPx3-3b, PyGPx1, PyGPx3-1) was mainly present at the larval stages, especially after the D stage veliger, and in the following umbo larvae and juvenile stages (Figure 3a), implying the involvement of these GPxs in antioxidation or detoxification in scallop metamorphosis and post-larval development, which may be due to the elevation of oxygen consumption to meet high demands of energy reserve utilization during organ initiation and structural re-modelling [32]. Moreover, both CfGPx4 and PyGPx4 were expressed from the very beginning of fertilization and multicellular stages to the larvae and juvenile stages, indicating their maternal origin to play protective roles and to maintain a redox balance during development.
In adult individuals, the identified scallop GPx genes were expressed in all the examined tissues/organs, except that CfGPx1-2 transcript was not detected in any tissue (Figure 3b). Notably, higher expression of GPxs was mainly detected in the hepatopancreas of both scallops. These highly expressed GPxs included CfGPx3-1, CfGPx3-2, CfGPx3-3a, CfGPx3-4, and CfGPx4 in C. farreri, and PyGPx1, PyGPx3-1, PyGPx3-2, PyGPx3-3, PyGPx3-4, and PyGPx4 in P. yessoensis, most of which were from the GPx3 sub-family. The dominant GPx3s expression in scallop hepatopancreas was consistent with its role as the main metabolic and defense organs against oxidative stress caused by excessive ROS [27], which suggested that scallop GPx3s might play crucial roles in molluscan metabolism and antioxidative defense system. The high expression of SeGPx in hepatopancreas was also reported in abalone H. discus discus [21] and H. discus hannai Ino [27], freshwater bivalve U. tumidus [17,45] and D. polymorpha [40], and clam M. meretrix [22] and V. philippinarum [14,20]. In addition, GPx4s of both scallops, as well as PyGPx1, were transcribed in almost all the examined tissues. Similar expression patterns were also reported in other invertebrates. For example, the GPxs of M. galloprovincialis were mainly expressed in the gills, hemocytes, and digestive gland, moderately expressed in adductor muscle and gonad [16], and the GPx of L. vannamei was widely distributed in the hemocytes, hepatopancreas, gills, muscles, and intestine tissues [46]. may be due to the elevation of oxygen consumption to meet high demands of energy reserve utilization during organ initiation and structural re-modelling [32]. Moreover, both CfGPx4 and PyGPx4 were expressed from the very beginning of fertilization and multicellular stages to the larvae and juvenile stages, indicating their maternal origin to play protective roles and to maintain a redox balance during development.
In adult individuals, the identified scallop GPx genes were expressed in all the examined tissues/organs, except that CfGPx1-2 transcript was not detected in any tissue ( Figure  3b). Notably, higher expression of GPxs was mainly detected in the hepatopancreas of both scallops. These highly expressed GPxs included CfGPx3-1, CfGPx3-2, CfGPx3-3a, CfGPx3-4, and CfGPx4 in C. farreri, and PyGPx1, PyGPx3-1, PyGPx3-2, PyGPx3-3, PyGPx3-4, and PyGPx4 in P. yessoensis, most of which were from the GPx3 sub-family. The dominant GPx3s expression in scallop hepatopancreas was consistent with its role as the main metabolic and defense organs against oxidative stress caused by excessive ROS [27], which suggested that scallop GPx3s might play crucial roles in molluscan metabolism and antioxidative defense system. The high expression of SeGPx in hepatopancreas was also reported in abalone H. discus discus [21] and H. discus hannai Ino [27], freshwater bivalve U. tumidus [17,45] and D. polymorpha [40], and clam M. meretrix [22] and V. philippinarum [14,20]. In addition, GPx4s of both scallops, as well as PyGPx1, were transcribed in almost all the examined tissues. Similar expression patterns were also reported in other invertebrates. For example, the GPxs of M. galloprovincialis were mainly expressed in the gills, hemocytes, and digestive gland, moderately expressed in adductor muscle and gonad [16], and the GPx of L. vannamei was widely distributed in the hemocytes, hepatopancreas, gills, muscles, and intestine tissues [46].

Expression Regulation of Scallop GPxs in Response to PST-Producing Dinoflagellates
As filter-feeders, scallops have a high ability to accumulate PSTs [37]. Previous studies have suggested that the scallop hepatopancreas and kidneys were the two major organs involved in PST bioaccumulation and transformation, respectively [38,47,48]. Moreover, dinoflagellates of the genus Alexandrium are the major PST producers containing varied toxic analogs among species and strains. Therefore, to understand the scallop defense mechanism against toxic algae, RNA-Seq analysis was performed to determine the expression profiles of GPx genes in scallop kidneys and hepatopancreas in response to the ingestion of PST-producing algae, A. minutum (AM-1 strain) and A. catenella (ACDH strain) (Tables S2 and S3).
In C. farreri hepatopancreas, the major organ for PST accumulation from algae [38,48], most CfGPxs' expression was not significantly altered after toxic algae exposure, with only CfGPx7 down-regulated at the 1 st day after A. catenella exposure (log2FC: −2.06) ( Figure  4a). This result implied that CfGPxs mediated antioxidant defense in hepatopancreas was

Expression Regulation of Scallop GPxs in Response to PST-Producing Dinoflagellates
As filter-feeders, scallops have a high ability to accumulate PSTs [37]. Previous studies have suggested that the scallop hepatopancreas and kidneys were the two major organs involved in PST bioaccumulation and transformation, respectively [38,47,48]. Moreover, dinoflagellates of the genus Alexandrium are the major PST producers containing varied toxic analogs among species and strains. Therefore, to understand the scallop defense mechanism against toxic algae, RNA-Seq analysis was performed to determine the expression profiles of GPx genes in scallop kidneys and hepatopancreas in response to the ingestion of PST-producing algae, A. minutum (AM-1 strain) and A. catenella (ACDH strain) (Tables S2 and S3).
In C. farreri hepatopancreas, the major organ for PST accumulation from algae [38,48], most CfGPxs' expression was not significantly altered after toxic algae exposure, with only CfGPx7 down-regulated at the 1st day after A. catenella exposure (log 2 FC: −2.06) (Figure 4a). This result implied that CfGPxs mediated antioxidant defense in hepatopancreas was not sensitive to PST challenge, and the highly expressed CfGPx3s might provide stable antioxidant capacity for redox homeostasis during toxic algae ingestion.
However, the regulation of CfGPxs in scallop kidneys exhibited different patterns between the two toxic algae exposure (Figure 4a). For example, the induction of CfGPx3s was mostly chronic (5th-10th days, log 2 FC: 2.18-5.61) with A. catenella challenge, whereas acute (1st-3rd days, log 2 FC: 2.19-5.45) with A. minutum exposure. Moreover, the induction of CfGPx1-2 and the down-regulation of CfGP3xs were only observed in the kidney of scallops exposed to A. minutum (Figure 4a and Table S3). Since the two micro-algae produce different PST analogs, with A. minutum mainly synthesizing gonyautoxins (GTXs, mainly GTX1-4) and A. catenella synthesizing N-sulfocarbamoyl toxins (C1 and C2) [38,48], it therefore suggested that the difference in time dependent pattern of GPx3s response between A. catenella and A. minutum challenge in C. farreri kidneys was probably due to the different toxin species accumulated ( Figure S2). This diverse expression profiles of CfGPxs indicated that, with the challenges of different PST analogs, the CfGPxs functioning in PST detoxification and antioxidation during PST transformation were similar in scallop kidneys, but the gene stimulation or inhibition procedure/speed might vary.
To further compare the regulation of GPxs between scallop species, we examined PyGPxs transcripts in both hepatopancreas and kidneys of P. yessoensis ingesting A. catenella (Figure 4b). There were less regulated GPxs than those identified in C. farreri, with only PyGPx3-4 being up-regulated in P. yessoensis kidney (10th day, log 2 FC: 2.88) and three PyGPx3s down-regulated (3rd and 5th days, log 2 FC: −2.42-−2.95) (Figure 4b). In P. yessoensis hepatopancreas, only PyGPx3-5 was down-regulated at the 3rd day (log 2 FC: −2.60). The inhibition of PyGPxs might be resulted from the decline of GSH contents under cellular redox imbalance, to maintain the stable GSH/GSSG ratio which is crucial in preventing oxidative damage [49]. Moreover, the species-specific and tissue-biased expression of genes from the expanded GPx3 sub-family may also indicate the different sensitivity of scallops to varied PSTs ( Figure S2) and the possible association of GPx3 expansion in the adaptation of scallops to the environment with different toxic algae.
The antioxidant metabolism is crucial for preventing the cellular oxidative damage in bivalves exposed to toxic dinoflagellates, such as Alexandrium sp. [30]. Stress response is usually modulated at the molecular level and shows significant changes in the transcripts expressed in the target tissues [50]. Many studies have described the impact of PST-producers on gene expression of diverse bivalve organs, majorly in C. gigas. For example, both repression and activation of genes in C. gigas gills by A. minutum induction was involved in key metabolic pathways such as oxidative stress, mitochondrial metabolism, endogenous clock, immunity, and detoxification processes, including SODs, GSTs, and GPxs [30,51]. The expression of defense-related genes modulated in the hemocytes of oyster after exposure to A. minutum could demonstrate the impact of direct contacting of some Alexandrium species on different cell types due to the production of extracellular compounds [52]. Expression of genes involved in antioxidant stress, detoxification, immune, and stress defense of C. gigas juveniles increased in a short exposure time by the toxic dinoflagellate, G. catenatum [53,54] and P. lima exposure [55]. Therefore, the high expression level of CfGP3xs maintained in scallop hepatopancreas, as well as the significantly regulated expression of CfGPxs in kidneys reasonably supported the fact that these two organs were the main centers involved in detoxification and antioxidant defense, but with different roles. These findings also indicated that the expression of GPx genes varied among scallop species and tissues, dinoflagellate strains, as well as the exposure time. The species-and organ-specific regulation of GPxs imply the diverse function of scallop GPxs in the defense against the harmful effects of PST accumulation and metabolism, which might contribute to the scallop adaptation to diverse marine environments. The heatmaps were drawn based on log 2 FC of the exposed group to the control group (0 day) (Table S3). Exposure time (1, 3, 5, 10, or 15 days) was displayed above each heatmap. Significant regulation (|log 2 FC| > 2 and p < 0.05) was indicated with "*", and very significant regulation (|log 2 FC| > 2 and FDR < 0.05) was indicated with "**".

Conclusions
In conclusion, a total of nine CfGPx and eight PyGPx genes were identified in scallops C. farreri and P. yessoensis, respectively, with the expansion of GPx3 sub-family in both species. GPx3s showed higher expression than other GPx genes, and hepatopancreas was the major organ for scallop GPx expression. After ingesting PST-producing algae, no GPx was induced in hepatopancreas of either scallop, but most GPxs were up-regulated in C. farreri kidneys, with acute and chronic regulation by A. minutum and A. catenella exposure, respectively. Meanwhile, only one GPx3 was induced in kidneys of P. yessoensis exposed to A. catenella. These results indicated that the response of scallop GPxs to PST-producing dinoflagellates was dependent on scallop species and tissues, as well as dinoflagellates strains. Our data revealed the involvement of GPxs in scallops against the toxic stress and highlighted their functional importance, which provided a better understanding of the detoxification systems in scallops and their regulation mechanism.

Screening GPx Genes from the C. farreri and P. yessoensis Genomes
The GPx genes were identified using the C. farreri and P. yessoensis transcriptome and genome data [38,44]. The coding sequences were translated using the ORF finder (https://www.ncbi.nlm.nih.gov/orffinder/), and the predicted ORFs were validated using the NCBI non-redundant (Nr) protein sequence database. The protein sequences of GPxs were further verified with BLASTP in NCBI (http://blast.ncbi.nlm.nih.gov/) and UniProt (http://www.uniprot.org/) with the setting of e-value to 1E-05. All possible GPx proteins from the BLASTP results were then tested using SMART (http://smart.emblheidelberg.de/). Protein sequences longer than 100 amino acids were used for further analysis. The translated sequences were submitted to the SMART program for identification of the signal peptide and other conserved GPx structures. Compute pI/MW tool (https://www.expasy.org/) was used to predict the pI value and molecular weight (kDa).

Multiple Sequence Alignment and Phylogenetic Analysis of the GPx Gene Family
Multiple sequence alignment of GPx sequences was performed using Clustal W [56] and then was edited by Genedoc software (version 2.7.0) [57]. The amino acid similarity matrix was made using the Clustal Omega software (http://www.ebi.ac.uk/Tools/ msa/clustalo/) with GPx protein sequences of both scallops and other selected species. The GPx amino acid sequences of selected species included Homo sapiens, Mus musculus, Bos taurus, Rattus norvegicus, Xenopus tropicalis, Anolis carolinensis, Gallus gallus, Danio rerio, and mollusk species such as Pinctada fucata, U. tumidus, H. discus hannai, Dreissena polymorpha, Ruditapes philippinarum, C. gigas, and M. galloprovincialis, all obtained from Ensembl (http://www.ensembl.org/) and NCBI (https://www.ncbi.nlm.nih.gov/) ( Table S4). The phylogenetic tree was constructed using the GPx amino acid sequences of C. farreri and P. yessoensis, as well as other selected species using MEGA 7.0 with the maximum-likelihood (ML) method [58] and the tested best model of LG + G. The bootstrap analysis with 1000 replicates was applied to test the relative support for the branches produced by ML method.

Expression Analysis of GPxs During Scallop Development and in Adult Organs/Tissues
C. farreri and P. yessoensis were collected from Xunshan Group Co., Ltd. (Rongcheng, Shandong Province, China) and Zhangzidao Group Co., Ltd. (Dalian, Liaoning Province, China), respectively. The RNA-Seq data for scallop embryo/larva including zygote, 2-8 cells, blastula, gastrula, trochophore, D stage larvae, early umbo larvae, mid umbo larvae, late umbo larvae, creeping larvae, and juvenile scallops, and the adult organs/tissues including mantle, gill, eye, kidney, hepatopancreas, hemocyte, female gonad, male gonad, smooth muscle, and adductor muscle, were analyzed. The expression profiles of CfGPxs and PyGPxs during development and in adult organs/tissues were normalized and represented in the form of RPKM (Reads Per Kilobase of exon model per Million mapped reads) using the data from our previous studies [38,44]. To visualize the expression patterns of GPx genes on both scallops, the heatmap was generated via heatmap package under the R software environment with the log 10 (RPKM + 1) value.

Expression Analysis of GPx Genes in Scallops Exposed to Toxic Dinoflagellates
The scallops C. farreri were exposed to PST-producing dinoflagellates, A. minutum (AM-1 strain) and A. catenella (ACDH strain) [38,48]. Briefly, 2-year-old adult scallops were collected and acclimated in filtered and aerated seawater at 12-13 • C for three weeks for depuration by feeding the non-toxic algae, Isochrysis galbana (7.5 × 10 5 cells/mL) as the control group [59]. 18 scallops were separated into independent tanks with aeration, and 6 groups with 3 individuals for each were taken at different test time points (0, 1, 3, 5, 10, and 15 days) during the challenging experiments. The two Alexandrium dinoflagellate strains were cultured independently in F/2 medium using artificial seawater, and were harvested at the late exponential growth phase through centrifugation of 2500 g/10 min [54,60]. Then, each scallop was fed once a day with 3 L (2.5 × 10 3 cells/mL) A. minutum or A. catenella. For comparison, 2-year-old P. yessoensis were also exposed to A. catenella (ACDH strain) using the same procedure described above. Scallops were sampled at 0 day (control), 1 and 3 (acute response) days, 5, 10, and 15 (sub-chronic reaction) days with three individuals at each time point after AM-1 and ACDH exposure. The kidney and hepatopancreas of these scallops were dissected, washed with sterile seawater, and immediately frozen at −80 • C for subsequent RNA extraction. PST concentrations in the tissues of sampled scallops were determined by high-performance liquid chromatography with tandem mass spectrometry (HPLC-MS/MS) method [38].
Total RNA was isolated from scallop kidneys and hepatopancreas, and the RNA quantity and quality were checked by Nanovue Plus and gel electrophoresis before the following reverse transcription and sequencing analysis. RNA-Seq data obtained from these tissues were used to examine the expression profiles of GPx genes in response to different toxic dinoflagellates. Briefly, the RNA-Seq libraries of these tissues from C. farreri and P. yessoensis fed with AM-1 and ACDH were constructed independently using the NEB Next mRNA Library Prep Kit following the manufacturer's instructions and were subjected to PE125 sequencing on the Illumina HiSeq 2000 platform. The RNA-Seq reads were mapped to the C. farreri and P. yessoensis genomes, respectively, using Tophat (ver 2.0.9), and the expression of all GPx genes were normalized and represented in the form of RPKM (Table S2). The fold change (FC) between the toxin-exposed and control groups was calculated and the significant differences were determined using the edgeR package with cutoff of |log 2 FC| > 2 and p value < 0.05, and the very significant difference with cutoff of |log 2 FC| > 2 and corrected FDR value < 0.05. The heat maps were drawn using pheatmap package with the log 2 FC values (Table S3).

Supplementary Materials:
The following are available online at https://www.mdpi.com/2072-6 651/13/1/21/s1, Table S1. The homology percentages of CfGPx and PyGPx amino acid sequences identity compared with other animals, Table S2. RPKM values of GPxs from RNA-Seq analysis after toxic A. catenella and A. miutum exposure in hepatopancreas and kidneys at different time points, Table S3. Different expression of GPxs in hepatopancreas and kidneys after A. catenella and A. minutum exposure analyzed using edgeR package, Table S4. Information of GPxs used in this study from vertebrates and mollusk species, Figure S1. Nucleotide and deduced amino acid sequences of (1) CfGPxs and (2) PyGPxs, Figure S2. The temporal abundance of PSTs in scallop kidneys and hepatopancreas after exposure to the toxic alga A. minutum and A. catenella.

Institutional Review Board Statement:
This study was conducted in accordance with the Institutional Animal Care and Use Committee of Ocean University of China, and it does not contain any studies with human participants.

Data Availability Statement:
The RNA-Seq data for exposure experiments in this study were deposited in NCBI Sequence Read Archive, with the accession numbers of SRX2445405-SRX2445440, and the RPKM values and differential expression results of GPxs were in supplementary Tables S2  and S3.