The oysters used in pigmentation pattern analysis were acquired from three farms located in Yantai, Rushan, and Penglai in Shandong Province, China, and produced in hatchery. Three hundred oysters were collected from each location, and their average shell heights in three locations were 9.4 ± 0.5, 8.4 ± 0.6, and 8.4 ± 0.4 cm, respectively.

The oysters, in the experiment of antioxidant physiology indexes detection, were obtained from market (Zhifu, Yantai, China). They were dissected and the color of the adductor muscle scars were observed. The muscles attached to white adductor scar (called “white muscle” for short) and the muscles attached to black adductor scar (called “black muscle” for short) were sampled and stored at −80°C.

The oysters in RNA-seq experiment, 6–8 cm in shell length, were obtained from a half-sib family (Changdao, Yantai, China). They were dissected and divided into a pigmented group and an unpigmented group, according to both the color of the adductor muscle scars and outer shell surface (Figure 1). Three oysters were included in each group. The adductor muscles and mantles were cut, snap frozen in liquid nitrogen and stored at −80°C until later extraction of RNA. Total RNA was extracted using the guanidinium thiocyanate-phenol-chloroform extraction method (TRIzol, Invitrogen) according to manufacturer's protocol.

The oysters used in real-time Q-PCR experiment were obtained from market (Zhifu, Yantai, China), and divided into pigmented scar VS unpigmented scar and pigmented shell VS unpigmented shell groups. Twelve individuals in each group. The adductor muscles and mantles were cut, snap frozen in liquid nitrogen and stored at −80°C until later extraction of RNA. Total RNA was extracted using TRIzol (Invitrogen) according to manufacturer's protocol.

No specific permissions were required for the above sampling locations. The Pacific oyster is not an endangered or protected species and is not a vertebrate. The oysters used in this study were farmed.

The outside surface of each oyster shell was brushed to remove any attached objects and washed to ensure that it was completely clean. The oysters were then boiled for 30 min in a pan on a hot plate. The shells were then opened and the soft bodies removed.

The soft bodies were then placed into an electric blast-drying oven for 12 h. The left shells of the oysters from each farm were marked 1–300 and the right shells 1′–300′ with a pencil. The colors of the outer surface of the shell and the adductor muscle scars of all the oysters were recorded and divided into three types: completely pigmented, partly pigmented, and unpigmented (Brake et al., 2004; Batista et al., 2008). The quantity of oysters whose outer surface of the left shell and left adductor muscle scar differed in color was calculated in each farm; that of oysters whose outer surface of the right shell and right adductor muscle scar differed in color was also done in each farm. The independence of the shell color and the adductor muscle scar color was analyzed through Chi-square test based on the quantities of oysters whose outer surface of the shell and adductor muscle scar in the same side differed in color.

The dried soft bodies of the oysters with an unpigmented or completely pigmented outer shell surface were weighed and recorded, and those with unpigmented or completely pigmented adductor muscle scar were also weighed and recorded.

The correlations between the color of the adductor muscle scar or the outer surface of the shell and the dried soft-body weight were analyzed using unpaired Student's t-test in SPSS 20.0. The cut off for significance for the p-values was 0.05.

Twelve white muscles homogenates and 12 black ones were prepared, respectively. Hydroxyl free radical scavenging capacity (HFRSC) of these homogenates were determined based on Fenton reaction (Glick, 2006) using a commercially available detection kit (Nanjing Jiancheng Bioengineering Institute) according to the manufacturer's instructions. The malondialdehyde (MDA) content, the total antioxidant capacity (TAC) and the protein concentration (using the dying method of Coomassie brilliant blue) were also detected with the assay kits from the same company.

Total RNA was extracted from each of the three “black muscles,” three “white muscles,” three “black mantles,” and three “white mantles.” These three samples combined together comprised the “black muscle,” “white muscle,” “black mantle,” and “white mantle” samples, respectively.

Poly-A RNA was isolated with oligo-dT-coupled beads from 20 μg total RNA from each sample and then sheared. The isolated RNA samples were used for first strand cDNA synthesis, which was performed with random hexamers and Superscript II reverse transcriptase (Invitrogen). The second strand was synthesized with Escherichia coli DNA PolI (Invitrogen). Double stranded cDNA was purified with Qiaquick PCR purification kit (Qiagen, Germantown, MD, USA). After end repair and addition of a 3′ dA overhang, the cDNA was ligated to Illumina paired-end adapter oligo mix, and size selected to ~200 bp fragments by gel purification. After 15 PCR cycles the libraries were sequenced using Illumina sequencing platform and the paired-end sequencing module.

The transcriptomes of “black muscle,” “white muscle,” “black mantle,” and “white mantle” were sequenced using 90 bp paired-end RNA-seq. The gene expression levels were determined in different tissues based on their reads per kilobase (RPKM) values of gene model per million mapped reads. After log2 conversion of RPKM values, upregulated and downregulated genes between “black muscle” and “white muscle” or between “black mantle” and “white mantle” were statistically analyzed with T-test and false discovery rate (FDR) correction. Differential gene expression was considered significant at an adjusted P < 0.01 (Xu et al., 2016; Yang et al., 2016).

Based on functional annotation using Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG), we extracted GO terms or KEGG pathways that were enriched in differently expressed genes between “black muscle” and “white muscle” or between “black mantle” and “white mantle” with EnrichPipeline (Chen et al., 2010).

Because no biological replicates were performed in RNA-seq, the further validation for the RNA-seq results by real-time quantitative RT-PCR were carried out (Xu et al., 2016; Yang et al., 2016). For reverse transcription, the first-strand cDNA was synthesized according to M-MLV RT Usage information (Promega, USA) using oligo (dT)-adaptor (5′-CTCGAGATCGATGCGGCCGCT17-3′) as primer and the DNase I-treated (Promega) total RNA as template. We verified the difference in the expression levels of three tyrosinase genes, two retinol dehydrogenase genes and two Cytochrome P450 genes between the black and white mantles. A constitutive expression gene, ribosomal protein S18 (RS18), was used as endogenous control (Miyamoto and Kajihara, 2005; Wei et al., 2015a,b). All primers used in this assay were listed in Table 1. The SYBR Green real-time PCR assay was carried out in the BIO-RAD CFX Connect™ Real-Time PCR Detection System. Data were analyzed with the CFX Manager™ SoftwareVersion 3.1 (BIO-RAD, USA). The PCR amplification was performed in a 20 μL volume, with 10 μL SYBR Green PCR Master Mix, 2 μL diluted cDNA, 0.8 μL primers (5 μmol/L), and 7.2 μL DEPC-treated water. The thermal profile was 50°C for 2 min and 94°C for 2 min, followed by 40 cycles of 95°C for 5 S, 60°C for 15 S and 72°C for 20 S. The relative expression levels of these genes were analyzed by the 2^−ΔΔCT method described previously (Livak and Schmittgen, 2001; Bustin et al., 2009). Statistical analysis was performed by one-way analysis of variance (one-way ANOVA) using SPSS 16.0 statistical software. The p-values less than 0.05 were considered statistically significant. The primers used in this experiment were listed as follows (Table 1).

We drew the location schematic of the upregulated tyrosinase genes and their neighboring tyrosinase genes in oyster genome scaffolds referred to the database of oyster genes and omics (http://www.oysterdb.com). Based on the pathway of retinol metabolism in animals in KEGG pathway database (http://www.genome.jp/kegg-bin/show_pathway?map00830) and the results of this study, the hypothetical pathway of retinol dehydrogenase genes and Cytochrome P450 genes involved in melanogenesis was proposed.

The number and percentage of oysters with inconsistent colors of the outer surface of the left shell and the left adductor muscle scar from the three oyster farms are detailed in Table 2 and Table S1, and illustrated using an example in Figures 2A,B. The data show that the color of the outer surface of the left shell was independent of that of the left adductor muscle scar (P > 0.05). The number and percentage of oysters with inconsistent colors of the outer surface of the right shell and the right adductor muscle scar from the three oyster farms are detailed in Table 2 and Table S1, and illustrated using an example in Figures 2A',B'. The statistical data show that the color of the outer surface of the right shell was also independent of that of the right adductor muscle scar (P > 0.05).

Though there were three types: completely pigmented, partly pigmented and unpigmented muscle scar or outer shell, but in order to observe the obvious dissimilarity, only the oysters with completely pigmented and unpigmented muscle scar or outer shell were applied to the correlation analysis between pigmentation and soft-body dry weight. Oysters with a black outer shell surface were not found to be heavier in terms of their dried soft-body weight at any of the three farms (P > 0.05, Table 3). However, oysters with black left adductor muscle scars were heavier in terms of their dried soft-body weight across all three farms (P < 0.05, Table 3); similarly, oysters with black right adductor muscle scars were also heavier in terms of their dried soft-body weight across all three farms (P < 0.05, Table 3). Above results hinted that the adductor muscle scar color was correlated with the soft-body dry weight. Thus, the black adductor muscle scar may be one new potential breeding character, which could be used to obtain a larger soft-body dry weight.

The HFRSC, MDA, and TAC in the homogenates of white or black muscles were detected and compared. As shown in Figure 3 and Table 4, the HFRSC in black muscles was higher than that in white ones (p < 0.01), but the MDA content and the TAC content were found no different between the black muscles and the white ones.

The four C. gigas transcriptomes of black muscle, white muscle, black mantle, and white mantle were constructed and sequenced using Illumina Hiseq 2500, and the raw data were submitted to the NCBI SRA database with accession numbers of SRR5638635 (black muscle), SRR5638636 (white muscle), SRR5638638 (black mantle), and SRR5638651 (white mantle). Then, the differentially expressed genes in black VS white muscle and in black VS white mantle were listed in Tables S2, S3, respectively. GO terms enriched based on the upregulated genes in “black muscle” compared to “white muscle” are presented in Table S4. GO terms enriched based on the upregulated genes in “black mantle” compared to “white mantle” are presented in Table S5. The results showed that the enriched GO terms related to adductor muscle scar pigmentation were different to those related to outer shell surface pigmentation.

According to the Tables S2, S3, the differentially expressed tyrosinase genes only existed in the Up-Regulation gene sets of black_Amu/white_Amu or black_Man/white_Man. Other trosinase genes weren't differentially expressed between the “black muscle” and the “white muscle,” or between the “black mantle” and the “white mantle.” We found that two tyrosinase genes (CGI_10011916 and CGI_10012743) were significantly upregulated in the “black muscle” compared to the “white muscle” (Figure 4A; Table 5). We also found that three tyrosinase genes (CGI_10013418, CGI_10017214 and CGI_10021076) were significantly upregulated in the “black mantle” compared to the “white mantle” (Figure 4B; Table 5).

As presented in Table S7, Cytochrome P450 genes were found upregulated not only in the black muscle (CGI_10016640, CGI_10011491, CGI_10017766, and CGI_10028005) and but also in the black mantles (CGI_10011065 and CGI_10022185). We noted that the upregulated cytochrome P450 genes in black muscle were also different with ones in black mantle.

We found the “Retinol metabolism” pathway enriched based on the upregulated genes in “black muscle” compared to “white muscle” (P < 0.01), and also found the same pathway enriched based on the upregulated genes in “black mantle” compared to “white mantle” (P < 0.01), but the upregulated genes in the above two groups were complete different (Table S7). In the muscle group, the upregulated genes were Cytochrome P450 1A2, Cytochrome P450 1A2, Cytochrome P450 26A1, Cytochrome P450 3A24, 17-beta-hydroxysteroid dehydrogenase 13 and Retinoic acid receptor responder protein 3, but in the mantle group, those were Cytochrome P450 1A1, Retinal dehydrogenase 1, Retinal dehydrogenase 1, Cytochrome P450 3A29, Diacylglycerol O-acyltransferase 1, and Omega-crystallin.

As shown in Figure 5 and Table S6, the two upregulated tyrosinase genes in the “black muscle,” CGI_10011916 and CGI_10012743, were located in scaffold43702 and scaffold1792, respectively; the three upregulated tyrosinase genes in the “black mantle,” CGI_10013418, CGI_10017214, and CGI_10021076, were located in scaffold1630, scaffold248, and scaffold203, respectively. Therein CGI_10011916 was among one tyrosinase gene cluster, CGI_10021076 among another one (Figure 5 and Table S6).

Based on the results of RNAseq, three tyrosinase genes, two retinol dehydrogenase genes, and two Cytochrome P450 genes were selected to perform qRT-PCR in black and white mantles. It was found that all the seven genes were upregulated in the “black mantle” compared to the “white mantle” (Table 6, Figure S1). Especially, two tyrosinase genes (CGI_10013418 and CGI_10021076), two retinol dehydrogenase genes (CGI_10026867 and CGI_10026868), and one Cytochrome P450 gene were significantly upregulated.

We also decided the expression levels of two tyrosinase genes, two retinol dehydrogenase genes and four Cytochrome P450 genes by qRT-PCR in the black and white adductor muscles. It was found that the eight genes were upregulated in the “black adductor muscle” compared to the “white adductor muscle” (Table 7, Figure S2). Especially, two tyrosinase genes (CGI_10011916 and CGI_10012743), and two Cytochrome P450 genes (CGI_10011491 and CGI_10017766) were significantly upregulated.

A comparison of qRT-PCR and RNA-seq data of the five tyrosinase genes was conducted. The results showed that the expression patterns of CGI_10011916, CGI_10012743 in muscle and CGI_10013418 and CGI_10021076 in mantle agreed well between RNA-seq and qRT-PCR, and CGI_10017214 had a similar trend in change of expression pattern between RNA-seq and qRT-PCR (Table 8).

It has been studied that the water-soluble melanin fractions from squid ink was more efficient than carnosine as the free radical scavenger (Xiao yan et al., 2003; Vate and Benjakul, 2013; Guo et al., 2014). The black muscles may contain more melanin that could scavenge hydroxyl free radicals. Bivalve adductor muscles are composed of semi-translucent and white opaque muscles. The former is thought to be responsible for the quick closure of shells, and the latter for catch contraction that keeps shells tightly closed for many hours and maintains this tension for long periods with little energy consumption (Funabara et al., 2013).

According to common sense, the shells of dead oyster is completely open because of the relaxation of its muscle, which should need the least energy consumption. So, it is possible that the physiological cost of maintaining shells open at an intermediate position should be larger and the physiological metabolization in the adductor muscle is likely to be intensive when oysters open their shells to filter algae. This would result in the production of many free radicals that could impair the muscle cells of oyster, just as human (Konczol et al., 1998; Kerksick and Zuhl, 2015; Pal et al., 2017). Adductor muscles attached to black adductor muscle scars might contain melanin that could eliminate these free radicals and reduce the damage to the adductor muscle (Sarna et al., 1986; Korytowski et al., 1987; Rozanowska et al., 1999). Compared with oysters with white adductor muscle scars, those with black scars might enable them to open their shells for longer and filter more algae, which could promote an accelerated growth rate.

It has been found that the pigmented adductor muscle scar to which the “black muscle” was attached contained melanin (Hao et al., 2015) and a tyrosinase gene was involved in melanin production (Kumar et al., 2011). So, the results of this study suggests that these two tyrosinase genes may play an important role in melanin deposition on adductor muscle scar in C. gigas. Adductor muscle scar may not only function to open/close the shell, but also to produce melanin. Previous work has shown that the pigmented shell, to which the “black mantle” corresponds, also contains melanin (Yu et al., 2015), and tyrosinase genes are expressed in bivalve mantles (Zhang et al., 2012; Aguilera et al., 2014), but didn't correlate gene expression with melanin deposition in mollusks. However, the results of this study hinted that these three tyrosinase genes were likely involved in melanin deposition on the outer surface of C. gigas shells.

Two steps of melanin synthesis are catalyzed by tyrosinase (Goodwill et al., 1998); therefore, tyrosinase is the rate-limiting enzyme in the production of melanin (Sanchez-Ferrer et al., 1995). It was easy to understand that the difference in pigmentation of the adductor muscle scars or the outer surfaces of the shell could be due to the differential expression of the genes encoding tyrosinase. However, we noted that the differentially expressed tyrosinase genes in the up-regulation gene set of black muscle/white muscle were completely different with those of black mantle/white mantle, hinting that the different members of the tyrosinase gene family (Zhang et al., 2012; Yu et al., 2015) could play a role in the independent pigmentation of different organs.

It has usually been observed that the co-expression of neighboring genes occurred in gene cluster (Boutanaev et al., 2002; Lercher et al., 2002, 2003; Birnbaum et al., 2003; Fukuoka et al., 2004). But, the co-expression of neighboring tyrosinase genes wasn't observed in this study, on the contrary, the non-neighboring tyrosinase genes co-expressed, hinting that the co-expression of these non-neighboring tyrosinase genes was probably caused by the requirement of physiologic function not the adjacent location in genome.

Different members of Cytochrome P450 gene family were found upregulated not only in the black muscle (CGI_10016640, CGI_10011491, CGI_10017766, and CGI_10028005) and but also in the black mantles (CGI_10011065 and CGI_10022185). The upregulated cytochrome P450 genes in black muscle were also different with ones in black mantle, and Cytochrome P450 can involve in the melanin biosynthesis in Streptomyces griseus (Funa et al., 2005), which hinted that the different members of Cytochrome P450 gene family were also involved in the independent pigmentation of different organs of oyster.

Neuroglobin and cytoglobin were two members belonging to the vertebrate globin superfamily, and they had similar physiologic function (oxygen transport and storage) in different organs (Burmester et al., 2004). In this study, it was also found that the different members of the same gene family performed the same function (pigmentation) in different organs of oyster.

Retinol (Vitamin A1) is one of the animal forms of vitamin A and retinol dehydrogenases can transfer vitamin A into acidum vitamin A (Liden and Eriksson, 2006). Acidum vitamin A (also called as retinoic acid) could induce the melanocyte maturation and promote basal levels of melanogenesis (Lotan and Lotan, 1980) by increasing the tyrosinase activity (Li and Zhu, 2001), it has also been reported that retinoic acid is critical in establishing asymmetric pigmentation of flatfish (Shao et al., 2017). Cytochrome P450 can also involve in the melanin biosynthesis by catalyzing retinoate into retinoic acid (Funa et al., 2005). The more important was that the “Retinol metabolism” pathway was enriched based on the upregulated genes in “black muscle” compared to “white muscle” or in “black mantle” compared to “white mantle.”

Combining the results of this study and the pathway information of “retinol metabolism in animals” acquired from KEGG pathway database (http://www.genome.jp/kegg-bin/show_pathway?map00830), we proposed the hypothetical pathway of retinol dehydrogenase genes and Cytochrome P450 genes involved in melanin biosynthesis of oyster (Figure 6). In this hypothetical pathway, retinal dehydrogenase and Cytochrome P450 affected retinoic acid production, more retinoic acid increased the tyrosinase activity, thus more melanin was generated.

In conclusion, it was found that the pigmentation of oyster adductor muscle scar was unrelated with that of its corresponding shell outer surface; interestingly, a relationship between the color of the adductor muscle scars and the dried soft-body weight; the different members of the tyrosinase or Cytochrome P450 gene families could play a role in the independent pigmentation of different organs. These findings could provide some suggestions for further pigmentation study of mollusks.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.