RNA Interference by Ingested Dsrna-Expressing Bacteria to Study Porphyrin Pigmentation in Crassostrea gigas

Porphyrins are a widespread group of pigments in nature which are believed to contribute to shell colors in mollusks. Previous studies have provided candidate genes for porphyrin shell coloration, however, the linkage analysis between functional genes and porphyrin pigmentation remains unclear in mollusks. RNA interference is a powerful molecular tool for analyzing the loss of functions of genes in vivo and alter gene expression. In this study, we used unicellular alga Platymonas subcordiformis and Nitzschia closterium f. minutissima as vectors to feed oysters with Escherichia coli strain HT115 engineered to express double-stranded RNAs targeting specific genes involved in porphyrin synthesis. A strain of Crassostrea gigas with orange shell was used to target key haem pathway genes expression using the aforementioned approach. We show here that feeding the oysters with E. coli, containing dsRNA targeting pigmentation genes, can cause changes in the color of the newly deposited shell. For example, the RNAi knockdown of CgALAS and CgPBGD resulted in the loss of uroporphyrin pigmentation from the shell due to the accumulation of the pigment in the oyster’s mantle. The study probed the crucial role of ALAS and PBGD genes potential functions of uroporphyrin production and shell color pigmentation in C. gigas.


Introduction
The diversity of mollusks shell colors has been well reported [1][2][3][4]. Pigments deposited by mantle determined the shell color [5]. The outer fold of the mantle contributes to shell formation [6]. The dorsal mantle epithelium secretes the mollusk's shell and controls pigmentation [7,8]. The formation of shell color has been studied in mollusks such as marine snails, Pacific oyster, pearl mussel and black-lipped pearl oyster [9][10][11][12]. Moreover, several lncRNA and mRNA transcripts associated with shell pigmentation have been identified which influence pigment biosynthesis including melanin, carotenoid and tetrapyrrole [13].
Porphyrins are a widespread group of pigments in nature which are believed to contribute to shell colors in mollusks [14], often leading to red, brown or purple shell pigmentation [3]. Among porphyrins, the uroporphyrin I and III are found in mollusks [15][16][17][18]. It has been suggested that porphyrin in shell pigmentation is produced de novo by the animal through the haem pathway, the production of uroporphyrin is derived from the oxidation of uroporphyrinogen I and III which belong to non-enzymatic side paths [17,19]. The regulation of the haem pathway initially occurs through the rate-limiting enzyme aminole-vulinic acid synthase (ALAS), by downregulating its transcription, upregulating mRNA breakdown, blocking its uptake into mitochondria, and increasing the breakdown of the protein in mammals [20]. The second gene in the haem pathway aminole-vulinic acid dehydratase (ALAD) has been proven to catalyze the formation of porphobilinogen by the dimerization of two molecules of 5-aminolaevulinic acid (ALA). Then, the porphobilinogen deaminase (PBGD) condenses two molecules of porphobilinogen to form

Impact of RNAi on Phenotypic Change
Oysters were fed with alga mixed with six different types of E. coli culture expressing distinct types of dsRNAs. Oysters were classified as growing obviously newly developed shell or not (Table 1 and Figure 2). Seven, nine and eight of the oysters fed on bacteriaexpressing CgALAS, CgALAD and CgPBGD dsRNA displayed the newly deposited shell growth (Table 1). At the same time, the newly deposited shell growth was detected in nine, ten and ten oysters from the CgUROS, CgUROD and EGFP groups, respectively. CgALAS fragment sizes and total plasmid sizes; (b) Plasmid construction of CgALAD-L4440 with CgALAD fragment sizes and total plasmid sizes; (c) Plasmid construction of CgPBGD-L4440 with CgPBGD fragment sizes and total plasmid sizes; (d) Plasmid construction of CgUROS-L4440 with CgUROS fragment sizes and total plasmid sizes; (e) Plasmid construction of CgUROD-L4440 with CgUROD fragment sizes and total plasmid sizes; (f) Plasmid construction of EGFP-L4440 with EGFP fragment sizes and total plasmid sizes.

Impact of RNAi on Phenotypic Change
Oysters were fed with alga mixed with six different types of E. coli culture expressing distinct types of dsRNAs. Oysters were classified as growing obviously newly developed shell or not (Table 1 and Figure 2). Seven, nine and eight of the oysters fed on bacteriaexpressing CgALAS, CgALAD and CgPBGD dsRNA displayed the newly deposited shell growth (Table 1). At the same time, the newly deposited shell growth was detected in nine, ten and ten oysters from the CgUROS, CgUROD and EGFP groups, respectively.
shell orange pigmentation, and there was no significant difference in the L*a*b* value compared with the control group. It is noteworthy that oysters fed on the PBGD dsRNA displayed an obvious orange pigmentation loss. The newly deposited shell of PBGD group has higher lightness value (L*) and greater orange pigmentation loss than the control group (a*b* value of PBGD group was significantly lower than the control group). In contrast, the UROS UROD group showed more visible orange pigmentation and higher a*b* value than the EGFP group but was not significant.   Compared with the images at the beginning of the experiment, the newly deposited shells of the ALAS and PBGD group represented a remarkable loss of orange pigmentation than the EGFP group ( Figure 2). The L*a*b* value of the newly deposited shells was calculated to show the variations of the orange shell color phenotype ( Figure 3). Oysters fed on E. coli-expressing dsRNA EGFP showed no significant changes in orange shell pigmentation (Figure 2c). In contrast, a number of oysters in the experiment groups displayed the orange pigmentation gain or loss phenotype. According to the L*a*b* value analysis, the a* and b* value of the ALAS group was significantly lower than the control group ( Figure 3). As for the ALAD group, there were no significant phenotype changes in the shell orange pigmentation, and there was no significant difference in the L*a*b* value compared with the control group. It is noteworthy that oysters fed on the PBGD dsRNA displayed an obvious orange pigmentation loss. The newly deposited shell of PBGD group has higher lightness value (L*) and greater orange pigmentation loss than the control group (a*b* value of PBGD group was significantly lower than the control group). In contrast, the UROS UROD group showed more visible orange pigmentation and higher a*b* value than the EGFP group but was not significant.  The p values from a t-test demonstrate ** p < 0.01, *** p < 0.0005 or **** p < 0.0001; and ns demonstrates no significance.

Detection of RNAi Product in Mantle Tissues
The effectiveness of dsRNA delivery to oyster mantle tissues by feeding was determined by analyzing dsRNA in oysters fed on bacteria containing the L4440-EGFP by qPCR. EGFP dsRNA expression was detected in oysters fed on bacteria containing the L4440-EGFP plasmid but not in oysters fed on CgALAS-L4440, CgALAD-L4440, CgPBGD-L4440, CgUROS-L4440, CgUROD-L4440. The detection of RNAi products (CgALAS, CgALAD, CgPBGD, CgUROS, CgUROD dsRNA) in oyster mantles was assessed as a function of the ratio of the interfering CgALAS, CgALAD, CgPBGD, CgUROS, CgUROD dsR-NAs to the endogenous CgALAS, CgALAD, CgPBGD, CgUROS, CgUROD mRNA levels ( Figure 4). A significant amount of RNAi products was found in CgALAS and CgUROS dsRNA oysters compared to EGFP dsRNA oysters at day 15 and at day 30. Significant amount of RNAi products was also found in CgALAD, CgPBGD and CgUROD dsRNA oysters at day 15. This is consistent with the dsRNA entrance in mantle tissues during the phase of interference treatment. At day 15, the levels of dsRNAs were 2.3, 2.1, 2.1, 3.9 and 2.1 times more abundant than endogenously expressed mRNAs in the CgALAS, CgALAD, CgPBGD, CgUROS, CgUROD dsRNA oysters, respectively. At day 30, the levels of dsR-NAs were 2.1 and 2.3 times more abundant than endogenously expressed mRNAs in RNAi-treated oysters expressing the CgALAS dsRNA and CgUROS dsRNA, respectively. Probabilities for two-tailed, paired t tests for differences in gene expression levels among control and experiment groups of C. gigas are shown in Supplementary Table S2. The p values from a t-test demonstrate ** p < 0.01, *** p < 0.0005 or **** p < 0.0001; and ns demonstrates no significance.

Detection of RNAi Product in Mantle Tissues
The effectiveness of dsRNA delivery to oyster mantle tissues by feeding was determined by analyzing dsRNA in oysters fed on bacteria containing the L4440-EGFP by qPCR. EGFP dsRNA expression was detected in oysters fed on bacteria containing the L4440-EGFP plasmid but not in oysters fed on CgALAS-L4440, CgALAD-L4440, CgPBGD-L4440, CgUROS-L4440, CgUROD-L4440. The detection of RNAi products (CgALAS, CgALAD, CgPBGD, CgUROS, CgUROD dsRNA) in oyster mantles was assessed as a function of the ratio of the interfering CgALAS, CgALAD, CgPBGD, CgUROS, CgUROD dsRNAs to the endogenous CgALAS, CgALAD, CgPBGD, CgUROS, CgUROD mRNA levels ( Figure 4). A significant amount of RNAi products was found in CgALAS and CgUROS dsRNA oysters compared to EGFP dsRNA oysters at day 15 and at day 30. Significant amount of RNAi products was also found in CgALAD, CgPBGD and CgUROD dsRNA oysters at day 15. This is consistent with the dsRNA entrance in mantle tissues during the phase of interference treatment. At day 15, the levels of dsRNAs were 2.3, 2.1, 2.1, 3.9 and 2.1 times more abundant than endogenously expressed mRNAs in the CgALAS, CgALAD, CgPBGD, CgUROS, CgUROD dsRNA oysters, respectively. At day 30, the levels of dsRNAs were 2.1 and 2.3 times more abundant than endogenously expressed mRNAs in RNAi-treated oysters expressing the CgALAS dsRNA and CgUROS dsRNA, respectively. Probabilities for two-tailed, paired t tests for differences in gene expression levels among control and experiment groups of C. gigas are shown in Supplementary Table S2.

Efficiency of RNAi on the Target Gene Transcription
qPCR results showed that in oysters fed on bacteria containing CgALAS-L4440, CgALAD-L4440, CgPBGD-L4440, CgUROS-L4440, CgUROD-L4440, dsRNA constructs showed reduced CgALAS, CgALAD, CgPBGD, CgUROS, CgUROD expressions compared to the control group. As shown in Figure 5, CgALAS, CgPBGD and CgUROD mRNA levels were significantly diminished in C. gigas fed on bacteria-expressing CgALAS, CgPBGD and CgUROD dsRNA as compared to control group at day 15. In the ALAS group, these results were significant and the relative CgALAS mRNA expression levels dropped down to ~62% of the control values. Thus, the average knockdown was ~38%, with a maximum knockdown value of ~37.5% (p = 0.0476) at day 15. A significant decrease (97%) was observed in CgPBGD transcripts at day 15 compared to the EGFP dsRNA oysters (p < 0.0001). Remarkable decrease in CgUROD mRNA expression levels dropped down to ~98% of the control values (p < 0.0001). There was no significant difference in CgALAD/CgUROS transcripts neither between the ALAD group or UROS group and the control at both the middle and the end of interference exposure. At day 15, individual analysis revealed that 80% of individuals in the PBGD group, and all the individuals in UROD group showed decreased gene expression by more than 70% compared to the control ( Supplementary Figure S1). At day 30, no individuals in experiment groups displayed decreased gene expression by more than 70% compared to the control group (Supplementary Figure S1).

Efficiency of RNAi on the Target Gene Transcription
qPCR results showed that in oysters fed on bacteria containing CgALAS-L4440, CgALAD-L4440, CgPBGD-L4440, CgUROS-L4440, CgUROD-L4440, dsRNA constructs showed reduced CgALAS, CgALAD, CgPBGD, CgUROS, CgUROD expressions compared to the control group. As shown in Figure 5, CgALAS, CgPBGD and CgUROD mRNA levels were significantly diminished in C. gigas fed on bacteria-expressing CgALAS, CgPBGD and CgUROD dsRNA as compared to control group at day 15. In the ALAS group, these results were significant and the relative CgALAS mRNA expression levels dropped down to~62% of the control values. Thus, the average knockdown was~38%, with a maximum knockdown value of~37.5% (p = 0.0476) at day 15. A significant decrease (97%) was observed in CgPBGD transcripts at day 15 compared to the EGFP dsRNA oysters (p < 0.0001). Remarkable decrease in CgUROD mRNA expression levels dropped down to~98% of the control values (p < 0.0001). There was no significant difference in CgALAD/CgUROS transcripts neither between the ALAD group or UROS group and the control at both the middle and the end of interference exposure. At day 15, individual analysis revealed that 80% of individuals in the PBGD group, and all the individuals in UROD group showed decreased gene expression by more than 70% compared to the control (Supplementary Figure S1). At day 30, no individuals in experiment groups displayed decreased gene expression by more than 70% compared to the control group (Supplementary Figure S1).

Discussion
Oysters are filter-feeding animals, and their open vascular system makes traditional injection RNA interference methods unable to efficiently interfere the expression of genes. The bacteria-feeding RNAi method used in this study revealed the function of genes related to uroporphyrin pigmentation in C. gigas. The effect of porphyrin pigments on shell color has been reported in mollusks, and uroporphyrin I and III in oyster shells have also been identified [3,17,18]. Genes associated with the uroporphyrinogen synthesis have been characterized in marine snails and pacific oysters [9]. The first five genes were identified from C. gigas genome and were revealed to have significant expression levels among different shell-color oysters [36]. Previous studies provide candidate genes for porphyrin pigmentation in mollusks. The effect of dsRNA interference usually reached the maximum level at 15 to 30 days after the post-induction of dsRNA in C. gigas [26,37,38]. According to previous studies on C. gigas, feeding oysters with dsRNA for 35 days conducted the successful knockdown of the corresponding genes and observed oysters with newly deposited shell growth [31]. A total 30-day knockdown experiment was designed to obtain the maximum RNAi effect and RNAi phenotype of the newly deposited shell [26,31].
The data in this study revealed that the knockdown of the five haem pathway genes has different effects on orange pigmentation. To compare the level of color change in the newly deposited shell, the color parameters of L* (brightness), a* (red) and b* (yellow) between the experimental group and the control group were obtained. Results showed that the L* of the newly deposited shell of C. gigas in the CgPBGD group was significantly higher than that of the EGFP group, which means the newly formed shell was whiter. At the same time, the a* and b* values of the CgALAS group and the CgPBGD group were significantly lower than those of the control group, which means that the newly deposited shells in the experimental groups had less red and yellow. Through the RNAi bacterial feeding process, compared with the control group, on the fifteenth day, the average level of transcripts in the CgALAS and CgPBGD groups significantly decreased by 38% and 97%, which indicates that the loss of orange pigmentation has remarkable relations with the decreasing transcription of CgALAS and CgPBGD. In contrast, the CgUROS group newly deposited shells had a higher a* and b* value than the control group. It is noteworthy that the average level of transcripts in the CgUROS group was significantly reduced

Discussion
Oysters are filter-feeding animals, and their open vascular system makes traditional injection RNA interference methods unable to efficiently interfere the expression of genes. The bacteria-feeding RNAi method used in this study revealed the function of genes related to uroporphyrin pigmentation in C. gigas. The effect of porphyrin pigments on shell color has been reported in mollusks, and uroporphyrin I and III in oyster shells have also been identified [3,17,18]. Genes associated with the uroporphyrinogen synthesis have been characterized in marine snails and pacific oysters [9]. The first five genes were identified from C. gigas genome and were revealed to have significant expression levels among different shell-color oysters [36]. Previous studies provide candidate genes for porphyrin pigmentation in mollusks. The effect of dsRNA interference usually reached the maximum level at 15 to 30 days after the post-induction of dsRNA in C. gigas [26,37,38]. According to previous studies on C. gigas, feeding oysters with dsRNA for 35 days conducted the successful knockdown of the corresponding genes and observed oysters with newly deposited shell growth [31]. A total 30-day knockdown experiment was designed to obtain the maximum RNAi effect and RNAi phenotype of the newly deposited shell [26,31].
The data in this study revealed that the knockdown of the five haem pathway genes has different effects on orange pigmentation. To compare the level of color change in the newly deposited shell, the color parameters of L* (brightness), a* (red) and b* (yellow) between the experimental group and the control group were obtained. Results showed that the L* of the newly deposited shell of C. gigas in the CgPBGD group was significantly higher than that of the EGFP group, which means the newly formed shell was whiter. At the same time, the a* and b* values of the CgALAS group and the CgPBGD group were significantly lower than those of the control group, which means that the newly deposited shells in the experimental groups had less red and yellow. Through the RNAi bacterial feeding process, compared with the control group, on the fifteenth day, the average level of transcripts in the CgALAS and CgPBGD groups significantly decreased by 38% and 97%, which indicates that the loss of orange pigmentation has remarkable relations with the decreasing transcription of CgALAS and CgPBGD. In contrast, the CgUROS group newly deposited shells had a higher a* and b* value than the control group. It is noteworthy that the average level of transcripts in the CgUROS group was significantly reduced by 98% compared to the control group on the 15th day. ALAS is the rate-limiting enzyme of the haem pathway [39]. Studies have shown that the increasing expression levels of ALAS could simultaneously increase the production of uroporphyrin [22]. The rising activity of PBGD has been suggested to be associated with uroporphyrin I production [40,41]. Whiter shell appears when the activity of CgALAS and CgPBGD is reduced, and the non-enzymatic side paths fail to generate the uroporphyrins at regular bases. This phenotypic change revealed the functions of CgALAS and CgPBGD on uroporphyrin pigments synthesis. Accordingly, it could be proposed that the downregulation of CgALAS and CgPBGD gene affects the accumulation of uroporphyrin pigments in mantles and affects the coloration of the newly deposited shell.
The RNAi response was highly variable among individuals in this study. It had been inferred that the change in the amount of dsRNA that eventually enters the mantle cell may largely contribute to the variability of the RNAi response [37,38]. It should be noted that the expression level of CgALAD, CgPBGD, CgUROS and CgUROD mRNA on the 30th day was lower than that on the 15th day. However, several individuals on the 30th day showed obvious high expression, such as No. 5 in the CgUROD group. These expression levels, which are significantly different from other individuals, affect the overall expression trend. The irregular expression level was also detected in the CgTyr knockdown and it has been speculated that the long-term feeding of bacteria may affect individual health and normal gene expression [31]. Although the significant downregulation of target genes was observed in this study, there are still parameters that need to be optimized. For example, the concentration of daily algae/bacterial co-inoculum, treatment duration and total number of individuals need to be optimized [26]. At the same time, when conducting research on shell-related interference, the selection of C. gigas at proper circadian rhythms is crucial. Higher shell growth could be obtained at water temperature between 18 • C and 24 • C, and shell development is increased in summer and autumn. Furthermore, the phenotypic change could take place in oyster mantles. With the expression modulation of the haem pathway genes, the accumulation of uroporphyrins could be affected. Therefore, the concentration of uroporphyrin between the experiment and control group could be analyzed by HPLC to make the data more detailed [17].

Oyster and Algal Culture
One-year-old Pacific oysters with the whole orange shell (120 individuals) were collected from the culture population in Weihai, Shandong, China (Oct 2020). Oysters were randomly stuck with labels, divided into six groups of 20 oysters, and placed into six tanks of 40 L aquarium equipped with an open circulatory system of natural seawater (30 psu) at a flow rate of 12 mL min −1 . They were acclimated 7 days before the start of the experiment and continuously fed with Platymonas subcordiformis (25,000 cells mL −1 ) and Nitzschia closterium f. minutissima (35,000 cells mL −1 ). Alga P. subcordiformis and N. closterium f. minutissima cultures were grown in f/2 medium using sterilized seawater. Cultures were maintained at 19 ± 1 • C, under a light:dark (L:D) 14:10 cycle. P. subcordiformis has four flagellates and is 16-30 µm long; N. closterium f. minutissima has no flagellate and is 12-23 µm long.

Experiment Primer Names Primer Sequences Product Length
Plasmid

RNAi Feeding Procedures
During the interference phase, six tanks of orange shell oysters were continuously exposed to the alga/dsRNA-producing bacteria co-inoculum. Six dsRNA-producing bacteria were prepared by inducing E. coli strain HT115 (DE3) bacteria transformed with three constructed plasmids (CgALAS-L4440, CgALAD-L4440, CgPBGD-L4440, CgUROS-L4440, CgUROD-L4440, EGFP-L4440) with isopropylβ-D-thiogalactoside (IPTG). Algae/bacteria co-inoculum was produced by mixing algal culture and bacterial suspension at a ratio of 100 bacteria per algal cell, with a final p. subcordiformis concentration of 25,000 cells mL −1 and N. closterium f. minutissima concentration of 35,000 cells mL −1 . The ratio was optimized by increasing algal concentration according to Payton et al. [26]. Food reserves were renewed with fresh algae/bacteria co-inoculum every 24 h. Specifically, E. coli strain HT115 bacteria containing recombinant plasmid were grown overnight with shaking in LB with ampicillin (50 µg mL −1 ) and tetracycline (12.5 µg mL −1 ) at 37 • C. Five milliliters of overnight culture was diluted 100-fold in 500 mL of fresh LB medium containing ampicillin (50 µg mL −1 ) and tetracycline (12.5 µg mL −1 ) and allowed to grow to OD 595 = 0.4. dsRNA production was induced by 0.4 mM IPTG for 4 h at 37 • C under agitation. Then, induced bacterial cultures were centrifuged and bacterial pellets were resuspended in 500 mL P. subcordiformis and 125 mL N closterium f. minutissima algae culture liquid. Bacteria adsorption rate on the algae was evaluated by preliminary experiments [31]. The adsorption rate was assessed as the ratio between the number of colonies from filtrated suspension and from total inoculum, which was calculated as 99%. Concentrations of algae and bacteria were monitored daily with 2800 UV-visible spectrophotometer (Unico, USA) for the duration of the experiment. Mantle tissues from 10 oysters per condition were individually sampled at the end of the interference phase (day 15 and day 30). No oyster died during the 15-day and 30-day exposure experiments. A total of 120 oysters were dissected. All experiments were performed in accordance with our institutional guidelines.

Phenotypic Change Statistical Analyses
Computer vision was used to obtain the objective and non-destructive assessment of color patterns in non-uniformly colored surfaces. L*a*b* international standard for color measurements was used to measure the shell color changes though experiments. L* is the lightness component, which ranges from 0 to 100, and parameters a* (from green to red) and b* (from blue to yellow) range from −120 to 120 [42]. The photographs of oyster shells were taken by a digital camera (Nikon D4). By comparing with the photo at the start of the RNAi experiment, individual oysters from the six groups were classified to either be having the newly deposited shell or not, and the details could be found in Table 1. Oysters which had a newly deposited shell were used to analyze the L*, a* and b* values. The lightness, a, and b values obtained in Photoshop CS6 (Adobe System Incorporated, USA) were converted into L*a*b* values using the following formulas [43]: Differences between groups (the color parameter) were investigated using a two-tailed Student's t test. For all statistical results, a probability of p < 0.05 was considered significant. Analyses were performed using IBMSPSS Statistics 20.0 (SPSS, USA). Probabilities for two-tailed, paired t tests for differences in L*a*b* value among the control and experimental groups of C. gigas could be found in Supplementary Table S1.

Total RNA Extraction and qPCR Analyses
RNA extraction was performed on the individual sample using the RNA-easy isolation reagent (Vazyme, Nanjing, China), following the manufacturer's instruction. The total RNA of 1 µg was subjected to reverse transcription using HiScript III 1st Strand cDNA Synthesis Kit with gDNA wiper (Vazyme, Nanjing, China). Relative quantification of gene expression was estimated for each gene using the ∆∆ Ct method [44]. Primer sets were designed from full cDNA sequences of CgALAS, CgALAD, CgPBGD, CgUROS, CgUROD, EGFP [26], and elongation factor I (Ef1) which were used as reference genes ( Table 2). A dissociation curve was generated in each case to check that only a single band was amplified. The primer sets (CgALAS-Q-O) for qPCR analysis of endogenous ALAS mRNA levels were derived from the region outside the cloned fragment of CgALAS to prevent the amplification of RNAi products (CgALAS dsRNA). The primer sets (CgALAS-Q-I) were designed in the region inside the cloned fragment. Similarly, the primer sets CgALAD-Q-O, CgPBGD-Q-O, CgUROS-Q-O, CgUROD-Q-O and CgALAD-Q-I, CgPBGD-Q-I, CgUROS-Q-I, CgUROD-Q-I were designed outside and inside of the cloned fragment of CgALAS, CgALAD, CgPBGD, CgUROS, CgUROD, respectively. The qPCR was carried in triplicate for each sample on a LightCycler 480 real-time PCR instrument (Roche Diagnostics, Burgess Hill, UK) using ChamQ SYBR Color qPCR Master Mix (Vazyme, Nanjing, China). Cycling parameters were 95 • C for 5 min and then 40 cycles at 95 • C for 5 s and 60 • C for 20 s. Melting curve analyses were performed following the PCR to verify specific amplification. PCR efficiency (E) was assessed for each primer pair by determining the slope of standard curves obtained from cDNA serial dilution analyses of different experimental samples. Relative gene expression data were analyzed using the comparative threshold cycle (Ct) 2 −∆Ct method, where the ∆Ct = Ct (interest gene)−Ct (housekeeping gene). Interest gene values were normalized in each sample with Ef1 housekeeping gene level, as no significant differences in Ct values were observed in Ef1 among the conditions. Results were expressed as the number of copies of gene per copy of Ef1. Statistical analyses were performed using the GraphPad Software 8 with an independent t-test. Differences were deemed statistically significant at p < 0.05.The detection of RNAi products (interest gene dsRNA) in oyster mantles was assessed as the mean ± SE of individuals 2 −(∆Ct(CgInside)-∆Ct(CgOutside)) , i.e., the ratio of the CgALAS mRNA level measured with primer sets CgALAS-I (quantification of both endogenously expressed CgALAS mRNAs and CgALAS dsRNAs) to the CgALAS mRNA level obtained with primer sets CgALAS-O (to quantify the endogenously expressed CgALAS mRNAs). Therefore, this ratio is equal to (gene dsRNAs + gene mRNAs)/gene mRNAs = (gene ds RNAs/gene mRNAs) + 1.

Conclusions
In conclusion, this study is further evidence of the bacteria-feeding RNAi with algae as the carrier of E. coli HT115 that produces dsRNA, providing an effective long-lasting method to trigger specific RNAi responses of C. gigas. When genes related to the accumulation of uroporphyrin (e.g., CgALAS, CgPBGD) were knocked down, the newly deposited shell of C. gigas showed more whiteness and less orange. Conversely, when the genes involved in uroporphyrinogen degradation were interfered (e.g., CgUROS) with, the newly deposited shell exhibited more orange. The study indicated that key haem pathway genes possess a vital role in porphyrin pigmentation, and thus influenced the shell color formation of C. gigas.