1 INTRODUCTION

The clam Cyclina sinensis is naturally distributed along the coastline of China, Korea, and Japan. For the nutritious and economic importance of this species, C. sinensis has become one of the main maricultural shellfish in eastern China since it was developed in the 1980s (Liu et al., 2002). As an intertidal clam, salinity stress is one of the abiotic stresses frequently suffered by C. sinensis. Although C. sinensis belongs to the euryhaline organism, salinity changes can also affect the growth, survival, embryonic development, and energy metabolism in clams (Ni et al., 2020). Additionally, Vibrio infection is another main stress suffered by clams. The clams live in shallow water that is rich in pathogens, and research suggested that Vibrio is the main pathogen that causes the mass mortality of clams (Yue et al., 2010; Ni et al., 2020). In recent years, the adverse effects of salinity stress and Vibrio infection on the aquaculture industry of C. sinensis are becoming serious. Therefore, researches on salinity tolerance and immune mechanism in C. sinensis have been increasing gradually (Lin et al., 2013; Ren et al., 2015; Ni et al., 2021; Sun et al., 2021). In most of these studies, gene expression analysis is one of the important ways to study gene function.

Quantitative real-time PCR (qRT-PCR) has been widely used to quantify gene expression levels owing to its high accuracy, sensitivity, and throughput. qRT-PCR including absolute and relative quantitative methods, of which relative quantitation has widely been used in the analysis of differentially expressed genes. In relative quantitation, the expression level of target gene is usually normalized according to stably expressed genes, known as reference genes. The most frequently used endogenous references are the housekeeping genes, which are always expressed at the more or less constant level that is essential to the basic processes of cell living (Livak and Schmittgen, 2001; Stephenson, 2010). Numerous studies suggested that housekeeping genes were not consistently expressed under different tissue types (Jin et al., 2021), developmental stages (Huan et al., 2016), and experiment treatments (Chen et al., 2020). Therefore, the selection and validation of appropriate reference genes are critical for the accuracy of gene expression analysis by qRT-PCR.

To date, an increasing number of housekeeping genes have been evaluated to identify reference genes for qRT-PCR normalization under specific conditions in bivalve mollusks. For instance, the Elongation factor 1-α (EF1-α), Glyceraldehyde-3-phosphate dehydrogenase (GAPDH), Tubulin-α (TUB-α), and 40S ribosomal protein S20 (RPS20), in the various combinations suggested by geNorm, NormFinder, and BestKeeper, were suitable for the normalization in clam Ruditapes philippinarum under the copper treated hemocytes (Volland et al., 2017). In oyster Crassostrea gigas, Ribosomal protein L7 (RL7) and Ribosomal protein S18 (RPS18) could be the best reference genes for normalization of qRT-PCR expression data in OsHV-1 infected larvae (Du et al., 2013). In king scallop Pecten maximus, NADH dehydrogenase (ubiquinone) 1 alpha subcomplex subunit 7 (ndufa7), RPSα, and EF1-α were the most stable genes in the ovary, and the 18s rRNA, ndufa7, and GAPDH were the best-ranked reference genes in testis (Mauriz et al., 2012). Therefore, there was no perfect reference gene that was appropriate for use in all experimental conditions, and the stability of reference genes should be validated for the conditions of studies.

For clam C. sinensis, systematic evaluation of reference genes that can be used among different tissues or experiment treatments is lacking, and β-actin is most frequently used as a qRT-PCR reference gene without validation (Ni et al., 2021; Sun et al., 2021). To improve this situation, we selected five commonly used housekeeping genes, including β-actin, EF1-α, GAPDH, RPS18, and TUB-α, and evaluated their expression stability in different adult tissues and under low salinity stress and Vibrio parahaemolyticus infection using three frequently used programs, geNorm, NormFinder, and BestKeeper. To our knowledge, this is the first systematic analysis of reference genes for normalization of qRT-PCR data in C. sinensis. These results will facilitate more accurate and reliable gene expression studies of this species.

2 MATERIAL AND METHOD 2.1 Sample collection

The one-year-old clams C. sinensis with a mean weight of 8.23±0.89 g used in this study were obtained from a local market in Yancheng, China. All clams were acclimated at 26 ℃ in continuous aerated artificial seawater (salinity 25) for one week before the experiments. Five tissues from four clams under normal physiological conditions, including hepatopancreas, mantle, gill, foot, and adductor muscle were dissected, immediately frozen in liquid nitrogen, and then stored at -80 ℃ until tissue distribution analysis of candidate reference genes could be conducted.

As an intertidal clam, salinity stress and Vibrio infection are the main environmental stresses suffered by C. sinensis. Therefore, two experimental treatments were conducted in this study, including low salinity stress and V. parahaemolyticus immersion challenge. For the low salinity stress, clams were transferred from the seawater with salinity 25 to 5. To obtain a global view of the expression of candidate reference genes during salinity stress, we collected samples at 0 h, 12 h, 1 d, 3 d, 5 d, and 7 d. Five clams were randomly collected at each time point, the hepatopancreases of clams were dissected and preserved in liquid nitrogen for RNA extraction.

For the Vibrio challenge, clams were immersed in the seawater (salinity 25) with ~1×10⁷ colony forming units (CFU)/mL of V. parahaemolyticus. The Vibrio challenge test was conducted as described in Jiang et al. (2018) with minor modification. During the Vibrio challenge, very low mortality was observed during the first 5 days, while the mortality increased drastically around 10 d, exceeding 50%. Five clams were randomly collected at 0 h, 12 h, 1 d, 2 d, 3 d, and 5 d post-infection respectively, the gills of clams were dissected and preserved in liquid nitrogen for RNA extraction.

2.2 RNA extraction and reverse transcription

Total RNA was extracted from samples using an E.Z.N.A Total RNA Kit II (Omega Bio-Teck, USA) according to the manufacturer's instructions. RNA quality was verified by agarose gel electrophoresis and spectrophotometry. About 1-μg total RNA was reverse-transcribed in a 20-μL reaction using the PrimeScript RT reagent kit with gDNA Eraser (TaKaRa, Japan) according to the manufacturer's protocol. To detect residual genomic DNA contamination in the cDNA samples, RNA samples were pooled to prepare an RT-minus negative control in which reverse transcriptase was excluded from the reaction mixture. No influence from genomic DNA was confirmed by the results that there was no amplification or high enough cycle threshold value (C_t) values (e.g., > 35) for all primer pairs when the RT-minus negative control was used as a template in the subsequent qRT-PCR assay.

2.3 Candidate reference genes selection

We collected five housekeeping genes as candidate reference genes (Table 1), including beta-actin (β-actin), Elongation factor 1-alpha (EF1-α), Glyceraldehyde-3-phosphate dehydrogenase (GAPDH), 40S ribosomal protein S18 (RPS18), and Tubulin alpha-1 (TUB-α). The gene sequences of candidate reference genes were obtained from the genome and transcriptome databases of the clam C. sinensis (Wei et al., 2020). The fragments amplified by primers of these sequences were verified and then were provided in Supplementary File 1.

2.4 Primer specificity and amplification efficiency

Primers were designed using Primer premier 5.0 with the following criteria: primer Tm, 60±5 ℃; primer length, 20–25 bp; expected products, 100–250 bp. Primer specificity and amplification efficiency of each pair primer were examined before they were used to estimate gene expression level. Amplicons of each primer pair were tested by 1.5% agarose gel electrophoresis to verify the product's size and specificity. Amplification specificity was also certified when only a single peak was observed in melting curve analysis in a qRT-PCR assay.

The amplification efficiency (E) of each primer pair was determined by qRT-PCR, using a serial dilution of pooled cDNA samples (×1, ×4, ×16, ×64, and ×256 dilutions). Amplification efficiency was calculated based on the slope of a standard curve generated by plotting C_t values against the relative quantity of templates, the calculation formula is E=4^-1/slope. Primer sequences were adjusted until all primers had good amplification specificity and had amplification efficiencies of 0.90–1.10 (Stephenson, 2010). Primer information is provided in Table 2.

2.5 Quantitative real-time PCR

The qRT-PCR was conducted in a 96-well plate using CFX Connect Real-Time PCR Detection System (Bio-Rad, USA). Each reaction was performed in triplicate and a 10-μL volume containing 2×TB Green Premix Ex Taq II (TaKaRa, Japan), 0.4 μmol/L of each primer, and 0.5 μL of cDNA. The qRT-PCR program was 95 ℃ for 30 s, and 40 cycles of 95 ℃ for 5 s, 55 ℃ for 30 s, and 72 ℃ for 30 s. A melting curve analysis was performed after each reaction to confirm specificity.

2.6 Gene expression stability analysis

Gene expression stability was evaluated using three frequently used programs, BestKeeper V1, NormFinder, and geNorm V3.5. Raw C_t values and amplification efficiency were directly submitted to BestKeeper. BestKeeper calculates a BestKeeper Index and evaluates gene expression stability based on correlation analysis between a gene and the BestKeeper Index, genes with high Pearson's correlation coefficient (r) and low P values are considered to be stably expressed (Pfaffl et al., 2004). For NormFinder and geNorm, data were transformed to relative quantification data by normalizing to the sample with the highest C_t value using the formula: relative quantification data=(1+E)^ΔCt, where ΔC_t=the highest C_t value–C_t value of the sample, and E represents the amplification efficiency. After transforming the raw C_t values into relative quantification data, the NormFinder calculates intra-and intergroup variations which are then combined into a stability value, the gene with the lowest stability value, was the best reference gene (Andersen et al., 2004). The geNorm calculates the value of gene expression stability (M) for all candidate genes, and a lower M value indicated a more stable expression. For determining the optimal number of internal parameters required for qRT-PCR, geNorm calculates the pairwise variation V_n/V_n+1 which represents the variation between using n most stable genes and using n+1 most stable genes. This evaluation uses 0.15 as the threshold value, when V_n/V_n+1 < 0.15, n reference genes are sufficient for accurate normalization, otherwise, n+1 internal reference genes are required (Vandesompele et al., 2002).

2.7 Quantification of target genes using different normalization strategies

The AMP-activated protein kinase α2 gene (AMPKα2, F: 5′-GAGTTCCAAATCAAGGAG-CCG-3′, R: 5′-GGCATTCTTTCAGGGTGGG-3′) and Interleukin17-1 gene (IL17-1, F: 5′-TCGTT-GAGACTTTTGACACCAGC-3′, R: 5′-GGACACG-AGGACATTTCACTTGC-3′) were selected as the target genes to test the effect of using different normalization strategies. The expression levels of AMPKα2 and IL17-1 were determined in clams under salinity stress and Vibrio infection respectively. The relative expressions of the target genes were calculated through the 2^-ΔΔCt method using either commonly used β-actin or the optimal gene combination as reference (Livak and Schmittgen, 2001). When multiple reference genes were recommended for normalization, the geometric mean of their C_t values was applied (Vandesompele et al., 2002). One-way analysis of variance (ANOVA) was used to determine whether the relative expression patterns derived from the two normalization strategies were different.

3 RESULT 3.1 Primer specificity and amplification efficiency

The performance of each primer pair was validated for both amplification specificity and efficiency. The amplification products of each primer pair were confirmed by agarose gel electrophoresis, appeared as a single band with the expected size (Fig. 1a). Then the specificity of the products from each primer pair was also validated as only a single peak was observed in melting curve analysis (Fig. 1b–f). The amplification efficiency of all the primer pairs varied from 0.99 for β-actin to 1.07 for TUB-α (Table 2). Therefore, the primers designed were acceptable for further qRT-PCR assays.

3.2 The expression of candidate reference genes

The boxplot graphs depicting the distributions of C_t values were shown in Fig. 2, and all C_t values are provided in Supplementary Tables S1–S3. In the different tissue samples tested, the C_t values of the five genes ranged from 14.60 to 29.92 (Supplementary Table S1). As shown in Fig. 2a, the C_t value of EF1-α was lower than those of the other four genes, indicating that EF1-α has a relatively higher expression level than the other genes. Among these five candidate reference genes, the C_t value distribution of TUB-α was the most discrete, indicating that the expression of TUB-α was unstable in different tissues of C. sinensis (Fig. 2a).

In the low salinity stress, the C_t values of the five genes in the hepatopancreas ranged from 13.24 to 21.79 (Supplementary Table S2). As shown in Fig. 2b, the gene expression levels of GAPDH and β-actin were relatively lower than the others, and the most highly expressed gene was EF1-α. In the Vibrio immersion challenge, the C_t values of the candidate genes in gills ranged from 14.34 to 27.05 (Supplementary Table S3). As shown in Fig. 2c, the most highly expressed gene was also EF1-α, and the gene expression level of GAPDH was lower than the others. Therefore, the expression levels of candidate reference genes might be more affected by tissue types and experimental treatments, a simple comparison of the raw C_t values for the candidate reference genes could not provide sufficient information. So, the following analysis was conducted using three different statistical algorithms for reference genes identification.

3.3 Expression stability analysis of candidate genes in different tissues

The expression level data of the candidate reference genes were analyzed using three programs, geNorm, NormFinder, and BestKeeper. For different tissues of C. sinensis under normal physiological conditions, geNorm analysis showed that EF1-α and GAPDH were the most stable genes, which were followed by RPS18, β-actin, and TUB-α (Table 3, Fig. 3a). The optimal number of reference genes required for accurate normalization in qRT-PCR was also calculated by geNorm, the result showed that there was no V_n/n+1 value below 0.15, the lowest value was V_3/4 (0.345), indicating that using three genes could be preferred for normalization (Fig. 3b). NormFinder calculated gene expression stability values, which revealed that RPS18 was the most stably expressed gene, which was followed by β-actin, GAPDH, EF1-α, and TUB-α. The five candidate reference genes were further evaluated using BestKeeper. The top three ranked genes were RPS18, β-actin, and GAPDH, these results were similar to those from NormFinder (Table 3).

While the results from different programs varied, they share many similarities, especially for NormFinder and BestKeeper. TUB-α was generally poorly ranked by all programs, ranked fifth by NormFinder and BestKeeper, and ranked fourth by BestKeeper. EF1-α was highly ranked by geNorm but poorly ranked by NormFinder and BestKeeper (Table 3). As a result, GAPDH, RPS18, and β-actin were identified as the optimal gene combination for qRT-PCR analysis in different tissues of clam C. sinensis under normal physiological conditions.

3.4 Expression stability analysis of candidate genes under low salinity stress

Salinity stress is one of the environmental stresses often suffered by C. sinensis in tidal flats. To find out a suitable reference genes combination, the expression stability of the candidate reference genes was assessed in hepatopancreas under low salinity stress. As shown in Fig. 4a, geNorm analysis indicated that EF1-α and RPS18 were the most stable genes, which were followed by GAPDH, β-actin, and TUB-α. The NormFinder recommended GAPDH as the most stable reference gene, the ranking of the other four candidate genes were EF1-α, β-actin, RPS18, and TUB-α. The BestKeeper identified EF1-α as the most stable gene, which was followed by GAPDH, β-actin, RPS18, and TUB-α (Table 3). For the number of reference genes required for data normalization, geNorm analysis showed that when the reference genes number increased from two to four, the V_n/n+1 value decreased from 0.157 to 0.122. The V_3/4 value (0.141) was less than the cut-off value of 0.15, thus the most three stable genes were required for accurate normalization (Fig. 4b). According to the comprehensive rank of these five candidate reference genes, EF1-α, GAPDH, and RPS18 could be recommended as the reference gene combination in the qRT-PCR analysis of clam C. sinensis under salinity stress.

3.5 Expression stability analysis of candidate genes after Vibrio challenge

Vibrio infection is one of the main biotic stresses suffered by C. sinensis. To find out appropriate reference genes, we tested five candidate housekeeping genes for expression stability in the gill of Vibrio infected clams. As shown in Fig. 5a, geNorm analysis showed that EF1-α and RPS18 were the most stable genes. Meanwhile, both NormFinder and BestKeeper showed that EF1-α and RPS18 were the most stable genes, although the ranking was different. For GAPDH, the ranking orders generated by the three programs were the same, ranking third. Both geNorm and NormFinder showed that TUB-α was the least stable gene, while the BestKeeper recommended β-actin as the least (Table 3). All the three V_n/V_n+1 values for the five candidate genes generated by geNorm were higher than the cut-off value of 0.15, shown that no optimal gene combination could be provided by geNorm (Fig. 5b). Here the V_2/3 value (0.191) was the lowest and a little bit higher than 0.15, indicating that using two stable genes could be preferred. Still, considering that the two most stable genes had the same function and were both involved in the translation process, we included GAPDH in the gene combination for normalization to avoid errors caused by gene co-regulation. Thus EF1-α, RPS18, and GAPDH could be recommended as the reference genes combination in the qRT-PCR analysis of clam C. sinensis after Vibrio challenge.

3.6 Comparison of different normalization strategies

To test whether different normalization strategies might influence the quantification results of gene expression, we analyzed the expression patterns of AMPKα2 and IL17-1 genes in different tissues under normal physiological conditions, salinity stress, and Vibrio infection of C. sinensis respectively. The results of the above reference genes validation indicated that the gene combination of three housekeeping genes should be used for normalization. Therefore, we used either β-actin or the gene combination GAPDH-RPS18-β-actin for normalization in different tissues, another gene combination EF1α-GAPDH-RPS18 for normalization in specific tissue under salinity and Vibrio stress. The results showed that the expression patterns of target genes mostly varied using different normalization strategies.

For AMPKα2, the expression patterns obtained from the two normalization strategies were various. In detail, the expression analysis normalized by β-actin showed that the expression of AMPKα2 in hepatopancreas was significantly higher than that in the other four tissues. On the contrary, the expression analysis normalized by gene combination showed that AMPKα2 also had relatively high expression in the gill, and no significant difference was found between hepatopancreas and gill (Fig. 6a). Then, the expression pattern of AMPKα2 was further analyzed in the hepatopancreas of clams under salinity stress. As shown in Fig. 6c, the expression of AMPKα2 was significantly up-regulated at 12 h, and then gradually recuperated to the basal level. However, no significant change was observed in the expression of the AMPKα2 gene normalized by a single β-actin gene.

For IL17-1, the expression patterns of IL17-1 in different tissues obtained from the two normalization strategies were slightly different. The result normalized by the gene combination GAPDH-RPS18-β-actin showed that the expression level of IL17-1 in gill was the highest in comparison with the other tissues. However, the expression analysis normalized by β-actin showed that IL17-1 also had relatively high expression in the hepatopancreas, and no significant difference was found between the hepatopancreas and gill (Fig. 6b). The expression patterns of IL17-1 in clams post-Vibrio infection obtained from the two normalization strategies were similar, expression significantly increased at 2 d post-treatment (Fig. 6d). Besides, it should be noted that most of the standard deviation of normalization by a single reference was larger than that of normalization by multiple genes. Hence, we recommend that multiple reference genes should be used for the normalization of gene expression analysis.

4 DISCUSSION

qRT-PCR is one of the most widely used techniques for gene expression analysis in biological research. In relative quantitation, the stable reference gene is a prerequisite for accurate normalization of gene expression levels (Guénin et al., 2009). Previous studies have shown that there is no ideal reference gene that is appropriate for use in all experimental conditions (Vandesompele et al., 2002; Chapman and Waldenström, 2015). As a consequence, it is necessary to verify the expression stability of reference genes under specific experimental conditions. To date, numerous researches have been conducted on reference gene selection in aquatic animals, such as fishes (Li et al., 2020; Dharmaratnam et al., 2021; Szczygieł et al., 2021), crustaceans (Hu et al., 2018; Zhou et al., 2018), and mollusks (Huan et al., 2016; Volland et al., 2017; Zhao et al., 2018), and research on reference gene selection in clam C. sinensis is vacant. For the economic and ecological importance, the first genome sequencing for C. sinensis was performed in 2020, and 27 344 protein-coding genes were annotated, providing massive gene resources for gene expression studies and systematic reference gene validation in this clam (Wei et al., 2020). In the present study, five commonly used reference genes were selected from the genome and transcriptome datasets of C. sinensis, and their expression stabilities were evaluated. We revealed that the expression stabilities of these housekeeping genes were not consistent in different tissues and experimental treatments, and hence highlighted the need to verify the expression stability of reference genes for normalization.

There is no consensus on how many control genes were required for reliable normalization. In the past, the vast majority of studies used a single reference gene for normalization. As exact normalization is such a vital component of gene expression analysis, it has been strongly advocated that at least two validated reference genes be employed for normalization (Chapman and Waldenström, 2015). In this study, we selected the algorithm of geNorm to define the optimal number of reference genes required for accurate normalization in qRT-PCR. Among the five different tissues, the result showed that all V_n/n+1 values were higher than 0.15. This suggested that all the five candidate reference genes were not sufficient for accurate normalization according to the cut-off value of 0.15 proposed by Vandesompele et al. (2002). Similar to our results, all V_n/n+1 values over 0.15 were also observed in different tissues of other species, such as the prawn Macrobrachium olfersii (Jaramillo et al., 2017) and the octopus Octopus minor (Whang et al., 2020). This may be different tissues have different cell types and perform different functions, so various housekeeping genes showed variation in expression. Thus, it is not surprising that all V_n/n+1 values exceeded 0.15. The number of reference genes used for normalization is a trade-off between practical considerations and accuracy. When only a few target genes need to be studied or minimal amounts of cDNA are obtained, it is impractical to use many reference genes for normalization. Literature showed that, in most cases, three proper control genes are sufficient for a valid normalization strategy (Vandesompele et al., 2002). We noticed that the lowest V_n/n+1 value of different adult tissues was V_3/4 (0.345), indicating that using three selected control genes (GAPDH-RPS18-β-actin) for normalization could be acceptable.

The optimal reference genes should not be regulated through common upstream effectors or should not directly regulate each other, that is, co-regulated genes should not be used as reference genes (Kadegowda et al., 2009). This criterion is particularly important when the multiple reference genes were used for normalization. When evaluating gene expression stability, co-regulated genes would lead to distorted results (Huan et al., 2016). In this study, the five selected housekeeping genes were involved in three functional classes, including cytoskeletal protein (β-actin, TUB-α), translation (EF1-α, RPS18), and glycolysis (GAPDH). The geNorm analysis showed that the V_2/3 value (0.191) was the lowest, indicating that using two stable genes for normalization could be accepted in clams after Vibrio infection. Nevertheless, the top two genes ranked by all the three programs (EF1-α, RPS18) belong to the same functional class, and they might be co-regulated. In oyster C. gigas, Similarly, EF1-α and RPS18 were found to be stable during the development of oyster C. gigas, and these two reference genes were unlikely to be co-regulated (Huan et al., 2016). Even so, to be cautious, we also included GAPDH in the gene combination (EF1α-GAPDH-RPS18) for normalization in clam's post-Vibrio challenge. Consequently, if it is not determined to exclude the co-regulated genes, the selected multiple reference genes should belong to at least two functional categories to avoid errors caused by gene co-regulation.

Among these selected housekeeping genes in this study, β-actin is the most commonly used reference gene in aquatic organisms. Although some studies have shown that the expression level of β-actin is stable and could be used as an internal reference gene (Hu et al., 2018; Geng et al., 2019; Li et al., 2020), an increasing number of studies have shown that β-actin is a less stable housekeeping gene (Morga et al., 2010; Feng et al., 2013; Zhou et al., 2018). In this study, β-actin was highly ranked by NormFinder and BestKeeper among different adult tissues but poorly ranked in clams under salinity stress and Vibrio infection. Hence, it is likely that β-actin is stably expressed in normal tissues and has varied expression under experimental stress. In contrast to β-actin, EF1-α was poorly ranked by NormFinder and BestKeeper in different tissues but highly ranked in clams under salinity stress and Vibrio infection. EF1-α encodes the α subunit of eukaryotic elongation factor 1 and has been commonly used to study gene expression levels in bivalves (Morga et al., 2010; Moreira et al., 2014; Huan et al., 2016; Yan et al., 2017). Similar to our results, EF1-α was a less stable reference gene when analyzing adult tissues of C. gigas (Dheilly et al., 2011). RPS18 takes part in ribosome biogenesis in all cell types and was used as a reference gene for normalization in bivalve species (Mateo et al., 2010; Du et al., 2013). Our study validated that RPS18 could be selected as a control gene to study target gene expression in clams. GAPDH encodes a key enzyme in glucose metabolism and is a frequently used reference gene for normalization. We found the expression of GAPDH was relatively stable in our test and could be used as a reference gene for normalization in clams. Likewise, GAPDH and EF1-α have been combined as reference genes when studying expression levels in hemocytes of flat oyster Ostrea edulis (Morga et al., 2010). Taken together, the gene combinations GAPDH-RPS18-β-actin and EF1α-GAPDH-RPS18 were recommended to use for normalization in future studies for normal tissues and experimental stress, respectively.

To test the effect of reference gene choice on gene expression analysis, we analyzed the expression patterns of AMPKα2 and IL17-1 genes using two normalization strategies. AMPKα2 is a highly conserved protein kinase that plays an important role in cellular stress response. Several studies have shown that salinity stress can cause significant changes in AMPKα2 expression level (Xu et al., 2016; Zeng et al., 2016; Xiao et al., 2018). In the present study, the expression patterns in clams under salinity stress obtained from the two normalization strategies were different. The significant change was observed only in the expression of AMPKα2 normalized by reference gene combination. IL17-1 is a proinflammatory cytokine that plays a key role in innate immune responses. In the Pacific oyster C. gigas, IL17-1 was highly expressed in the gill tissues, and the expression of the IL17-1 was significantly up-regulated post pathogen challenge (Li et al., 2014). Similar results were also found in this study, although the expression patterns of IL17-1 in different tissues obtained from the two normalization strategies were slightly different. As a result, this study indicated that the normalization strategy could influences quantification results, especially when it comes to which genes are subtle changes in expression. Combined with the results of gene stability analysis, we recommend that the geometric mean of three reference genes be used for normalization in future gene expression study of C. sinensis.

5 CONCLUSION

Using three stable reference genes with different functional classes was most appropriate for the normalization strategy. We proposed to use the gene combination GAPDH-RPS18-β-actin for normalization among normal adult tissues, and another gene combination EF1α-GAPDH-RPS18 for normalization under salinity and Vibrio stress. This is the first study that provides the validated reference gene combinations for accurate normalization for specific experimental conditions of C. sinensis, which is the key to functional studies of target genes in this marine bivalve in the future.

6 DATA AVAILABILITY STATEMENT

All data generated and/or analyzed during this study are included in this published article and its supplementary information files.

Electronic supplementary material

Supplementary material (Supplementary File 1 and Supplementary Tables S1–S3) is available in the online version of this article at https://doi.org/10.1007/s00343-021-1335-z.