Chemosphere 243 (2020) 125415

Contents lists available at ScienceDirect

Chemosphere

journal homepage: www.elsevier.com/locate/chemosphere

CO2-driven acidification weakens shell defense capacity and induces global molecular compensatory responses

* Xinguo Zhao a, b, c, Yu Han c, Bijuan Chen a, b, Bin Xia a, b, Keming Qu a, Guangxu Liu c, a Fisheries Research Institute, Chinese Academy of Fishery Sciences, Qingdao, 266071, PR b Laboratory for Marine Ecology and Environment Science, Qingdao National Laboratory for Marine Science and Technology, Qingdao, 266237, PR China c College of Sciences, Zhejiang University, Hangzhou, 310058, PR China highlights graphical abstract

OA will damage shell structure, and reduce shell strength and shell closure strength. OA will lead to extracellular acidosis þ and Ca2 deficiency. OA will significantly alter gene expression profile in mantle tissue. OA will weaken ’ shell de- fense capacity, and thus reduce their fitness. The findings of this study have sig- nificant ecological and economic implications. article info abstract

Article history: Oceanic uptake of atmospheric CO2 is reducing seawater pH and shifting carbonate chemistry within, a Received 23 September 2019 process termed as ocean acidification (OA). Marine mussels are a family of ecologically and economically Received in revised form significant bivalves that are widely distributed along coastal areas worldwide. Studies have demon- 6 November 2019 strated that OA greatly disrupts mussels’ physiological functions. However, the underlying molecular Accepted 18 November 2019 responses (e.g., whether there were any molecular compensation mechanisms) and the extent to which Available online 19 November 2019 OA affects mussel shell defense capacity remain largely unknown. In this study, the thick shell mussels Handling Editor: Jim Lazorchak coruscus were exposed to the ambient pH (8.1) or one of two lowered pH levels (7.8 and 7.4) for 40 days. The results suggest that future OA will damage shell structure and weaken shell strength and Keywords: shell closure strength, ultimately reducing mussel shell defense capacity. In addition, future OA will also þ þ Ocean acidification disrupt haemolymph pH and Ca2 homeostasis, leading to extracellular acidosis and Ca2 deficiency. Mussel Mantle transcriptome analyses indicate that mussels will adopt a series of molecular compensatory Calcification responses to mitigate these adverse effects; nevertheless, weakened shell defense capacity will increase Acid-base status mussels’ susceptibility to predators, parasites and pathogens, and thereby reduce their fitness. Overall, Defense capacity the findings of this study have significant ecological and economic implications, and will enhance our Mantle transcriptome sequencing understanding of the future of the mussel aquaculture industry and coastal ecosystems. © 2019 Elsevier Ltd. All rights reserved.

1. Introduction

Anthropogenic activities (e.g., fossil fuel burning) emit large * Corresponding author. quantities of carbon dioxide (CO ) into the atmosphere. E-mail address: [email protected] (G. Liu). 2 https://doi.org/10.1016/j.chemosphere.2019.125415 0045-6535/© 2019 Elsevier Ltd. All rights reserved. 2 X. Zhao et al. / Chemosphere 243 (2020) 125415

Approximately one third of the CO2 is eventually absorbed by the show the physiological mechanisms underlying the observed al- global ocean (Sabine et al., 2004), leading to surface seawater pH terations in shell defense capacity. Importantly, owing to mantle’s reduction and carbonate chemistry shifts, a phenomenon known as critical roles in shell formation and growth, gene expression pro- ocean acidification (OA) (Caldeira and Wickett, 2003). Since the files of mantle tissue were investigated through transcriptome Industrial Revolution, surface seawater pH has already decreased sequencing to uncover the underlying molecular responses to OA. by 0.1 units, from approximately 8.21 to the current level of about 8.10 (Caldeira and Wickett, 2003). According to the Representative Concentration Pathway (RCP) 8.5 scenario of the Intergovern- 2. Materials and methods mental Panel on Climate Change (IPCC), the average surface seawater pH will further decrease by 0.3e0.4 and 0.7e0.8 units by 2.1. Animal collection and acclimation the end of the 21st and 23rd centuries, respectively (IPCC et al., 2014). Previous studies have suggested that OA poses a threat to Thick shell mussels M. coruscus (shell length of 21.91 ± 2.34 mm) a wide variety of marine organisms, especially calcifying organisms were collected from an intertidal site on Dongtou Island, Wenzhou, (Hofmann et al., 2010; Kerr, 2010). However, the responses of ma- China (121.22 E, 27.75 N), where the natural seawater pH ranges rine organisms to OA have been found to be -specific and to from 8.0 to 8.2. The mussels were directly transported to the vary with life stage (Kroeker et al., 2010, 2013; Ries et al., 2009). For Qingjiang Station of Zhejiang Mariculture Research Institute, gently example, OA leads to decreases in metabolic rates in the blood clam cleaned of epibionts without damaging the shells, and acclimated Tegillarca granosa (Zhao et al., 2017b), wherease it leads to increases for two weeks in filtered and UV-irradiated natural seawater with in metabolic rates in the blue mussel Mytilus edulis (Thomsen and þ continuous aeration prior to the experiments. The seawater pH was Melzner, 2010). Similarly, the Ca2 content of the red king crab controlled at 8.10 ± 0.06, the temperature was controlled at Paralithodes camtschaticus is increased at the larvae stage, but re- 23 ± 0.4 C, and the salinity was controlled at 23 ± 0.5‰. The mains unchanged at the juvenile stage upon OA exposure (Long mussels were fed twice daily with the microalgae Platymonas et al., 2013a, 2013b). Regarding the species-specific and life stage subcordiformis at a rate of ~5% dry tissue weight, according to variations, the current understanding of biological responses to OA O’Donnell et al. (2013). Excess food and feces were removed daily cannot be simply applied to other species and/or life stages. through seawater changes, in which seawater was removed by Therefore, more research is needed to increase our understanding siphoning, followed by refilling with seawater pre-equilibrated to of the effects of OA on marine organisms and ecosystems. the desired pH values. Marine mussels are a family of ecologically and economically significant bivalve species that are widely distributed along coastal areas worldwide. By aggregating into beds, marine mussels create habitats for other organisms, and are thus recognized as important 2.2. Experimental design and seawater parameters marine ecosystem engineers (Borthagaray and Carranza, 2007). Many of them are important aquaculture species and have been Following acclimation, the mussels were randomly assigned to human food for thousands of years (Ponce Oliva et al., 2019). Ac- an ambient pH (8.1) group and two lowered pH (7.8 and 7.4) groups. cording to a report by the Food and Agriculture Organization of the The ambient seawater pH level (8.1) served as the control, while the United Nations (FAO), the mussel aquaculture industry was worth lowered pH levels of 7.8 and 7.4 were set to mimic the oceanic approximately 4.0 billion USD in 2016 (FAO yearbook, 2018). surface pH conditions projected for the years 2100 and 2300, Therefore, increasing attention has been paid to understanding respectively (IPCC et al., 2014). The seawater pH levels were ach- how OA affects marine mussels. Current evidence demonstrates ieved and maintained by continuously bubbling with CO2 gas that OA will exert significant negative effects on various physio- mixture, obtained by mixing CO2-free air and pure CO2 gas at logical processes in marine mussels, especially leading to the controlled flow rates, according to Zhao et al. (2017a). All pH levels reduction in calcification and even to shell damage (Asplund et al., were conducted with 10 replicate chambers that contained 10 in- 2014; Fitzer et al., 2015; Li et al., 2015a; Melzner et al., 2011; Sadler dividuals each and filled with approximately 10 L of filtered and et al., 2018; Thomsen et al., 2010, 2013). However, previous studies UV-irradiated natural seawater with the desired pH value have mainly focused on the physiological effects of OA, leaving the (Table S1). The mussels were fed with the microalgae underlying molecular responses (e.g., whether there were any P. subcordiformis as described above. The seawater was maintained molecular compensation mechanisms) largely overlooked. In at 23 C using temperature regulators, and was changed daily. addition, previous studies have mostly been performed with the Seawater chemistry parameters, including the pH, salinity, total blue mussel M. edulis. Considering the species-specific variations in alkalinity (TA) and carbonate system parameters, were monitored responses to OA mentioned above, whether findings in other spe- daily to ensure that no substantial fluctuations occurred cies also apply to mussel species such as M. coruscus needs to be throughout the experiment period. Seawater pH levels were investigated. Moreover, although shell defense capacity is crucial measured using a Sartorius PB-10 pH meter (Sartorius, Germany) for the survival of individual mussels, the extent to which OA affects calibrated with standard NBS buffers. Salinity was determined with this capacity remains largely unknown. a Multi 3410 conductivity meter (WTW, Germany). TA was assessed The present study was aimed to minimize these knowledge through potentiometric titration with an SM-Titrino 702 automatic gaps. It was performed with the thick shell mussel M. coruscus,an titrator system (Metrohm, Switzerland). The carbonate system ecologically and economically important species that is widely parameters, including CO2 partial pressure (pCO2), dissolved inor- distributed along the coastal areas of China, and Japan (Qi ganic carbon (DIC), and aragonite saturation state (Uara) and calcite et al., 2019; Shang et al., 2019). Following 40-day exposure of M. saturation state (Ucal), were calculated using the open-source coruscus to ambient condition (pH 8.1) or one of two acidified program CO2SYS (Pierrot et al., 2006) with the pH, salinity, tem- conditions (pH 7.8 and 7.4), the shell structure, shell strength and perature, and TA values and the established constants (Dickson, relative size of the posterior adductor muscle (a proxy for shell 1990; Dickson and Millero, 1987; Mehrbach et al., 1973). The closure strength) were measured to investigate potential alter- experiment lasted for 40 days. During the entire experimental ations in shell defense capacity upon OA exposure. In addition, the period, no mussel mortality was observed. The seawater parame- þ haemolymph pH and Ca2 concentration were also determined to ters of the experimental trials are summarized in Table S2. X. Zhao et al. / Chemosphere 243 (2020) 125415 3

2.3. Shell integrity and structure analysis concentrations were measured using a WFX-130A flame atomic spectrophotometer (Beijing Rayleigh Analytical Instruments Co, Following the 40-day exposure, the mussels were carefully Ltd, China) according to Shi et al. (2016). dissected without damaging the shell surfaces. The shells were carefully cleaned and dried at room temperature. The internal and 2.7. Mantle transcriptome sequencing, analysis and validation external shell surfaces were checked and photographed using a digital camera. To assess the severity of corrosion, the percentages One mussel from each replicate chamber was randomly selected of corroded areas for both the internal and external shell surfaces and dissected (for a total of ten mussels per pH level). The whole were quantified by image analysis using the open-access software mantle tissue of the mussel was isolated and immediately frozen in ImageJ version 1.46r (http://imagej.nih.gov/ij/). The curvature of liquid nitrogen. Total RNA was extracted using TRIzol Reagent the shell was neglected in image analysis, according to Melzner (Invitrogen, 15,596,018) following the manufacturer’s protocol. The et al. (2011). The severity of the external shell surface corrosion RNA sample was further treated with DNase I (Invitrogen, was further assessed by quantifying the occurrences of each type of 18047019) to remove DNA contamination. The quality of the RNA corrosion, including periostracum discoloration, periostracum sample was checked by 1.0% formaldehyde-denatured agarose gel breakage and lifting, and prismatic layer dissolution. For micro- electrophoresis. The concentration of the RNA sample was quanti- structure analysis, target shell regions (Fig. S1) were fragmented fied using a NanoDrop 1000 spectrophotometer (Thermo Scienti- along defined trajectories. Sections were mounted on pedestal fic). For each pH level, the ten mantle RNA samples were pooled in stubs, coated with gold-palladium, and observed by a scanning equal proportions to obtain a mixture. The standard Illumina pro- electron microscopy (SEM; SU8010, Hitachi, Japan). tocol was followed for cDNA synthesis and library construction. Sequencing was performed by Shanghai OE Biotech Co, Ltd. 2.4. Shell strength measurement (Shanghai, China) on an Illumina HiSeq2000 platform. The raw sequence reads were deposited in the Sequence Read Archive (SRA) The left and right valves of five mussels from each replicate at the National Center for Biotechnology Information (NCBI) with chamber (i.e., fifty mussels per pH level) were randomly chosen, accession number PRJNA543748. The raw sequence reads were and then shell strength measurement was performed following trimmed by removing adapters and low quality sequences. The Burnett and Belk (2018). Briefly, a universal material-testing ma- quality of sequence reads was verified using the software FASTQC chine (AGS-J, Shimadzu, Japan) was used to determine shell (http://www.bioinformatics.babraham.ac.uk/projects/fastqc/). De strength. Each shell valve was placed between the horizontal jaws novo transcriptome assembly was carried out using the software of the testing machine, as shown in Fig. S2. The valve was com- Trinity (Grabherr et al., 2011). The longest transcript of each locus pressed at a constant loading rate of 10 mm/min until failure was defined as an unigene and was used for downstream analyses. occurred, and the applied force was continuously recorded by a The unigenes were annotated by alignment to the Non-redundant computer. Shell strength was measured as the force required to (Nr), Swiss-Prot, Clusters of Orthologous Groups of proteins/ break the valve and is expressed in Newtons (N). In other words, euKaryotic Orthologous Groups (KOG/COG), Gene Ontology (GO) shell strength is the maximum force that the shell valve could and Kyoto Encyclopedia of Genes and Genomes (KEGG) databases endure. For each mussel, the mean of the strengths of the left and using BlastX (E value < 10 5). The expression level of each unigene right valves was recorded as the shell strength. was calculated using the fragments per kilo bases per million mapped reads (FPKM) method (Trapnell et al., 2010). Differentially 2.5. Adductor muscle size analysis expressed unigenes (DEGs) between the control (pH 8.1) and acidified (pH 7.4) conditions were identified using the DESeq R Owing to the significant positive linear correlation between the package (Anders and Huber, 2010). Only unigenes with an absolute relative size of the posterior adductor muscle and the shell closure fold change ˃ 2 and a p value < 0.05 were considered true DEGs. strength (Christensen et al., 2012; Thomas, 1976). The relative size Finally, the DEGs were subjected to GO enrichment and KEGG of the posterior adductor muscle was used a proxy for the shell pathway analyses as described by Altermann and Klaenhammer closure strength of the mussel, and was determined following (2005) and Ashburner et al. (2000) to determine the overall im- Christensen et al. (2012). Briefly, the posterior adductor muscle was pacts of the experimental manipulation. The p values were adjusted severed with a scalpel along the plane of the shell edge, and its by the Benjamini-Hochberg false discovery rate (FDR). An FDR- diameter was measured using a digital Vernier calliper with a adjusted p value < 0.01 was selected as the threshold to identify precision of 0.01 mm. The relative size of the posterior adductor the most representative enriched KEGG and GO terms. In addition muscle was calculated as the ratio of the posterior adductor muscle to GO and KEGG terms/pathways, the DEGs were also classified into diameter to the shell length. three specific categories, including the ion and acid-base regula- tion, calcification, and adductor muscle categories, to reveal the þ 2.6. Determination of haemolymph pH and Ca2 concentration specific responses of M. coruscus to OA, according to previous studies (Li et al., 2016a; Liao et al., 2015; Moya et al., 2016; Mussel haemolymph samples were collected by pericardial Nagasawa, 2013; Zhang et al., 2012). puncture with gas-tight disposable syringes, and directly trans- To validate the transcriptome expression data, the expression ferred to 1.5 mL sterile tubes. Haemolymph samples of five mussels levels of fifteen representative DEGs were determined through from each replicate chamber were pooled in equal proportions and real-time quantitative PCR (RT-qPCR) using the same RNA samples used as a biological replicate (i.e., ten biological replicates were as those used for transcriptome sequencing. The degree of agree- tested per pH level). The haemolymph samples were centrifuged at ment between the transcriptome expression data and the RT-qPCR 5000g for 5 min. For haemolymph sampling and handling, the data was estimated through linear regression, according to Li et al. temperature was controlled at 23 C (i.e., the seawater temperature (2016a). RT-qPCR was performed on a CFX96™ Real-Time System during mussel incubation period). The supernatants (cell-free (Bio-Rad). The relative expression levels of the fifteen representa- þ DD haemolymph) were used to measure the pH values and Ca2 con- tive DEGs were calculated using the 2 Ct method (Livak and centrations following Asplund et al. (2014). The pH values were Schmittgen, 2001) with the 18S rRNA gene as the internal refer- þ determined with a pH microelectrode (WTW, Germany). The Ca2 ence (Hüning et al., 2013; Zhao et al., 2017a). Detailed information 4 X. Zhao et al. / Chemosphere 243 (2020) 125415 of the fifteen representative DEGs and the gene-specific primers is Person’sr¼ 0.79, R2 ¼ 0.61, p < 0.0001) (Fig. 2c), indicating that listed in Table S3. internal and external shell surface corrosion occurred concurrently.

2.8. Statistical analysis 3.2. Shell strength and the relative size of the adductor muscle

One-way ANOVA was performed to show the effects of OA on OA significantly weakened shell strength (one-way ANOVA, the percentage of corroded shell surface area, shell strength, rela- df ¼ 2, p < 0.01). Shell strength was significantly decreased from tive size of the posterior adductor muscle, haemolymph pH and 86 N in the control group to 56 N and 50 N in the pH 7.8 and pH 7.4 þ haemolymph Ca2 concentration. Post-hoc Tukey’s multiple tests groups, respectively (Fig. 3a). OA also significantly decreased the were conducted to compare differences among groups. The relative size of the posterior adductor muscle (one-way ANOVA, normality assumption and homogeneity of variance were verified df ¼ 2, p < 0.01). The relative size was significantly decreased from prior to analysis using Shapiro-Wilk’s test and Bartlett’s test, 0.154 at pH 8.1 to 0.147 and 0.133 at pH 7.8 and 7.4, respectively respectively. A Chi-square test was performed to compare the oc- (Fig. 3b). currences of the different types of external shell surface corrosion þ at the three pH levels. A p value < 0.05 was considered to indicate 3.3. Haemolymph pH and Ca2 concentration statistically significance. The Chi-square test was conducted using the software GraphPad Prism 5, while the other statistical analyses The haemolymph pH level was significantly decreased by OA were performed with OriginPro 8.0. (one-way ANOVA, df ¼ 2, p < 0.01). It was significantly reduced from 7.52 in the control group to nearly 7.38 and 7.24 in the pH 7.8 3. Results and pH 7.4 groups, respectively (Fig. 4a). Similarly, the haemolymph þ Ca2 concentration was significantly reduced by OA (one-way þ 3.1. Shell integrity and structure ANOVA, df ¼ 2, p < 0.01). In control mussels, the haemolymph Ca2 concentration was maintained at 297 mg/L, while in mussels For mussels under ambient condition (pH 8.1), the internal shell exposed to pH 8.1 and 7.8, the concentration was reduced to surface was intact with a typical glossy appearance, and the 283 mg/L and 278 mg/L, respectively (Fig. 4b). external shell surface showed only periostracum discoloration (Fig. 1a). In contrast, for mussels under acidified conditions (pH 7.8 3.4. Mantle transcriptomic responses to OA and 7.4), the typically glossy internal shell surface had become dull and white (Fig. 1a). In addition, the periostracum was broken, lifted Transcriptome sequencing of mantle tissues yielded 36, 539, or even absent extending from the umbo region to the shell margin 606 clean reads with a Q30 of 90.12% and 47, 196, 236 clean reads on the external surface (Fig. 1a). SEM showed that the aragonite with a Q30 of 90.29% for the control (pH 8.1) and acidified (pH 7.4) tablets of the nacreous layer and the calcite crystals of the prismatic groups, respectively (Table S4). De novo assembly generated layer on the normal internal shell surface had uniform structural 205,936 transcripts with an average length of 1078 bp and an N50 orientations, while those of mussels under acidified conditions of 2058 bp, which represented 109,823 unigenes with an average appeared to be disorientated or dissolved (Fig. 1b). The intact per- length of 700 bp and an N50 of 1237 bp (Table S5). These results iostracum was purple-black to black, while the discolored perios- suggested that the results of transcriptome sequencing and de novo tracum had become beige (Fig. 1a). SEM illustrated that the intact assembly for the mantle tissues of M. coruscus were of high quality region of periostracum had a smooth appearance, while the dis- and were reliable. The mantle gene expression profiles of colored region had become rough, suggesting slight damage to the M. coruscus were significantly altered by OA. A total of 2448 unig- periostracum surface microstructure (Fig. 1b). Additionally, SEM enes were found to be differentially expressed in acidified (pH 7.4) revealed that periostracum breakage and lifting ultimately surfaced condition relative to control condition (pH 8.1), including 1624 up- the prismatic layer, and led to dissolution of calcite crystals on regulated DEGs and 824 down-regulated DEGs (Fig. S3). Linear external surfaces (Fig. 1b). fitting indicated that there was a strong positive linear correlation OA significantly increased the corroded area of the internal shell between the expression data generated by transcriptome surface (one-way ANOVA, df ¼ 2, p < 0.01). The corroded area of sequencing analysis and that generated by RT-qPCR (R2 ¼ 0.9233, inner shell surface (%) was significantly increased from 0% in the p < 0.01), validating the reliability and accuracy of the tran- control group to 14% and 43% in the pH 7.8 and pH 7.4 groups, scriptome sequencing data (Fig. S4). respectively (Fig. 2a). Similarly, the corroded area of the external GO enrichment analyses revealed that among the up-regulated shell surface was significantly affected by OA (one-way ANOVA, DEGs, a total of 15 GO terms were significantly enriched, including df ¼ 2, p < 0.01); the corroded area (%) was approximately 26% at pH three terms in the “Cellular Component” category, seven terms in 8.1, while it had increased to 31% and 47% at pH 7.8 and pH 7.4, the “Molecular Function” category, and five terms in the “Biological respectively (Fig. 2b). Additionally, the severity of external shell Process” category (Table 1). In contrast, only eight GO terms were surface damage (measured as the occurrence of each type of significantly enriched among the down-regulated DEGs, including external shell surface corrosion) was also siginifcantly increased by two terms in the “Cellular Component” category, two terms in the OA (c2 ¼ 189.9, df ¼ 4, p < 0.0001). As shown in Fig. 2d, most (58%) “Molecular Function” category, and four terms in the “Biological of the analyzed mussels at pH 7.8 showed prismatic layer dissolu- Process” category (Table 1). KEGG enrichment analyses suggested tion (the most severe type of external shell surface damage), and that seven pathways were significantly enriched among the up- the others (42%) showed periostracum breakage and lifting; regulated DEGs, including “focal adhesion”, “apoptosis”, “tight furthermore, 100% of the analyzed mussels at pH 7.4 showed junction”, “ECM-receptor interaction”, “NF-kappa b signaling prismatic layer dissolution. In contrast, the mussels at pH 8.1 pathway”, “cardiac muscle contraction”, and “toll-like receptor showed only periostracum discoloration (the least severe type of signaling pathway” (Table 2). However, no pathway was signifi- external shell surface damage). These results suggest that the cantly enriched among the down-regulated DEGs (Table 2). Among severity of shell damage increased with increasing OA. Moreover, the DEGs, 26 were potentially involved in ion and acid-base regu- there was a signifcant positive linear correlation between inner and lation, including 23 up-regulated DEGs and three down-regulated external shell surface corrosion (linear fitting, y ¼ 0.46x þ 25.92, DEGs (Table 3 and Table S6). Additionally, a total of 35 DEGs X. Zhao et al. / Chemosphere 243 (2020) 125415 5

Fig. 1. Inner and external shell surfaces images of M. coruscus after 40 days treatment. (a) Representative stereomicroscopic images; (b) Representative SEM images. Capital letters indicate where the SEM images were taken. A: the normal nacreous layer on the internal shell surface; B: the corroded nacreous layer on the internal shell surface; C: the normal prismatic layer on the internal shell surface; D: the corroded prismatic layer on the internal shell surface; E: the normal region of periostracum on the external shell surface; F: the discoloration region of periostracum on the external shell surface; G: the breakage and lifting region of periostracum on the external shell surface; H: the periostracum loss and prismatic layer dissolution region on the external shell surface. putatively participating in calcification were identified; 24 and 11 of suggested that microstructures of the periostracum, prismatic layer them were up- and down-regulated in acidified condition (pH 7.4), and nacreous layer were damaged by OA. Notably, the external shell respectively (Table 3 and Table S7). Finally, a total of 28 DEGs surfaces of M. coruscus were also slightly damaged at the umbo potentially involved in adductor muscle function were identified; region under ambient condition, displaying periostracum dis- all of them (28) were up-regulated (Table 3 and Table S8). colouration (a change from purple-black/black to beige), which commonly occurs among field mussels. A previous study revealed that periostracum discolouration is caused by the mutual friction of 4. Discussion individuals in a wave-swept environment (Thomsen et al., 2010). The mutual friction of mussels in the field therefore hampers the 4.1. OA impaired shell formation and maintenance capacity, and protective function of the periostracum and acts as an accelerator eventually damaged shell structure for external shell surface damage. Nevertheless, this study suggests that OA markedly impairs the shell formation and maintenance The results revealed that OA simultaneously corroded the in- capacity of M. coruscus and ultimately damages shell structure. ternal and external shell surfaces of M. coruscus, and that the These findings are consistent with those reported in other marine severity increased with increasing seawater acidity. The internal bivalve species, such as the blood clam T. granosa (Zhao et al., shell surfaces lost their typical glossy appearances and became 2017b), the striped venus clam Chamelea gallina (Bressan et al., white, while the external shell surfaces displayed obvious breakage 2014), the Mediterranean mussel M. galloprovincialis (Gazeau and lifting of the periostracum and even prismatic layer dissolution et al., 2014) and the blue mussel M. edulis (Gazeau et al., 2007; extending from the umbo region to the shell margin. SEM analysis 6 X. Zhao et al. / Chemosphere 243 (2020) 125415

Fig. 2. Effects of ocean acidification on shell intergity of M. corucus. (a) Corroded area of inner shell surface (%); (b) Corroded area of external shell surface (%); (c) Linear positive correlation between the corroded area of inner shell surface (%) and corroded area of external shell surface (%) of mussels at pH 7.8 and 7.4; (d) Occurrence (%) of each type of external shell surface corrosions (50 samples per pH level, c2 ¼ 189.9, df ¼ 4, p < 0.0001). Means not sharing the same superscript are significantly different (n ¼ 10, Tukey’s HSD, p < 0.05). Error bar represents SD.

Fig. 3. Effects of ocean acidification on (a) shell strength and (b) the relative size of the posterior adductor muscle of M. coruscus. Means not sharing the same superscript are significantly different (n ¼ 10, Tukey’s HSD, p < 0.05). Error bar represents SD. X. Zhao et al. / Chemosphere 243 (2020) 125415 7

þ Fig. 4. Effects of ocean acidification on (a) haemolymph pH and (b) haemolymph Ca2 concentration of M. coruscus. Means not sharing the same superscript are significantly different (n ¼ 10, Tukey’s HSD, p < 0.05). Error bar represents SD.

Table 1 GO enrichment table (pH 7.4 vs pH 8.1). CC: Cellular Component. MF: Molecular Function. BP: Biological Process. Up: terms enriched in the set of up-regulated DEGs in pH 7.4. Down: terms enriched in the set of down-regulated DEGs in pH 7.4.

regulation category GO ID GO term No. of DEGs adjusted p-value

Up CC GO:0016021 Integral component of membrane 126 8.06E-03 GO:0005576 Extracellular region 42 3.91E-06 GO:0005938 Cell cortex 12 1.03E-03 MF GO:0005509 Calcium ion binding 98 6.45E-16 GO:0005516 Calmodulin binding 15 1.18E-04 GO:0004713 Protein tyrosine kinase activity 12 1.11E-04 GO:0004197 Cysteine-type endopeptidase activity 12 5.99E-03 GO:0002020 Protease binding 10 3.14E-04 GO:0008061 Chitin binding 10 1.30E-02 GO:0051015 Actin filament binding 7 6.55E-03 BP GO:0007155 Cell adhesion 20 9.48E-03 GO:0007275 Multicellular organismal development 15 6.94E-03 GO:0043123 Positive regulation of I-kappa b kinase/NF-kappa b signaling 12 1.91E-03 GO:0051592 Response to calcium ion 8 5.79E-05 GO:0048085 Adult chitin-containing cuticle pigmentation 2 3.40E-03 Down CC GO:0005576 Extracellular region 19 1.36E-07 GO:0005578 Proteinaceous extracellular matrix 7 1.84E-05 MF GO:0003714 Transcription corepressor activity 4 1.55E-03 GO:0008134 Transcription factor binding 4 9.02E-03 BP GO:0006030 Chitin metabolic process 6 1.63E-04 GO:0051260 Protein homooligomerization 5 2.22E-03 GO:0045892 Negative regulation of transcription, DNA-templated 5 3.25E-03 GO:0006310 DNA recombination 5 4.42E-03

Table 2 KEGG enrichment table (pH 7.4 vs pH 8.1). Up: pathways enriched in the set of up-regulated DEGs in pH 7.4. Down: pathways enriched in the set of down-regulated DEGs in pH 7.4.

regulation KEGG ID KEGG term No. of DEGs adjusted p-value

Up ko04510 Focal adhesion 29 1.52E-05 ko04210 Apoptosis 16 1.20E-03 ko04530 Tight junction 13 2.93E-03 ko04512 ECM-receptor interaction 12 1.34E-02 ko04064 NF-kappa b signaling pathway 11 4.69E-03 ko04260 Cardiac muscle contraction 9 6.52E-03 ko04620 Toll-like receptor signaling pathway 9 9.90E-03

þ Thomsen et al., 2010). It is highly likely that shell formation and 4.2. OA disrupted haemolymph pH and Ca2 homeostasis maintenance capacity impairment and therefore shell structural damage are widespread responses of marine bivalves to OA. In the present study, the haemolymph pH levels of M. coruscus were markedly reduced by OA, indicating disruption of 8 X. Zhao et al. / Chemosphere 243 (2020) 125415

Table 3 Summary of the genes responsive to acidified condition (pH 7.4).

category gene family No. of unigenes regulation

þ þ Ion- and acid-base regulation Ca2 /Mg2 -permeable cation channels 5 up þ Ca2 -modulated nonselective cation channel 1 up þ L-type voltage-dependent Ca2 channel 3 up Calmodulin 5 up Calcium-binding protein 3 up Carbonic anhydrase 1 down Sulfate/bicarbonate/oxalate exchanger SAT-1 1 down Electroneutral sodium bicarbonate exchanger 1 1 down þ þ Na /K ATPase, beta subunit 1 up Potassium channel subfamily K member 18 1 up Sodium/glucose cotransporter 2 up Sodium/myo-inositol cotransporter 1 up Solute carrier organic anion transporter family member 2B1 1 up Calcification Asparagine-rich protein 1 down Calponin-like protein 1 up Caltractin 1 up Cartilage matrix protein 1 up Chitin synthase 2 up Calmodulin 5 up Calcium-binding protein 3 up Carbonic anhydrase 1 down Lectin 9 6 up, 3 down Mucin 3 2 up, 1 down Nacre protein 1 down Perlucin 2 down Perlwapin 1 down Pif 1 down Tyrosinase 2 up Shell matrix protein 1 down Adductor muscle Dynein 1 up Filamin-A 1 up Filamin-C 2 up Myosin 9 up Paramyosin 1 up Titin 13 up Transgelin 1 up

extracellular acid-base homeostasis and therefore induction of et al., 2017; Sun et al., 2016) and C. gigas (Ginger et al., 2013)is extracellular acidosis upon OA exposure. Similar effects have also significantly suppressed by OA. In addition, such inference is sup- been detected in other marine bivalve species, such as the Pacific ported by studies on the pearl oyster Pinctada fucata, in which OA oyster Crassostrea gigas (Lannig et al., 2010), the blood clam conversely elevates filter-feeding behavior (Liu and He, 2012) while þ T. granosa (Zhao et al., 2017b), the blue mussel M. edulis (Mangan increasing haemolymph Ca2 concentration (Li et al., 2015b) and et al., 2017; Ramesh et al., 2017; Thomsen et al., 2013) and the shell CaCO3 crystal dissolution (Li et al., 2016a; Liu et al., 2017). We þ Mediterranean mussel M. galloprovincialis (Michaelidis et al., 2005). therefore suggest that OA inhibits Ca2 uptake and thereby de- þ Notably, a linear relationship between haemolymph pH and creases the haemolymph Ca2 concentration in M. coruscus. seawater pH has even been found in the blue mussels M. edulis (Thomsen et al., 2013). It seems that extracellular acidosis is also a widespread response of marine bivalves to OA. 4.3. OA induced potential mantle tissue injury, but also triggered Previous studies have reported that extracellular acidosis sub- compensatory responses sequently dissolves CaCO3 crystals, leading to the observed corro- sion on inner shell surfaces (Lindinger et al., 1984; Michaelidis et al., Genes involved in the “apoptosis” KEGG pathway were signifi- 2005). In addition to HCO- 3, dissolution of shell CaCO3 crystals also cantly up-regulated, suggesting that OA induced mantle cell death þ þ releases free Ca2 , which is predicted to increase the Ca2 level in in M. coruscus. This finding was supported by GO enrichment an- haemolymph (Michaelidis et al., 2005). However, the haemolymph alyses, in which the “cysteine-type endopeptidase activity” and þ Ca2 concentrations of M. coruscus were found to be significantly “protease binding” terms in the “Molecular Function” category reduced in this study with OA. Similar effects have been observed in were significantly enriched for the up-regulated DEGs, and the the blood clam T. granosa (Zhao et al., 2017b). A paradox between “protein homooligomerization” and “DNA recombination” terms in theoretical speculation and empirical observation has also been the “Biological Process” category were significantly enriched for the revealed in studies on the blue mussel M. edulis (Asplund et al., down-regulated DEGs. Endopeptidase and protease are critical 2014; Thomsen et al., 2010) and the Pacific oyster C. gigas (Lannig enzymes catalyzing protein degradation, which is an important et al., 2010), in which OA caused obvious dissolution of shell process of cell death (Lecker et al., 2006). Therefore, the up- 2þ “ CaCO3 crystals without significantly changing haemolymph Ca regulation of genes involved in cysteine-type endopeptidase ac- þ concentrations. This paradox may be attributable to reduced Ca2 tivity” and “protease binding” indicates that OA triggered protein uptake from food and seawater, because it has been demonstrated degradation and thereby cell death. Additionally, the down- that the filter-feeding behavior of M. coruscus (Sui et al., 2016; regulation of genes involved in “protein homooligomerization” Wang et al., 2015), T. granosa (Zhao et al., 2017b), M. edulis (Stapp suggests that OA inhibited the formation of protein quaternary structures, leading to protein dysfucntion and ultimately cell death. X. Zhao et al. / Chemosphere 243 (2020) 125415 9

Furthermore, the down-regulation of genes related to “DNA formation, including calponin-like protein, caltractin, cartilage recombination” indicates that OA hindered homologous recombi- matrix protein and calmodulin, were significantly up-regulated, nation repair of DNA damage, which in turn increased the accu- indicating increased intracellular amorphous CaCO3 formation. mulation of DNA damage and eventually resulted in cell death We thus conclude that M. coruscus enhanced intracellular amor- (Zada et al., 2019). These data suggest that OA induced mantle phous CaCO3 formation to alleviate the negative effects on shell tissue injury in M. coruscus. formation. Notably, GO enrichment analyses revealed the up-regulation of Generally, the formed amorphous CaCO3 is transported onto the genes associated with the “integral component of membrane”, growing shell and eventually transforms into CaCO3 crystals (i.e., “extracellular region” and “cell cortex” terms in the “Cellular crystalline aragonite and calcite) (Radha et al., 2010; Weiss et al., Component” category, and with the “cell adhesion” term in the 2002). However, the gene expression levels of shell matrix pro- “Biological Process” category under OA conditions. These findings teins regulating CaCO3 crystal formation, growth and orientation are in agreement with the results of KEGG analyses, which revealed (Nagasawa, 2013; Suzuki et al., 2009), including asparagine-rich that the “focal adhesion”, “tight junction” and “ECM-receptor protein, nacre protein, perlucin, perlwapin, pif and shell matrix interaction” pathways were significantly enriched for the up- protein, were significantly down-regulated, suggesting that the regulated genes. Up-regulation of genes related to cell adhesion transformation of amorphous CaCO3 and subsequent orientation of has also been detected in the pearl oyster P. fucata (Li et al., 2016a). CaCO3 crystals were hampered by OA. This conclusion is in accor- These results indicate the enhancement of cell-matrix adhesion, dance with our SEM data and with the findings of previous studies which plays essential roles in important biological processes on the blue mussels M. edulis (Fitzer et al., 2014a, 2014b, 2016) and including cell motility, cell proliferation, cell differentiation, cell the pearl oyster P. fucata (Li et al., 2016a), showing disorientated survival and cell communication. Additionally, GO enrichment aragonite tablets and calcite crystals on internal shell surfaces. analyses also revealed the up-regulation of genes involved in Therefore, the down-regulation of these shell matrix proteins “multicellular organismal development” term in the “Biological might be a possible explanation for the observed corrosion on the Process” category, suggesting positive effects of OA on cell prolif- internal shell surface. eration and differentiation. Furthermore, the GO terms “transcrip- This study revealed significant up-regulation of tyrosinase in tion corepressor activity” and “transcription factor binding” in the M. coruscus under OA conditions. Similar effects have also been “Molecular Function” category, and “negative regulation of tran- observed in the blue mussel M. edulis (Hüning et al., 2013) and the scription, DNA-templated” in the “Biological Process” category Mediterranean pteropod Heliconoides inflatus (Moya et al., 2016). were significantly enriched for the down-regulated DEGs, indi- Previous studies have also demonstrated the distinctive roles of cating the promotion of DNA transcription and protein synthesis by tyrosinase in forming the shell periostracum (Nagasawa, 2013). OA, with positive effects on cell proliferation. Therefore, these data Thus, the up-regulation of tyrosinase could be a compensatory suggest that there were compensatory responses against mantle response to attenuate periostracum damage. We also observed tissue injury. significant up-regulation of chitin synthase, which catalyzes chitin Together, the findings of this study reveal that OA induced synthesis, consistent with the results of previous studies on the mantle tissue injury and compensatory responses. In addition to blue mussel M. edulis (Hüning et al., 2013) and the Chilean scallop calcification, the mantle also plays essential roles in reproduction Argopecten purpuratus (Ramajo et al., 2016). GO enrichment ana- and sensory processes (Gosling, 2015). Structural damage in mantle lyses showed that “chitin binding” in the “Molecular Function” tissue could lead to global physiological and behavioral alterations. category and “adult chitin-containing cuticle pigmentation” in the We therefore suggest that future work should endeavor to deter- “Biological Process” category were significantly enriched among mine the extent to which OA affects mantle tissue. the up-regulated DEGs, while “chitin metabolic process” in the “Biological Process” category was significantly enriched among the 4.4. The molecular mechanism and compensatory responses down-regulated DEGs. These results suggest the promotion of underlying shell damage chitin synthesis and the inhibition of chitin degradation in OA- exposed M. coruscus. Chitin plays critical roles in creating the þ þ þ Genes encoding Ca2 channels, including Ca2 /Mg2 -permeable organic framework of mollusk shells (Schonitzer and Weiss, 2007). þ cation channel, Ca2 -modulated nonselective cation channel and L- Therefore, alterations in these processes and functions might be þ type voltage-dependent Ca2 channel, were significantly up- adaptive responses of M. coruscus to cope with OA. þ regulated, suggesting that extracellular Ca2 influx was evoked It should be noted that among the shell matrix proteins iden- þ under OA conditions. Additionally, genes involved in Ca2 regula- tified in this study, the respective transcripts encoding lectin and tion, including calmodulin and calcium-binding protein, were also mucin (two families of shell matrix proteins) (Moya et al., 2016; þ up-regulated, indicating that much of the internalized Ca2 was Nagasawa, 2013), were differentially regulated. Some were down- temporarily stored and released upon use. This conclusion is sup- regulated, while others were up-regulated by OA. In addition to ported by the results of GO enrichment analyses, in which the palying important roles in the calcification process, these two terms “calcium ion binding” and “calmodulin binding” in the protein families also participate in cell-cell adhesion, immune re- “Molecular Function” category, and “response to calcium ion” in the sponses and signal transduction (Drickamer, 1999; Strous and “Biological Process” category were significantly enriched for the Dekker, 1992). It is therefore likely that different isoforms of the þ up-regulated DEGs. For bivalves, intracellular Ca2 not only par- two protein families played different roles in the responses of ticipates in intracellular signal transduction but also drives the M. coruscus to OA and thus were differentially regulated. intracellular formation of amorphous CaCO3, which is the precursor material for shell formation (Li et al., 2016b; Mount et al., 2004; 4.5. Energy-consuming compensatory responses were mounted to þ Xiang et al., 2014). Therefore, Ca2 influx might promote the for- maintain haemolymph ion and acid-base homeostasis mation of amorphous CaCO3. This possibility is consistent with the þ þ þ þ findings of a previous study (DeCarlo et al., 2018), showing that the The gene expression levels of Na /K -ATPase, K channel, Na - þ coral Pocillopora damicornis increases the transport of Ca2 into related cotransporters and solute carrier organic anion transporter calcification sites and thereby increases CaCO3 formation to resist were significantly up-regulated, which might have facilitated the OA. Interestingly, the expression levels of genes regulating CaCO3 establishment of a transmembrane electrochemical potential 10 X. Zhao et al. / Chemosphere 243 (2020) 125415 gradient and in turn promoted the transmembrane transport of a snail Nucella lapillus (Sherker et al., 2017). variety of ions and small molecules. This result indicates a potential Generally, bivalves close their shells tightly upon encountering compensatory response of M. coruscus to maintain ion and acid- predators and/or other threats. Therefore, shell closure strength base homeostasis. This conclusion is in agreement with previous (i.e., the force required to open the shell) has been adopted as studies on the pearl oyster P. fucata (Li et al., 2016a) and the yesso another indicator of bivalves’ shell defense capacity (Aoki et al., scallop Patinopecten yessoensis (Liao et al., 2019) reporting a similar 2010). In this study, shell closure strength was estimated using compensation mechanism. However, the gene expression level of the relative size of the posterior adductor muscle as a proxy, carbonic anhydrase was significantly down-regulated by OA. because there is a significant positive linear correlation between Down-regulation of carbonic anhydrase has also been observed in these two parameters (Christensen et al., 2012; Thomas, 1976). For the blue mussel M. edulis under OA conditions at 22 C and 25 C(Li mussels exposed to pH 7.8 and 7.4, the relative size of the posterior et al., 2015a). Carbonic anhydrase is an enzyme that catalyzes the adductor muscle was significantly decreased to approximately 95% inter conversion between CO2 and HCO- 3 and thus plays important and 86% of that in control mussels, respectively. We therefore roles in acid-base regulation (Wang et al., 2017). Down-regulation deduced that the shell closure strength of M. coruscus was signifi- of carbonic anhydrase should therefore lead to decreased inter- cantly decreased under OA conditions. This finding is consistent conversion between CO2 and HCO- 3, thereby weakening the acid- with that of a previous study on the same mussel species base regulation capacity. This possibility is consistent with the M. coruscus (Sui et al., 2017). The weak shell closure strength should down-regulation of the two HCO- 3 exchangers (i.e., sulfate/bicar- greatly decrease mussels’ resistance to parasites, such as the pea bonate/oxalate exchanger sat-1 and electroneutral sodium bicar- crab Pinnotheres sinensis (Sun et al., 2006); and to shell-entering bonate exchanger 1). We therefore conclude that M. coruscus predators, such as the starfish Asterias rubens (Reimer and unsuccessfully compensated for OA-induced impairment of acid- Tedengren, 1996). base regulation capacity and that ineffective compensation even- These results suggest that along with damaging the shell tually resulted in extracellular acidosis. structure, OA also weakened the shell strength and shell closure þ þ Notably, Na /K -ATPase must consume ATP energy to drive strength of M. coruscus. In other words, the projected OA will þ þ active transport of Na and K , implying that the compensatory dramatically decrease mussels’ shell defense capacity and thereby response of M. coruscus is an energy-consuming process. Although elevate their vulnerability to predators, parasites and pathogens. it was not investigated in this study, M. coruscus metabolism has Fortunately, the transcriptome analyses revealed that the GO terms been found to be suppressed by OA (Shang et al., 2018; Sui et al., “protein tyrosine kinase activity” in the “Molecular Function” 2016; Wang et al., 2015). It is widely accepted that temporal category, “positive regulation of I-kappa b kinase/NF-kappa b metabolic depression is a main adaptive strategy of marine in- signaling” in the “Biological Process” category, and the KEGG vertebrates to survive abiotic environmental stress (Guppy, 2004). pathways “NF-kappa b signaling pathway” and “Toll-like receptor We therefore propose that M. coruscus adopted metabolic depres- signaling pathway” were significantly enriched among the up- sion to decrease overall energy demand and to provide sufficient regulated DEGs. Given the essential roles of these functions, pro- energy only for essential biological processes to enable survival cesses and pathways in immunity (Liu et al., 2016), these findings þ þ under OA conditions. However, the up-regulation of Na /K -ATPase indicate the elevation of immune responses against pathogens. indicates that additional energy was required for ion and acid-base Additionally, the GO term “actin filament binding” in the “Molec- regulation, which are essential biological processes. Under these ular Function” category, and the KEGG pathway “cardiac muscle circumstances, relatively less energy was available for other contraction” were significantly enriched among the up-regulated energy-consuming biological processes. This scenario suggests that DEGs, suggesting positive effects of OA on adductor muscle func- there were trade-offs among energy-consuming biological pro- tion and shell closure strength. These results are consistent with cesses in M. coruscus, such as ion/acid-base status regulation and the up-regulation of adductor muscle-related genes, including calcification. Therefore, the ability to differentially reallocate en- dynein, filamin-A, filamin-C, myosin, paramyosin, titin and trans- ergy among essential biological processes may determine the gelin. As discussed above, there were also compensatory responses sensitivity of marine mussel species to OA (Pan et al., 2015; related to shell formation and maintenance. The mounting of these Wittmann and Portner,€ 2013). compensatory responses indicated that the negative effects of OA on shell defense capacity had been partially mitigated; otherwise, 4.6. OA weakened shell defense capacity by damaging shell they would have been more severe. structure, and decreasing shell strength and closure strength 5. Conclusion Since shell structure was markedly damaged, it was reasonable to hypothesize that shell strength should be weakened by OA. To In conclusion, this study demonstrates that OA weakens mussel test this hypothesis and further elucidate the impacts of OA on shell shell defense capacity. Although mussels will adopt a series of defense capacity, shell strength was analyzed in this study. We molecular compensatory responses to resist OA, weakened shell found that for mussels exposed to pH 7.8 and 7.4, shell strength was defense capacity will increase mussels’ susceptibility to predators, significantly reduced to approximately 65% and 58% of that of parasites and pathogens, and thereby reduce their fitness. Mussels control mussels, respectively. These results suggest that the shell are important aquaculture bivalves and marine ecosystem engi- strength of M. coruscus was indeed weakened by OA. Similar effects neers. The findings of this study thus have significant ecological and have also been observed in other mussel species, such as the Cali- economic implications. Overall, this study sheds light on the effects fornia mussel M. californianus (Gaylord et al., 2011) and the blue of OA on mussel defense and the underlying molecular responses. It mussel M. edulis (Li et al., 2015a; Mackenzie et al., 2014). Hence, a will also improve our understanding of the future of the mussel decrease in shell strength seems to be a common response of ma- aquaculture industry and coastal ecosystems. rine mussels to OA. A fragile shell structure and reduced mechan- ical strength should tremendously elevate mussels’ vulnerability to Author contribution pathogens, such as the bacterium Vibrio tubiashii (Asplund et al., 2014); to shell-crushing predators, such as the crab Carcinus mae- X.Z. and G.L. conceived and designed this study. X.Z. and Y.H. nas (Edgell et al., 2008); and to shell-drilling predators, such as the performed the whole experiments and collected the data. All X. Zhao et al. / Chemosphere 243 (2020) 125415 11 authors contributed to data analysis and interpretation. X.Z. wrote Functional impacts of ocean acidification in an ecologically critical foundation e the manuscript, B.C., B.X., K.Q. and G.L. revised the manuscript. G.L. species. J. Exp. Biol. 214, 2586 2594. Gazeau, F., Alliouane, S., Bock, C., Bramanti, L., Lopez Correa, M., Gentile, M., et al., provided oversight of the project. All authors gave final approval for 2014. Impact of ocean acidification and warming on the Mediterranean mussel publication. (Mytilus galloprovincialis). Frontiers in Marine Science 1, 62. Gazeau, F., Quiblier, C., Jansen, J.M., Gattuso, J.-P., Middelburg, J.J., Heip, C.H.R., 2007. Impact of elevated CO2 on shellfish calcification. Geophys. Res. Lett. 34, L07603. Declaration of competing interest Ginger, K.W.K., Vera, C.B.S., D, R., Dennis, C.K.S., Adela, L.J., Yu, Z., et al., 2013. Larval and post-larval stages of Pacific oyster (Crassostrea gigas) are resistant to The authors declare no competing interests. elevated CO2. PLoS One 8, e64147. Gosling, E., 2015. Marine Bivalve Molluscs, second ed. Wiley-Blackwell. Grabherr, M.G., Haas, B.J., Yassour, M., Levin, J.Z., Thompson, D.A., Amit, I., et al., Acknowledgements 2011. Full-length transcriptome assembly from RNA-Seq data without a refer- ence genome. Nat. Biotechnol. 29, 644. Guppy, M., 2004. The biochemistry of metabolic depression: a history of percep- This work was funded by National Key R & D Program of China tions. Comp. Biochem. Physiol. B Biochem. Mol. Biol. 139, 435e442. € (No. 2018YFD0900603), National Natural Science Foundation of Hüning, A.K., Melzner, F., Thomsen, J., Gutowska, M.A., Kramer, L., Frickenhaus, S., et al., 2013. Impacts of seawater acidification on mantle gene expression pat- China (No. 31672634) and China Postdoctoral Science Foundation terns of the Baltic Sea blue mussel: implications for shell formation and energy (No. 2017M622323). The authors greatly thank Yichen Wang and metabolism. Mar. Biol. 160, 1845e1861. Xingguan Lin for their assistance with animal incubation. Hofmann, G.E., Barry, J.P., Edmunds, P.J., Gates, R.D., Hutchins, D.A., Klinger, T., et al., 2010. The effect of ocean acidification on calcifying organisms in marine eco- systems: an organism-to-ecosystem perspective. Annu. Rev. Ecol. Evol. Syst. 41, Appendix A. Supplementary data 127e147. IPCC, 2014. In: Core Writing Team, Pachauri, R.K., Meyer, L.A. (Eds.), 2014: Climate Change 2014: Synthesis Report. Contribution of Working Groups I, II and III to Supplementary data to this article can be found online at the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. https://doi.org/10.1016/j.chemosphere.2019.125415. IPCC, Geneva, Switzerland, p. 151. Kerr, R.A., 2010. Ocean acidification unprecedented, unsettling. Science 328, 1500e1501. References Kroeker, K.J., Kordas, R.L., Crim, R., Hendriks, I.E., Ramajo, L., Singh, G.S., et al., 2013. Impacts of ocean acidification on marine organisms: quantifying sensitivities Altermann, E., Klaenhammer, T.R., 2005. PathwayVoyager: pathway mapping using and interaction with warming. Glob. Chang. Biol. 19, 1884e1896. the Kyoto Encyclopedia of genes and Genomes (KEGG) database. BMC Genomics Kroeker, K.J., Kordas, R.L., Crim, R.N., Singh, G.G., 2010. Meta-analysis reveals 6, 60. negative yet variable effects of ocean acidification on marine organisms. Ecol. Anders, S., Huber, W., 2010. Differential expression analysis for sequence count data. Lett. 13, 1419e1434. € Genome Biol. 11, R106. Lannig, G., Eilers, S., Portner, H.O., Sokolova, I.M., Bock, C., 2010. Impact of ocean Aoki, H., Ishikawa, T., Fujiwara, T., Atsumi, T., Nishikawa, H., Okamoto, C., et al., 2010. acidification on energy metabolism of oyster, Crassostrea gigasechanges in Utility of shell-closing strength as the indicator of good health in breeding and metabolic pathways and thermal response. Mar. Drugs 8, 2318e2339. culture management of Japanese pearl oyster Pinctada fucata. Aquaculture 308, Lecker, S.H., Goldberg, A.L., Mitch, W.E., 2006. Protein degradation by the S115eS118. ubiquitineproteasome pathway in normal and disease states. J. Am. Soc. Ashburner, M., Ball, C.A., Blake, J.A., Botstein, D., Butler, H., Cherry, J.M., et al., 2000. Nephrol. 17, 1807e1819. Gene Ontology: tool for the unification of biology. Nat. Genet. 25, 25. Li, S., Huang, J., Liu, C., Liu, Y., Zheng, G., Xie, L., et al., 2016a. Interactive effects of Asplund, M.E., Baden, S.P., Russ, S., Ellis, R.P., Gong, N., Hernroth, B.E., 2014. Ocean seawater acidification and elevated temperature on the transcriptome and acidification and host-pathogen interactions: blue mussels, Mytilus edulis, biomineralization in the pearl oyster Pinctada fucata. Environ. Sci. Technol. 50, encountering Vibrio tubiashii. Environ. Microbiol. 16, 1029e1039. 1157e1165. Borthagaray, A.I., Carranza, A., 2007. Mussels as ecosystem engineers: their Li, S., Liu, C., Huang, J., Liu, Y., Zheng, G., Xie, L., et al., 2015a. Interactive effects of contribution to species richness in a rocky littoral community. Acta Oecol. 31, seawater acidification and elevated temperature on biomineralization and 243e250. amino acid metabolism in the mussel Mytilus edulis. J. Exp. Biol. 218, Bressan, M., Chinellato, A., Munari, M., Matozzo, V., Manci, A., Marceta, T., et al., 3623e3631. 2014. Does seawater acidification affect survival, growth and shell integrity in Li, S., Liu, Y., Liu, C., Huang, J., Zheng, G., Xie, L., et al., 2015b. Morphology and bivalve juveniles? Mar. Environ. Res. 99, 136e148. classification of hemocytes in Pinctada fucata and their responses to ocean Burnett, N.P., Belk, A., 2018. Compressive strength of Mytilus californianus shell is acidification and warming. Fish Shellfish Immunol. 45, 194e202. time-dependent and can influence the potential foraging strategies of preda- Li, S., Liu, Y., Liu, C., Huang, J., Zheng, G., Xie, L., et al., 2016b. Hemocytes participate tors. Mar. Biol. 165, 42. in calcium carbonate crystal formation, transportation and shell regeneration in Caldeira, K., Wickett, M.E., 2003. Oceanography: anthropogenic carbon and ocean the pearl oyster Pinctada fucata. Fish Shellfish Immunol. 51, 263e270. pH. Nature 425, 365-365. Liao, H., Yang, Z., Dou, Z., Sun, F., Kou, S., Zhang, Z., et al., 2019. Impact of ocean Christensen, H.T., Dolmer, P., Petersen, J., Tørring, D., 2012. Comparative study of acidification on the energy metabolism and antioxidant responses of the yesso predatory responses in blue mussels (Mytilus edulis L.) produced in suspended scallop (Patinopecten yessoensis). Front. Physiol. 9, 1967. long line cultures or collected from natural bottom mussel beds. Helgol. Mar. Liao, Z., Bao, L-f, Fan, M-h, Gao, P., Wang, X-x, Qin, C-l, et al., 2015. In-depth pro- Res. 66, 1e9. teomic analysis of nacre, prism, and myostracum of Mytilus shell. Journal of DeCarlo, T.M., Comeau, S., Cornwall, C.E., McCulloch, M.T., 2018. Coral resistance to Proteomics 122, 26e40. ocean acidification linked to increased calcium at the site of calcification. Proc. Lindinger, M.I., Lauren, D.J., Mcdonald, D.G., 1984. Acid-base balance in the sea R. Soc. Biol. Sci. 285, 20180564. mussel, Mytilus edulis. III. Effects of environmental hypercapnia on intra- and Dickson, A.G., 1990. Thermodynamics of the dissociation of boric-acid in potassium- extracellular acid-base balance. Mar. Biol. Lett. 5, 371e381. chloride solutions form 273.15 K to 318.15 K. J. Chem. Eng. Data 22, 113e127. Liu, S., Shi, W., Guo, C., Zhao, X., Han, Y., Peng, C., et al., 2016. Ocean acidification Dickson, A.G., Millero, F.J., 1987. A comparison of the equilibrium constants for the weakens the immune response of blood clam through hampering the NF-kappa dissociation of carbonic acid in seawater media. Deep-Sea Res. 34, 1733e1743. b and toll-like receptor pathways. Fish Shellfish Immunol. 54, 322e327. Drickamer, K., 1999. C-type lectin-like domains. Curr. Opin. Struct. Biol. 9, 585e590. Liu, W., He, M., 2012. Effects of ocean acidification on the metabolic rates of three Edgell, T.C., Brazeau, C., Grahame, J.W., Rochette, R., 2008. Simultaneous defense species of bivalve from southern coast of China. Chin. J. Oceanol. Limnol. 30, against shell entry and shell crushing in a snail faced with the predatory 206e211. shorecrab Carcinus maenas. Mar. Ecol. Prog. Ser. 371, 191e198. Liu, W., Yu, Z., Huang, X., Shi, Y., Lin, J., Zhang, H., et al., 2017. Effect of ocean FAO yearbook, F.A.O., 2018. Fishery and Aquaculture Statistics 2016. acidification on growth, calcification, and gene expression in the pearl oyster, Fitzer, S.C., Chung, P., Maccherozzi, F., Dhesi, S.S., Kamenos, N.A., Phoenix, V.R., et al., Pinctada fucata. Mar. Environ. Res. 130, 174e180. 2016. Biomineral shell formation under ocean acidification: a shift from order Livak, K.J., Schmittgen, T.D., 2001. Analysis of relative gene expression data using DDC to chaos. Sci. Rep. 6, 21076. real-time quantitative PCR and the 2 t method. Methods 25, 402e408. Fitzer, S.C., Cusack, M., Phoenix, V.R., Kamenos, N.A., 2014a. Ocean acidification Long, W.C., Swiney, K.M., Foy, R.J., 2013a. Effects of ocean acidification on the em- reduces the crystallographic control in juvenile mussel shells. J. Struct. Biol. 188, bryos and larvae of red king crab, Paralithodes camtschaticus. Mar. Pollut. Bull. 39e45. 69, 38e47. Fitzer, S.C., Phoenix, V.R., Cusack, M., Kamenos, N.A., 2014b. Ocean acidification Long, W.C., Swiney, K.M., Harris, C., Page, H.N., Foy, R.J., 2013b. Effects of ocean impacts mussel control on biomineralisation. Sci. Rep. 4, 6218. acidification on juvenile red king crab (Paralithodes camtschaticus) and tanner Fitzer, S.C., Zhu, W., Tanner, K.E., Phoenix, V.R., Kamenos, N.A., Cusack, M., 2015. crab (Chionoecetes bairdi) growth, condition, calcification, and survival. PLoS Ocean acidification alters the material properties of Mytilus edulis shells. J. R. One 8, e60959. Soc. Interface 12, 20141227. Mackenzie, C.L., Ormondroyd, G.A., Curling, S.F., Ball, R.J., Whiteley, N.M., Gaylord, B., Hill, T.M., Sanford, E., Lenz, E.A., Jacobs, L.A., Sato, K.N., et al., 2011. Malham, S.K., 2014. Ocean warming, more than acidification, reduces shell 12 X. Zhao et al. / Chemosphere 243 (2020) 125415

strength in a commercial shellfish species during food limitation. PLoS One 9, Shi, W., Zhao, X., Han, Y., Che, Z., Chai, X., Liu, G., 2016. Ocean acidification increases e86764. cadmium accumulation in marine bivalves: a potential threat to seafood safety. Mangan, S., Urbina, M.A., Findlay, H.S., Wilson, R.W., Lewis, C., 2017. Fluctuating Sci. Rep. 6, 20197. seawater pH/pCO2 regimes are more energetically expensive than static pH/ Stapp, L.S., Thomsen, J., Schade, H., Bock, C., Melzner, F., Portner, H.O., et al., 2017. pCO2 levels in the mussel Mytilus edulis. Proc. R. Soc. Biol. Sci. 284, 20171642. Intra-population variability of ocean acidification impacts on the physiology of Mehrbach, C., Culberson, C.H., Hawley, J.E., Pytkowicz, R.M., 1973. Measurement of Baltic blue mussels (Mytilus edulis): integrating tissue and organism response. apparent dissociation constants of carbonic acid in seawater at atmospheric J. Comp. Physiol. B 187, 529e543. pressure. Limnol. Oceanogr. 18, 897e907. Strous, G.J., Dekker, J., 1992. Mucin-type glycoproteins. Crit. Rev. Biochem. Mol. Biol. Melzner, F., Stange, P., Trübenbach, K., Thomsen, J., Casties, I., Panknin, U., et al., 27, 57e92. 2011. Food supply and seawater pCO2 impact calcification and internal shell Sui, Y., Kong, H., Huang, X., Dupont, S., Hu, M., Storch, D., et al., 2016. Combined dissolution in the blue mussel Mytilus edulis. PLoS One 6, e24223. effects of short-term exposure to elevated CO2 and decreased O2 on the phys- Michaelidis, B., Ouzounis, C., Paleras, A., Portner,€ H.O., 2005. Effects of long-term iology and energy budget of the thick shell mussel Mytilus coruscus. Chemo- moderate hypercapnia on acid-base balance and growth rate in marine mus- sphere 155, 207e216. sels Mytilus galloprovincialis. Mar. Ecol. Prog. Ser. 293, 109e118. Sui, Y., Liu, Y., Zhao, X., Dupont, S., Hu, M., Wu, F., et al., 2017. Defense responses to Mount, A.S., Wheeler, A.P., Paradkar, R.P., Snider, D., 2004. Hemocyte-mediated shell short-term hypoxia and seawater acidification in the thick shell mussel Mytilus mineralization in the eastern oyster. Science 304, 297e300. coruscus. Front. Physiol. 8, 145. Moya, A., Howes, E.L., Lacoue-Labarthe, T., Foret,^ S., Hanna, B., Medina, M., et al., Sun, T., Tang, X., Zhou, B., Wang, Y., 2016. Comparative studies on the effects of 2016. Near-future pH conditions severely impact calcification, metabolism and seawater acidification caused by CO2 and HCl enrichment on physiological the nervous system in the pteropod Heliconoides inflatus. Glob. Chang. Biol. 22, changes in Mytilus edulis. Chemosphere 144, 2368e2376. 3888e3900. Sun, W., Sun, S., Wang, Y., Yang, B., Song, W., 2006. The prevalence of the pea crab, Nagasawa, H., 2013. Mollusk shell structures and their formation mechanism. Can. J. Pinnotheres sinensis, and its impact on the condition of the cultured mussel, Zool. 91, 349e366. Mytilus galloprovincialis, in Jiaonan waters (Shandong Province, China). Aqua- O’Donnell, M.J., George, M.N., Carrington, E., 2013. Mussel byssus attachment culture 253, 57e63. weakened by ocean acidification. Nat. Clim. Chang. 3, 587e590. Suzuki, M., Saruwatari, K., Kogure, T., Yamamoto, Y., Nishimura, T., Kato, T., et al., Pan, T.-C.F., Applebaum, S.L., Manahan, D.T., 2015. Experimental ocean acidification 2009. An acidic matrix protein, Pif, is a key macromolecule for nacre formation. alters the allocation of metabolic energy. Proceedings of the National Academy Science 325, 1388e1390. of Sciences of the United States of America 112, 4696e4701. Thomas, R.D.K., 1976. Constraints of ligament growth, form and function on evo- Pierrot, D., Lewis, E., Wallace, D.W.R., 2006. MS Excel Program Developed for CO2 lution in the Arcoida (: ). Paleobiology 2, 64e83. System Calculations. ORNL/CDIAC-105a. Carbon Dioxide Information Analysis Thomsen, J., Casties, I., Pansch, C., Kortzinger,€ A., Melzner, F., 2013. Food availability Center. Oak Ridge National Laboratory, US Department of Energy, Oak Ridge, outweighs ocean acidification effects in juvenile Mytilus edulis: laboratory and Tennessee. field experiments. Glob. Chang. Biol. 19, 1017e1027. Ponce Oliva, R.D., Vasquez-Lavín, F., San Martin, V.A., Hernandez, J.I., Vargas, C.A., Thomsen, J., Gutowska, M.A., Saphorster,€ J., Heinemann, A., Trübenbach, K., Gonzalez, P.S., et al., 2019. Ocean acidification, consumers’ preferences, and Fietzke, J., et al., 2010. Calcifying invertebrates succeed in a naturally CO2 market adaptation strategies in the mussel aquaculture industry. Ecol. Econ. enriched coastal habitat but are threatened by high levels of future acidifica- 158, 42e50. tion. Biogeosciences 7, 3879e3891. Qi, P., Huang, H., Guo, B., Liao, Z., Liu, H., Tang, Z., et al., 2019. A novel interleukin-1 Thomsen, J., Melzner, F., 2010. Moderate seawater acidification does not elicit long- receptor-associated kinase-4 from thick shell mussel Mytilus coruscus is term metabolic depression in the blue mussel Mytilus edulis. Mar. Biol. 157, involved in inflammatory response. Fish Shellfish Immunol. 84, 213e222. 2667e2676. Radha, A.V., Forbes, T.Z., Killian, C.E., Gilbert, P.U.P.A., Navrotsky, A., 2010. Trans- Trapnell, C., Williams, B.A., Pertea, G., Mortazavi, A., Kwan, G., van Baren, M.J., et al., formation and crystallization energetics of synthetic and biogenic amorphous 2010. Transcript assembly and quantification by RNA-Seq reveals unannotated calcium carbonate. Proceedings of the National Academy of Sciences of the transcripts and isoform switching during cell differentiation. Nat. Biotechnol. United States of America 107, 16438e16443. 28, 511. Ramajo, L., Marba, N., Prado, L., Peron, S., Lardies, M.A., Rodriguez-Navarro, A.B., Wang, X., Wang, M., Jia, Z., Qiu, L., Wang, L., Zhang, A., et al., 2017. A carbonic et al., 2016. Biomineralization changes with food supply confer juvenile scallops anhydrase serves as an important acid-base regulator in Pacific oyster Cras- (Argopecten purpuratus) resistance to ocean acidification. Glob. Chang. Biol. 22, sostrea gigas exposed to elevated CO2: implication for physiological responses of 2025e2037. mollusk to ocean acidification. Mar. Biotechnol. 19, 22e35. Ramesh, K., Hu, M.Y., Thomsen, J., Bleich, M., Melzner, F., 2017. Mussel larvae modify Wang, Y., Li, L., Hu, M., Lu, W., 2015. Physiological energetics of the thick shell calcifying fluid carbonate chemistry to promote calcification. Nat. Commun. 8, mussel Mytilus coruscus exposed to seawater acidification and thermal stress. 1709. Sci. Total Environ. 514, 261e272. Reimer, O., Tedengren, M., 1996. Phenotypical improvement of morphological de- Weiss, I.M., Tuross, N., Addadi, L., Weiner, S., 2002. Mollusc larval shell formation: fences in the mussel Mytilus edulis induced by exposure to the predator Asterias amorphous calcium carbonate is a precursor phase for aragonite. J. Exp. Zool. rubens. Oikos 75, 383e390. 293, 478e491. Ries, J.B., Cohen, A.L., McCorkle, D.C., 2009. Marine calcifiers exhibit mixed re- Wittmann, A.C., Portner,€ H.-O., 2013. Sensitivities of extant animal taxa to ocean sponses to CO2-induced ocean acidification. Geology 37, 1131e1134. acidification. Nat. Clim. Chang. 3, 995e1001. Sabine, C.L., Feely, R.A., Gruber, N., Key, R.M., Lee, K., Bullister, J.L., et al., 2004. The Xiang, L., Kong, W., Su, J., Liang, J., Zhang, G., Xie, L., et al., 2014. Amorphous calcium oceanic sink for anthropogenic CO2. Science 305, 367e371. carbonate precipitation by cellular biomineralization in mantle cell cultures of Sadler, D.E., Lemasson, A.J., Knights, A.M., 2018. The effects of elevated CO2 on shell Pinctada fucata. PLoS One 9, e113150. properties and susceptibility to predation in mussels Mytilus edulis. Mar. En- Zada, D., Bronshtein, I., Lerer-Goldshtein, T., Garini, Y., Appelbaum, L., 2019. Sleep viron. Res. 139, 162e168. increases chromosome dynamics to enable reduction of accumulating DNA Schonitzer, V., Weiss, I., 2007. The structure of mollusc larval shells formed in the damage in single neurons. Nat. Commun. 10, 895. presence of the chitin synthase inhibitor Nikkomycin Z. BMC Struct. Biol. 7, 71. Zhang, G., Fang, X., Guo, X., Li, L., Luo, R., Xu, F., et al., 2012. The oyster genome Shang, Y., Wang, X., Kong, H., Huang, W., Hu, M., Wang, Y., 2019. Nano-ZnO impairs reveals stress adaptation and complexity of shell formation. Nature 490, 49e54. anti-predation capacity of marine mussels under seawater acidification. Zhao, X., Guo, C., Han, Y., Che, Z., Wang, Y., Wang, X., et al., 2017a. Ocean acidification J. Hazard Mater. 371, 521e528. decreases mussel byssal attachment strength and induces molecular byssal Shang, Y., Lan, Y., Liu, Z., Kong, H., Huang, X., Wu, F., et al., 2018. Synergistic effects of responses. Mar. Ecol. Prog. Ser. 565, 67e77. nano-ZnO and low pH of sea water on the physiological energetics of the thick Zhao, X., Shi, W., Han, Y., Liu, S., Guo, C., Fu, W., et al., 2017b. Ocean acidification shell mussel Mytilus coruscus. Front. Physiol. 9, 757. adversely influences metabolism, extracellular pH and calcification of an Sherker, Z.T., Ellrich, J.A., Scrosati, R.A., 2017. Predator-induced shell plasticity in economically important marine bivalve, Tegillarca granosa. Mar. Environ. Res. mussels hinders predation by drilling snails. Mar. Ecol. Prog. Ser. 573, 167e175. 125, 82e89.