Effects of anthropogenic sound on digging behavior, metabolism, Ca2+/Mg2+ ATPase activity, and metabolism-related gene expression of the bivalve Sinonovacula constricta

Anthropogenic sound has increased significantly in the past decade. However, only a few studies to date have investigated its effects on marine bivalves, with little known about the underlying physiological and molecular mechanisms. In the present study, the effects of different types, frequencies, and intensities of anthropogenic sounds on the digging behavior of razor clams (Sinonovacula constricta) were investigated. The results showed that variations in sound intensity induced deeper digging. Furthermore, anthropogenic sound exposure led to an alteration in the O:N ratios and the expression of ten metabolism-related genes from the glycolysis, fatty acid biosynthesis, tryptophan metabolism, and Tricarboxylic Acid Cycle (TCA cycle) pathways. Expression of all genes under investigation was induced upon exposure to anthropogenic sound at ~80 dB re 1 μPa and repressed at ~100 dB re 1 μPa sound. In addition, the activity of Ca2+/Mg2+-ATPase in the feet tissues, which is directly related to muscular contraction and subsequently to digging behavior, was also found to be affected by anthropogenic sound intensity. The findings suggest that sound may be perceived by bivalves as changes in the water particle motion and lead to the subsequent reactions detected in razor clams.

Effects of sound type, frequency, and intensity on the digging behavior of razor clams. As shown in Fig. 1 and Table 2, the digging depths of the razor clams exposed to the 500 Hz sine wave (aeration background noise plus anthropogenic sine wave sound input) were not significantly different from those exposed to 1000 Hz of the same sound type. P values were 0.21 and 0.96 from the ANOVA for the ~80 dB re 1 μPa and ~100 dB re 1 μPa trials, respectively. Similarly, the digging behavior was not affected much by various anthropogenic input sound types, where no significant difference in digging depths was detected between the "white noise" (broad frequency sound generated with anthropogenic white noise input) and sine wave sound input groups, with trials carried out at the same underwater sound intensity. Interestingly, the three sound intensities tested led to a significant difference in the digging depths (p < 0.05) of the clams. Generally, digging depth increased with underwater sound intensity. The digging depths of razor clams exposed to anthropogenic sound of "white noise", 500 Hz sine wave, and 1000 Hz sine wave at ~80 dB re 1 μPa were 1.05, 1.05, and 1.13 times deeper respectively than those exposed to the control (~60 dB re 1 μPa of ambient aeration noise). When the underwater sound  intensity was increased to ~100 dB re 1 μPa, the digging depths of razor clams exposed to anthropogenic sound of "white noise", 500 Hz sine wave, and 1000 Hz sine wave were 1.20, 1.15, and 1.14 times deeper, respectively, than those exposed to the control. These results indicated that the digging behavior of the razor clams was significantly affected by underwater sound intensity.
Effects of anthropogenic sound exposures on oxygen consumption, ammonia excretion, O:N ratio, and Ca 2+ /Mg 2+ ATPase activity of razor clams. The oxygen consumption rates, ammonium excretion rates, and O:N ratios of razor clams exposed to the ambient aeration sound (control) or to ~80 dB re 1 μPa or ~100 dB re 1 μPa of "white noise" for one week were listed in Table 3. Results from ANOVA showed that while both the rates of oxygen consumption and ammonia excretion were not significantly different among trials, the O:N ratios were significantly affected by underwater sound intensity. The highest and lowest O:N ratios were detected in the samples exposed to the ~80 dB re 1 μPa and ~100 dB re 1 μPa of underwater sound respectively. As shown in Fig. 2, the intensity of exposed sound exerted a significant effect on the activity of Ca 2+ / Mg 2+ -ATPase, with the highest (8.97 ± 0.49 U/mg protein) and lowest activity (2.95 ± 0.56 U/mg protein) observed in clams exposed to ~80 dB re 1 μPa and ~100 dB re 1 μPa of underwater sound, respectively.
Effects of sound exposures on the relative expression of metabolic genes. As shown in Fig. 3a,b, the 6-phosphofructokinase-1 and pyruvate kinase genes, both function in the glycolysis pathway, showed similar relative gene expression patterns. In clams exposed to ~80 dB re 1 μPa of underwater sound, both genes showed significantly higher expression than those of control or exposed to ~100 dB re 1 μPa of sound. Between the groups exposed to quiet ambient sound or ~100 dB re 1 μPa of sound, gene expression was found to be higher in clams exposed to quiet ambient sound, but the difference was not statistically significant.
The relative expression of acetyl-CoA carboxylase and arylformamidase genes in the fatty acid biosynthesis and tryptophan metabolism pathways showed a similar trend (Fig. 3c,d) as well. The only difference was found in a lower relative expression of the arylformamidase gene in clams exposed to the quiet ambient sound, albeit not significant when compared to that in the clams exposed to ~100 dB re 1 μPa of underwater sound.
All six genes in the TCA cycle ( Fig. 4) displayed similar relative expression patterns, where highest values were detected in clams exposed to ~80 dB re 1 μPa of sound. Relative gene expression was significantly lower in clams exposed to ~100 dB re 1 μPa of sound than those of the control, with the exception of the oxoglutarate dehydrogenase gene.

Discussion
Theoretically, in the far field of an acoustic source, the pressure (p) and velocity components (v) are related as p = v × z, where z is the parameter that indicates the impedance of the medium 12 . Therefore, with the same sound type and frequency, higher intensity sound, such as the ~80 re 1 μPa and ~100 dB re 1 μPa of underwater sound tested in the present study, will give rise to a more intense particle motion and subsequently lead to a higher sound pressure level in the seawater. Though sensory hearing has not been reported in bivalve species, both the mantle and gills of a bivalve possess sensory palps that are sensitive to environmental disturbances such as water proton movement. In S. constricta, in addition to the mantle and gills, foot, exhalent siphon, inhalant siphon, and the palps around these organs are sensitive to the motion of surrounding water as well. Few studies have suggested that sound may exert influences on the embryonic development and settlement of marine bivalves 12,44,45 . With field experiments, it has been shown that the addition of replayed habitat-related sounds significantly increased the settlement of free-swimming larvae of the eastern oyster (Crassostrea virginica), compared to no-sound controls 44,45 . Whereas a laboratory study carried out in New Zealand scallop (Pecten novaezelandiae) has shown that the noise generated by a seismic air gun significantly increased the percentage of body malformations and retarded the growth and development of the larvae 12 , probably owing to the increase in sound pressure. Similarly, the variations detected in the present study in digging behavior, metabolism, and expression of metabolic genes in response to noise exposure at different intensities could attribute to the changes in water particle motion as well. The digging behavior of the razor clam, a bottom burrower that lives in the intertidal zone, is an adaptation to its surroundings and is closely related to water movement. With alterations in the ebb and flow, the razor clam exhibits corresponding digging or emerging behaviors. Importantly, the digging behavior of the razor clam is   Table 3. Oxygen consumption rates, ammonium excretion rates, and O:N ratios of Sinonovacula constricta after one week of exposure to the ambient control, ~80 or ~100 dB re 1 μPa of underwater sounds (mean ± SE) (Different superscripts indicate significant differences between trials by Tukey's test).
closely related to its metabolism status. On the one hand, as a result of muscular movement of the foot, digging, or emerging behavior is an energy-consuming process that relies on the energy produced through pathways such as glycolysis and the TCA cycle. On the other hand, digging behavior is associated with active feeding because shallow digging increases the contact with oxygen and food supplies in seawater, albeit accompanied by an increased risk of encountering predators and environmental stressors.
In the present study, exposure to the intensified ~100 dB re 1 μPa of underwater sound induced an avoidance response in the razor clams, resulting in a significantly more active digging activity. Similar to other  Relative gene expression levels (mean ± SE) of 6-phosphofructokinase-1 (a), pyruvate kinase (b), acetyl-CoA carboxylase (c), and arylformamidase (d) in razor clams exposed to one week of ambient sound or to ~80 dB re 1 μPa or ~100 dB re 1 μPa of underwater sounds (* and **indicate a significant and an extremely significant difference by t-test, respectively, relative to the control of ambient sound). environmental stressors such as unfavorable temperature, salinity, and pH, exposure to ~100 dB re 1 μPa of underwater sound also led to a significant decrease in the O:N ratio, indicating a higher rate of protein catabolism relative to lipid and carbohydrate catabolism 31,32 . This observation may be due to the reduction in oxygen availability associated with deeper digging, a behavior that may occur in response to stress, as previously reported in other species 46,47 . When razor clams were exposed to the intensified ~100 dB re 1 μPa of underwater sound, the expression of all tested genes from the glycolysis 48 , fatty acid biosynthesis 49 , tryptophan metabolism 50,51 , and TCA cycle 52,53 pathways were repressed, suggesting a slowing of metabolic activity as the individual retreat deeper into the mud. To give a brief summary, the intensified water particle movement brought about by ~100 dB re 1 μPa of underwater sound rendered the clams digging deeper into the mud and entering an inactive state, as a means to avoid environmental disturbance.
Unlike the ambient aeration sound from indoor cultivation and the ~100 dB re 1 μPa of underwater sound, the ~80 dB re 1 μPa of sound rendered the clams more active, as indicated by the significantly higher O:N ratio 46,47,54 and increased expression of genes from the glycolysis 48 , fatty acid biosynthesis 49 , tryptophan metabolism 50,51 , and TCA cycle pathways 52,53 . Anthropogenic sound at ~80 dB re 1 μPa is comparable to the natural ambient sound level in the intertidal zone when the bottom substrates are covered by seawater 55,56 . Therefore, exposure to ~80 dB re 1 μPa of broad frequency underwater sound probably created a similar water particle movement condition indicating high food and oxygen availability for razor clams and subsequently induced shallow digging behavior for feeding. The expression of genes involved in glycolysis and the TCA cycle in clams exposed to ~80 dB re 1 μPa of underwater sound was in accordance with the O:N data obtained, which was significantly higher than those in clams exposed to ambient control and ~100 dB re 1 μPa of intensified underwater sound. Upon activation by ~80 dB re 1 μPa of underwater sound, the clams exhibited more catabolism of carbohydrates and lipids relative to protein to meet the new energy demands associated with the induced active feeding behavior. Similarly, as tryptophan catabolism will produce a series of biologically active substances that play important roles in physiological metabolism 51,57 , the arylformamidase gene in the tryptophan catabolism pathway was found to be activated. The increase in active feeding behavior also triggered the storage of energy as fatty acids, which was indicated by a significantly higher expression of the acetyl-CoA carboxylase gene from the fatty acid synthesis pathway 49 .
Less digging depth was observed when razor clams were exposed to the relatively quiet ambient aeration sound. This indoor cultivation sound was quieter than the natural ambient sound in the intertidal zone. As Relative gene expression levels (mean ± SE) of citrate synthase (a), isocitrate dehydrogenase (NAD + ) (b), isocitrate dehydrogenase (NADP + ) (c), oxoglutarate dehydrogenase (E1) (d), dihydrolipoamide succinyltransferase (E2) (e) and dihydrolipoamide dehydrogenase (E3) (f) in razor clams after exposure for one week to ambient sound or to ~80 dB re 1 μPa or ~100 dB re 1 μPa of "white noise" underwater sounds (* and **indicate a significant and an extremely significant difference by t-test, respectively, relative to control of ambient sound). discussed above, sound is probably perceived by bivalve species as a variation in water particle movement, which delivers information about the magnification of flow or tides. Therefore, less digging was probably exhibited because the quiet ambient sound was regarded as "risk-free" by the individual razor clams. Moreover, the mild water particle movement driven by the quiet ambient sound neither activated nor inhibited the metabolism of the razor clams, which was observed with the ~80 dB re 1 μPa and ~100 dB re 1 μPa underwater sound treatments.
The digging and/or emerging behaviors of bottom-burrowing bivalves are directly controlled by foot muscular movement through an energy-consuming process 30 . The digging behavior has been reported to be Mg 2+ -concentration dependent in the hard clam Meretrix lusoria 58,59 . The Ca 2+ /Mg 2+ -ATPase is an Mg 2+ -dependent enzyme and directly related to muscular contraction [60][61][62][63][64] ; therefore, the status of Ca 2+ / Mg 2+ -ATPase in the foot can be used as an indicator of digging and/or emerging activities. In the present study, the highest enzyme activity of the Ca 2+ /Mg 2+ -ATPase was found in razor clams exposed to ~80 dB re 1 μPa of underwater sound when active feeding was induced. The high enzyme activity could be accounted by the increase in foot movement during feeding. Though with deeper digs, exposure to the ~100 dB re 1 μPa of underwater sound led to the lowest activity of the Ca 2+ /Mg 2+ -ATPase. As the clams has already dug into the substrate and entered an inactive state to avoid the disturbance before the sampling time-point, the low enzyme activity reflected the resting condition of the individuals at the time of sampling. Similarly, the clams were neither activated nor inhibited by exposure to the ambient sound, with moderate Ca 2+ /Mg 2+ -ATPase activity found in samples under these conditions.

Methods
Animal collection and maintenance. Adult razor clams were collected from Yueqing Bay (28.28°N and 121.11°E), Wenzhou, China, before the spawning season from late May to early July of 2014. Once transported to the lab, the clams were acclimated in a 2000 L tank with 500 L aerated flowing seawater (temperature 23.9 ± 1.0 °C, pH 7.95 ± 0.40, salinity 20.0 ± 0.5%, dissolved oxygen 8.01 ± 0.3 mg/L, and total alkalinity 1.91 ± 0.40 mmol/L) for a week before the commencement of experiment. Clams were fed with microalgae Platymonas subcordiformis twice a day, at 8:00 a.m. and 8:00 p.m. Healthy individuals with no shell damage and of regular size (shell length at 5.3 ± 1.1 cm) were used for the experiments.
Sound treatment and record. Referring to preliminary survey results and published literatures, underwater sound levels of ~80 dB re 1 μPa and ~100 dB re 1 μPa were selected to simulate normal sound level in the intertidal zone when covered by tide and under situations of intensive anthropogenic sound, respectively 55,56 . An ambient aeration sound level of the culture system without any addition of anthropogenic sound input was used as control. To find out the factor of the anthropogenic sound affecting digging behavior, the various combinations of two types of sound-input signals (white noise and sine wave) and two frequencies of sine wave sound (500 Hz and 1000 Hz) were tested. In total, the following seven sound exposures with basal ambient aeration noise were investigated: 1) no additional anthropogenic sound (control), 2) ~80 dB re 1 μPa of "white noise" , 3) ~100 dB re 1 μPa of "white noise", 4) ~80 dB re 1 μPa 500 Hz sine wave, 5) ~100 dB re 1 μPa 500 Hz sine Primers for the genes analyzed Sequence (5′ to 3′)   The hydrophone was deployed in the sediment (the hydrophone was completely buried in the mud) and at the interface between seawater and sediment at 5 cm, 15 cm, 25 cm away from the center of the concentric circle located vertically below the submersible loudspeaker (Fig. 5). Acoustic data were analyzed by Soundscape Analysis Software SACS V1.0 (Register number: 2014SR216788) in MATLAB R2013a. The spectral density measurement with bandwidth of 86 Hz was applied to calculate the sound pressure levels of the control and experimental trials and the mean of sound intensity was subsequently obtained by taking the average of each 30 seconds acoustic data. In this calculation method, the sound signals in a very short period of time (10~30 ms) were deemed as smooth and underwent Fourier transformation as following.
Let The discrete time Fourier transformation (DTFT) of x(n) was then obtained via equation (2), in which, m is the time serial number which synchronizes with n and N is the length of given acoustic signal. The discrete Fourier transformation (DFT) was then estimated using equation (3), where X n k ( , ) is the short-time magnitude of spectra.
The power spectral density at time m was calculated through equation (4). = = × P n k X n k X n k conj X n k ( , ) ( , ) ( ( , )) ( ( ( , ))) (4) 2 The sound pressure level at each bandwidth was subsequently measured via equation (5), where P(k) is short-time Fourier spectral density at center frequency and P 0 is the reference pressure (1μpa in water).
The sound pressure level was calculated through equation (6). Since the frequency bandwidth is determined by the signal sampling rate and the parameter n of the Fourier transformation, the bandwidth analyzed in the present study is therefore 11025 Hz/128 = 86 Hz (the frequency range after Fourier transformation and n are 0~11025 Hz and 128, respectively).
Digging behavior trials. Three replicates, each containing 40 experimental individuals, were performed for each trial. A fishing line (0.18 mm in diameter and 1.5 m in length), with a label identifying the clam, was attached with waterproof glue to the shell of each razor clam in a longitudinal direction to facilitate estimations of the digging depth. Upon a gentle straight pull, the change in length of the attached fishing line before and after experiment was defined as the digging depth. With preliminary experiments indicating that all digging behavior of the razor clams was completed within 24 hours after upload onto the bucket, an experimental period of 24 hours was monitored for all the digging behavior trials. The seawater of each trial was aerated throughout the entire experiment, serving as the main ambient sound.
Oxygen consumption rate, ammonia excretion rate, O:N ratio and Ca 2+ /Mg 2+ ATPase activity assays. According to the results obtained in the digging behavior trials, digging behavior was mainly affected by sound intensity. Therefore, the effects of underwater sounds with anthropogenic "white noise" input at ~80 dB re 1 μPa and ~100 dB re 1 μPa on oxygen consumption rate, ammonia excretion rate, and Ca 2+ /Mg 2+ -ATPase activity were investigated in the present study.
Apart from exposure timings and fishing line attachment, similar methods described in Digging behavior trials were carried out for the sound exposure experiments. During the one-week exposure, one-half of the seawater volume was replaced daily with fresh filtered seawater, and the clams were fed with microalgae P. subcordiformis at 8:00 a.m. and 8:00 p.m every day. Upon introduction to the ambient sound of control, ~80 dB re 1 μPa or ~100 dB re 1 μPa of underwater sounds for a week, individual razor clams were used to analyze oxygen consumption, ammonia excretion, and Ca 2+ /Mg 2+ -ATPase activity.
Twenty-five individuals from each trial were divided equally into 5 respiratory chambers (2 L) filled with seawater. In total, each sound level tested consisted of 5 replicates containing 5 individuals in each chamber, and a blank trial was conducted with no individual assigned. After buffering the clams in still water for 1 hour, the oxygen consumption rate and the ammonia excretion rate were analyzed. To obtain the oxygen consumption rate, the dissolved oxygen concentrations before and after the experiment were determined by an oxygen meter (Multi 3410 SET4, WTW, Germany). To obtain the ammonia excretion rate, seawater ammonia concentrations before and after the experiment were measured using the standard indophenol blue photometric method 67 . After measurement, the soft tissue of each individual was peeled off carefully with a scalpel and then dehydrated in an 80 °C oven for 10 days. Dry weights of soft tissues were estimated using a Sartorius electronic balance (BSA2245). The oxygen consumption and ammonia excretion rates were calculated using the equation (7): where R(E) is the oxygen consumption (or ammonium excretion) rate, t 0 and t 1 represent oxygen consumption (or ammonium excretion) before and after the experiment, t is the total respiration (or excretion) time, w is the dry weight of the soft tissues, and V is the volume of the respiratory (or excretion) chamber. The atomic ratio of oxygen to nitrogen (O:N) was obtained by dividing the oxygen consumption rate by the ammonia excretion rate. After one week of sound exposure, another 5 individuals of each experimental trial were used to carry out the Ca 2+ /Mg 2+ -ATPase activity assay. Once carefully peeled off on ice, the foot of each razor clam was used in enzyme activity analysis, following the protocol of the Minim ATP enzyme test kit (Ca 2+ /Mg 2+ ATPase) from Nanjing Jiancheng Bioengineering Institute ® . The enzyme Ca 2+ /Mg 2+ ATPase decomposes ATP into ADP and inorganic phosphate. The concentrations of the inorganic phosphate produced from ATP hydrolysis were determined using a spectrophotometer (UV-2100, Shanghai Jinghua Instruments) at 636 nm. The Ca 2+ /Mg 2+ -ATPase activity was then estimated according to the equation (8). One Ca 2+ /Mg 2+ -ATPase activity unit was defined as the amount of enzyme decomposing 1 μmol ATP per milligram tissue protein per hour. where OD measured , OD control , OD standard , and OD blank indicate the OD values for the tested sample, control, standard sample, and the blank respectively; C indicates the concentration of the standard protein; and t indicates the reaction time.
Quantitative real time PCR of metabolic genes. After being exposed to the ambient control, ~80 dB re in the glycolysis pathway; acetyl-CoA carboxylase (JZ897769) in the fatty acid biosynthesis pathway; arylformamidase (JZ897770) in the tryptophan metabolism pathway (tryptophan is the first limiting amino acid in razor clams); and citrate synthase (JZ897771), isocitrate dehydrogenase (both NAD + (JZ897772) and NADP + (JZ897773) dependent types) and α -oxoglutarate dehydrogenase (with all the three parts, E1, E2, and E3, representing oxoglutarate dehydrogenase (JZ897774), dihydrolipoamide succinyltransferase (JZ897775) and dihydrolipoamide dehydrogenase (JZ897776), respectively) in the TCA (citrate cycle) pathway respectively. The 18S ribosomal RNA gene was used as an internal reference 68 . The sequences of all primers were listed in Table 4. Quantitative PCR was performed in a CFX 96 TM Real-Time System. Amplifications were carried out in triplicate in a total volume of 10 μL containing 5 μL SsoFast EvaGreen Supermix (Bio-Rad ® ), 3 μL PCR-grade water, 1 μL cDNA, and 1 μL primers (100 μM). The reaction conditions included an initial denaturation at 95 °C for 5 min; 39 cycles of 95 °C for 20 sec, 61 °C for 20 sec; and 72 °C for 20 sec. After the reaction, the Bio-Rad CFX Manager was used for melting-curve analysis to confirm that a specific PCR product was amplified. The 2 −△△CT method was applied to analyze the relative gene expression of target genes.
Statistics. The intensities of sounds in the sediment and at the interface between seawater and sediment at 5 cm, 15 cm, and 25 cm away from the center of concentric circle were analyzed using a one-way ANOVA.
To investigate the effect of sound frequency on the digging depth of S. constricta, the digging depths of clams exposed to the sine wave anthropogenic sound input with different noise frequencies and constant intensity (~80 or ~100 dB re 1 μPa) were analyzed using a one-way ANOVA. Similarly, a one-way ANOVA was performed on data of the digging depths of razor clams exposed to different types (white noise versus sine wave anthropogenic sound input) and intensities (ambient control, ~80, or ~100 dB re 1 μPa) of sounds to estimate the effects of sound types and intensities on the digging behavior of S. constricta. One-way ANOVA followed by Tukey's post hoc test was conducted to compare the oxygen consumption rates, ammonia excretion rates, O:N ratios, and Ca 2+ / Mg 2+ -ATPase activities among experimental trials. The expression level of each target gene was compared against the control using t-test. All statistics were performed with the statistical package "R" 69,70 , and a p-value < 0.05 was accepted as statistically significant.