Role of Calanus sinicus ( Copepoda , Calanoida ) on dimethylsulfide production in Jiaozhou Bay

The role of copepod Calanus sinicus on the production of dimethylsulfide (DMS)/dimethylsulfoniopropionate (DMSP) in Jiaozhou Bay was evaluated in field measurements and laboratory experiments. Samples at 10 sites in the bay were collected monthly from June 2010 to May 2011 (except for March 2011), and zooplankton species composition was analyzed. Effects of C. sinicus grazing on DMS/DMSP production at different conditions (i.e., algal diets, food 15 concentrations, and salinities) were assessed in the laboratory. Data from the field experiment showed that C. sinicus was the dominant copepod in Jiaozhou Bay (up to 123 individuals m−3 in May 2011) and preferred to graze on diatom. DMS and DMSP concentrations not only depend on phytoplankton abundance, but also phytoplankton species and other factors. In the laboratory experiment, compared with Gymnodinium sp. or Emiliania huxleyi, C. sinicus feeding on Isochrysis galbana and Chaetoceros curvisetus exhibited increased DMS concentration, whereas high salinity inhibited DMS production. This study 20 indicated that DMSP was transferred from phytoplankton to copepod body, fecal pellet, and seawater through copepod grazing. Our results provided important information to understand the biogeochemical cycle of DMSP in Jiaozhou Bay.


Introduction
Dimethylsulfide (DMS) is the most abundant biogenic sulfur gas that may influence planetary climate by forming cloud 25 condensation nuclei that alters global radiation balance (Charlson et al., 1987). The biogeochemical cycling of dimethylsulfoniopropionate (DMSP), which is the main precursor of DMS. The marine environment is the major source of DMSP, and DMSP is synthesized by many marine phytoplankton species as an osmolyte . Therefore, the dynamics of DMSP in the ocean will have an important influence on global DMS production. The conversion of DMSP to DMS is regulated by complex trophic processes in the water column, e.g., algal senescence (Nguyen et al., 1988), 30 phytoplanktonic enzyme catalysis (Niki et al., 2000), bacterial activity (Kiene and Linn, 2000), and zooplankton grazing (Dacey and Wakeham, 1986;Wolfe et al., 2000;Yu et al., 2015). Currently, most studies have focused on field research concerning the spatial and temporal distributions and fluxes of DMS/DMSP (Turner et al., 1996;Wong et al., 2005). Many field studies indicated a weak correlation between DMS and parameters directly related to primary producers (Kettle et al., 1999). Numerous biotic factors (i.e., phytoplankton, zooplankton, bacteria, and virus) and abiotic factors (i.e., temperature, salinity, light, and nutrient), e.g., factors other than phytoplankton, might play an important role in DMS and DMSP 5 dynamics. Zooplankton are known to consume for a large fraction (10% to 25%, or higher) of the daily oceanic primary production (Lancelot and Billen, 1985). Copepods, a trophodynamic link between primary and tertiary productions, play a key role in the cycling of materials and energy in marine ecosystems (Kiørboe, 1997). Dacey and Wakeham (1986) were the first to show that copepod grazing stimulated DMS production in a laboratory experiment. Other studies have also documented that zooplankton grazing induced DMS production (Belviso et al., 1990;Christaki et al., 1996;Daly and 10 DiTullio, 1996;Leck et al., 1990) because of sloppy feeding (Dacey and Wakeham, 1986;Tang et al., 2000) or enhanced DMSP lyase activities (Wolfe and Steinke, 1996). The DMSP ingested by zooplankton was accumulated in the body and transferred to the upper food chain, and a portion of the DMSP ingested was transferred to fecal pellets, which were subsequently uncoupled as DMS and DMSP productions.
The calanoid copepod Calanus sinicus is distributed in the East China Sea, the Yellow Sea, and the coastal waters of 15 Japan (Brodsky, 1965) and is one of the dominant zooplankton in the East China Sea, the Yellow Sea, and the Jiaozhou Bay.
In this study, we performed field measurements and laboratory experiments to investigate the effects of copepod on DMS

Abundance and taxonomic composition of zooplankton
Zooplankton samples were collected by vertical tows from the bottom to the surface (the depth varied from 3 m to 28 m according to different stations) using a conical-cylindrical plankton net with a 50 cm mouth opening and a 160 µm mesh size.
A flow meter was used to estimate the amount of filtered water. We retrieved the net at 0.3 m s −1 to 0.5 m s −1 at each station.
Samples were then rinsed into cod-end buckets, concentrated, and preserved in 5% formalin for qualitative and quantitative 5 analyses of in situ zooplankton. Taxonomy was determined by optical microscopy. In situ abundance was calculated using tow volume estimates, which were determined by net dimensions and flow meter values.

Dilution experiments
Dilution experiments were set up according to the methods of Landry and Hassett (1982) and Wolfe et al. (2000), which are widely used to estimate microzooplankton grazing and phytoplankton growth rates. The technique assumes that 10 increasingly diluted treatments reduce grazer-prey encounter and therefore grazing rates (g), without changing specific growth rates (µ) of prey. Net production of a prey biomaker B is thus given by B(t) = B(0)e (µ~dg)t , where d is the fraction of unfiltered water; regressing 1/t ln (B(t)/B(0) vs. d yields µ as the Y-intercept and g as the negative of the slope (Wolfe et al., 2000).
Depending upon the time allowed for water collection, we conducted dilution experiments aboard at three stations (C3: 15 120.26°E, 36.13°N; D4: 120.30°E, 36.09°N; E3: 120.30°E, 36.03°N) on each cruise from June 2010 to May 2011 in Jiaozhou Bay (Fig. 1). A dilution series was prepared, consisting of 100, 80, 60, 40, and 20% unfiltered water in 1 L polycarbonate bottles (washed with 10% HCl and distilled water, and rinsed with seawater prior to use). The water was collected using 12 L Niskin bottles on a CTD rosette. Water for dilutions was filtered using gravity through Gelman Suporcap 0.2 µm capsule filters into an acid-washed carboy. These bottles were incubated for 24 h under simulated in situ 20 conditions in a water-bath deck incubator with neutral density screening. Nutrients were added in the bottles to ensure constant phytoplankton growth in the dilution series when the final concentrations of nitrate and phosphate were lower than 0.5 µM and 0.03 µM. We did not pre-screen water to remove mesozooplankton grazers, because copepod numbers were usually low (< 1 L −1 ) compared with microzooplankton and our preliminary trials showed that mesograzers had a negligible grazing impact at natural densities compared to microzooplankton. Specific growth and grazing rates of Chl a were 25 calculated by regressing the time-normalized, log-transformed ratio of final and initial concentrations vs. fraction unfiltered water (Landry and Hassett, 1982).

Abiotic parameters and analyses of chlorophyll a (Chl a) and bacteria
The shipboard measurements of temperature and salinity of surface water were obtained with the Sea-Bird 911. The surface water samples were collected by a Niskin sampler and then they were filtered using Whatman GF/F filters (nominal pore 30 size of 0.7 µm) to determine Chl a. Chl a concentrations were measured using a Hitachi F4500 fluorometer according to the 4 methods of Parsons et al. (1984). Based on the procedures of Porter and Feig (1980), bacteria were counted by epifluorescence microscopy (Hitachi F4500; total magnification of ×1,000).

Phytoplankton and zooplankton cultures
Phytoplankton species (Isochrysis galbana, Chaetoceros curvisetus, Emiliania huxleyi, and Gymnodinium sp.) were provided 5 by the Marine Microalgae Research Center, Ocean University of China in different sizes and DMSP contents. They are widespread species in Jiaozhou Bay and Chinese coastal waters (Li et al., 2005;Zhong et al., 2001). These algae are described in Table 1, and the sizes of algal cells are obtained by measuring at least 200 cells using a calibrated ocular micrometer with a light microscope at a magnification of × 400. Cell volume was estimated by approximating the shape of each species to an elliptic sphere, which subsequently determined the biovolume according to the method of Verity et al. 10 (1992). In this study, we determined the cellular carbon contents in phytoplankton according to the method of Strathmann (1967).
All algae were cultured with f/2 medium (Guillard, 1975) at 60 µmol m −2 s −1 under a dark/light cycle of 12 h:12 h at a temperature of 15 ± 1°C. According to the methods of Wolfe and Steinke (1996), the algal culture was detected by epifluorescence microscopy following staining with acridine orange and by plating on 1% peptone agar plates to check for 15 bacterial growth. No bacterial contamination was found in any of the experimental cultures. Algae were fed to copepods during their exponential growth phase. Copepods C. sinicus were collected from Jiaozhou Bay, Qingdao, China (120°8′ E, 36°8′ N) in April and May 2011 using a 0.5 m standard ring net equipped with a 160 µm mesh and a solid cod end. The copepods were grown at 15 ± 1°C in 30 PSU sterilized seawater. In the incubation experiment, adult copepods were utilized as experimental animals. 20

Ingestion configuration
Four species of algae (C. curvisetus, I. galbana, Gymnodinium sp., and E. huxleyi) were utilized to assess dietary effects on changes in ingestion rate (IR), clearance rate (CR), and DMS/DMSP production in a salinity of 30 PSU. I. galbana was used as a single food source, which was a DMSP-poor and favorite food for C. sinicus, and four concentrations of I. galbana (10,15,20, and 25 × 10 4 cells mL −1 ) were set as concentration contrasts to obtain ingestion and DMS/DMSP information from 25 food concentration. Laboratory studies have confirmed that salinity can significantly influence the ingestion rates of copepod and the S-compounds production of phytoplankton (Tang et al., 1999;Yu et al., 2015). Therefore, four levels of salinities (20,25,30,and 35 PSU) were set up to investigate the effects of salinity on copepod grazing and the variations of DMSP and DMS concentrations. Single-factor experiments were performed, and all parameters were run individually.

IR and CR
In order to ensure that the copepods were hungry to graze the diets, copepods were starved for 24 h before they were fed with phytoplankton. Copepods were rinsed with sterilized seawater before the beginning of the grazing experiment. Batches of 10 individual copepods in each of 3 replicates were sorted out and transferred to 250 mL polycarbonate bottles. Three bottles containing no grazers were used as controls. All bottles were topped off with suspensions of algae, sealed with 5 parafilm, and fastened onto a spinning plankton wheel (2 rpm) at 15 ± 1°C in darkness for 24 h. After a 24-h incubation, copepods and fecal pellets were harvested according to the methods of Tang (2001). Water samples were used to analyze cell density, IR, CR, DMS, DMSPp, and DMSPd. The IRs and CRs were calculated according to the equations of Frost (1972).
Because no significant differences were found in algae concentrations between the initial and final control bottles, the growth constant (k) for algal growth was eliminated from the equations, thus yielding: Where CR is the clearance rate (mL ind −1 h −1 ), IR is the ingestion rate (cells ind −1 h −1 ), C1 is the initial algal concentration in control bottles (cells mL −1 ), C2 * is the final algal concentration in the experimental bottles (cells mL −1 ), t is the duration of the experiment (h), V and N are the volume (mL) and number of copepods in the experimental bottles (ind), 15 respectively.
Duplicates of 10 mL aliquots of each algal suspension were placed in 40 mL serum bottles that contained 2 mL of 10 mol L −1 KOH solution for DMSP detection. Ten copepods from each bottle were placed individually in serum bottles for DMSPz (DMSP in the copepod body) measurement before gut clearance. Fecal pellets were separated from detritus with a mouth micropipette, rinsed with filtered seawater, and concentrated onto a 47 mm Whatman GF/F filter by gravity filtration to 20 determine DMSPf (DMSP in the fecal pellets).
We ran preliminary experiment to check the effects of the bacteria on DMS concentration. According to the methods of Agostini et al. (2016), the treatment with antibiotics (0.025 g L −1 penicillin G potassium + 0.08 g L −1 streptomycin sulphate + 0.04 g L −1 neomycin sulphate) were used to inhibit the bacteria in the algal culture. When C. sinicus were fed on the four diets (I. galbana, C. curvisetus, E. huxleyi, and Gymnodinium sp.), no significant differences were found between DMS 25 concentrations in the control (without antibiotics) and those in the treatment (with antibiotics) (data not shown). Therefore, the copepod cultures were not treated with antibiotics in our laboratory experiment to obtain axenicity. Yost and Mitchelmore (2009) reported that antibiotic treatment negatively affected algal growth, which was the other reason for not using antibiotics. 6

DMS and DMSP determinations
According to the methods of Andreae and Barnard (1983), DMS and DMSP concentrations were determined using the purge-and-trap technique by a gas chromatograph, which was equipped with a flame photometric detector (Shimadzu GC-14B). For DMS measurement, 10 mL seawater was directly introduced into a glass purge chamber. Gravity filtering of samples for DMSPd was obtained according to the method of Kiene and Slezak (2006) with minor modifications. The filtrate 5 and unfiltered seawater for DMSPd and total DMSP (DMSPt, DMSPd + DMSPp) measurements were transferred to a 40 mL serum bottle containing 2 mL of 10 mol L −1 KOH solution and kept at 4°C for at least 24 h to complete cleavage. DMSPd and DMSPp concentrations were determined by the total DMS subtracting the DMS in original seawater. The DMSPf and DMSPz concentrations were calculated by subtracting DMS and DMSPd in filtered water.

Statistical analysis
10 Data were expressed as mean ± standard deviation. Student's t test and one-way ANOVA were used to determine the differences between control and treatment samples. Pearson correlations were utilized to assess the relationships between DMS/DMSP and IRs.
A total of 74 species of zooplankton were identified and were categorized into 5 phyla, 6 classes, 17 orders, 44 families, 25 and 55 genera. Among these species, 31 are Arthropoda, 22 are Coelenterata, 17 are planktonic larvae, 2 are Chaetognatha, 1 are Protozoa, and 1 are Tunicata ( Table 2). The species composition of zooplankton varied with months, and the number of species ranged from 19 to 35. The lowest and highest numbers of species occurred in February 2011 (winter) and June 2010 (summer), respectively. In total, 22 copepods were identified, and Acartia pacifica, Acartia bifilosa, Centropages abdominalis, Eurytemora pacifica, and C. sinicus were the dominant copepods (Table 2 and   At station D4, a nearshore at the east of Jiaozhou Bay, µ = 0.49 d −1 , and g = 0.11 d −1 . In the present example for station E3, the Jiaozhou Bay mouth, µ = 0.23 d −1 , and g = 1.38 d −1 , indicating that mortality in excess of growth (µ < g). Phytoplankton 15 growth rates for the nutrient addition treatments were higher than those for the no-nutrient treatments at the three stations.
The results of dilution experiment at three stations (C3, D4 and E3) are presented in Table 3

Relationships with biotic and abiotic parameters
In Jiaozhou Bay, many factors (biotic and abiotic parameters) could affect DMS and DMSP concentrations, and five factors (salinity, temperature, Chl a concentrations, bacterial abundance, and copepod abundance) were selected to investigate their effects on DMS and DMSP in this study. Salinity was negatively correlated with temperature and DMSPp (p < 0.05). In 30 comparison, Chl a concentrations were positively correlated with the copepod abundance (p < 0.01), and the copepod 8 abundance was positively correlated with C. sinicus abundance (p < 0.05). Furthermore, positive correlations were observed among DMS, DMSPd, and DMSPp (p < 0.01). In addition, no significant correlation was observed between DMS/DMSP and zooplankton, copepod, C. sinicus abundance or Chl a concentration (p > 0.05).

10
The ingestion of C. sinicus of four species of algae resulted in different DMS productions (Fig. 6B). With regard to I. galbana and C. curvisetus, C. sinicus grazing promoted DMS production in the treatments compared with the controls, e.g., DMS in C. curvisetus treatment by C. sinicus grazing was 1.7-fold of DMS in the controls. In comparison, DMS production has an opposite changing trend for E. huxleyi and Gymnodinium sp., in which the treatments showed lower DMS production than that in the controls.

15
For four algae species, DMSPp concentrations in the controls ranged from 48.78 nmol L −1 to 7,165 nmol L −1 and those in the treatments ranged from 27.25 nmol L −1 to 11,055 nmol L −1 . In the control and treatment groups, DMSPp concentrations of E. huxleyi and Gymnodinium sp. were two to three orders of magnitude higher than those of I. galbana and C. curvisetus  increased. DMSPd in the treatments were higher than those in the controls (p > 0.05). DMSPd concentrations in the controls ranged from 10.60 nmol L −1 to 6,595 nmol L −1 . In comparison, DMSPd concentrations ranged from 11.39 nmol L −1 to 10,848 nmol L −1 in the treatments (Fig. 6D). DMSPd in the treatments were higher than those in the controls (p > 0.05).
Significant differences for the DMSPz and DMSPf contents of the four diets were observed (Fig. 6E). DMSPz contents decreased according the following order: E. huxleyi > Gymnodinium sp. > I. galbana > C. curvisetus. DMSPz contents of C.

25
sinicus fed on E. huxleyi were 23.51-fold of those fed on C. curvisetus (Fig. 6E). When compared with DMSPz contents, DMSPf of four diets had different contents with the following order: Gymnodinium sp. > E. huxleyi > I. galbana > C.

25
In this study, Fig. 8B shows that increased salinities restrained the production of DMS. Moreover, DMS contents in the treatment decreased from 84.54 nmol L −1 to 26.54 nmol L −1 when salinities increased, which was consistent with the DMS changes in the control (reduction from 41.11 nmol L −1 to 25.90 nmol L −1 ). The DMSPd changes indicated a trend similar to those of DMS contents, that is, increasing salinities decreased DMSPd contents. Fig. 8D shows that DMSPd contents decreased from 481.41 nmol L −1 to 72.35 nmol L −1 in the treatment. In comparison, DMSPd contents in the control reduced 30 from 423.99 nmol L −1 to 66.80 nmol L −1 . DMS contents of the treatments were higher than those of the controls (p > 0.05) and increased by 106%, 51%, 4.7%, and 2.4% at the salinities of 20,25,30,and 35 PSU,respectively. The increasing salinities facilitated the accumulation of DMSPp in the controls and treatments (Fig. 8C). DMSPp contents in the treatments increased from 92.20 nmol L −1 to 371.49 nmol L −1 , and those in the controls rose from 121.57 nmol L −1 to 532.16 nmol L −1 . DMSPp contents in the controls and treatments at 30 PSU were 2.9-and 1.7-fold of those at 25 PSU, respectively. DMSPp contents in the treatments were lower than those in the controls because of grazing activity (p > 0.05).
DMSPd contents in the treatments were higher than those in the controls (p > 0.05) (Fig. 8D). The relationship between IR and the increment of DMS + DMSPd proved to be positive (r = 0.662, p = 0.338, n = 4). In comparison, the correlation between the increments of DMS and DMSPd was negative (r = −0.955, p = 0.045, n = 4).
DMSPz ranged from 0.21 nmol L −1 to 5.38 nmol L −1 , reached the maximum content at 35 PSU, and reached the minimum 10 content at 20 PSU. When salinities increased, DMSPz initially increased, subsequently declined, and finally increased (

Effects of C. sinicus on DMS/DMSP in the field study
In Jiaozhou Bay, C. sinicus was the dominant copepod in 74 species of zooplankton identified from years 2010 to 2011. In 25 this study, no significant correlations were observed between zooplankton, copepod, and/or C. sinicus abundance and DMS/DMSP concentrations in the field study, illustrating that Jiaozhou Bay was a complex ecosystem with different abundances and types of phytoplankton, zooplankton, and/or copepod in the natural environment. Many kinds of copepods inhabited Jiaozhou Bay, and their IRs depended on copepod species, e.g., IRs of Harpacticus sp. (Yu et al., 2015) were 10fold of those of C. sinicus in this study. The gut contents of C. sinicus were checked, and the results showed that the C.

30
sinicus preferred to graze on diatom Chaetoceros curvisetus and Thalassiosira nordenskioldi (data not shown), suggesting that diatom (DMSP-poor algae) were the preferable diet for C. sinicus. Zheng et al (2014) and Luo et al (2016) have investigated the species composition and abundance of phytoplankton in 2010 and 2011 Jiaozhou Bay, respectively.
According to their data, the changing trends of species composition, abundance of phytoplankton and dinoflagellate/diatom ratio from June 2010 to May 2011 were presented in Table 4 and Fig. 9. The predominancy of dinoflagellate Ceratium fusus was 0.10 in September 2010 (Table 4). The dinoflagellate/diatom ratios in the three months (July, August and September) were high among the whole year, and the abundances of dinoflagellate and diatom in September were the highest among the 5 three months (Fig. 9), what is more, the bacterial abundance in September was the highest among the year (Fig. 2). Therefore, the occurrence of high abundances of dinoflagellate and bacteria might were the reason of high DMS and DMSP in September 2010. In February, April and May 2011, the dominant phytoplankton were diatom Rhizosolenia delicatula, Skeletonema costatum, and Skeletonema costatum, and the predominancies were 0.7, 0.99 and 0.68, respectively (Table 4).
Although the phytoplankton abundances and Chl a contents were high during January 2011 to May 2011, the DMSPp and 10 DMSPd concentrations were lower than those in September 2010, suggesting that DMSP concentration not only depend on phytoplankton abundance, but also phytoplankton species and other factors. We evaluated the effects of several agents (i.e., food, diet concentration, and salinity) on DMS and DMSP productions in the laboratory study. Our incubation data showed that copepod grazing increased DMS production, which was consistent with previous investigations on the effects of copepod grazing on DMS production (Dacey and Wakeham, 1986;Yu et al., 2015). Consistent with our field study results, 15 no significant correlations between mesozooplankton abundance and the distribution of DMS or DMSPd were also observed in the Gulf of Maine and St. Lawrence in previous studies (Cantin et al., 1996;Matrai and Keller, 1993). Cantin et al. (1996) concluded that mesozooplankton grazing played a minor role in DMS and DMSPd productions in the Gulf of St. Lawrence.

Dietary effects on copepod grazing and DMS/DMSP production
Different diets contain various DMSP contents, which affected DMS production induced by copepod grazing. DMSP-rich algae (E. huxleyi and Gymnodinium sp.) were not the preferable food for copepod C. sinicus in the incubation study because the released acrylic acid from DMSP in algae prevents them from copepod grazing. Based on the species of the diets, algae produced the corresponding DMS and DMSP (DMSPd and DMSPp) contents, which might be different. For example, when 25 comparing Gymnodinium sp. and E. huxleyi with C. curvisetus and I. galbana, we determined that cellular DMSPp in Gymnodinium sp. and E. huxleyi were one to two orders of magnitude higher than those in C. curvisetus and I. galbana (see Table 1). In terms of I. galbana, our detection results on cellular DMSPp production were consistent with those reported by Niki et al. (2000), indicating that cellular DMSPp production in given algae was approximately invariable.
When Gymnodinium sp. and E. huxleyi (DMSP-rich phytoplankton) were grazed by C. sinicus, DMSP released from algae 30 protected them from being grazed and stimulated the increase in DMSPp. In the natural environment, acrylic acid or DMSP depressed copepod appetites and forced them to ingest other DMSP-poor phytoplankton. Thus, DMSP-poor C. curvisetus became the favorite diet of copepods among four algae species, elucidating that the C. sinicus fed on C. curvisetus, which evidently promoted DMS production in this study. Further studies have revealed that DMSPd was a feeding inhibitor, and DMSPp, DMSPd, DMS, and acrylic acid constituted a cellular antioxidant system involved in the scavenging of hydroxyl radicals (Strom et al., 2003;Sunda et al., 2002). Non-DMS-producing phytoplankton species in a mixture of prey were preferentially selected by grazers, and single DMS/DMSP-rich diet decreased the food intake of copepods, e.g., copepod Harpacticus sp. had inferior IRs and PPRs when fed on DMS/DMSP-rich alga Prymnesium parvum (Wolfe et al., 1997; 5 Wolfe and Steinke, 1996;Yu et al., 2015).
Our results showed that DMSP in copepod bodies and fecal pellets accounted for 0.035% to 4.5% and 0.13% to 3.3% of DMS and DMSP in this study, illustrating that the ingestion of C. sinicus transferred DMSP from phytoplankton to the copepod bodies and fecal pellets. When compared with Harpacticus sp., lower DMSPf and DMSPz contributions in C.
sinicus attributed to the lower IRs of C. sinicus.

10
Based on the C. sinicus abundance on May 2011, the maximum DMSPz (0.02-7 nmol L −1 ) and DMSPf (0.1-12 nmol L −1 ) of C. sinicus in the seawater of Jiaozhou Bay were achieved. The results of our study and that of Tang (2001) showed that DMSPp in copepod bodies and fecal pellets was an essential part of DMSP flux in the ocean. In addition, another sink of DMSP in the ocean was achieved by microbial processes. The results of Tang et al. (2001) and Dong et al. (2013) indicated that copepods and their pellets harbored a dense population of DCB (DMSP-consuming bacteria), which played an important 15 role in DMSP degradation.
DMSPz and DMSPf contents were strongly associated with C. sinicus grazing. Their changing trends were consistent with that of IR, which depended on the critical concentration of phytoplankton. IRs of C. sinicus increased steadily with the increase in algal concentration below the critical concentration and decreased above the critical concentration. Critical concentration differed depending on the copepod and algal species, which were confirmed by this study and other 20 investigations (Yu et al., 2015).

Effects of salinity on copepod grazing and DMS/DMSP
In this study, low salinity induced high DMS production in seawater, whereas high salinity increased DMSPp (intracellular DMSP) and decreased DMSPd, which were consistent with the conclusion of Variamuthy et al. (1985). Our results were in accordance with the observation of a benthic diatom documented by van Bergeijk et al. (2003) and Skeletonema costatum 25 documented by Yang et al. (2011). Intracellular DMSP, as an osmotically active compound, was accumulated or released to help algal cells adjust their osmotic potential when salinity increased or decreased . C. sinicus grazing promoted DMS production in this study, which was consistent with the results obtained by many investigators (Belviso et al., 1990;Christaki et al., 1996;Daly and DiTullio, 1996;Leck et al., 1990). The decrement of DMSPp changed into DMS and DMSPd by copepod grazing, which was consistent with the positive relationships between IR and the increment of DMS + DMSPd 30 and/or the decrement of DMSPp. The increment of DMS was significantly negatively correlated with the increment of DMSPd, indicating that the cleavage of DMSPd was the source of DMS by DMSP lyase. Many reports on the location of DMSP lyase in the cell have been published. Wolfe and Steinke (1996) indicated that the DMSP lyase location of E. huxleyi 13 CCMP 370 was in the membrane bound inside cells. Stefels and Dijkhuizen (1996) reported that DMSP lyase of Phaeocystis was membrane-bound and located extracellularly. Cellular locations and functions of DMSP lyase might differ depending on algal species.
This study confirmed that DMS production was affected by primary producers, e.g., algae. DMS/DMSP in algae transferred to the food web by predation, and several researchers investigated the effects of zooplankton grazing on DMS 5 production with copepods and krill, indicating that the breakage of algal cells through sloppy feeding may increase DMSPd production (Dacey and Wakeham, 1986;Kaamatsu et al., 2004). When DMSPp concentrations changed, DMSPf and DMSPz concentrations were altered correspondingly, which was verified by other results in which the DMSP defecation rate of copepod Acartia tonsa feeding on Tetraselmis impellucida (prasinophyte) was closely related to food concentration and DMSPz content (Tang, 2001). Salinity changes altered the osmotic pressure surrounding copepod and algae cells, which in 10 turn adjusted DMSP (intracellular DMSPp, DMSP in tissues, and DMSP in gut content). Copepods contained more DMSP at high salinity, indicating the osmoregulatory function of DMSP (Tang et al., 1999).

Conclusions
In the present study, field and incubation experiments were performed to investigate the effects of C. sinicus grazing on DMS production in Jiaozhou Bay. Copepods (C. sinicus) was the dominant copepod in Jiaozhou Bay and preferred to graze 15 on diatom. Appropriate diets and salinities facilitated DMS/DMSP production, e.g., C. sinicus feeding on I. galbana and C. curvisetus exhibited increased DMS production at 30 PSU in the laboratory experiment. C. sinicus grazing promoted the productions of DMS and DMSPd, and DMS was released mainly from DMSPd and low salinity increased DMS production.
Copepods and fecal pellets supplied substantial DMSPp into the water column and were important to the biogeochemical cycling of DMS.

20
Competing interests. The authors declare that they have no conflict of interest.