Feeding response of the tropical copepod Acartia erythraea to short-term thermal stress: more animal-derived food was consumed

Copepod collection and gut clearance

Zooplankton were collected at dusk on April 23, 2014, using surface plankton net-tows (diameter 50 cm, mesh size 505 µm) in the coastal waters of Luhuitou, Sanya Bay (109°28′E, 18°12′N). Live copepods were kept in a 1-L plastic container and transferred to the laboratory for gut evacuation. For the feeding experiment, surface seawater was also collected using a clean 25-L bucket in the same region.

Copepods were immediately transferred into several 500-mL beakers containing 0.45 µm-filtered seawater for gut evacuation that was performed in the dark for >48 h. Every 4 h during the clearance period, the copepods were gently collected using 200-µm mesh and rinsed 3–5 times with fresh seawater (FSW) to remove any attached detritus, and then transferred to FSW to fully empty their gut contents. After this procedure, healthy adult A. erythraea females of similar size were selected for the feeding experiment.

Feeding experiment

According to the ambient field temperature, the feeding experiment was performed at three different temperatures: 29 °C, 33 °C, and 35 °C. The temperature of the sampling site was 29 °C, therefore this temperature treatment was used as the control. According to previous studies, a 6 °C temperature increase was observed in the waters of thermal discharge-affected areas in tropical zones (Jiang et al., 2009); therefore, 33 °C and 35 °C were used as the experimental temperatures, with the goal of investigating the rapid feeding response of copepods to acute heat shock.

Twenty liters of collected seawater were gently filtered through 0.45-µm polycarbonate membranes and then re-suspended in 5 L of FSW to concentrate. After thorough mixing, a 50 mL sample of concentrated seawater was removed and immediately fixed with Lugol’s solution for cell counting as the prey concentration before the experiment. Another 200 mL seawater was taken and centrifuged at 3,000 × g for 15 min, and the concentrated samples were pelleted by centrifugation at 12,000 × g for 5 min. The cell pellets were re-suspended in 0.5 mL DNA buffer for DNA extraction. The concentrated seawater was then aliquoted into twelve 200-mL polycarbonate bottles to serve as a source of prey. For each treatment, a control group, without the addition of copepods, was prepared to monitor the growth of phytoplankton during the short exposure period. Each treatment had three replicates. Fifty adult individuals of A. erythraea were added to each experimental bottle and then incubated, at different temperatures, for 4 h in tanks with flowing seawater culture systems. After a 4 h incubation period, individual copepods were gently collected using 200 µm mesh and immediately fixed with neutral Lugol’s solution to avoid defecation. Neutral Lugol’s (no acetic acid added) was found to be effective in preserving DNA of phytoplankton and other copepod associated organisms (Zhang & Lin, 2002; Hu et al., 2014). There was no mortality of copepods during the short-term experiments. At the end of the experiments, 50 mL water samples were collected from each treatment and preserved with Lugol’s solution as the prey concentration after the experiment. There was no measurable growth of phytoplankton during this period.

Microscopic analysis of phytoplankton in seawater

Seawater samples were allowed to settle for 2 d in the dark, and then concentrated to 1 mL; they were then identified and enumerated in a Sedgwick-Rafter counting chamber under an Olympus BX51 microscope. The cell density of the seawater before and after the feeding experiment was obtained for the calculation of feeding parameters.

Grazing rate and filter rate based on the phytoplankton cells

The grazing rate (GR, cells ind ⁻¹h⁻¹) was calculated using the following equation: ${\displaystyle \mathrm{GR}=V\times \frac{C_0-C_t}{N\times t}.}$ The filter rate (FR, mL ind ⁻¹h⁻¹) was calculated using the following equation: ${\displaystyle \mathrm{FR}=V\times \frac{ln{C}_0-ln{C}_t}{N\times t}}$ where V (ml) is the volume of seawater in the experimental bottle, N is the number of copepods in the bottle, and t (h) is the incubation time. C ₀ and C_t (cells mL⁻¹) are the prey concentrations (phytoplankton cells in the experimental seawater) before and after the feeding experiment, respectively.

Molecular analysis of copepod diet

A previously reported procedure was followed (see Hu et al., 2014). Briefly, the sorted fixed copepod samples were thoroughly rinsed >3 times, with autoclaved 0.45-µm-filtered seawater and again with sterilized water several times. Copepods were examined under stereomicroscopy to ensure no attachment of other visible organisms on the body surface and then homogenized in a microfuge tube using a disposable micro pestle (Eppendorf, Hamburg, Germany). DNA from all copepod and seawater samples was extracted following the CTAB protocol. The DNA samples were PCR amplified using a PCR protocol based on a CEEC primer set (Hu et al., 2014). The PCR amplifications were concentrated using a Zymo DNA Clean & Concentrator TM-25 Kit. The target bands (∼0.8 kb) were recovered from a 1% agarose gel using Zymoclean™ Gel DNA Recovery Kit (Zymo Research, Irvine, CA, USA). The purified PCR products were ligated into a PMD18-T vector (Takara Bio Inc., Kusatsu, Japan) and transformed to DH-5 α competent cells (Takara Bio Inc., Kusatsu, Japan). Approximately 50–80 clones for each copepod sample and 150–200 clones for each seawater sample were randomly picked and sequenced (Invitrogen sequencing company).

Phylogenetic analysis of sequence data

Sequences obtained were trimmed of primer sequences and then searched against the GenBank database using the Basic Local Alignment Search Tool. The identified sequences were represented by the highest homologous sequences. Identified clones were categorized into plant-origin and animal-origin, and then a heat-map analysis was conducted using R (R Core Team, 2017). Our sequences were then aligned with the best hits using CLUSTAL W (1.8). The resulting alignment was imported into MEGA 6.0 to identify the best-fit nucleotide substitution model. The best-fit model Kimura 2, with gamma distribution (K2+G), was then employed for maximum likelihood (ML) analysis. The reliability of the tree topology was evaluated using bootstrap analysis with 1,000 replicates for ML analysis (Hu et al., 2014). Diversity indices (Chao-1, Simpson, and Shannon) were estimated using Past 3.05 to evaluate the sequencing depth of all the samples.

Omnivory index and Ivlev’s selectivity index

A percentage of each sequences from the whole clone library of each copepod sample was used to estimate the relative diet proportion. The trophic level and percentage were recorded to calculate the omnivory index (OI). This index was first introduced in the initial version of the Eco-path software and is calculated as the variance of the trophic level of a consumer’s prey groups: ${\displaystyle O{I}_i=\sum _{j=1}^n{\left(T{L}_j-\left(T{L}_i-1\right)\right)}^2\cdot D{C}_{ij}}$ where TL_j is the trophic level of prey j, TL_i is the trophic level of the predator i, and, DC_ij is the proportion prey j constitutes to the diet of predator i. When the omnivory index value is zero, the consumer in question is specialized, i.e., it feeds on a single trophic level. A large value indicates that the consumer feeds on many trophic levels, or in other words is more omnivorous (Pauly, Soriano-Bartz & Palomares, 1993).

Ivlev’s selectivity index (Ivlev, 1961) was also used to determine the selectivity pattern of A. erythraea in the incubation experiment: ${\displaystyle E_i=\left(r_i-P_i\right)∕\left(r_i+P_i\right)}$ where E_i is the Ivlev’s selectivity index, r_i is the relative abundance of prey i in the diet of A. erythraea and P_i is the relative abundance of the prey in ambient seawater. Observed values ranged from −1 to 1, where −1 indicates prey avoidance, 0 indicates that a prey species is being ingested at the same proportion as it is found in the environment, and 1 indicates a preference for a determined prey.

Grazing rate and filter rate of A. erythraea in different temperature treatments

Generally, the grazing rate and filter rate of A. erythraea showed an overall decrease with increasing temperature (Table 1). Compared with the control group (29 °C), the feeding of A. erythraea was increased slightly at 33 °C, with the grazing rate increasing from 40.04 ± 0.07 to 40.14 ± 0.11 cells ind⁻¹ h⁻¹ and the filter rate increasing from 2.59 ± 0.04 to 3.34 ± 0.11 mL ind⁻¹ h⁻¹, respectively. At 35 °C, the grazing rate of A. erythraea dramatically decreased to 28.20 ± 0.14 cells ind⁻¹ h⁻¹, and the filter rate decreased to 0.44 ± 0.11 mL ind⁻¹ h⁻¹. Increased temperature influenced the feeding behavior of A. erythraea; their ability to filter feed, especially phytoplankton, was limited when the temperature was higher than 33 °C.

Dietary composition of A. erythraea from different temperature treatments

18S rDNA clone libraries were constructed for A. erythraea fed at different temperatures, as well as the plankton community of the concentrated field seawater used in the experiment. Based on randomly sequenced clones, a total of 12, 10, and 10 taxa of prey were detected for A. erythraea at 29 °C, 33 °C, and 35 °C, respectively (Table 2). Although Chao-1 estimates indicated that the actual numbers of taxa were likely to be 19, 10, and 11, respectively, the prey taxa detected in the experimental groups (33 °C and 35 °C) were similar to the Chao-1 estimates, indicating an adequate coverage of diversity to demonstrate the effects of elevated temperature.

Despite the incomplete recovery of the entire dietary diversity in the control group (29 °C), prey taxa found in the samples consisted of a wide phylogenetic range of organisms, including diatoms, metazoans, protozoans, and fungi, as well as land plants (Fig. 1). In the 29 °C treatment, land plants (99% identical to Hordeum jubatum AF168852.1) and diatoms (98% identical to Cylindrotheca closterium LC054954.1) dominated the diet, accounting for 46% and 36%, respectively. Other numerically minor constituents such as fungi, protozoans, and Synurophyceae were also detected. In the 33  °C treatment, the same diatom species (Cylindrotheca closterium) was the most abundant prey taxa, comprising almost 58% of the diet. Other prey items belonged to the metazoans, protozoans, dinoflagellates, Synurophyceae, and fungi. The dietary spectrum of A. erythraea at 29 °C and 33 °C was similar. However, in the 35 °C treatment, a less diverse array of prey taxa was detected, with the contribution of metazoan taxa increasing up to 70%, but diatoms decreased to 16% (Fig. 1).

Feeding preference changes with increasing temperature

Although the dietary spectrum was similar at different temperatures, the species numbers and relative proportions of each prey group (e.g., diatoms, metazoans, and protozoans) differed (Fig. 1). The omnivory index of A. erythraea at 29 °C and 33 °C was 0.1 and 0.24, respectively, while a remarkable increase (as high as 2.84) was observed in the 35 °C treatment. This indicates a more carnivorous feeding habit under temperature stress. The feeding habit of A. erythraea showed a likely shift from herbivory to carnivory when the temperature exceeded 33 °C. Furthermore, we divided all the prey items of A. erythraea from the different temperature treatments into three groups, metazoa, protozoa, and vege-prey (includes phytoplankton and plants), and calculated their Ivlev’s selectivity index. Acartia erythraea showed positive feeding on vege-prey at 29 °C and 33 °C, but negative feeding on vege-prey at 35 °C. For metazoan prey, A. erythraea showed negative feeding at 29 °C but strong positive feeding in the 35 °C treatment, and the Ivlev’s index increased from −0.62 to 0.75 (Fig. 2).