Article pubs.acs.org/est
Ocean Acidification and Fertilization in the Antarctic Sea Urchin Sterechinus neumayeri: the Importance of Polyspermy Mary A. Sewell,*,† Russell B. Millar,‡ Pauline C. Yu,§ Lydia Kapsenberg,§ and Gretchen E. Hofmann§ †
School of Biological Sciences and ‡Department of Statistics, University of Auckland, Private Bag 92019, Auckland 1142, New Zealand § Department of Ecology, Evolution and Marine Biology, University of California, Santa Barbara (UCSB), Santa Barbara, California 93106-9620, United States S Supporting Information *
ABSTRACT: Ocean acidification (OA), the reduction of the seawater pH as a result of increasing levels of atmospheric CO2, is an important climate change stressor in the Southern Ocean and Antarctic. We examined the impact of OA on fertilization success in the Antarctic sea urchin Sterechinus neumayeri using pH treatment conditions reflective of the current and nearfuture “pH seascape” for this species: current (control: pH 8.052, 384.1 μatm of p CO 2 ), a high CO 2 treatment approximating the 0.2−0.3 unit decrease in pH predicted for 2100 (high CO2: pH 7.830, 666.0 μatm of pCO2), and an intermediate medium CO2 (pH 7.967, 473.4 μatm of pCO2). Using a fertilization kinetics approach and mixed-effect models, we observed significant variation in the OA response between individual male/female pairs (N = 7) and a significant population-level increase (70−100%) in tb (time for a complete block to polyspermy) at medium and high CO2, a mechanism that potentially explains the higher levels of abnormal development seen in OA conditions. However, two pairs showed higher fertilization success with CO2 treatment and a nonsignificant effect. Future studies should focus on the mechanisms and levels of interindividual variability in OA response, so that we can consider the potential for selection and adaptation of organisms to a future ocean.
■
INTRODUCTION
able to regulate the pH in specific cells (e.g., calcifying primary mesenchyme cells18). Studies of the effects of OA on fertilization success in a variety of marine invertebrates has, however, produced a broad spectrum of results; with some species showing negative effects on fertilization success at pH decreases expected for the year 2100 (0.4 units), but with the majority of studies showing no significant effects (review in ref 6, 19−23). In sea urchins specifically, only a few studies show a negative effect on fertilization success,24−28 with others showing no significant effects (e.g., 29−33), negative effects only at extreme pH,34−36 or considerable individual variability in the response.37 In this paper, we examine the effects of OA on fertilization in the Antarctic sea urchin Sterechinus neumayeri. Continuous pH time series data from coastal depths in McMurdo Sound have shown that the pH is particularly stable during the early spawning season of S. neumayeri.38,39 Consequently, we chose pH treatment conditions that are reflective of the current and
Ocean acidification (OA), the reduction of the seawater pH as a result of increasing levels of atmospheric CO2, is one of the climate change stressors currently impacting marine ecosystems (reviewed in refs 1−6). Since the industrial age, anthropogenic atmospheric CO2 has reduced the pH of the world’s oceans by 0.1 unit (a 30% elevation in the H+ concentration), and by 2100, the pH is predicted to decrease by another 0.3−0.4 units.7 Lower pH also lowers carbonate concentrations, and the action of low carbonate and higher H+ impedes calcium carbonate skeleton formation and perturbs acid−base regulation of marine organisms.4,7−10 Because the CO2 solubility is greater in cold water,8 organisms inhabiting polar ecosystems, such as the Southern Ocean and Antarctic, may be particularly vulnerable to the effects of OA.5,11−15 The early developmental stages of marine organisms are predicted to be more susceptible to OA than the adults because gametes are generally thought to lack specialized ion-regulatory epithelia and, with the ocean as their extracellular space, have limited ability to regulate their acid−base environment,16 although embryos may have cellular defenses against other environmental challenges such as UV and pathogens17 and be © 2013 American Chemical Society
Received: Revised: Accepted: Published: 713
June 26, 2013 November 18, 2013 December 3, 2013 December 3, 2013 dx.doi.org/10.1021/es402815s | Environ. Sci. Technol. 2014, 48, 713−722
Environmental Science & Technology
Article
near-future “pH seascape” for this species39 to provide an environmentally relevant test of the impact of OA on fertilization success. The experiments were designed to address the mechanism whereby OA might be affecting fertilization success by using a fertilization curve approach40,41 to determine if OA was impacting maximum fertilization success and/or the efficiency of polyspermy blocks, as previously noted in temperate sea urchins.25
ately before use. Three replicate (A−C) eight-point sperm dilution series (vials 1−8) were prepared using 2.3 mL of the initial stock (100 μL of dry sperm in 9.9 mL of the CO2 treatment water), the vials were gently mixed three times by inversion, and 2.3 mL was transferred to the next vial in the dilution series. In the final dilution, 2.3 mL of the mixed solution was discarded to allow space for the eggs. The vials were incubated in the seawater table at a subzero temperature (range over 7 days of −0.9 to −1.3 °C) until the eggs were added within minutes of completion of the sperm dilution series. Eggs were preincubated in CO2 treatment water for 20 min, while the sperm dilution series was being prepared, at a concentration so that 1000 eggs were added in the 2.3 mL of water required to completely fill each scintillation vial. Replicate sperm dilutions received eggs sequentially (e.g., vials 1A, 1B, 1C, 2A, etc.), and the vials were gently inverted to mix the eggs and sperm and returned to the seawater table for incubation. After the eggs had been in the sperm dilution for 30 min (approximately the time taken to add the eggs to all 24 vials at 1 vial/min), approximately 90% of the sperm/treatment water was removed by reverse filtration, and the vial was refilled with treatment water at the incubation temperature. This process was repeated for the remaining two CO2 treatments; the order of CO2 treatments was randomized each day. Scintillation vials were left to incubate in the seawater table at approximately −1.0 °C for 18−24 h to allow embryos to undergo several cleavages48 and allow assessment of the levels of polyspermy.41,49 At least 200 eggs/embryos from each vial were classified into four categories (embryo normal development, embryo abnormal development, unfertilized egg, and incompletely elevated fertilization membrane) by running transects across a Sedgewick-Rafter cell. Embryos were generally at the 4−8 cell stage when development was assessed. A total of 72 scintillation vials (3 CO2 treatments × 8 dilutions × 3 replicates) were involved in each experimental trial, which used the gametes from one male/female pair. Each trial was repeated seven times (N = 7 biological replicates). To ensure that there were no changes in the pH during the course of the fertilization trial, scintillation vials containing CO2 treatment water only were incubated for the 30-min exposure period, and the pH was measured using the spectrophotometric methods described above. The pH was not measured again at the end of the experiments when the percentage of normal development was quantified. Sperm concentrations were determined from hemocytometer counts (N = 3 replicate counts from an equivalent of Dilution 2 made separately from the treatment stock). Egg diameters were determined from digital photographs of unpreserved eggs taken on a Zeiss Axioskop microscope after completion of the third treatment of that male/female pair, within 3−4 h of the start of the experiment (mean N = 37, range 29−42). Fertilization Curves. Fertilization curves were fitted to individual replicate experiments (resulting in a total of 63 fits: 7 pairs × 3 treatments × 3 replicates) using the Millar−Anderson model.41 Fitting of the curves was implemented using nonlinear least squares via SAS PROC NLMIXED, after the variance stabilizing arcsin-square-root transformation was first applied to the observed fertilization. The Millar−Anderson model is a formal extension of the fertilization kinetics model of Vogel et al.50 to incorporate the effects of polyspermy. Models of fertilization kinetics require specification of a collision rate constant, β0, given by the product of the sperm swimming
■
MATERIALS AND METHODS S. neumayeri was collected from Cape Evans, Ross Island, Antarctica (S 77° 38.059′, E 166° 26.905′), by divers in approximately 12 m of seawater on a cobble bottom. Animals were returned in coolers to the Crary Laboratory at McMurdo Station and maintained in flow-through seawater tanks at a temperature of ∼1 °C and under continuous light conditions until use. For each experiment, gametes were obtained from one male and one female sea urchin after injection of ca 1 mL of 0.5 M KCl into the coelomic cavity. Eggs were collected in a 100 mL beaker filled with 0.35 μm filtered seawater (FSW; pHTOTAL mean ± SD = 7.991 ± 0.003, N = 7 days); sperm was collected “dry” using a 10 μL pipettor with a cut tip directly from the gonopore and stored on ice until use. Gamete quality was examined on a compound microscope to check for egg roundness, the absence of accessory cells, and sperm motility. Gametes were not used in the experiments unless a test fertilization showed >90% fertilization and normal elevation of fertilization membranes. Preparation of Treatment Waters. Seawaters of differing pCO2, referred to here as CO2 treatment waters, were prepared using the system described in detail by Fangue et al.42 with minor modifications. Briefly, dried CO2-free air is mixed with CO2 using mass flow controllers, and the mixed gas at the target pCO2 is bubbled into FSW in reservoir buckets through a Venturi injector until equilibrium.42 Seawaters were prepared to target the natural baseline pH recorded near McMurdo Station (8.019−8.04538) as a control, a high-CO2 treatment (target 700 μatm of pCO2) approximating a 0.2−0.3 unit decrease in pH predicted for 2100, 7 and an intermediate medium-CO2 treatment (target 500 μatm of pCO2). Treatment seawater for the fertilization experiments was taken directly from the reservoir buckets immediately before use. Measurements of the pH and total alkalinity (AT) in the reservoir buckets were taken daily using the standard operating procedures (SOP 43 ) for pH TOTAL (SOP 6b) using a spectrophotometer (Shimadzu Instruments) and total alkalinity by open-cell titration (SOP 3b) using a Mettler-Toledo T50, as described by Fangue et al.42 The temperature and salinity were recorded daily with a calibrated wire thermocouple (Omega) and a 3100 conductivity probe (Conductivity Meter, YSI), respectively. CO2SYS software for MS Excel44 was used to calculate the carbonate chemistry parameters based on the dissociation constants of Mehrbach et al.45 as refit by Dickson and Millero.46 Assessment of Fertilization Success. Fertilization experiments were conducted in 20-mL glass scintillation vials using established methods,25,47 with the modification that the full volume (23 mL) was used in the preparation of the dilution series to ensure that there was no headspace during the period of fertilization. For each fertilization trial, the vials were prerinsed with the appropriate CO2 treatment water immedi714
dx.doi.org/10.1021/es402815s | Environ. Sci. Technol. 2014, 48, 713−722
Environmental Science & Technology
Article
Table 1. Measured and Calculated (CO2SYS, Pierrot et al., 2006) Seawater Parameters for CO2 Treatment Waters for the Seven Male/Female Pairsa control
medium CO2
high CO2
pHTOTAL salinity temperature (°C) alkalinity (μmol/kg of SW)
8.052 ± 0.006 34.8 ± 0.04 −1.07 ± 0.11 2336.76 ± 12.57
pCO2 (μatm)
384.1 ± 4.4
Measured Parameters 7.967 ± 0.005 7.830 ± 0.009 34.8 ± 0.04 34.8 ± 0 −1.10 ± 0.06 −1.08 ± 0.04 2331.81 ± 14.10 2334.93 ± 12.44 Calculated (CO2SYS) 473.4 ± 4.5 666.0 ± 14.4
HCO3− (μmol/kg of SW) CO32−(μmol/kg of SW) ΩCa ΩAr
2104.41 ± 10.79 91.73 ± 1.29 2.20 ± 0.03 1.38 ± 0.02
2137.45 ± 12.21 76.64 ± 1.03 1.84 ± 0.03 1.16 ± 0.02
2189.72 ± 11.58 57.24 ± 1.26 1.38 ± 0.03 0.86 ± 0.02
one-way ANOVA results F2,18 = 2378.79, p < 0.0001 C < M < H no test performed F2,18 = 0.46, NS F2,18 = 0.24, NS F2,18 = 2299.57, p < 0.0001 C < M < H F2,18 F2,18 F2,18 F2,18
= = = =
90.82, p < 0.0001 C < M < H 1624.49, p < 0.0001 C > M > H 1606.56, p < 0.0001 C > M > H 1620.82, p < 0.0001 C > M > H
All values are mean ± SD (N = 7). The results of one-way ANOVA tests are shown (see the Materials and Methods section for details). When oneway ANOVA tests were significant, Tukey’s (α =0.05) groupings are shown: control (C), medium (M), high (H). NS = nonsignificant.
a
speed (mm/s) and egg cross-sectional area (mm2). The average cross-sectional area of S. neumayeri eggs for the seven females in this experiment was 0.029 mm2 (N = 260). Independent measurement of the sperm swimming speed was not possible because we did not have access to a tb, as is the case here. Analysis of the residuals from fits of the Millar−Anderson model41 (generated by SAS PROC NLMIXED) showed a lack of fit in some replicates when the maximum observed fertilization was substantially below unity. This indicated that imperfect fertilization could not simply be explained by polyspermy. Thus, the Millar−Anderson model41 was modified by assuming that only proportion Pmax of the eggs was compatible, where Pmax is an additional parameter to be estimated. That is, the fitted fertilization curve was given by multiplying the Millar−Anderson curve in eq 2 by Pmax. Each fitted fertilization curve provided an estimate of tb (the blocking time), F (the fertilizable fraction), and Pmax (the proportion of compatible eggs). Using these values, the maximum fertilization (Fmax) and corresponding sperm concentration (Smax) were obtained by numerical maximization of the curve. Statistical Analysis. One-way analysis of variance was used to confirm that there were significant differences in the seawater chemistry between the three CO2 treatments and to test for differences in the egg size between females (SAS, version 9.2). 715
dx.doi.org/10.1021/es402815s | Environ. Sci. Technol. 2014, 48, 713−722
Environmental Science & Technology
Article
Figure 1. (a−g) Percentage of normal development (± SE, N = 3) in control and medium- and high-CO2 treatment waters for male/female pairs 1− 7 (see Table 1 for full seawater chemistry) plotted against the sperm concentration (sperm/mL). Dilutions 1−8 run from the far right on the X axis toward the origin.
with a grand mean (±SD) of 191.97 ± 6.10 μm. Female 1 had significantly larger eggs than all other females (203.95 μm; Tukey HSD, α = 0.05), with the smallest eggs in females 3 and 7 with mean diameters of 185.02 and 186.48 μm, respectively. Eggs/embryos were assessed for development after 18−24 h, when normally developing S. neumayeri embryos had undergone two cell divisions.48 Fertilization curves (as a percentage of normal development) in S. neumayeri were generally of typical shape, with a reduction in fertilization success at high sperm concentrations that are probably due to polyspermy (Figure 1). Abnormal development under OA conditions fell into two distinct categories: (1) abnormal cleavage patterns that resulted in uneven cell divisions (e.g., three cells, five cells, etc.) and/or unequal blastomere sizes and (2) tight fertilization membranes (TFMs; sensu Tyler and Scheer54), where the fertilization membrane failed to elevate to a normal extent and appeared to be “soft” and easily deformed (Figure 2). An apparent consequence of TFM was that embryo cleavage, if it occurred, was spatially constrained and often resulted in abnormal development, as described in category 1. Embryos that showed no cleavage after the 18−24 h incubation period, with either a normal fertilization or TFM, were also considered to be abnormal. Unfertilized eggs were present in all male/ female pairs at low sperm concentrations. TFM was present in all male/female pairs, generally at a low percentage. At dilution 4, the sperm concentration at which the
Figure 2. Light microscope images of (a) normal fertilization and (b) TFM. Scale bar = 100 μm.
maximum percentage of normal development generally occurred (Figure 1), the mean TFM was
