Effect of reduced pH on physiology and shell integrity of juvenile Haliotis iris (pāua) from New Zealand

Experimental set up

Haliotis iris physiological response

We assessed survival at weekly intervals, and growth and physiological condition at the end of the experiment. Survival assessments included all individuals in each replicate, while all other responses were made on one or two randomly selected individuals per replicate. The absolute growth of live pāua, identifiable from the blue coloured shell, was measured to the nearest mm (from photographs with a scale-bar included, using Image-J) and expressed as a daily growth rate (μm d⁻¹). The relative growth rate (% change in SL relative to the initial SL) was also calculated, using the formula of Hopkins (1992). Physiological condition was calculated using the ratio of dry flesh weight (FW; dried at 60 °C for 48 h) to dry SW (air dried for 48 h), following Lucas & Beninger (1985; FW/SW*100). The resultant condition index (CI_FW:SW) provides an integration of the present metabolic state of the organism (Mann, 1978; Roper et al., 1991). Simple allometric relationships between SW and SL were also examined. One individual per replicate was available for the evaluations involving FW, and two individuals per replicate for those involving SW.

Shell morphology and integrity

To evaluate changes in morphology and integrity of the pāua shells over the course of the experiment, shells from one pāua from each of the six treatment replicates were examined (with the exception of the pH 8.00/15 °C treatment: only five replicates were analysed). Whole shells were first cleaned in MilliQ water in an ultrasonic bath, to remove detritus from the shell surfaces, and to partially hydrate the organic matrix in the shell (Sun & Bhushan, 2012) making it less prone to shattering during sectioning. Precision cuts were then made on the surface (Fig. 1B) that did not fully penetrate the shell, in order to isolate the three specific portions for subsequent sampling: new shell grown during the experiment (blue), pre-experiment shell (brown), and the spire (oldest part of the shell). Once dry the shells were fractured by bending, using forceps with thick soft material on the inner faces to prevent physical damage to the surfaces. The different shell sections were analysed using X-ray diffractometry (XRD), scanning electron micrographs (SEM) and energy-dispersive X-ray spectroscopy (EDS), as described below.

Shell ultrastructure, thickness and surface characteristics were assessed using SEM. Each shell section was coated with gold-palladium and photographed in a JEOL SEM-2200FS Cryo-TEM on the cross section, dorsal and ventral surfaces at multiple scales. The SEM cross section for the experimental (new) growth portion always included the growing margin of the shell, to ensure consistent representative measurements of layer thicknesses along the shell. Cross-sections at 30× magnification were used to assess total shell thickness, as well as thicknesses of the (inner) nacre and (outer) prismatic layers (Figs. 2A and 2C). For each shell, layer thickness was assessed at eight positions across the shell—four across the experiment phase growth (labelled 1–4) and four positions in the pre-experiment acclimation phase (labelled 5–8), using ImageJ/Fiji (Schindelin et al., 2012) (Fig. 1C). Imaged positions were approximately 1.0–1.3 mm apart, starting at the growing edge (i.e. position 1). At the growing edge, the nacre had not yet formed; hence the nacre thickness is zero μm.

The degree of etching of the outer (dorsal) shell surface was assessed on 500× magnification images of the newly deposited and pre-experiment shells, each incorporating an area of 4,610 μm² of shell surface. The proportion of the shell that was unetched, lightly etched or heavily etched was quantified using ImageJ (Fig. 1D), as a percentage of the total imaged area. Finally, crystal ultrastructure, size and formation were examined using SEM micrographs of cross sections and surfaces taken at 3,000× magnification.

Shell composition (mineralogy)

Changes in mineral composition of shells were examined (one individual per replicate). Bulk skeletal composition (calcite:aragonite ratio, and wt% MgCO₃ in calcite) was determined using XRD, while a higher spatial resolution EDS analysis was used to determine quantitative elemental composition (wt% calcium and wt% magnesium in the carbonate). For XRD analysis samples were powdered to crystallites, spiked with 0.1 g halite (NaCl) as an internal standard, and smeared onto a glass slide (per Gray & Smith, 2004). The position of the calcite peak (near 29.4 °2θ) was corrected using the known position of the halite peak at 31.72 °2θ, and wt% MgCO₃ in calcite was calculated using the machine-specific calibration equation of Gray & Smith (2004), after the methods of Chave (1952, 1954). Wt% calcite in carbonate was calculated using the Peak Height Ratio method (Gray & Smith, 2004). For EDS analysis, shell fragments were coated in Au/Pd and a four mm² area of the shell surface examined. The same locations on each specimen were selected: (a) based on the number of counts/second recorded (>1,500 counts/s is required to provide an accurate sample), and (b) to avoid clustering of readings on the surface. Each scan ran for 100 s. Percentages of the target elements calcium and magnesium present in the sampling area were generated from comparisons to factory standards.

Statistical analyses

Differences in each response variable between treatments were assessed using two-way crossed ANOVA after assumptions were satisfied, with pH, temperature and a pH*temperature interaction term included. Assumptions were checked by examining the residual distribution plots and residuals vs predicted values and quantiles and using the Shapiro–Wilk test for normality. Any variables for which these suggested non-normality or heterogeneity of variance were rank transformed prior to analysis. Actual differences among treatment means were determined using Duncan’s multiple comparisons tests, where no significant interaction was detected. If a significant interaction term was found, multiple comparisons determined the effect of pH at each temperature separately, and temperature at each pH separately. All analyses were conducted using SAS software (SAS Institute), with p < 0.05 used to indicate a statistically significant response.

Haliotis iris physiological response

The majority of the pāua survived the experiment, with only five mortalities across all treatments. These included three individuals from the lowered pH, 15 °C treatment and one each from the ambient pH treatments at each temperature. All the pāua grew measurably in the experimental conditions (Table 2). The average amount of new growth ranged from 5,273.5 ± 860.8 μm in the lowered pH, 13 °C treatment (a 21% increase in SL over the experiment; 44 μm d⁻¹), to 8,837.5 ± 578.5 μm in the ambient pH, 15 °C treatment (a 34% increase; 73 μm d⁻¹). Although growth was lower at pH_T 7.66 than at ambient pH, at each temperature (Table 2), this was not statistically significant. Both the amount of new growth and the percentage growth were significantly greater at 15 °C than at 13 °C (Table 2; new growth: pH p = 0.1844, temp p = 0.0003, pH*temp p = 0.4486; % growth: pH p = 0.3290, temp p = 0.0002, pH*temp p = 0.5337). Physiological condition (CI_FWSW) was reduced at 15 °C, but was not affected by pH (Table 2; pH p = 0.0684, temp p = 0.0367, pH*temp p = 0.5733). Dry FW (normalised by SL) of the pāua from the ambient pH/13 °C treatment was slightly higher than those in the other treatments, although not significantly so (Fig. 2A; CI_FWSL pH p = 0.6272, temp p = 0.8583, pH*temp p = 0.1202).

Shell morphology and integrity

The relationship between SW and SL in the juvenile H. iris is shown in Fig. 2B. Individuals from the lowered pH treatments had lighter shells for a given SL, a pattern that was true at both 15 and 13 °C (Fig. 2B). CI_SW:SL was significantly lower at pH_T 7.66 than ambient pH_T 8.00, and also at 13 °C than 15 °C (pH p = 0.0129, temp p = 0.0127, pH*temp p = 0.3623; Table 2).

The total thickness of the shells, measured from the SEM images, ranged from 300 to 600 μm on average, and did not differ significantly between treatments at any of the eight positions assessed (p > 0.05 in all cases). Figures 3A and 3B shows the prismatic and nacre layer thickness across the experimental shell. Comparisons revealed variable effects of pH and/or temperature at positions 4, 3 and 1. At position 4 the prismatic layer was significantly thinner at pH_T 7.66 than at ambient pH_T 8.00 (pH p = 0.0004, temp p = 0.8796, pH*temp p = 0.0567). At position 3, individuals from lower pH also had a thinner prismatic layer (pH p = 0.0315, temp p = 0.7088, pH*temp, p = 0.0883), as well as a thicker nacre layer (pH p = 0.0203, temp p = 0.0431, pH*temp p = 0.7338). The nacre at position 3 was also thinner at 13 °C than at 15 °C. At the growing edge (position 1) where the nacre layer had not yet formed, the prismatic layer was thicker at 13 °C than at 15 °C but was not affected by pH (pH p = 0.8537, temp p = 0.0124 pH*temp p = 0.0506; Figs. 3A and 3B). In the pre-experiment shell at position 8 (Figs. 3C and 3D) the prismatic layer was thicker at ambient pH than at pH_T 7.66 (pH p = 0.0192, temp p = 0.9019; pH*temp = 0.2956). At position 6, at ambient pH only, the nacre layer was thicker at 13 °C (pH*temp p = 0.0326; temp at pH_T 8.00 p = 0.0324, temp at pH_T 7.66 p = 0.3473; pH at 13 °C p = 0.1548, pH at 15 °C p = 0.1187).

In the experiment shell, the amount of heavy etching on the dorsal surface was not consistent across temperatures or pH levels, respectively (pH*temp, p < 0.0001; Fig. 4A). Considerably more heavy etching was noted at pH_T 7.66 than pH_T 8.00 at 13 °C (by around 70%; p = 0.0002), but the opposite effect was found at 15 °C (by around 20%; pH_T 8.00 > pH_T 7.66; p = 0.0291). At ambient pH only, the percentage of heavily etched shell was greater at 15 °C than 13 °C (by around 70%; p = 0.0001) while at lowered pH the opposite effect was noted (13 > 15 °C p = 0.0254) (Fig. 4A). All experimental shells were lightly etched regardless of treatment (Fig. 4A; pH p = 0.3542, temp p = 0.0533, pH*temp p = 0.6517). In the pre-experiment shell, the percentage of heavily etched shell was significantly greater at the lower pH regardless of temperature (Fig. 4B; pH p < 0.001, temp p = 0.1056, pH*temp p = 0.0903). The percentage of lightly etched shell was influenced by both pH and temperature, being considerably greater at pH_T 7.66 and also at 15 °C (Fig. 4B; pH p = 0.0005; temp p = 0.0117, pH*temp p = 0.7358).

Our examination of the many SEM cross sectional and surface images did not reveal any differences in, or disruption to the ultrastructure of the shell.

Shell composition (mineralogy)

The percent of calcite (and thus aragonite, as pāua are bimineralic) in the shell, determined from XRD, differed with the age of the shell. The spire (oldest shell) had the least calcite (8–23%; Table 3), whereas the experiment (new) shell and the pre-experiment shell contained appreciable amounts of very low-magnesium calcite (68–78% and 58–65%, respectively). In the experiment shell the wt% calcite was 5–6% higher at ambient pH_T 8.00 than at pH_T 7.66, and was not affected by temperature (Table 3; pH p = 0.0313, temp p = 0.0628, pH*temp p = 0.7464). There was no effect of treatment on the wt% calcite in either the pre-experiment shell (pH p = 0.6355, temp p = 0.2661, pH*temp p = 0.6355) or in the spire shell (pH p = 0.2604, temp p = 0.1674, pH*temp p = 0.5294).

The effect on wt% calcium in skeletal carbonate of the experiment shell, determined using SEM-EDS, was not consistent across treatments (pH*temp = 0.0010), with an effect of temperature observed only at ambient pH (15 > 13 °C; pH_T 8.00 temp p = 0.0002; pH_T 7.66 temp p = 0.4081), and an effect of pH only at 13 °C (pH_T 7.66 > 8.00; temp 13 °C pH p = 0.0029; temp 15 °C pH = 0.0519). The same was noted for the pre-experiment shell (pH*temp p = 0.0277; pH_T 8.00 temp p = 0.0422; pH_T 7.66 temp p = 0.2017; temp 13 °C pH p = 0.0001; temp 15 °C pH = 0.1931). Wt % calcium in the spire shell was significantly greater at the lowered pH and the higher temperature (pH_T 7.66 > 8.00 p = 0.0466, temp 15 > 13 °C p = 0.0055, pH*temp = 0.5779).

X‐ray diffractometry results showed that the magnesium content in our juvenile pāua shell calcite is very low: always less than 2 wt% MgCO₃ and mostly less than 1%. EDS samples, too, showed very low magnesium content. In both cases, the values were close to the detection limit for this method (0.5 in XRD, 0.25 in EDS); consequently, we did not conduct any statistical analyses on these data.

Haliotis iris physiological response

While we may have anticipated that exposure to lower pH would result in higher mortality, slower growth and reduced physiological condition in juvenile pāua, this was not observed. Survival over the 4 months was very high and was not affected by our experiment treatments. Although growth was not influenced by pH, it was clearly greater at 15 than 13 °C (73 vs. 44 μm d⁻¹ SL), and physiological condition (CI_FWSW) was greater at the lower temperature (Table 2). These growth rates are in the range of those predicted based on tag and recapture across a wide range of sites around New Zealand (Naylor, Andrew & Kim, 2006), and, along with the high survival rates, confirm the suitability of our experimental system for these investigations.

In contrast to our findings, Cunningham, Smith & Lamare (2016) found a negative influence of pH on survival and growth of juvenile pāua (5–13 and 30–40 mm SL) after 100 days (pH_NBS 8.1 > pH 7.6). One reason for these different findings may be the lower carbonate saturation states experienced by the pāua in the Cunningham, Smith & Lamare (2016) experiments: for example, at pH_NBS 7.6 and 16.5 °C, the Ω_Ar and Ω_Ca were 0.77 and 1.20, respectively. In our pH_T 7.66/15 °C treatment, Ω_Ar and Ω_Ca were considerably higher, at 1.07 and 1.66, respectively (Table 1). Additionally, the pāua were fed differently (Ramajo et al., 2016), and had different environmental histories (wild caught in Cunningham, Smith & Lamare (2016) vs hatchery spawned and reared from wild caught parents (F₁) in this study). Reduced growth rates were noted for small (~15 mm SL) Pacific abalone (H. discus hannai) after 3 months at pH_NBS 7.9 and 7.7 (range of Ω_Ca = 1.27–1.77 and 1.90–2.63, respectively; Ω_Ar not reported; Li et al., 2018), and their CI_FWSW was reduced at pH_NBS 7.7 compared with pH_NBS 8.1 control animals (by 14.7%; Li et al., 2018). These differences illustrate the importance of considering seawater carbonate conditions when comparing between experiments, and the need for caution when generalising effects of one study (or species) to other situations.

Shell morphology and integrity

Allometric relationships in pāua are quite consistent (see, Sainsbury, 1982), with SL strongly correlated to age, FW, and other measures. SWs were lighter and CI_SW:SL was significantly reduced at pH_T 7.66 compared with pH_T 8.00 (Fig. 2B; Table 2), with CI_SW:SL also lower at 13 °C than at 15 °C (Table 2). Cunningham, Smith & Lamare (2016) too found that pāua juveniles had lighter SWs for a given size in lower pH seawater. This pH effect may indicate that calcification is less effective in lower pH conditions, or that biomineralization has been disrupted (Fitzer et al., 2015; Lu et al., 2018). While we noted some differences in calcite content which might support the former (discussed below), our SEM investigations did not reveal any disruption to the shell ultrastructure.

Abalone shells have an inner nacreous aragonitic layer overlain by a prismatic layer, which is in turn overlain by a thin organic periostracum that protects the shell. The outer prismatic layer may be calcite (e.g. H. kamtschatkana, H. rufescens), aragonite (e.g. H. asinina, H. glabra), or a combination (H. rubra). In our H. iris, the prismatic layer is low-Magnesium calcite, and the nacre layer is aragonite (Table 3). Natural populations of adult H. iris exhibit variable calcification rates, shell layer thicknesses (Gray & Smith, 2004), and growth rates (Naylor, Andrew & Kim, 2006), which have been attributed to local differences in environmental conditions such as water temperature and wave exposure (McShane & Naylor, 1995; Naylor, Andrew & Kim, 2006). In our study, total shell thickness was not influenced by experimental treatment. We did, however, detect differences in thicknesses of the layers. At lower pH, the pāua had significantly thinner prismatic layers at two positions in the newly generated shell, and at one position in the pre-experiment shell (positions 3, 4 and 8; Figs. 3A and 3B). This layer was affected by temperature only at one position, where it was thinner at 15 °C than 13 °C (position 1; Figs. 3A and 3B). The nacre thickness in the newly deposited shell was significantly thicker at pH_T 7.66, and also at 15 °C, at one position only (position 3; Figs. 3A and 3B). Polar brachiopods also generated a thicker internal shell layer when grown under ocean acidification conditions (pH_NIST 7.54) for 7 months, demonstrating a plasticity in the calcification process that may help their response to future change (Cross, Harper & Peck, 2019). This nacreous layer can be repaired or augmented by the animal anywhere from the inside (see Cusack et al., 2013) and, because it is not (mostly) in direct contact with sea water, it is less vulnerable to mechanical breakage, chemical dissolution, or biological encrustation/erosion than the outer prismatic layer (although dissolution of the inner nacre layer has been noted for some species at reduced pH; Michaelidis et al., 2005; Melzner et al., 2011).

The prismatic layer’s outer surface is protected from direct contact with seawater by the overlying periostracum (Ries, 2011). However, abalone periostracum (thin among molluscs at about 0.2 μm thick) is often damaged in the high-energy environments they inhabit, and can also be affected by encrustation and biological borings of the shell. Visible erosion is often observed in wild adults (after 3 years in H. fulgens; Shepherd, Avalos-Borja & Ortiz Quinatilla, 1995). Greater surface etching overall (heavy and light) at lower pH was noted for the pre-experiment shell (Fig. 4B). The most striking effect was noted in the experiment shell, where heavy etching was minimal (~10%) in the ambient pH/13 °C treatment, but ranged from 60% to 80% in the remaining treatments, reflecting greater heavy etching at the lower pH and the higher temperature (Fig. 4A); the latter possibly reflecting the principle that dissolution rates are greater at higher temperature (Ries et al., 2016). Other studies have reported surface dissolution and thinning of the outer calcite layer with reduced pH. In small Pacific abalone incubated at pH_NBS 7.7 large areas of shell were without periostracum and extensive dissolution of the exposed calcite layer occurred (Li et al., 2018). Brachiopods (Liothyrella uva and Calloria inconspicua) also showed dissolution of their outer shell surface (Cross, Harper & Peck, 2019). Shells of Mytilus edulis were significantly thinner, had more injuries at the outer surface, and showed dissolution of the outer prismatic layer at pH 7.4 after 6 months (Ω_Ca = 1.15–1.21 and Ω_Ar 0.72–0.76; Bressan et al., 2014). In M. edulis shell properties have been shown to be altered under ocean acidification (increasing pCO₂ from 380 to 1,000 μatm), through a reduced ability of the animal to control the ultrastructure (crystallographic orientation) of the calcite shell (Fitzer et al., 2014). These mussels produced calcite that was stiffer, harder and more brittle under ocean acidification conditions (Fitzer et al., 2015), which may potentially contribute to greater erosion.

In our low pH treatments, the average rate of thinning of the outer calcitic layer over the 4 month experiment was 0.1 mm. Since the layer is on average 0.4 mm thick, that rate of dissolution could remove the calcite altogether in about 16 months, exposing the aragonitic nacre to seawater. Farm-reared abalone exposed to high pCO₂ levels can lose the prismatic layer altogether, even when small (Wright, 2011). In nature, where low pH seawater combines with abrasion of the periostracum and bioerosion, major shell thinning with consequent loss of strength and resilience to predation and wave action could become the norm for the adult life of these long-lived molluscs.

Shell composition (mineralogy)

The shells of all the pāua in this experiment were composed of low magnesium calcite and aragonite (Table 3). The experimental shell contained 68–78% calcite (22–32% aragonite) on average, the pre-experiment shell 58–65% calcite (35–42% aragonite), and the spire shell 8–23% calcite (77–92% aragonite) (Table 3). This variable carbonate composition at different ages is to be expected (Auzoux-Bordenave et al., 2015). As both nacreous aragonite and prismatic calcite are stable minerals that retain their composition after deposition (at least in a time frame of months) we did not expect any difference in carbonate composition between individuals in the shell generated prior to the start of the experiment (Table 3). Indeed, only the shell which had grown during our experiment showed any treatment effect: 5–6% more wt% calcite was found at ambient pH than at pH_T 7.66, across both temperatures (Table 3). This likely reflects the significant erosion of the outer calcitic layer discussed above, possibly due to interaction with the other individuals crawling around in the chambers, and/or dissolution (although the seawater was oversaturated with respect to calcite in all treatments; Table 1). Whilst one might expect greater dissolution of aragonite at the saturation levels in our treatments (see Table 1), as discussed above, the aragonite nacre layer was not directly exposed to seawater. The wt% calcium in skeletal carbonate generally showed an opposite pattern to that of the wt% calcite (pH_T 7.66 > 8.00, 15 > 13 °C), in the experiment, pre-experiment and spire shells (albeit with some pH*temp interactions). Although we note that calcium can originate from both the aragonite and calcite shell layers, this pattern with pH seems counter to expectations of an effect on dissolution.

Temperature and pH synergies

While significant temperature treatment effects varied with response parameter, higher growth, CI_SW:SL, % calcium content and etching were noted at the slightly warmer temperature (15 > 13 °C). In particular, the influence of this relatively small difference in temperature on juvenile growth over only 4 months was striking (i.e. >2,000 μm more growth at 15 °C; Table 2). In several cases, a temperature effect was not apparent across both of the pH levels tested (i.e. when pH*temp < 0.05), with significant influences of temperature found only at ambient pH (e.g. % calcium, heavy etching of the experiment shell). Similarly, effects of pH were sometimes significant only at one temperature (e.g. only at 13 °C was % calcium content of the experimental and pre-experiment shell higher at pH_T 7.66 than ambient pH). This may in part be because of the differences in Ω_Ar in our lower pH treatments at the two temperatures—slight undersaturation at 13 °C and slightly above saturation at 15 °C (i.e. 0.992 and 1.066, respectively; Table 1). However, we do note that living organisms can be negatively affected at Ω_Ar > 1 (Waldbusser et al., 2015), and that calcification clearly occurred in all of our treatments, regardless of Ω_Ar (Ries, Cohen & McCorkle, 2009). The subtleties of effects of interactions between such small variations in temperature and ocean acidification (see Fitzer et al., 2014, 2015; Lu et al., 2018) and their relevance in a natural, more variable coastal environment remain to be determined. When temperature was elevated by 2 °C above ambient, disruption of crystallographic orientation by ocean acidification in M. edulis was greater (Fitzer et al., 2014), while effects on the shell stiffness, hardness and brittleness were reduced (Fitzer et al., 2015). Clearly, ocean acidification affects shell integrity, and the modifying influences of elevated temperature are complex and require more investigation. Indeed, Lu et al. (2018) showed that ocean acidification decreases the percentage of seawater DIC synthesised into shell carbonate in M. edulis, a response which is slightly greater at elevated temperature; they conclude that ongoing acidification and warming might interfere with calcification physiology through disrupting its ability to efficiently extract DIC from seawater.