Effects of Low-ph Stress on Shell Traits of the Dove Snail, Anachis Misera, Inhabiting Shallow-vent Environments off Kueishan Islet, Taiwan

The effects of naturally acidified seawater on shell traits were quantified through the comparison of dove snails (Family: Columbellidae) Anachis misera from vent environments with Euplica sp. from non-vent sites in northeastern Taiwan. Samples of A. misera were collected around a shallow vent (24.8341 • N, 121.96191 • E), which included the east, south, southwest, and northwest sites. An absence of Anachis snails was found in the most acidic north site (pH 7.19–7.25). Based on the similarities of protein expression profiles, the Anachis snails were classified into two groups, i.e., V-South (pH 7.78–7.82) and V-Rest (pH 7.31–7.83). Comparing their shell traits to the non-vent Euplica sp. from Da-xi (DX) and Geng-fang (GF) (pH 8.1–8.2), a difference in shell shape (shell width : shell length) was found, with the populations having more globular shells than the non-vent ones. The means of shell width were significantly different among sites (p < 0.01), with a descending order of GF > DX > V-South and V-Rest. The relationships of shell length to total weight were curvilinear for both Anachis and Euplica snails. The logarithmically transformed slopes differed significantly among sites, and the mean body weight of the GF population was greater than that of the others (p < 0.01). Positive correlations between shell length and shell thickness of body whorl (T1) and penultimate whorl (T2) were only observed in non-vent GF and DX populations. Anachis snails from vent sites were thinner in T1 and T2 compared to the Euplica snails from non-vent sites (p < 0.05). Within each vent group, shell thickness between T1 and T2 was insignificantly different. Between vent groups, T1 and T2 from V-Rest showed a decrease of 10.6 and 10.2 %, respectively, compared to V-South ones. The decrease of T1 and T2 between vent Anachis snails and non-vent Euplica snails was as great as 55.6 and 29.0 %, respectively. This was the first study to compare snail's morphological traits under varying shallow-vent stresses with populations previously classified by biochemical responses. Overall, the shallow-vent-based findings provide additional information from subtropics on the effects of acidified seawater on gastropod snails in natural environments.


Introduction
Although current evidence indicates that organisms with a CaCO 3 skeleton, e.g., mollusks, echinoderms, and corals, are likely to be among the most susceptible to ocean acidification (e.g., Fabry et al., 2008;Sokolov et al., 2009), specific information obtained from field investigations has been limited, particularly with regard to gastropod snails (Gazeau et al., 2013).Thus, the current study was performed to address this issue within an extreme hydrothermal environment.
The shallow hydrothermal vents of interest are located east of Kueishan (KS) Islet, Taiwan, near the southern end of the Okinawa Trough (Fig. 1).The vents emit yellow or white plumes, with temperature and pH varying in the ranges of 78-116 • C and 1.52-6.32and 30-65 • C and 1.84-6.96,respectively.The gas bubbles are comprised of 90-99 % CO 2 , 0.8-8.4% H 2 S, < 0.03 % SO 2 , and < 50 ppm HCl (Chen et al., 2005).The diffusive plumes are affected by the wind, sea waves, and tides (Chen et al., 2005;Han et al., 2014).Based on the observed data, the emitted fluids diffuse mainly from north to south due to ebb tide and move from southeast to northwest during the spring tide.In addition, the fluids are also directed by the Kuroshio Current flowing along the coast of Kueishan Islet to the north throughout the year.Because the diffusion is closely correlated with diurnal tides, benthic organisms would face the lowest pH twice per day but for no more than 4 hours each time.
Near the yellow vents, the crab Xenograpsus testudinatus is the only benthic macrofauna (Jeng et al., 2004).In contrast, around the white vents, benthic invertebrates include the crab X. testudinatus, two kinds of sea anemones, hexacoral Tubastraea aurea, serpulid polychaete, a chiton, the snail Nassarius sp., and the dove snail Anachis misera.These vent organisms naturally inhabit acidic and toxic environments.High concentrations of trace metals in various tissues of the crab X. testudinatus are reported, and the levels are not above those found in other crabs collected from different habitats (Peng et al., 2011).
We herein test the hypothesis that populations of A. misera distributed around vents exposed to varying degrees of plumes would exhibit a different ecophysiological performance compared to the non-vent dove snail, Euplica sp., a common species in coastal waters off northeastern Taiwan.
A proteomic-based method was used to classify the samples of A. misera collected around the vent-based environments.This approach involves measuring changes in many proteins.Through the comparison of the protein expression profile of each snail by cluster analysis, similarities among samples can be determined and classified.This method has been applied to laboratory and field pollution studies, such as studies on blue mussels exposed to polyaromatic hydrocarbons and heavy metals (Knigge et al., 2004) and on Sydney rock oysters inhabiting an acid sulfate runoff estuary (Amaral et al., 2012).

Sampling sites and collection of snails
Anachis misera was collected around a shallow-water vent in Kueishan Islet, Taiwan (Fig. 1), including the north (N), east (E), south (S), southwest (SW), and northwest (NW) sites during the period of 28 June to 1 July 2011.The sampling vent emitted white plumes, and another vent with yellow plumes was nearby to the northeast.The distance of the collection sites to the vent center was 10-16 m, and the water depth was in the range of 14.5-17.5 m.Snails of Euplica sp. were sampled from Da-xi (DX) and Geng-fang (GF), northeastern Taiwan, between July and September 2012.
Sampling locations were identified by scuba divers equipped with GPS.Temperature was determined by a thermometer inserted into the seawater samples.The flow rate was measured by a Hydro-Bios digital flowmeter (Model 438 110).The pH was measured by a pH meter (Radiometer, Copenhagen, Denmark).Each environmental parameter was determined with one or three replicates, and the results are shown in Table 1.The collected snails were preserved in dry ice in the field.Upon returning to the laboratory, they were deep-frozen at −70 • C for later use.

Measurements of snail morphological traits
Shell traits, i.e., shell length and width, shell thickness of the body whorl (T1) and the penultimate whorl (T2), as well as the total weight of the intact individual, were measured (Fig. 2).Shell thickness was determined through enlarged X-ray radiographs which were produced by exposing snail shells to X-rays at 80 kVp and 1 mA for 116.7 ms.The distance between the X-ray source and the objects was 50 cm.The shell images were further drawn with outlines using GIMP version 2.8, which is an open source imaging system (http://www.gimp.org/).For statistical analysis, an ANCOVA (analysis of covariance) was used to compare the least square means (LSM) for each variable (i.e., shell width, total weight, shell thickness T1 and T2) among sites with shell length as the covariate.The relationships of shell length to shell width and the shell thickness of T1 and T2 were calculated using linear regression analysis.If the relationship of total weight and shell length was curvilinear, linear regression slopes were obtained and compared after data were logarithmically transformed.

Proteomic study
The protein expression profiles of Anachis snails were determined by one-dimensional sodium dodecyl sulfatepolyacrylamide gel electrophoresis (1-D SDS-PAGE).The foot tissue was taken and homogenized with lysis buffer (0.5 M Tris-HCl, pH 7.4, 10 % SDS, 0.5 M DTT) for proteomic analysis.Homogenates were centrifuged at 13 000 g for 10 min at 4 • C. The homogenous supernatant was collected, and the protein concentration was determined by Bradford assay, using bovine serum albumin as the standard.
The stacking and resolving gels were prepared with percentages of 5 and 12 % (Hoefer SEM 260 system, Amersham Pharmacia).After loading 25 µg protein in each sample lane, electrophoresis was run for 30 min at 120 V and then for 4 h at 180 V.The gels were stained with Coomassie blue G-250 (Candiano et al., 2004).
Stained gels were scanned and transformed into digitalized images using Image Scanner (Amersham Pharmacia).The Multi Gauge software v2.2 (Fujifilm) was utilized for protein quantification.The protein bands were assigned band numbers, and their intensity levels were calculated as their relative area to the total protein area on the gel.A cluster analysis of the Bray-Curtis similarity (BCS) indices (Primer 6.0) was employed to compare the expression of overall protein patterns among snail individuals (Clarke and Warwick, 2001).In addition, the contribution of each protein band was further estimated by principal component analysis (PCA).

Morphological traits of Anachis snails from vent sites
The temperature ranges of the sampling sites were from 26 to 27 • C (  clearly observed, with the lowest pH being 7.22 ± 0.03 at the north site (p < 0.01).Anachis snails were found around the vent, except at the most acidic north site.Shell lengths of the snails ranged from 6.88 to 11.01 mm.Several snails with an eroded apex were observed at the east and northwest sites (Fig. 2).

Protein expression profiles of Anachis snails from vent sites
Based on the protein expression results, 16 protein bands were selected for further Bray-Curtis similarity (BCS) analysis (Fig. 3).The classification of snails fell into three clusters (Fig. 4).Snails from the high-pH south site were all within one cluster.In contrast, snails from the remaining sites were indistinguishable from other clusters.In the process of the further determination of the contribution of each protein variable, the data were characterized by principle component analysis (PCA).The first to the fifth principal components accounted for 35.4,28.5, 13.2, 8.8, and 4.2 % of the total variance, respectively.The separation was mainly contributed by the first (i.e., bands 8, 1, 15, and 12) and second (i.e., bands 15, 13, 12, 1, and 11) principal components.Based on the cluster results, the Anachis snails were classified into groups of V-South (pH 7.78-7.82)and V-Rest (pH 7.31-7.83).Their shell traits were subsequently compared to non-vent Euplica snails (pH 8.10-8.20).

Comparison of the shell traits of dove snails among vent and non-vent sites
Shell traits of the Anachis and Euplica snails were listed in Table 3.A positive correlation between shell length and shell width was observed in all populations (Fig. 5).A difference in shell shape (shell width : shell length), with vent populations having more globular shells, was also found, as shown by the significant difference in regression slopes.AN-COVA with shell length as the covariate showed the mean values of shell width to be significantly different among sites (p < 0.01), with a descending order of GF > DX > V-South and V-Rest (Table 3).
The relationships of shell length to total weight were curvilinear for both Anachis and Euplica snails (Fig. 6).The slopes among sites were significantly different, and the mean body weight of the GF population was significantly greater than that of the others (Table 3 and Fig. 6).
Positive correlations between the shell length and shell thickness of body whorl (T1) and penultimate whorl (T2) were only observed in non-vent GF and DX populations.Their slopes were significantly different for T1 only (Fig. 7).The mean shell thickness of T1 and T2 varied among sites (Table 3).Anachis snails from vent sites were thinner in T1 compared to the non-vent Euplica snails (p < 0.001), with a descending order of GF > DX > V-South > V-Rest.A similar trend was also found in T2.Within each vent site, shell thickness between T1 and T2 was insignificantly different (paired t test, p> 0.05).By comparison, T1 and T2 of the snails from V-Rest were 89.4 and 89.8 %, respectively, of those of the V-South snails.In the comparison of vent and non-vent sites, T1 and T2 of the Anachis snails from V-Rest were 44.4 and 71.0 %, respectively, of those of the Euplica snails from GF.
Clearly, both measurements of shell thickness decreased under acidic environments.

Application of proteomic-based approach to shallow-vent snails
The proteomic-based method has been used in environmental toxicology to characterize an organism's responses to specific treatments with various gradients (Bradley et al., 2002;Jackson et al., 2002).It has been applied to laboratory and field studies, such as studies in which the blue mussel Mytilus edulis was exposed to polyaromatic hydrocarbons and heavy metals (Knigge et al., 2004), to crude oil (Mi et al., 2006), to polychlorinated biphenyls, and polycyclic aromatic hydrocarbons extracted from Baltic Sea sediments (Olsson et al., 2004) and studies in which the mussel Mytilus galloprovincialis was exposed to a tributyltin-polluted area (Magi et al., 2008).The application of the proteomic approach to the vent mussel Bathymodiolus azoricus has been conducted with samples collected from three distinct hydrothermal vent fields in the Mid-Atlantic Ridge (Companya et al., 2011).The expression profiles of 35 proteins from the gill revealed a clear difference among sites, which indicates that specific adaptations of B. azoricus depend on local conditions.
It is known that large spatial and temporal variations in environmental parameters, such as temperature, pH, and hydrothermal fluid composition are detected around vent environments with regard to dissolved oxygen, methane, and sulfide concentrations, etc.The pH of the hydrothermal fluids within our sampling vent and the surrounding seawater was determined on 31 May 2011, with pH ranges from 2.29 to 5.11 and 5.51 to 6.15, respectively (Zeng et al., 2013).The diffusion activities of vent plumes were also evaluated through the environmental factors of temperature, pH, and Eh (Han et al., 2014).The diffusive plume is mainly affected by the wind, sea waves, and tides.If ocean currents in the east-west direction are not considered, sea currents around the vents are from north to south during ebb tide, whereas, in flood tide, the opposite direction dominates.Our proteomic results indicated that snails from the south site were distinguished from those at the rest of the sites; this is consistent with the diffusion of local vent fluids.

Comparison with other dove snail studies
Anachis avara is a common dove snail (Family: Columbellidae) living off the coast of the eastern United States (Scheltema, 1968;Hatfield, 1980).At Bear Cut, FL, the population of A. avara showed seasonal fluctuation in its size structure (Hatfield, 1980).It reached a mean terminal size of 10.50 mm (8.00-13.29 mm) and matured quickly at the age of 6-7 months.The estimated life span was less than 2 years.It is suggested that the fluctuation in size structure was primarily the result of seasonal recruitment, and the abundance was probably determined by predation.
In Anachis fluctuate, the regression equation of shell length (SL) (mm) and dry tissue weight or shell weight (g) has been reported as Y = −0.025+ 0.003 SL (R 2 = 0.88; N = 26) and Y = −2.39+ 1.04 lnSL (R 2 = 0.92), respectively (Bertness and Cunningham, 1981).By comparison, in this study, shell lengths of A. misera and Euplica sp.ranged from 6.88 to 11.01 and 3.40 to 7.56 mm, respectively (Tables 2 and 3).Variations in size structure among sites were also obvious.Positive correlations between shell length and shell width or total weight in both dove snails was present, but the R 2 values of the equations were low in A. misera (0.07-0.28) compared to Euplica sp.(0.94-0.98) (Figs. 5 and 6).Positive correlations between shell length and shell thickness of the body whorl (T1) and penultimate whorl (T2) were only found in non-vent populations with an R 2 of 0.43-0.64 (Fig. 7).Although differential recruitment and acidic stress are potential factors to account for low or even no correlation between the above shell traits in vent A. misera, further study is needed to address this question.

Comparison with other ocean acidification studies
To date, ocean acidification studies have been conducted mostly in the laboratory or controlled environments for a short period of time.The results indicate that exposure to future global change scenarios (Caldeira and Wickett, 2003;Sokolov et al., 2009) may alter the tolerance of calcifying species and, ultimately, their fitness and survival through complex physiological and ecological pathways.Based on data from the literature, it is concluded that the effects of acidified seawater on species growth occurred at higher pH than those on species reproduction (mean pH 10 was 7.73 vs. 7.63 and mean pH 50 was 7.28 vs. 7.11) (Azevedo et al., 2015).
Studies conducted in the natural vent system at Ischia, Italy, indicated that the settlement and colonization of mollusks and microfauna showed high reductions in recruitment in the acidified stations (Cigliano et al., 2010;Ricevuto et al., 2012;Milazzo et al., 2014).In experiments in which juvenile pen shells Pinna nobilis were transplanted to Ischia for 45 days, decreases in survival, growth, and oxygen consumption were found.A 22 % decrease in survival rate for specimens transplanted at pH 7.7 compared to those at pH 8.1 was reported (Basso et al., 2015).In studies on the limpet Patella caerulea within and outside the Ischia vent, shell formation and dissolution are both observed at a low-pH site where enhanced shell production counteracts shell dissolution (Hall-Spencer et al., 2008;Rodolfo-Metalpa et al., 2011;Langer et al., 2014).In contrast, shell dissolution is absent at normal-pH sites.The nassariid gastropods Nassarius corniculus and Cyclope neritea adapted to the Ischia vent were smaller than those found in normal-pH conditions and had a higher massspecific energy consumption but a significantly lower wholeanimal metabolic energy demand (Garilli et al., 2015).Compared with deep-sea vent studies, on the northwest of Eifuku Volcano, Mariana Arc, the vent mussel Bathymodiolus brevior, inhabiting low-pH environments (pH 5.36-7.29),exhibited a shell thickness and daily growth increments in shells of only about half of that of mussels living in environments with pH > 7.8 (Tunnicliffe et al., 2009).
Under low pH (7.7 vs. 8.0), the periwinkle Littorina littorea increased less in weight and were shorter than snails grown in normal conditions (Melatunan et al., 2013).Similar results have been obtained for other calcifying organisms, e.g., the reduction in shell growth of the oysters Crassostrea gigas (Lannig et al., 2010) and Crassostrea virginica (Beniash et al., 2010), of the larvae of the Mediterranean pteropods Cavolinia inflexa (Comeau et al., 2010), and of the mussels Mytilus edulis (Gazeau et al., 2010) and Mytilus californi-anus (Gaylord et al., 2011).Along a gradient of pH (5.78-8.30)and salinity (3.58-31.2p.s.u) in the Sungai Brunei Estuary, Malaysia, whelk Thais gradata exposed to acidified sites possessed heavier shells, and the degrees of erosion were negatively related to water pH and calcium concentration (Marshall et al., 2008).At low pH (7.7), a 2.45 % change in shell shape (shell width : shell length) towards a more globular shell and a decrease in the outer lip shell thickness of up to 27 % in Littorina littorea were observed (Melatunan et al., 2013).
In this study, the comparison of A. misera from V-South (pH 7.78-7.82)and V-Rest (pH 7.31-7.83)revealed that the change in shell ratio was 3.4 % and shells were more rounded in the V-Rest group.In addition, snails of V-Rest exhibited a 10.6 and 10.2 % decrease in shell thickness of the body whorl (T1) and penultimate whorl (T2), respectively, compared to the V-South snails.In the comparison of vent and non-vent sites, T1 and T2 of the Anachis snails from V-Rest were 44.4 and 71.0 %, respectively, of those of the Euplica snails from GF (pH 8.1-8.2).Our shallow-vent-based results were, in general, consistent with other laboratory, controlled, and field studies, i.e., shell organisms are susceptible to acidic environments.
It is known that vent systems are not entirely representative of future ocean changes, not only because of the temporal variability in pH but also because of the existence of other toxic elements.However, vents' acidifying environments are sufficiently large on spatial and temporal scales for a valid comparison.It is a naturally occurring system to assess the effects of ocean acidification on the whole life cycle and across multiple generations of target organisms.

Figure 2 .
Figure 2. Shell morphology and X-ray photos of Anachis misera around the vent off Kueishan Islet and Euplica sp. from non-vent control sites of Da-xi and Geng-fang.(a) A. misera from the east site; (b) A. misera from the south site; (c) A. misera from the southwest site; (d) A. misera from the northwest site; (e) Euplica sp. from Da-xi; (f) Euplica sp. from Geng-fang; (g) X-ray photos of A. misera from the south site(left) and the northwest site (right).Scale bar: 5 mm; SL: shell length; SW: shell width; T1: thickness of body whorl; T2: thickness of penultimate whorl.

Figure 5 .Figure 6 .
Figure 5. Relationship between shell length and shell width of Anachis and Euplica snails from different sites.Different letters following the values in the legend indicate that the regression lines differ significantly (p< 0.05).

Table 1 .
Sampling locations and environmental parameters of vent sites in the Kueishan Islet and northeastern Taiwan.Data are shown as mean ± SD (range of parameter).

Table 2 .
Environmental parameters and shell traits of Anachis misera around the vent off Kueishan Islet.Data are shown as mean ± SD (range of paramenter).SL: shell length; SW: shell width; TW: total weight; T1: thickness of body whorl; T2: thickness of penultimate whorl.Means that differ significantly from each other are indicated by different letters.

Table 3 .
Shell traits of Anachis misera around the vent off Kueishan Islet and Euplica sp. from non-vent control sites of Da-xi and Gengfang.Data are shown as mean ± SD (range of parameter): shell length; SW: shell width; TW: total weight; T1: thickness of body whorl; T2: thickness of penultimate whorl.Least square (LS) means that differ significantly from each other are indicated by different letters.
Figure 4. Results from the combined principal component analysis (PCA) and cluster analysis of Bray-Curtis similarity (BCS) indices using standardized overall protein expressions of Anachis snails from different sampling sites.E: east; S: south; SW: southwest; NW: northwest; 1-16: variable of protein bands.