Assessing skeleton and microbiome responses of a calcareous sponge under thermal and pH stresses

Abstract

Ocean higher temperature and acidity levels affect its carbonate chemistry, and calcifying organisms that depend on the calcium carbonate saturation state (ΩCaCO₃) to build their skeleton may be vulnerable to these changes. Despite their important roles in the marine environment, little is known about the vulnerability of calcareous sponges under these stressors. We performed an acute short-term experiment (9 days) with thermal and pH stresses to explore their effects on the skeleton and microbiome of the calcareous sponge Sycettusa hastifera. We observed that sponge spicules showed no corrosion and no Mg/Ca ratio variation when exposed to stress conditions. However, the outermost spicules were smaller in low pH treatment, although these effects were often diminished when higher temperatures were also applied. In general, the sponges synthesized normally shaped spicules under stress conditions, but the proportion of deformed spicules increased significantly in elevated temperature. In addition, we observed a stable host–microbiome association in which there is microbiome flexibility under thermal and pH stresses with a significantly increase in abundance of some associated bacteria. Our results suggest that S. hastifera has low vulnerability under the future ocean conditions here designed, since it showed indications of resistance that could help them adapt and survive.

bacterial community, calcifying organism, calcium carbonate, climate change, Porifera, Sycettusa hastifera

Introduction

In the last 260 years since the Industrial Revolution, the concentration of carbon dioxide (CO₂) in the Earth’s atmosphere has increased by 47%, reaching >400 parts per million (ppm) (WMO, 2020). The high emission of CO₂ is caused by anthropogenic activities that have caused ocean warming (OW) and ocean acidification (OA), two major threats to marine environment. According to the worst-case scenario, predicted by the Intergovernmental Panel on Climate Change (IPCC) for the end of the 21st century, the global temperature is expected to increase up to 4.8°C while the ocean pH is expected to reduce 0.3–0.5 units, which represents a 170% increase in acidity (RCP 8.5; IPCC, 2014). Both stressors affect the carbonate chemistry of the oceans and thus the calcium carbonate saturation state (ΩCaCO₃) by changing the rates of formation and dissolution of this mineral (Doney et al., 2009). Calcium carbonate (CaCO₃) is an essential mineral in the composition of shells and skeletons of calcifying organisms, thereby making them especially vulnerable to climate change stressors.

Besides these direct structural impacts, OW and OA may indirectly affect the homeostasis of marine organisms. Virtually all metazoans are considered holobionts, as they live in association with microorganisms, which may include members of bacteria, fungi, archaea, algae, and viruses. These associated microbes play essential roles to maintain the health and survival of the holobiont, such as providing organic carbon, protecting against pathogens, and contributing to nutrient cycling and calcification (Peixoto et al., 2017). A shift in the microbial community of marine organisms under a certain stress can define whether the host will acclimatize and survive or suffer dysbiosis and possibly die (Pita et al., 2018).

Thus, it is crucial to evaluate both the skeleton and the associated microbiome of calcifying organisms under OW and OA conditions to understand how the structural maintenance of these organisms may be affected in predicted scenarios, and how the microbes may help the host to adapt to the new conditions. Reef-building corals have been the main focus of climate change research, and the deleterious effects of OW and OA have been extensively explored in these organisms, which has led to specific investments in adaptation strategies (Bell et al., 2018; Rosado et al., 2019). Conversely, calcifying organisms like calcareous sponges are usually neglected even by spongologists, and the effects of climate change on these animals are poorly understood.

Sponges (phylum Porifera), in general, appear to have low vulnerability to climate change stressors (Bell et al., 2018), although some works have shown that in most cases responses are species specific. Studies published so far have shown no major negative effects in the skeleton of siliceous sponges under OW and OA (Goodwin et al., 2014; Vicente et al., 2016), but an increase in fragility and elasticity in the skeleton of juvenile glass sponges was observed in either stress condition (Stevenson et al., 2020). Most studies have shown that sponge microbiomes can be highly stable under a wide range of OW stress levels (e.g. Strand et al., 2017), but after reaching a certain temperature threshold, be that sub-lethal or lethal, microbiome stability can be affected (Webster et al., 2008; Ramsby et al., 2018; Vargas et al., 2020). On the other hand, some species can tolerate and even restructure their microbiome under OA, enhancing their chances to tolerate such environmental changes (Morrow et al., 2015; Ribes et al., 2016; Kandler et al., 2018). Most of these studies focus on sponges with skeletons made of silica and use a single climate change stressor only (either OW or OA) to test sponge responses (Bell et al., 2018). Thus, to have a broader projection of future ocean conditions, it is imperative to recognize the effects of both OW and OA on different species and mineral skeletons.

Among phylum Porifera, Calcarea is the only class that forms spicules (microscopic skeleton structures) of CaCO₃ crystallized as high-Mg calcite, which is the most soluble phase of this mineral (Jones and Jenkins, 1969; Andersson et al., 2008). However, these spicules are surrounded by organic sheaths (Jones, 1955) that may act as a protection against changes in the natural environment. Smith et al. (2013) raised concerns about the possible threat a high-CO₂ future world presents for calcareous sponges, but no experimental work has been done to address this issue. Peck et al. (2015) did perform an experiment on the effects of OA on a biofouling community, in which the abundance of the calcareous sponge Leucosolenia sp. increased in low pH conditions. However, detailed analysis on the biology of these sponges was not performed and temperature was not used as a stressor, leaving the question about the vulnerability of calcareous sponges in future oceans still open.

In this study, we evaluated the skeleton and microbiome responses of calcareous sponges under thermal and pH stresses, using the species Sycettusa hastifera (Row, 1909) as a model. Herein, we describe an acute short-term experiment that tested the effects of high temperature, low pH, and the combination of both stressors on adult S. hastifera. Under thermal and pH stresses, the skeleton and microbiome of these sponges were analysed by (i) evaluating possible corrosion and the variation in size, shape, and chemical composition (magnesium to calcium ratio) of the spicules and (ii) identifying response indicators in the associated bacterial community.

Sycettusa hastifera has a wide distribution in the Indian Ocean and is considered introduced in Brazilian coastal waters. Sycettusa hastifera has some bioinvader characteristics, such as the ability to colonize artificial substrates, high reproductive effort, a short life cycle, and early sexual maturity (Cavalcanti et al., 2013; Lanna and Klautau, 2018). However, even though S. hastifera is an exotic species, it is not a threat to native organisms, and understanding how this sponge may respond to climate change allows for predictions on its potential success of new invasions.

Materials and methods

Sampling of calcareous sponges

Adult specimens of S. hastifera were collected randomly in an area of ∼25 m² during May 2017 at ∼1 m depth by snorkeling in a mussel farm in Arraial do Cabo, Rio de Janeiro state, Brazil (22°58′02.9″S–42°00′25.3″W). The sponges were transferred to the facilities of the Rio de Janeiro Marine Aquarium Research Centre (AquaRio) in 20-l tanks with an air pump.

Experimental design

We developed an aquaria open system in AquaRio to perform an acute short-term experiment of thermal and pH stresses (Figure 1). The system comprised four aquaria (80 l each), with constant water flow and a turnover rate of 10% of the total aquarium water per hour (8 l h⁻¹). The tested temperature and pH values were set according to present and expected future conditions following RCP 8.5 (IPCC, 2014). Four treatments were performed with three temporal replicates: control (CO; 22°C and pH 8.1), high temperature (HT; 26°C and pH 8.1), low pH (LPH; 22°C and pH 7.6), and combined effects (CEF; 26°C and pH 7.6). Experiments lasted 10 days, with the first 24 h dedicated to acclimation (CO conditions). The CO temperature was chosen based on the mean temperature of a temporal series obtained at the sampling site (Fernandes et al., 2017) and in accordance with the temperature on the sampling day of each replicate (i.e. 23.8, 23 and 22°C;). After the acclimation period, sponges were randomly distributed into treatment aquaria (N = 5 per aquarium; totalling N = 15 per treatment, considering the three temporal replicates) where temperature was increased 1°C per hour in the HT and CEF tanks and pH levels were decreased 0.1 unit per hour in the LPH and CEF tanks, until they reached the desired conditions. Despite having five sponges per tank, only two were used in the subsequent analyses of skeleton and microbiome. All aquaria were fitted with air pumps (porous air stones diffusers) to keep the seawater flowing. Chillers (Mundo Sub—MS3000) and heaters (Glass Heater Via Aqua and Super Aquatic—200W) were used to reach the target temperatures. The pH was reduced by injecting CO₂ directly into the experimental tanks through an aquarium diffuser coupled to a pump (Sarlo Better 2000) to maximize the dissolution of the gas in the seawater. A pH controller (Analytical Instruments—Model PH-301—probe precision of 0.01 pH units) was used to maintain the acidic conditions (i.e. pH 7.6). Temperature and salinity (Hand-Held Analog Refractometer) were measured directly in the experimental tanks. For pH (Analyser—Model PH 300M—probe precision of 0.01 pH units) and alkalinity (Gran Method; Camourze-Mod, 1994) analysis, a sample of seawater of each tank was collected in a 250ml borosilicate bottle and then analysed at AquaRio laboratory. These four parameters were measured daily, and other parameters of the carbonate system were calculated using the software CO2Calc (Table 1, Supplementary Table S1, and Supplementary Figure S1; Robbins et al., 2010).

Figure 1.

Experimental design of the acute short-term experiment of thermal and pH stresses using Sycettusa hastifera as a model. Aquaria open system: CO: 22°C/pH 8.1; HT: 26°C/pH 8.1; LPH: 22°C/pH 7.6; and CEF: 26°C/pH 7.6. See also Supplementary Table S1 and Supplementary Figure S1.

Table 1.

Summary of seawater chemistry parameters measured (*) and calculated (**) for the short-term experiment of thermal and pH stresses, represented as the mean and the standard deviation (SD) of measurements taken daily (N = 8 days) for the three experimental replicates.

Treatment	Temperature (oC)*	pH (total)*	Salinity*	Alkalinity (µmol kg−1)*	pCO2 (µatm)**	Ω calcite**
CO	22.1 (0.1)	8.14 (0.03)	34.5 (1.0)	2 311 (48)	301 (30)	5.67 (0.31)
HT	27.1 (1.3)	8.16 (0.03)	34.9 (0.9)	2 317 (47)	278 (28)	6.78 (0.51)
LPH	22.2 (0.2)	7.63 (0.04)	35.0 (0.9)	2 308 (44)	1 182 (119)	2.13 (0.19)
CEF	26.4 (1.0)	7.64 (0.05)	34.8 (1.0)	2 315 (52)	1 156 (167)	2.54 (0.25)

Calcareous sponge skeleton analyses

Size and shape of the spicules

Sponges (N = 3 sponges per treatment; one sponge per tank) were randomly collected and fixed in 96% ethanol and slides of the spicules were prepared following standard procedures (Klautau and Valentine, 2003). The length and width of the actines (paired and unpaired) of each one of the five spicule categories present in S. hastifera skeleton (i.e. cortical diactines, cortical triactines, subcortical triactines, subatrial triactines, atrial triactines; 20 spicules per category, totalling 100 spicules per sponge, and 300 spicules per treatment) were measured with an optical microscope (Nikon Eclipse E200). To analyse possible corrosion in the spicules, they were also analysed by Scanning Electron Microscopy (SEM—Jeol 6510). As some deformed spicules were observed, the proportion of deformed spicules between treatments was calculated. A total of 300 randomly spicules per treatment were classified as “normal” or “deformed” in the optical microscope (Nikon Eclipse E200) using the same spicules slides that were used for measurements. We considered “deformed” by visual assessment, spicules with angles <90° between the paired actines or when the basal actines were not all in the same plane.

Chemical composition of the spicules

For the chemical composition analyses of calcium and magnesium, spicules from the same sponges cited above (N = 3 sponges per treatment; one sponge per tank) were pulverized and analysed by the Atomic Absorption Spectroscopy technique using a Shimadzu (AA-6800) Spectrophotometer operated with an air–acetylene flame atomizer. Each sample was analysed three times to evaluate the technique precision and to provide the standard deviation. The wavelengths used to detect calcium and magnesium were 422.7 and 285.2 nm, respectively.

Skeleton statistical analyses

A two-way permutational analysis of variance (PERMANOVA; Anderson, 2001) using 9999 permutations with Euclidean distance matrices with temperature and pH as factors were used to determine if the differences in the spicules sizes among treatments were statistically significant. The dependent variables were separated between length and width of the actines of each one of the five spicules categories, totalling 20 variables named from A to T (outer to inner spicules in skeleton position). This analysis was performed in the software PAST version 3 (Hammer et al., 2001). The normality and homogeneity of the skeleton data (size measurements, Mg/Ca ratio, and proportion of deformed vs. normal spicules) was checked based on a Shapiro–Wilk and Bartlett tests, respectively. When the data did not meet the assumptions of normality and homogeneity, a non-parametric analysis was used. The non-parametric Kruskal–Wallis test (K–W) was performed to compare the measurements of length and width of each spicule category among treatments, and for Mg/Ca ratio of the spicules, followed by a pairwise comparison using Dunn’s test with Bonferroni adjustment, when K–W P minuscule < 0.05. All analyses were performed in the software R version 3.6.3^® for Windows (R Core Team, 2020). A two-way analysis of variance (ANOVA) with logit transformation (Warton and Hui, 2011) using temperature and pH as factors was used to determine if the differences in spicules proportion (normal vs. deformed spicules) were statistically significant between factors and their interaction. This analysis was performed in the software R version 3.6.3^® for Windows (R Core Team, 2020), using the package car (Fox and Weisberg, 2011). The graphs were made in Microsoft Excel.

Calcareous sponge microbiome analyses

Characterization of the microbiome

To assess possible changes in the microbiome across treatments, sponges were randomly collected (N = 3 sponges per treatment; one sponge per tank), immediately frozen in liquid nitrogen, and then, macerated with a pestle. DNA extraction was performed using a DNeasy PowerSoil Kit (Qiagen) following the manufacturer’s protocol. DNA was quantified using a Qubit Fluorometer (Invitrogen). The V4 hypervariable region of the 16S rRNA gene was sequenced in the Illumina MiSeq platform using the bacterial primers 515F/806R (Caporaso et al., 2011) at the Argonne National Laboratory (Lemont, IL, USA).

Sequence data were processed using Qiime2 (https://qiime2.org/), where raw sequences were de-multiplexed, quality-filtered, and trimmed in the 5′ end with a maximum length of 218 bp. Reads were denoised, and amplicon sequence variants (ASVs) and chimeras were determined using the plugin DADA2, with the latter being excluded from the further analysis (Callahan et al., 2016, 2017). Representative sequences of a taxonomic group were determined, and taxonomic assignments were performed based on the SILVA 138 database (https://www.arb-silva.de/documentation/release-138/). The ASV table was built in Microsoft Excel from two Quiime2 output files (Taxonomy and Feature-Table). Sequences matching chloroplast, mitochondrial, archaea, eukaryote, and unassigned sequences were filtered and discarded. Pivot tables were made to visualize and condense the ASV table by taxonomic levels for graphical interpretation.

Microbiome statistical analyses

Alpha diversity indices (Shannon H′ for diversity and Chao1 for richness estimation) were calculated, and the normality and homogeneity of the data were checked and confirmed based on a Shapiro–Wilk and Bartlett tests, respectively, to perform the subsequent analysis. A two-way ANOVA with temperature and pH as factors was used to determine if the differences in alpha diversity indices among treatments were statistically significant. Diversity indices were calculated in the software PAST version 3 (Hammer et al., 2001). Shapiro–Wilk test, Bartlett test, and ANOVA were performed in the software R version 3.6.3^® for Windows (R Core Team, 2020).

For beta diversity, a non-metric Multidimensional Scaling (nMDS) was performed to compare the bacterial community submitted to temperature and pH as factors. A two-way PERMANOVA (Anderson, 2001) using 9999 permutations with Bray–Curtis distance matrices and temperature and pH as factors was used to determine if the differences in the bacterial community among treatments were statistically significant. Graphs were made in Excel, and statistical analyses were performed in the software PAST version 3 (Hammer et al., 2001).

An indicator analysis was performed to determine the ASVs that had abundances significantly correlated to one of the treatments. The multipatt function (func = “r.g”, 1000 permutations) from the indicspecies package (Cáceres and Legendre, 2009) in the software R version 3.6.3^® for Windows (R Core Team, 2020) was used to compare the ASV abundances between treatments. ASVs that significant correlated with one of the treatments (p < 0.05) were considered as indicator species. Indicator species with a total abundance above 0.1% were plotted using the ggplot2 package (Wickham, 2016) in the software R version 3.6.3^® for Windows (R Core Team, 2020).

Results

Thermal and pH stresses did not affect sponge survival

Three sponges presented signs of tissue degeneration at the base of the body in elevated temperature treatments at the end of the acute short-term experiment (9 days; Supplementary Figure S2). All the other sponges presented no mortality nor tissue degeneration under any of the treatments, resembling a healthy condition by visual assessment (Supplementary Figure S3).

pH stress affected the size, while temperature stress affected spicules shape

The spicules of S. hastifera presented no visual signs of corrosion in either light or scanning electron microscopy for any of the treatments (Figure 2a and b). However, the multivariate analysis showed that pH stress significantly affected the size of the spicules (two-way PERMANOVA, F = 16.22, p = 0.0001 for the factor pH; Table 2a andSupplementary Table S2). Four out of the five spicules categories of S. hastifera presented significant differences in size (length and/or width of the actines; K–W test, p < 0.05; Figure 3 and Supplementary Table S3). Only the atrial spicules did not show any significant differences. Subatrial spicules showed no differences only in the unpaired actine whereas the subcortical spicules presented no significant differences only in the length of the unpaired actines. In general, spicules were smaller in LPH, and often these effects were reduced when both stressors were applied together. HT showed significant results only in the thickness of the paired actine of the cortical triactines.

Figure 2.

Sycettusa hastifera skeleton. (a and b) Spicules with no signs of corrosion—(a) diactine; (b) cortical triactines; (c) subcortical triactine; (d) subatrial triactine; and (e) atrial triactine (u = unpaired actine; p = paired actine); and (c) proportion of normal vs. deformed spicules in the treatments with standard deviation bars for deformed spicules (N = 300 spicules per treatment). (*) Increase in deformed spicules under thermal stress (p = 0.0467). See also Supplementary Table S4A.

Figure 3.

Spicules sizes of Sycettusa hastifera from each treatment. (a and b) Diactines; (c and d) cortical triactines (unpaired); (e and f) cortical triactines (paired); (g and h) subcortical triactines (unpaired); (i and j) subcortical triactines (paired 1); (k and l) subcortical triactines (paired 2); (m and n) subatrial triactines (unpaired); (o and p) subatrial triactines (paired); (q and r) atrial triactines (unpaired); and (s and t) atrial triactines (paired). N = 60 actines for each spicule category, except for (m–p) of CEF (N = 53). The same letters at the top of each bar represent a significant difference between treatments. See also Table Supplementary S3.

Table 2.

Summary of the statistical analysis performed in this study to detect the effects on Sycettusa hastifera skeleton and microbiome in response to temperature, pH, and their interaction as fixed factors: (a) two-way PERMANOVA with Euclidean matrix; (b) two-way ANOVA with logit transformation; (c and d) two-way ANOVA; and (e) two-way PERMANOVA with Bray–Curtis matrix.

Source	df	MS	F	p-Value
(a) Spicules sizes
Temperature	1	2.86 × 109	27.559	0.0614
pH	1	1.68 × 1010	16.222	0.0001
Temperature × pH	1	62 242	0.599	0.5589
Residual	236	1.04 × 109
(b) Proportion of deformed vs. normal spicules
Temperature	1	3.760	5.519	0.0467
pH	1	0.443	0.650	0.4435
Temperature × pH	1	0.023	0.034	0.8578
Residual	8	0.681
(c) Alpha diversity—Shannon H′
Temperature	1	0.309	1.829	0.2132
pH	1	0.031	0.186	0.6780
Temperature × pH	1	0.072	0.426	0.5322
Residual	8	0.169
(d) Alpha diversity—Chao1
Temperature	1	11 224.1	0.685	0.4318
pH	1	8 910.75	0.544	0.4819
Temperature × pH	1	18 018.8	1.1	0.3249
Residual	8	16 380.2
(e) Beta diversity
Temperature	1	0.274	1.730	0.0948
pH	1	0.146	0.920	0.4662
Temperature × pH	1	0.144	0.911	0.4755
Residual	8	0.158

Statistically significant effects in bold (p < 0.05).

We found a significantly higher percentage of deformed spicules in sponges exposed to elevated temperature treatments (HT and CEF; two-way ANOVA, F = 5.519, p = 0.0467 for temperature factor; Figure 2c, Table 2b, andSupplementary Table S4A). These sponges had 2–15% deformed spicules in their skeletons, while sponges in CO treatment samples had only 1–2% deformed spicules (Figure 2c andSupplementary Table S4A).

The chemical composition of the spicules (Mg/Ca ratio) showed no significant differences among treatments (K–W test, χ² = 7.217, p = 0.0653; Figure 6 and Supplementary Table S4B), although those in LPH presented a lower Mg/Ca average than those submitted to other treatments (0.067 ± 0.009 wt.%; Figure 6).

Therefore, our results suggest that the outermost spicules have been more impacted by the treatments than the inner ones (atrial regions and surroundings) and that low pH seems to affect spicules size more than heat, while the opposite has been found for spicules shape.

Thermal and pH stresses did not affect microbial alpha and beta diversities

The bacterial community of S. hastifera in the experiment comprised 3526 ASVs from 39 phyla and candidate phyla (Supplementary Table S5). Proteobacteria was the dominant phylum (Alpha- and Gammaproteobacteria) in all treatments, ranging from 54 to 81% of the ASVs (Figure 4). Within Proteobacteria, Alphaproteobacteria comprised an average of 34% [±7% standard deviation (SD)], and Gammaproteobacteria comprised an average of 35% (±7% SD) of the bacterial microbiome (Figure 4). The phyla Bacteroidota and Firmicutes were also abundant, comprising an average of 10% (±4% SD) and 9% (±9% SD) of the bacterial community, respectively. Other less abundant phyla, but still representative (i.e. at least 1% abundance in at least one sample) were Verrucomicrobiota, Myxococcota, Acidobacteria, Bdellovibrionota, Cyanobacteria, Dadabacteria, Spirochaetota, and Planctomycetota (Figure 4 and Supplementary Table S5). The other 28 phyla (i.e. <1% abundance in at least one sample) and unassigned bacteria were classified as others (Figure 4 and Supplementary Table S5).

Figure 4.

Relative abundance of the ASVs present in the bacterial community of Sycettusa hastifera under thermal and pH stress at phylum level. (*) Classes of Proteobacteria phylum. R, replicate. See also Supplementary Table S5.

Alpha diversity of the samples was similar across the treatments with no significant differences and presented an average Shannon H′ of 4.9 ± 0.4 and Chao1 of 709 ± 124 (Supplementary Figure S4 and Table 2c and d). Beta diversity analysis also showed no significant differences related to temperature or pH treatments (Figure 5a and  b and Table 2e).

Figure 5.

nMDS ordination of the bacterial microbiome of Sycettusa hastifera in two dimensions using a Bray–Curtis matrix. Grey shape: (a) temperature factor and (b) pH factor. Each dot represents one replicate. Dashed lines represent the replicates of each treatment; and (c) abundant indicator ASVs (>0.1% of the total bacterial abundance) with abundances significantly correlated to one of the treatment conditions (p < 0.05). R, replica. See also Supplemenatry Table S6.

Both stressors impacted the abundance of some bacteria

The indicator analysis detected 52 treatment indicators, comprising a total of 12 ASVs for CO, 10 for HT, 23 for LPH, and 7 for CEF (p < 0.05; Supplementary Table S6). The CO indicators were in lower abundance in stress treatments (i.e. HT, LPH, CEF), while the stress indicators were in higher abundance in their respective groups when compared to CO (Figure 5c and Supplementary Table S6). The CO indicators comprised members of the phyla Proteobacteria (both Alpha- and Gammaproteobacteria classes), Myxococcota, Chloroflexi, and Verrucomicrobiota (Supplementary Table S6). The HT indicators belonged to the phyla Proteobacteria (Alpha and Gamma classes), Myxococcota, and Bdellovibrionota (Supplementary Table S6). The LPH indicators were members of the phyla Proteobacteria (Alpha and Gamma classes), Gemmatimonadota, Planctomycetota, Bdellovibrionota, Bacteroidota, Myxococcota, and Verrucomicrobiota (Supplementary Table S6). The CEF indicators comprised the phyla Proteobacteria (Alpha and Gamma classes), Planctomycetota, and Bacteroidota (Supplementary Table S6). Four taxa at the lowest taxonomic level were detected as indicators of different treatments: the class OM190 (phylum Planctomycetota) in LHP and CEF; the family Nannocystaceae (phylum Myxococcota) in CO, HT, and LPH treatments; the family Rhodobacteraceae (class Alphaproteobacteria) in HT, LPH, and CEF treatments; and the genus Woeseia (class Gammaproteobacteria) in LPH and CEF treatments (Supplementary Table S6).

Of the 52 treatment indicators, 13 comprised >0.1% of the total relative abundance of bacteria (Figure 5c) as follows: genera Halioxenophilus (class Gammaproteobacteria; ASV0170), Denitromonas (class Gammaproteobacteria; ASV0045), and Cerasicoccus (phylum Verrucomicrobiota; ASV0193) for CO treatment; genera Neptuniibacter (class Gammaproteobacteria; ASV0005), Glaciecola (class Gammaproteobacteria; ASV0013), and the families Nannocystaceae (phylum Myxococcota; ASV0061), Rhodobacteraceae (class Alphaproteobacteria), and Hyphomonadaceae (class Alphaproteobacteria; ASV0023) for HT treatment; genera Woeseia (class Gammaproteobacteria; ASV0044), and the families Rhodobacteraceae (class Alphaproteobacteria; ASV0105) and Saprospiraceae (phylum Bacteroidota; ASV0020) for LPH treatment; and genera Woeseia (class Gammaproteobacteria; ASV0086) for CEF treatment (Figure 5c).

Discussion

Effects of thermal and pH stresses on calcareous sponge skeleton

The skeleton of S. hastifera showed significantly smaller (shorter and thinner), and more deformed spicules when exposed to thermal and pH stresses in the acute short-term experiment. However, no signs of corrosion were observed in the spicules in any of the treatments (Figure 2a and b), which may be explained by the presence of the organic sheath that covers the spicules of calcareous sponges (Jones, 1955), also observed for many other calcifying organisms (e.g. crabs, lobsters, coralline red algae). Ries et al. (2009) showed that calcifying organisms with a high extent of organic cover, regardless of their CaCO₃ form, presented higher resistance with no dissolution in a 60-day experiment, even under extremely acidic conditions (pH ∼7.31). Therefore, it is possible that the organic sheath that covers the spicules of calcareous sponges protects the skeleton of S. hastifera from corrosion, at least in a short term, as already proposed by Peck et al. (2015) and Bell et al. (2018).

Despite not presenting signs of corrosion, spicules showed significant variation in their size under stress conditions and were smaller under LPH treatment ( Figure 3, Table 2a, andSupplementary Table S3). The mean saturation state of calcite (Ω) in LPH was the lowest observed among the treatment, which may explain the smaller size of the spicules under this treatment (Table 1 and Supplementary Table S1), as many calcifying organisms growing under LPH conditions also have reduced calcification due the reduction in the ΩCaCO₃ in the seawater (e.g. Orr et al., 2005; Ries et al., 2009). To explain smaller spicules under low calcite saturation state conditions showing no signs of corrosion, we must assume that spicules have been synthesized during the experiment. If these smaller spicules had been synthesized before the experiment, but had their sizes reduced during the experiment, then we should have seen signs of corrosion because corrosion is expected when spicules are submitted to decalcification procedures in acidic conditions.

Adult calcareous sponges seem to continuously renew their skeleton at a relatively quick rate, supporting the idea that spicules have been synthesized during the experiment (Ilan et al., 1996; Voigt et al., 2014). For example, Sycon sp. renews its diactines from the top (oscula) and central part of the body at a rate of ∼45 and 30%, respectively, within 55 h (Ilan et al., 1996), while Sycon ciliatum synthesizes new diactines and triactines within an 18-h experiment, but no renewal rate was calculated (Voigt et al., 2014). For S. hastifera, it is not known what its skeleton renewal rate is; however, preliminary results for this species showed higher rates of spicule renewal than those found for Sycon sp. during a 48-h experiment (B. Ribeiro, pers. obs.).

Our results also show that the outermost spicules suffered more impact from the treatments than the inner spicules (i.e. atrial region and surroundings; Figure 3). Four out of the five spicules categories presented significantly smaller sizes under LPH conditions (i.e. 7.6 pH), often with reduced effects when subjected to both pH and temperature stress conditions conditions (CEF: 26°C/7.6 pH; Figure 3 and Supplementary Table S3). The diactines, the largest and outermost spicules of this species, were the only spicule type that experienced a negative synergistic effect of pH and thermal stresses combined (Figure 3a and b). This spicule category, under LPH and CEF, presented the lowest mean sizes when compared to the CO and to natural populations from different sites (Supplementary Table S2; Azevedo and Klautau, 2007; Van Soest and de Voogd, 2018). Patterns of negative and positive synergistic effects of acidity with elevated temperature conditions have already been seen for other calcifying organisms. For example, juvenile corals of the species Acropora spicifera growing under low pH presented skeleton deformities that were attenuated in treatments of low pH and high temperature (Foster et al., 2016). In contrast, for the coralline algae Porolithon onkodes, mortality and dissolution were intensified under low pH plus warmer temperatures (Diaz-Pulido et al., 2012).

In our study, elevated temperature treatments (26°C), regardless of the pH, significantly raised the proportion of deformed spicules (from 2% to 15%; Figure 2c, Table 2b, andSupplementary Table S4A). Again, we assume that these spicules were synthesized during the experiment, as spicules do not change their shape after being synthesized even when subjected to temperatures up to 60°C, which is the case during slide preparation for morphological identification (Klautau and Valentine, 2003). Thus, these results also reinforce the hypothesis that those spicules were synthesized during the experiment.

Although deformed spicules have been observed mainly in sponges exposed to elevated temperature treatments, they were also found, in minor proportions, in the CO treatment (22°C/8.1 pH; Figure 2c). This suggests that the sclerocytes (cells responsible for spicule synthesis) present a natural error rate, which was intensified with the increase in temperature. As proteins are involved in spicules synthesis (Aizenberg et al., 1995; Voigt et al., 2014), elevated temperatures could be affecting these proteins’ functions, resulting in changes in the crystal elongation to specific crystallographic directions, possibly leading to the deformation observed.

The Mg/Ca ratio of the spicules in the experiment was not statistically different among any of the treatments (Figure 6 and Supplementary Table S4B). However, as average in LPH treatment was lower than the others, we cannot rule out that our sample of three sponges may have lowered our ability to detect significant statistical differences. The Mg/Ca ratio in organisms growing under these stressors frequently increases in high temperature and reduces in low pH, equilibrating each other when the two conditions are applied together (e.g. Glandon et al., 2018). Because the increase in Mg proportions in calcifying organisms increases the mineral solubility, growing in such conditions might represent a problem for their survival (e.g. Ragazzola et al., 2016).

Figure 6.

Mg/Ca ratio (± SE) in weight percent (wt.%) of the spicules of Sycettusa hastifera at the end of the experiment (N = 3 sponges per treatment). See also Supplementary Table S4B.

Thereby, our results indicate that pH and thermal stresses have multiple effects on the size and/or shape of the spicules of the calcareous sponge S. hastifera. The smaller size and deformities of the spicules may impair the support of the sponge body and water flux. Further studies are necessary to investigate this hypothesis. In addition, it is worth investigating the resistance of the spicules under such stress conditions in more detail, as our Mg/Ca ratio results indicate a possible increase in solubility under LPH conditions.

Effects of thermal and pH stresses on the calcareous sponge microbiome

In our acute short-term experiment of thermal and pH stresses, the temperature was raised 4°C and the pH decreased by 0.5 units compared to that in the CO aquaria (i.e. 22°C/8.1 pH; Table 1 Supplementary Figure S1 and Supplementary Table S1). Even in such conditions, Proteobacteria was the dominant phylum of the bacterial community in S. hastifera regardless of the treatment, followed by the phyla Bacteroidota and Firmicutes (Figure 4). Although the microbiome of sponges is species specific, Proteobacteria is usually dominant in the natural sponge-associated bacterial communities, especially in species with low microbial richness, like S. hastifera (Thomas et al., 2016; Moitinho-Silva et al., 2017). The other two most representative phyla in the S. hastifera microbiome in our study, Bacteroidota and Firmicutes, have also been previously reported among the main phyla found in sponges microbiome (Thomas et al., 2016; Moitinho-Silva et al., 2017).

Even though the changes in seawater parameters performed in our experiment are considerably different from the current ocean scenario, we believe that neither temperature nor pH thresholds (or even sub-lethal parameters) for the studied species seemed to have been reached, as no significant shifts in alpha and beta diversities were observed (Figure 5a and  b, Table 2c–e, and Supplementary Figure S4). In sponges, stress conditions at sub-lethal or lethal parameters can stimulate dysbiosis processes. Often, when these animals face such disequilibrium, alpha and/or beta diversities get higher than control conditions, with an increase in richness and diversity of opportunistic microbes (Pita et al., 2018). However, sponges microbiomes usually exhibit a wide range of tolerance to environmental changes, including changes in geographical ranges, nutrient inputs, pollutants, and also temperature and acidity higher levels (e.g. Webster et al., 2008; Ribes et al., 2016; Gantt et al., 2017; Strand et al., 2017; Turon et al., 2019; Easson et al., 2020). For instance, in low pH conditions, the sponges Coelocarteria singaporensis and Stylissa cf. flabelliformis do not significantly change their microbiomes when exposed to in situ acidification (LPH—0.2 pH units below control) for 7 months (Kandler et al., 2018). Likewise, even when subjected to the combination of both stressors, the sponge Xestospongia muta does not change its microbial community composition under elevated temperature (varying between 2 and 5.4°C above control) and low pH (0.2 pH units below control) during 12 days of stress and likewise seems to have a stable microbiome (Lesser et al., 2016). However, often when a threshold is reached, the microbiome abruptly changes. For example, the microbiome of the bioeroding sponge Cliona orientalis remains stable from 23 to 27°C for ∼2 months, but significant microbial changes occur when they are submitted to the sub-lethal temperature of 29°C (Ramsby et al., 2018). Similarly, the microbiome composition of the common aquarium sponge Lendenfeldia chondrodes changes significantly when submitted to sub-lethal temperatures (i.e. 5°C above control) within just one week, showing that these changes can occur very quickly once they reach their tolerance levels (Vargas et al., 2020).

In our study, although S. hastifera specimens were collected at a site presenting a mean surface temperature of 22°C, occasionally the maximum seawater temperature can reach up to 28.7°C (https://seatemperature.info/arraial-do-cabo-water-temperature.html). This statement corroborates our hypothesis that the stress threshold for this species was not reached during the experiment, indicating a more stable host–microbiome association under the tested conditions.

Despite the stability of the bacterial community, the indicator analysis identified 52 bacterial indicators for the tested treatments, suggesting flexibility of some members of the microbiome under stress conditions (Figure 5c andSupplementary Table S6). This flexibility was not taxon specific, as four indicators of the same taxa were detected responding differently among treatments. The microbiome flexibility, even with no significant changes to the whole bacterial community diversity, has been previously observed in sponges of the species Dysidea avara and Agelas oroides under acidic conditions (0.3 pH units below control; 66 days of exposure; Ribes et al., 2016). The capacity to switch members of the microbiome has been suggested to enhance the tolerance and survivorship of sponge species under stress conditions by providing an extra organic carbon source (Morrow et al., 2015).

Regarding the most abundant indicators for stress conditions at genus level, Glaciecola and Neptuniibacter (both phylum Proteobacteria) were identified as HT-treatment indicators (Figure 5c). Glaciecola genus is widespread in marine environments and has also been reported associated with corals in warm pools varying between 26 and 33°C (Ziegler et al., 2017). So far, its presence has not been associated with pathogenicity or parasitism (Qin et al., 2014). Some members of the Neptuniibacter genus have been previously reported as capable of assimilating taurine nitrogen by a process that releases sulphur compounds (Krejčík et al., 2008). Taurine is a compound commonly found in sponges and their associated symbionts are known for importing and using the taurine produced by the host (Engelberts et al., 2020). The genus Woeseia (phylum Proteobacteria), a monotypic genus containing Woeseia oceani, was identified as an indicator for LPH and CEF treatments (Du et al., 2016). This species has a wide distribution in marine sediments, and it is related to denitrification processes (Liu et al., 2020). However, any suggestions of possible effects in the functioning of the host–microbiome interaction in our study would be highly speculative and should be addressed in future works.

An acute short-term experiment might be a more feasible approach than a long-term one, depending on the target species, the location of the sampling sites, and the costs involved in the maintenance of the experiments. Voolstra et al. (2020) have showed that an acute short-term experiment of 18 h and a long-term experiment of 21 days have the same resolution in detecting corals thermal tolerance, showing how promising an acute approach can be for climate change studies. In our case, the challenge of identifying and collecting calcareous sponges in the field, and cultivating them in aquaria systems, makes an acute short-term experiment more feasible than a long-term one. Our study is the first to detail the skeleton and microbiome responses of calcareous sponges under thermal and pH stresses and can be used as a reference for further long-term investigation.

Conclusions

Here, we provide the first study on how the skeleton and the microbiome of an adult calcareous sponge respond to an acute short-term thermal and pH stresses, using the exotic species S. hastifera as a model (Figure 7). The spicules of S. hastifera did not show signs of corrosion nor significant differences in their Mg/Ca ratio under any of the treatments. However, the spicules were significantly smaller under LPH, but these effects were often diminished when subjected to the pH and thermal stresses. The sponges were capable of synthesizing normally shaped spicules in all treatments during the experiment. However, a significant increase in the number of deformed spicules in elevated temperature treatments was observed, but there is no indication if these increases could threaten long-term maintenance of the sponge skeleton. Alpha and beta diversities of the bacterial community associated with S. hastifera did not change significantly in response to the short-term stresses, showing a stable host–microbiome association under thermal and pH stresses. However, a restructuring of the microbiome under such conditions was observed with the increasing of some associated bacteria in the treatments. Although our results indicate a potential for resistance and adaptation in the calcareous sponge S. hastifera in ocean conditions predicted by the end of 21st century climate change experiments performing morphological and functional molecular analyses (e.g. transcriptomics) in different life stages, along with ecological and population genetics studies, are necessary to expand knowledge regarding the vulnerability of this species. This possible resistance, together with its opportunistic characteristics and high reproductive output (Cavalcanti et al., 2013; Lanna and Klautau, 2018), may contribute to the success of S. hastifera in different environments and reinforce its high potential to colonize and thrive in new sites with different environmental conditions.

Figure 7.

Schematic representation of the effects of acute thermal and pH stresses based on the predictions for the end of the 21st century (RCP 8.5; IPCC, 2014) on the skeleton and microbiome of the calcareous sponge Sycettusa hastifera.

Supplementary data

Supplementary material is available at the ICESJMS online version of the manuscript.

Funding

AP received a post-doc scholarship from Fundação Carlos Chagas Filho de Amparo à Pesquisa do Estado do Rio de Janeiro (E-26/202.677/2016). MK received grants from the Coordination for the Improvement of Higher Education Personnel (CAPES—Grant PROTAX, Edital 001/2015) and the Brazilian National Research Council (CNPq—Grant 476597/2013-7) and a Fellowship (CNPq—33 305451/2017-0). RP received grants from the Rio de Janeiro Marine Aquarium (AquaRio) Research Centre and the Brazilian National Research Council (CNPq—305325/2019-1).

Acknowledgements

We would like to thank Pingo and his team, AquaRio Research Centre, and AquaRio staff for all the support during this study. This paper is the Master thesis of Bárbara Ribeiro at the Biodiversity and Evolutionary Biology Graduate Program of the Federal University of Rio de Janeiro.

Ribeiro, B., Padua, A., Barno, A., Villela, H., Duarte, G., Rossi, A., Fernandes, F. d. C., Peixoto, R., and Klautau, M. 2020. Assessing skeleton and microbiome responses of a calcareous sponge under thermal and pH stresses. – ICES Journal of Marine Science, 00:000–000.