Apoptosis in gonadal somatic cells of scleractinian corals: implications of structural adjustments for gamete production and release

Abstract

Apoptosis is an evolutionarily conserved process of programmed cell death. Here, we show structural changes in the gonads caused by apoptosis during gametogenesis in the scleractinian coral, Euphyllia ancora. Anatomical and histological analyses revealed that from the non-spawning to the spawning season, testes and ovaries increased in size due to active proliferation, differentiation and development of germ cells. Additionally, the thickness and cell density of the gonadal somatic layer decreased significantly as the spawning season approached. Further analyses demonstrated that the changes in the gonadal somatic layer were caused by apoptosis in a subpopulation of gonadal somatic cells. The occurrence of apoptosis in the gonadal somatic layer was also confirmed in other scleractinian corals. Our findings suggest that decreases in thickness and cell density of the gonadal somatic layer are structural adjustments facilitating oocyte and spermary (male germ cell cluster) enlargement and subsequent gamete release from the gonads. In animal reproduction, apoptosis in germ cells is an important process that controls the number and quality of gametes. However, apoptosis in gonadal somatic cells has rarely been reported among metazoans. Thus, our data provide evidence for a unique use of apoptosis in animal reproduction.

1. Introduction

Apoptosis is an essential biological process during both embryonic development and adult life [1,2]. Apoptosis is involved in mechanisms for removing unnecessary or damaged cells, adjusting cell numbers, deleting structures and sculpting tissues and organs [3]. Generally, apoptosis is characterized by a set of morphological processes (e.g. cell shrinkage, nuclear fragmentation) and biochemical activities (e.g. caspase activity) [4]. Although there are some differences in apoptotic mechanisms, key biochemical components of the apoptosis machinery remain remarkably conserved throughout the animal kingdom [5–10]. In cnidarians (e.g. hydras, corals, sea anemones and jellies), apoptosis is involved in a wide range of biological processes [11,12]. For example, in the model cnidarian Hydra, apoptosis serves important functions in gametogenesis [13–15], survival mechanisms during starvation [16] and head regeneration [17,18]. Recent studies from scleractinian corals show that apoptosis is involved in coral bleaching and stress responses [19–22]. Further investigation of apoptosis in cnidarians will facilitate better understanding of cnidarian physiology.

Sexual reproduction of scleractinians has captured the attention of the general public since reports describing scleractinian mass spawning events in the Great Barrier Reef first appeared in the 1980s [23,24]. Various studies have been performed worldwide in recent decades, and fundamental aspects of scleractinian sexual reproduction (e.g. gametogenesis and spawning behaviour) have been characterized [25,26]. In general, scleractinian corals produce gametes in the gastrodermal (inner) tissue of polyps, the so-called mesentery [23–25]. The mesentery is composed of two layers of gastrodermis separated by a thin layer of mesoglea, a gelatinous, thin, extracellular matrix (electronic supplementary material, figure S1a–d). Scleractinian germ cells develop within the mesoglea at a specific site in each mesentery (electronic supplementary material, figure S1d), which is conventionally referred to as the gonad (testis/ovary). Histologically, the gonad can be observed as germ cells with surrounding mesoglea located between two layers of gastrodermal (gonadal somatic) cells (electronic supplementary material, figure S1e) [25–27]. Gametogenesis requires several months to a year. The gametes produced are eventually released to the gastrodermal cavity and used for external or internal (brooding) fertilization.

During periods of active gametogenesis, male germ cell clusters (called spermaries) and oocytes generally increase in volume from tens to hundreds of times in a sandwiched state between the two thick gonadal somatic cell layers. In addition, although tubule structures facilitating gamete transportation (e.g. seminiferous tubules like those found in mammalian testes) are not present, gametes are released from gonads to the gastrodermal cavity through the mesogleal and gonadal somatic cell layers. However, it is unclear how germ cell size increases and gamete release is accomplished without complications in gonads having such structures. Underlying cellular and molecular mechanisms remain largely unknown.

Our previous histological analysis of gonads of the gonochoric stony coral, Euphyllia ancora (family Euphyllidae), found that the thickness of the somatic cell layer in mature gonads appeared thinner than in immature gonads [28]. This led to the hypothesis that structural changes occur in the gonads during gametogenesis to facilitate spermary/oocyte enlargement and subsequent gamete release, and that apoptosis is involved in this process. To verify this hypothesis, we quantified changes in the sizes of gonads and spermaries/oocytes and thicknesses and cell densities of gonadal somatic layers by histological analysis during gametogenesis using the stony coral, E. ancora. In addition, we investigated the involvement of apoptosis in cellular events occurring in scleractinian gonads using molecular and cellular approaches.

2. Material and methods

(a). Sampling of experimental animals

Euphyllia ancora were collected by scuba divers at Nanwan Bay in southern Taiwan (21°57' N, 120°46' E). Our previous study revealed that a single gametogenic cycle of E. ancora in this region requires approximately half a year in males and a year in females [28,29]. Spawning occurs during a few nights in May (occasionally in April or June) every year [29]. To collect corals with gonads in different developmental phases, male and female E. ancora colonies were labelled, and parts of the colonies were collected at different times from 2015 to 2016. Gonads were isolated under a stereomicroscope from these coral fragments and were used for downstream analyses. Collecting was approved by the administration of Kenting National Park (issue number: 1010006545). Sampling of other scleractinian corals was conducted by snorkelling at Pitouchiao Park (25°07' N, 120°54' E) in northern Taiwan with permission of the Fisheries and Fishing Port Affairs Management Office of the New Taipei city government (issue number: 1063334179).

(b). Estimation of gonad sizes

Isolated gonads were fixed with 4% paraformaldehyde and 2% glutaraldehyde (Sigma-Aldrich, St Louis, USA) in 100 mM HEPES buffer (pH 7.4). To estimate gonad sizes, fixed gonads were photographed with a digital camera (RX100II, Sony, Tokyo, Japan) under a stereomicroscope. Since isolated gonads were basically spherical, we treated each gonad as a sphere and estimated the volume using the formula of a sphere (4/3r³) with ImageJ software (Wayne Rasband, National Institutes of Health, Bethesda, MD, USA; https://imagej.nih.gov/ij). Gonad radii (r) were calculated using the formula r = (d₁ + d₂)/2, where d₁ was the largest diameter of the gonad and d₂ was the second largest diameter at a right angle to d₁.

(c). Histological analysis

Sample fixation, decalcification and embedding were performed according to methodologies described in our previous study [29]. Serial sections at a thickness of 4 µm were prepared with a microtome (Thermo Shandon, Pittsburgh, USA), and rehydrated sections were stained with haematoxylin and eosin Y (Thermo Shandon) or immunohistochemically stained with an anti-E. ancora piwi (Eapiwi) antibody, as described elsewhere [28]. Since approximately 100–600 serial sections encompassed an entire isolated gonad, 10 to 60 sections (approximately 10% of the total sections/gonad) were randomly photographed for each gonad and subsequently used for downstream analyses.

Diameters of spermaries/oocytes were determined as the average of the two measurements (the largest dimension of the spermary/oocyte and the second largest dimension at a right angle to the first). Approximately 10 to 50 randomly chosen spermaries/oocytes were analysed from each gonad. The ratio of the area occupied by germ cells in each gonad was estimated by measuring the total gonadal area and the area occupied by germ cells (germ cell area) on the gonadal section. The area occupied by somatic cells was estimated by deducting the germ cell area from the total gonadal area. Approximately 10 sections were analysed to determine the ratio in each gonad, and average values were calculated. The thickness of the somatic cell layer adjacent to a spermary/oocyte was analysed using horizontal sections. The thickness of the thinnest part of the somatic cell layer adjacent to each spermary/oocyte (n = 10) was determined in each gonad, and the average value was calculated. Somatic cell densities were determined by counting somatic cell nuclei in randomly selected areas of the gonadal somatic layer adjacent to spermaries/oocytes on gonadal sections (electronic supplementary material, figure S2a–h). All haematoxylin-positive nuclei (including normal cells and possible apoptotic cells/bodies) were counted. Approximately, 10 areas from 10 sections were analysed in each gonadal sample. All analyses were performed with ImageJ. Nine testes and nine ovaries from three different colonies were analysed for each phase.

(d). TUNEL assay

To visualize apoptotic cells/bodies, a TUNEL assay was performed on the paraffin sections with a DeadEnd Chrometric TUNEL System (Promega, Madison, WI, USA) according to the manufacturer's instructions. Numbers of apoptotic signals in the selected somatic cell regions were determined on the micrographs using ImageJ. The average density of apoptotic signals (apoptotic cells/bodies) per 10 000 µm² was calculated after determining the densities in 12–15 randomly selected sections from each testis and ovary. Nine testes and nine ovaries from three E. ancora colonies were analysed for each phase.

(e). Transmission electron microscopy

Testes and ovaries in middle phase were isolated and fixed as described above. The samples were post-fixed in 1% OsO₄ for 4 h and then rinsed three times with the same buffer. After dehydration using an acetone series, samples were embedded in Spurr resin. Ultrathin sections (70–90 nm) were prepared with an EM UC6 ultramicrotome (Leica Microsystems, Wetzlar, Germany) and collected on copper grids. Thereafter, sections were stained with 5% uranyl acetate/50% methanol for 20 min and 0.4% lead citrate/0.1 N NaOH for 6 min. An FEI G2 Tecnai Spirit Twin transmission electron microscope at 80 kV was used for viewing, and images were obtained with a Gatan Orius CCD camera.

(f). Identification of apoptosis pathway-related genes

Apoptosis pathway-related genes were identified from our transcriptome assembly of E. ancora gastrodermal tissue, including late-phase gonads and mesenterial filaments. We performed two types of searches based on the following strategies: (i) full-length cDNA sequences from vertebrates and invertebrates, including scleractinians, were retrieved from GenBank, and then a local blast search (BLASTP, cut-off e-value of <1 × 10⁻⁵) was conducted against the E. ancora transcriptome database with CLC Main Workbench (CLC Bio, Aarhus, Denmark); and (ii) keyword searches of target categories were performed in the annotated database. Sequences of identified apoptosis pathway-related genes are listed in electronic supplementary material, table S1 and files S1–S3.

(g). Reverse transcription polymerase chain reaction

Isolated testes and ovaries in middle phase were analysed (four colonies for each sex). RNA isolation was performed with TRIzol reagent (Invitrogen, Carlsbad, USA), following the manufacturer's protocol. First-strand cDNA was synthesized from 1 µg of DNase-treated RNA using SuperScript III reserve transcriptase (Invitrogen). Expression of apoptosis pathway-related genes in testes and ovaries was detected by PCR with gene-specific primers. PCR conditions were as follows: 94°C for 5 min; 33 cycles of 94°C for 30 s, 60°C for 30 s, 72°C for 1 min; and a final elongation step of 72°C for 7 min. Two milligrams of cDNA were used for the analysis. Specific primer sequences and annealing temperatures for each gene are listed in electronic supplementary material, table S2. The transcript of β-actin (GenBank accession number JQ968408.1) was selected as the reference control. Negative controls for each primer set comprised reactions in which the reverse transcriptase or template was omitted.

(h). Caspase assay

Detection of caspase-3 activity was performed using a Caspase 3 Colorimetric Activity Assay Kit, DEVD (Merck Millipore, Billerica, MA, USA) according to the manufacturer's instructions. Briefly, testes (early and middle phases) and ovaries (middle and late phases) were microscopically isolated from E. ancora colonies (five colonies for each sex) collected in December 2015 and March 2016. Samples were lysed in lysis buffer provided with the kit, using a homogenizer on ice. Protein concentrations were determined by Bradford protein assay (Bio-Rad, Hercules, CA, USA) following the manufacturer's protocol. Protein samples (1 µg/sample for males, 2 µg/sample for females) were mixed with caspase-3 substrate (Ac-DEVD-pNA) in reaction buffer and incubated at 37°C for 2 h in the absence or presence of a caspase-3 inhibitor (Ac-DEVD-CHO). Caspase activity was monitored using a SynergyTM HT microplate reader (BioTek Instruments) at 405 nm. Enzymatic activity was estimated following the protocol described by the manufacturer.

(i). Statistics

All data are presented as means ± standard deviations (s.d.) or standard errors of the mean (SEMs). Statistical differences between two groups were determined using Student's t-test. For comparisons among more than three groups, statistical significance was determined using one-way ANOVA followed by Tukey's test with a statistical significance level of p < 0.05. All analyses were performed using Statistical Package for the Social Sciences (SPSS) software.

3. Results

(a). Changes in gonadal sizes from the non-spawning to spawning seasons

Euphyllia ancora annually releases gametes in spring (May, or occasionally in April or June) at reefs of southern Taiwan. In testes and ovaries isolated from wild E. ancora colonies from August 2015 (non-spawning season) to May 2016 (spawning season), the progression of gametogenesis could be observed histologically as the spawning season approached. Based on the germ cell types observed in gonads, we classified gonadal development into four major phases, early, middle, late and mature phases (table 1), and monitored germ cell development by determining the sizes of spermaries (clusters of male germ cells in testes) and oocytes (table 1). Gonadal sizes at each phase were also estimated. In testes, spermaries increased in size accompanied by the progression of active spermatogenesis (figure 1a). In ovaries, we confirmed oocyte growth histologically with accumulation of yolk protein and lipids (figure 1b). Sizes of testes and ovaries also increased from the early to mature phases along with the progression of gametogenesis (figure 1c,d).

Table 1.

Criteria for the classification of gonadal phase.

gonadal stage of collection	the most represented germ cells observed in the gonads	approximate timings
testis
early	spermatogonia	December
middle	spermatogonia and primary spermatocytes	March
late	secondary spermatocytes and spermatids	April
mature	sperm	May
ovary
early	oocytes with cytoplasmic polarization (<125 µm in diameter)	August, October
middle	oocytes with accumulation of yolk and other substances (126–200 µm in diameter)	December, January
late	oocytes with accumulation of yolk and other substances (201–275 µm in diameter)	February, March
mature	oocytes with semicircular or ‘U’-like germinal vesicles (>276 µm in diameter)	April, May

Figure 1.

Changes in E. ancora gonads from early to mature phases. (a,b) Changes in sizes of spermaries (a) and oocytes (b) in testes and ovaries, respectively (spermaries: early, n = 100; middle, n = 102; late, n = 103; mature, n = 111; oocytes: early, n = 88; middle, n = 298; late, n = 306; mature, n = 450). (c,d) Changes in sizes of testes (c) and ovaries (d) (testes: early, n = 130; middle, n = 159; late, n = 99; mature, n = 77; ovaries: early, n = 50; middle, n = 122; late, n = 100; mature, n = 47). Insets are micrographs of isolated testes and ovaries in early and mature phases. Data are shown as means ± SDs. (e–h) Immunohistochemical micrographs of an early-phase testis (e), a mature testis (f), an early-phase ovary (g) and a mature ovary (h). Germline cells (brown) were detected by immunohistochemical staining with an anti-E. ancora piwi (Eapiwi) antibody [28]. Sections were counterstained with haematoxylin. Note that sperm were deeply stained with haematoxylin (purple) in mature testes, because they have condensed nuclei and low expression of Eapiwi. Insets are schematic illustrations depicting areas occupied by germ cells (brown) and somatic cells (blue). (i and j) Changes in ratios of areas occupied by germ cells (brown) and somatic cells (blue) in testes (i) and ovaries (j) at different phases, as assessed by histological analyses. (k and l) Changes in the thicknesses of somatic cell layers adjacent to spermaries and oocytes in testes (k) and ovaries (l) at different phases, respectively (spermaries: early, n = 90; middle, n = 90; late, n = 100; mature, n = 100; oocytes: early, n = 110; middle, n = 100; late, n = 100; mature, n = 100). (m and n) Changes in the cell numbers in the somatic cell layers in testes (m) and ovaries (n) at different phases (the section numbers analysed for testes, early, n = 78; middle, n = 50; late, n = 100; mature, n = 63; the section numbers analysed for ovaries: early, n = 110; middle, n = 100; late, n = 100; mature, n = 100). For the analysis shown in (i–n), nine testes and nine ovaries from three different colonies were examined for each phase. Data are shown as means ± SEMs. Means with different letters within a group are significantly different (p < 0.05). (Online version in colour.)

(b). Changes in gonadal somatic cells during gametogenesis

Next, we examined cellular changes in gonadal somatic cell layers during gametogenesis using histological and immunohistochemical analyses. We first investigated ratios of areas occupied by germ cells or somatic cells in gonadal sections at each phase. Immunohistochemical observation revealed that the majority of the area was occupied by somatic cells in early-phase testes sections (figure 1e). By contrast, in sections of mature testes, areas occupied by germ cells were noticeably increased, and those of somatic cells were decreased (figure 1f). A similar tendency was also observed in sections of ovaries (figure 1g,h). Quantitative analysis showed that ratios of somatic cell areas decreased approximately 50% during testicular and ovarian development from early to mature phases (figure 1i,j). Further histological analysis revealed that the thickness of the somatic cell layer (figure 1k,l) and the density of somatic cells (cells per 10 000 µm²) (figure 1m,n) decreased significantly as maturation approached.

(c). Discovery of apoptosis in gonadal somatic cells by TUNEL assay

During the course of histological analysis, small particles appeared that were nuclear stain-positive (i.e. haematoxylin- and DAPI-positive) in somatic cell layers of testes and ovaries (electronic supplementary material, figure S2 and S3). These particles and decreases in somatic cell number and thickness led us to investigate the possibility of apoptosis. The TUNEL assay demonstrated a number of apoptotic cells in the somatic cell layer in both testes (electronic supplementary material, figure S4a-c) and ovaries (electronic supplementary material, figure S4d–f). In testes, apoptotic cells were observed in the somatic cell layer throughout spermatogenesis (figure 2a and b(i–iv)). Quantitative analysis revealed that significantly elevated numbers of apoptotic cells were present in middle- and late-phase testes possessing proliferating spermatogonia, primary spermatocytes, secondary spermatocytes and spermatids (figure 2c). Similarly, in ovaries, apoptotic cells were observed in the somatic cell layer throughout oogenesis (figure 2d–f). Elevated numbers of apoptotic cells were observed in early- and middle-phase ovaries (figure 2f). Fewer apoptotic cells were observed in ovaries close to maturation (figure 2e(i–iv) and f). Apoptotic cells were rarely detected in germline cells and other somatic tissues, such as tentacles (electronic supplementary material, figure S5a and b) and mesenterial filaments (electronic supplementary material, figure S5c and d).

Figure 2.

Detection of apoptosis in testes and ovaries at different developmental phases by TUNEL assay. (a) Schematic illustration depicting the positional relationship between the testicular somatic cell layer (tsc) and male germ cells (mgc) shown in micrographs (b(i)–b(iv)). (b(i)–b(iv)) TUNEL assay results for testes in early (b(i)), middle (b(ii)), late (b(iii)) and mature (b(iv)) phases. Developmental phases of testes are classified according to observed germ cell types ( table 1). (c) Quantified numbers of apoptotic cells/bodies (brown dots) in testicular somatic cell regions at different developmental phases. (d) Schematic illustration depicting the positional relationship between the ovarian somatic cell layer (osc) and oocytes (oc) shown in micrographs (e(i)–e(iv)). (e(i)–e(iv)) TUNEL assay results for ovaries in early (e(i)), middle (e(ii)), late (e(iii)) and mature (e(iv)) phases. Ovarian phases are classified according to oocyte diameters (table 1). Purple: nuclei stained with haematoxylin. Scale bars, 20 µm. (f) Quantified numbers of apoptotic bodies in ovarian somatic cell regions. Data are shown as means ± SEMs. Nine testes and ovaries from three different colonies were analysed at each phase. Means with different letters within a group are significantly different (p < 0.05). (Online version in colour.)

(d). Demonstration of apoptosis by transmission electron microscopy, reverse transcription polymerase chain reaction analysis and caspase-3 assay

Subsequently, we attempted to further confirm the occurrence of apoptosis in gonadal somatic cells. In the TEM analysis of testicular and ovarian somatic layers, we observed both normal cells possessing nuclei with dispersed and lightly stained nuclear chromatin and apoptotic cells characterized by aggregation of electron-dense nuclear material at the periphery of nuclear membranes, degradation of the nuclei, chromatin condensation and apoptotic bodies (figure 3a,b).

Figure 3.

Demonstration of apoptosis in gonads by TEM analysis, RT-PCR analysis and caspase-3 assay. (a and b) Transmission electron micrographs of the testicular somatic cell layer (a) and ovarian somatic cell layer (b) in E. ancora. Testes and ovaries in middle phase were used for this analysis. Arrows indicate apoptotic cells/bodies. (c) RT-PCR analysis of the expression of apoptosis pathway-related genes (7 genes) in testes (tes) and ovaries (ova) in middle phase. β-actin was used as the internal control. Reactions lacking either reverse transcriptase (RT-) or template (NC) were also included as negative controls. (d) Caspase-3 activity in protein extracts of isolated testes and ovaries. The activity of caspase-3, which recognizes the sequence DEVD, was determined by measuring free pNA released from the synthetic caspase-3 substrate DEVD-pNA. Caspase-3 activity (dark grey) was compared between testes in early and middle phases and between ovaries in middle and late phases. Reactions with a caspase inhibitor (DEVD-CHO, light grey) were also included as negative controls for each set of reactions. Data of testes and ovaries are shown as relative caspase-3 activity in early-phase testis (+inhibitor) and in middle-phase ovary (+inhibitor), respectively. Data are shown as means ± SEMs (n = 3 colonies). Significant differences were found between groups (Student's t-test, *p < 0.05, **p < 0.01). (Online version in colour.)

We identified apoptotic pathway-related genes Caspase-3-like, Caspase-3/9-like, Caspase-8, Bcl-2, Bax-like, Apaf, baculoviral IAP repeat-containing protein and Bax inhibitor from the transcriptome assembly of E. ancora gastrodermal tissues (including late-phase gonads and mesenterial filaments) (electronic supplementary material, table S1 and files S1 and S2). RT-PCR analyses confirmed expression of those genes in both middle-phase testes and ovaries (figure 3c).

The caspase-3 assay demonstrated that in agreement with results of the TUNEL assay, caspase activity in middle-phase testes was higher than in early-phase testes (figure 3d). Similarly, in agreement with the results of the TUNEL assay, caspase activity in middle-phase ovaries was higher than in late-phase ovaries (figure 3d).

(e). Apoptosis in gonadal somatic cells of other scleractinians

Finally, we investigated how common apoptosis is among gonadal somatic cells of scleractinians using a TUNEL assay. All scleractinians that we examined (family Agariciidae, Leptoseris sp.; family Poritidae, Porites sp.; family Agariciidae, Pachyseris sp.; family Pectiniidae, Echinophyllia sp.; family Acroporidae, Acropora hyacinthus) displayed apoptotic cells among gonadal somatic cells (electronic supplementary material, figure S6a–e).

4. Discussion

By anatomical and histological analyses, we demonstrated that gonads, including germ cells (spermaries and oocytes) increased in size during gametogenesis, while the thicknesses and densities of gonadal somatic cell layers decreased. We also demonstrated that apoptosis in a subpopulation of gonadal somatic cells caused these changes in gonadal somatic cell layers. These results support our hypothesis that gonadal structure is adjusted for germ cell development and gamete release. Although scleractinian sexual reproduction has been studied worldwide in more than 400 species in recent decades [26], to the best of our knowledge, no similar findings have been described in previous studies. Thus, the present study is the first to reveal one of the major cellular events occurring in scleractinian gonads during gametogenesis.

In reproduction, apoptosis controls the number and quality of gametes [30]. Generally, apoptosis occurs in germ cells. Damaged and excess germ cells are eliminated from gonads by apoptosis [30]. In some invertebrates, such as freshwater Hydra, Caenorhabditis elegans and Drosophila, apoptosis is associated with oocyte development [13–15]. For instance, in Hydra, oocytes increase in size primarily by cell fusion with germ cell-derived nurse cells. The nurse cells undergo apoptosis after cytoplasmic transfer to oocytes and are eventually phagocytosed by oocytes [13–15]. By contrast, the present study discovered apoptosis in gonadal somatic cells of all scleractinian species that we examined, suggesting that this is a common characteristic among scleractinians. Since apoptosis in gonadal somatic cells has rarely been reported among metazoans, our data provide evidence for a unique case of apoptosis in animal reproduction.

Gonadal somatic cells support germ cell development in both vertebrates and invertebrates, including scleractinians [31–38]. Why do such important cells need to be removed from the gonads? There are four possible reasons: (i) apoptosis of gonadal somatic cells may provide space that allows oocytes and spermaries to enlarge in gonads, (ii) apoptosis may remove excess/aged gonadal somatic cells from gonads, (iii) gonadal somatic cells may derive nutrition to support germ cell development by engulfing apoptotic somatic cells, and (iv) reducing the somatic cell layer thickness and cell number by apoptosis may modify gonadal structure for gamete release. To obtain a better understanding of the biological significance of apoptosis during scleractinian gametogenesis, investigation of the influence of apoptosis inhibition in gonadal somatic cells on germ cell development and spawning is required. In addition, migration and proliferation of gonadal somatic cells need to be investigated in future studies to understand cellular dynamics of the gonadal somatic population.

Our analysis revealed three findings. (i) Apoptotic cells were rarely detected in extragonadal tissues, such as tentacles and mesenterial filaments in E. ancora polyps. (ii) The timing of the apoptotic peak differed between testicular and ovarian somatic cells. The peak was in spring in testes (March and April, the middle and late phases) and winter in ovaries (December and January, the middle phase). (iii) Apoptosis was also detected in gonadal samples that had been fixed on a ship immediately after sampling (electronic supplementary material, figure S7). These findings suggest that apoptosis is likely to be induced by a programmed, reproduction-specific mechanism, rather than by environmental stressors. There may be gonad-/phase-specific mechanisms for controlling somatic cell numbers corresponding to the status of germ cell development.

In mammals, apoptosis is induced either by intrinsic stimuli affecting the integrity of the mitochondrial membrane or by extrinsic ligands binding to death receptors on the cell surface [1]. The presence of vertebrate-like apoptotic machinery has been demonstrated in scleractinians [10,19–22]. In E. ancora gonads, we also confirmed expression of some core genes involved in apoptotic machinery. Interestingly, expression of Caspase-8, a key player in the extrinsic pathway, was detected by RT-PCR in both testes and ovaries, implying the existence of ligands and death receptors for apoptosis. Tumour necrosis factor (TNF) is the most potent ligand to induce apoptosis in mammals [39] and corals [40]. Our E. ancora transcriptome database contained some sequences showing similarities to TNF ligands and TNF receptors (electronic supplementary material, file S3). Further identification and characterization of these genes, together with spatiotemporal expression analyses and functional assays, will provide new insight into mechanisms of apoptosis induction in scleractinian gonads.

While a number of gonadal somatic cells underwent apoptosis during gametogenesis in E. ancora, some gonadal somatic cells probably remain in the gonads without undergoing apoptosis, implying that gonadal somatic cells are heterogeneous, and that there may be differences in physiological and functional characteristics among the cell population. Remaining somatic cells probably preserve somatic cell lineages in gonadal tissue to proliferate and reconstruct gonads in mesenterial tissue during the next sexual reproductive cycle.

5. Summary

A schematic illustration summarizing major cellular events in gonads of E. ancora during gametogenesis is presented in figure 4. From the non-spawning to spawning seasons, testes and ovaries increase in size due to active spermatogenesis and oogenesis. Meanwhile, the thickness and cell density of the gonadal somatic cell layer decrease due to apoptosis occurring in a subpopulation of gonadal somatic cells. This structural change in the somatic cell layer facilitates spermary/oocyte enlargement in the gonads and subsequent gamete release.

Figure 4.

Schematic diagram of cellular changes in gonads during germ cell development. Testes and ovaries increase in size from the non-spawning to spawning seasons. In the gonads, spermaries (male germ cell clusters) and oocytes increased in size during the progression of gametogenesis from the early to mature phases. Meanwhile, gonadal somatic cell layer thickness and somatic cell number decreased due to somatic cell apoptosis. (Online version in colour.)

Disclosure statement

The authors have nothing to disclose.

Ethics

Experiments were carried out in accordance with principles and procedures approved by the Institutional Animal Care and Use Committee, National Taiwan Ocean University.

Data accessibility

This article has no additional data.

Authors' contributions

S.S. conceived and designed the experiments. C.-C.C., Y.-L.C., P.-H.T. and S.S. performed the experiments. C.-C.C., Y.-L.C., P.-H.T. and S.S. analysed the data. S.S. wrote the manuscript. C.-F.C. edited the manuscript.

Competing interests

We have no competing interests.

Funding

This research was supported by a grant from the Ministry of Science and Technology, Taiwan (MOST 107-2311-B-019-001 to S.S.) and partially supported by a grant no. MOST 104-2313-B019-MY3 to C.-F.C.