Introduction

Sperm flagellar motility is crucial for sexual reproduction, and the interaction between sperm and eggs is a prerequisite in various eukaryotes. Cnidarians, as phylogenetically early branching animals, exhibit diverse sexual reproduction, ranging from gonochoric to hermaphroditic, even within coral species (Siebert and Juliano 2017). For example, the reef-building coral Acropora is hermaphroditic (Baird et al. 2009), whereas mushroom corals, such as Fungia and Ctenactis, are gonochoric (Baird et al. 2009; Loya and Sakai 2008). Other cnidarians also exhibit diverse sexual modes; reproductive modes and regulatory mechanisms governing sperm flagellar motility leading to fertilization show diversity among cnidarians (Glass et al. 2023; Speer et al. 2021). However, the downstream mechanisms of this system require further exploration.

Upon activation of flagellar motility in the hermaphroditic Montipora capitata, a species of the Acropora family, and gonochoric Astrangia poculata corals, soluble adenylyl cyclase (sAC) produces cAMP, involving cAMP-dependent protein kinase A (PKA) (Glass et al. 2023; Speer et al. 2021). Notably, the Ac-PKA signal transduction is a highly conserved pathway for sperm flagellar motility activation in eukaryotes, contributing to activation in various animals such as sea urchins (Loza-Huerta et al. 2013; Nakajima et al. 2005), ascidians (Nomura et al. 2000; Shiba and Inaba 2014), teleosts (Inaba et al. 1999, 1998), and mammals (Baro et al. 2020; Vyklicka and Lishko 2020). Moreover, the sequences of the catalytic subunit of PKA are highly conserved (Soberg et al. 2013). However, the downstream phosphorylation via PKA remains unclear.

Despite the conservation of Ac-PKA signal transduction, the identification of PKA substrates is imperative. In salmon sperm, PKA phosphorylates the motor protein dynein light chain during flagellar motility initiation (Inaba et al. 1999). Other PKA-dependent protein kinase substrates include A-kinase anchoring protein 4 in human spermatozoa (Su et al. 2010), tyrosine kinase (Alvau et al. 2016), phosphatase, ion channels, and GTP binding proteins (Fujita et al. 2000; Leclerc and Kopf 1999). In sea urchins, several substrates have been identified, including ATP synthase, creatine kinase, NADH dehydrogenase, tubulin beta chain, and cAMP-dependent protein kinase type II regulatory subunit (PKA RII; Loza-Huerta et al. 2013).

In the evolution and speciation of animals, axoneme architecture has remained conserved (Inaba 2011), whereas flagellar morphology has changed. For example, mammalian sperm flagella possess a fibrous sheath (Turner 2006), contrasting with phylogenetically early branching animals such as corals, whose flagella consist of an axoneme and a plasma membrane without the accessory structure found in mammalian sperm (Gaino et al. 2008; Padilla-Gamiño et al. 2011; Steiner 1991). Consequently, the anchoring proteins mediating flagellar motility initiation and activation should also change during evolution.

The fertilization conditions in corals and other animals exhibit distinct characteristics. For example, fertilization can occur either internally or externally, even within coral species (Baird et al. 2009). Various ecological factors may have influenced the evolution of their motility features. For the broadcast spawning coral Acropora, sperm must locate eggs for fertilization after being released into the water. However, sperm is susceptible to dilution (Levitan and Petersen 1995; Yund 2000), and its quantity may be correlated with fertilization success (Kitanobo et al. 2022). From an ecological perspective, ejaculations by multiple colonies or males in other taxa can potentially lead to sperm competition, potentially influencing sperm evolution (Pizzari and Parker 2009). Additionally, the reproduction of corals is essential for coral reef maintenance (Knowlton 2001). However, their sexual reproduction, spawning, is limited to a few times a year (Baird et al. 2021a). Sperm initiate coral reproduction via mating with other conspecific eggs. Notably, the genetic diversity of the colonies releasing sperm and their sperm number denote the fertilization success and genetic diversity of the fertilized eggs in the ocean (Kitanobo et al. 2022). Moreover, ocean acidification and local stressors influence coral reproduction (Hagedorn et al. 2016), such as a decline in egg numbers (Leinbach et al. 2021). In the mating pathways, ocean acidification potentially suppresses the alkalinization of sperm intraflagellar region, which is essential for sAC activation for sperm flagellar motility initiation (Morita et al. 2010; Nakajima et al. 2005), resulting in lower fertilization success (Albright et al. 2010). Therefore, understanding sperm flagellar motility under PKA is crucial for coral reproduction. Nevertheless, comprehensive analyses of the identification of PKA substrates and their potential roles in motility diversity are yet to be conducted.

To address these gaps, we aimed to identify sperm proteins in the reef-building coral Acropora digitifera, identify potential PKA substrates from the sequence data, explore orthologs, and assess the sequence conservation of phosphorylation residues. We believe that our research holds implications for comprehending the adaptive evolution of fertilization-related proteins, highlighting potential adaptations to specific fertilization conditions, and offering valuable information for broader comparative studies in the field of reproductive biology.

Materials and Methods

Corals and Sperm Collection for Mass Spectrometry (MS) Analyses

The coral A. digitifera (Fig. 1a), Acropora tenuis, and Acropora florida were collected 7 days before spawning at Sesoko Island, Okinawa, Japan (26.629, 127.862), in May or June 2019, 2020, and 2022. All colonies were collected from shallow reefs at depths ranging 2–5 m. The collected colonies (five colonies each from A. digitifera, A. florida, and A. tenuis) were placed in a flowing water aquarium under natural conditions. The gamete setting was monitored from 20:30 to 21:00, and colonies ready to spawn (Fig. 1b) were transferred to seawater-filled 5 L buckets to collect gametes from each colony. Spawning commenced around 19:20 to 19:40 in A. tenuis and 21:40 to 22:30 in A. digitifera and A. florida, releasing gamete bundles. Collected gamete bundles were separated into sperm and eggs using a plankton mesh following the method described by Morita et al. (2006). The isolated sperms were counted using a Thoma cell counting chamber (Erma, Tokyo, Japan), and 40 mL of seawater (SW) containing 1 × 107 to 6 × 106 spermatozoa (cells/mL) were centrifuged at 12,000 × g for 10 min at 4 °C to obtain pelleted sperm cells. Isolated sperm samples were stored at − 30 °C until MS analyses.

Fig. 1
figure 1

Corals and analytical procedures. a Sperm protein analyses were conducted using the reef-building coral Acropora digitifera through mass analyses. b Acropora, which spawns once or a few times a year, exhibits the “setting” of gamete bundles on the surface of the colony. c Orthologs from the same clade of the identified protein were isolated

Full size image

Effect of Potassium Ionophore and Membrane-Permeable cAMP on Sperm Motility Activation

The sperm of A. florida (N = 4) were suspended in artificial seawater (ASW) containing 8-bromo-cAMP to confirm the effect of cAMP on sperm motility activation. A total of 0.2 µL of sperm suspension was added to 100 µL of artificial seawater (ASW, 450 mM NaCl, 10 mM KCl, 9 mM CaCl2, 30 mM MgCl2, 16 mM MgSO4, pH 8.0) containing 100 µM of 8-bromo-cAMP with DMSO (10 mM). We also examined the effect of potassium ionophore, valinomycin, which causes soluble adenylate kinase activation owing to membrane hyperpolarization (Izumi et al. 1999). The sperm of A. tenuis (N = 3) were suspended in ASW containing 10 or 100 µM valinomycin in DMSO (1 mM or 10 mM, respectively). An activation solution containing NH4Cl was used for the coral Acropora (0.5 M choline chloride, 20 mM NH4Cl, and 10 mM HEPES–NaOH pH 8.0) as the positive control. The recording of the sperm was initiated from the beginning of the dilution, and the sperm showed motility roughly 30 s later. Subsequently, we recorded these sperm movements with a BlackMagic design connected with a CCD camera (Mintron, Taiwan) on the microscope (Nikon Optiphoto) under dark field illumination or phase contrast. The objective lens was × 10 with a × 10 relay lens. Sperm numbers were counted from the recordings (about 200 sperm in the field), and the motile percentage of the sperm was calculated. The swimming velocities were measured with three recordings converted into TIFF files for 1 s for 30 frames with Premier 2022 (Adobe). The converted TIFF files were opened with Image J, and the sperm head with color profiles was selected. Heads were then copied and pasted to a new file for 15 frames. The 15 framed pasted files indicate the trajectories of the sperm head, and we measured the length of each trajectory to calculate sperm movement for 0.5 s to calculate sperm swimming velocity.

Mass Spectrometry of Coral Sperm

Sperm proteins from A. digitifera (N = 1) were subjected to liquid chromatography–tandem mass spectrometry (LC–MS/MS) analyses by Promega Japan and were identified using the NCBI GenBank dataset. All identified amino acid sequences were retrieved from the accession codes.

The eggs or sperm were treated with trypsin and subjected to nano-LC–MS analyses at Kazusa DNA Research Institute. The proteins were eluted with sonication in the elution solution [2 (w/v) % SDS and 100 mM Tris–HCl; pH 8.5]. The eluted protein concentrations were measured using the BCA assay and adjusted to 1 µg/µL with the elution solution. Tris(2-carboxyethyl)phosphine (TCEP) was added to 20 µg of proteins (the final concentration was 20 mM) and incubated for 10 min at 80 °C to cleave disulfide linkages. 2-Iodoacetamide (IAA) was then added for the alkylation of cysteine residues (final concentration of 30 mM) and incubated for 30 min at room temperature around 25 °C in the dark. A mixture of Sera-Mag SpeedBead Carboxylate-Modified Magnetic Particles (Hydrophilic) and those of Hydrophobic was added in a 1:1 (v/v) ratio and washed with distilled water three times. These alkylated samples were adjusted to 15 µg solids/µL with SP3 beads (20 µL), to which 2.5 vol of 99.5% EtOH was added. The mixture was then incubated for 20 min at room temperature. The incubated beads were washed twice with 80% EtOH and then mixed with 100 µL of 50 mM Tris–HCl (pH 8.0). Peptide fragments from the proteins were obtained by adding 500 ng trypsin/Lys-C mix (Promega, Madison, WI, USA) into the beads and Tris base (50 mM Tris–HCl; pH 8.0), followed by incubation at 37 °C overnight. Then, 20 µL of 5% trifluoroacetic acid (TFA) was added to the trypsin-treated mixture and sonicated. The mixture was desalinized using reversed-phased spin columns (GL-TIP SDD; GL Science, Tokyo, Japan), and the desalinated samples were evaporated using the rotary evaporator. The dried samples were diluted with 2% acetonitrile (ACN) and 0.1% TFA and then sonicated to dissolve the dried peptides. The concentration of the peptides was measured using the BCA assay and adjusted to 200 ng/µL with 2% ACN and 0.1% TFA. These peptide samples were analyzed via LC–MS (nano-LC: UltiMate 3000 RSLCnano LC system; Thermo Fisher Scientific, Waltham, MA, USA) with MS (Orbitrap Exploris 480) in ESI positive mode. The peptides were sprayed from a column (120 mm length and 75 µm diameter) at 40 °C. The flow rates were 750 nL/min for the first 4 min, and then 200 nL/min for the remaining 40 min. The mobile phase consisted of (A) 0.1% formic acid and (B) 0.1% formic acid and 80% ACN. The mixture rates of the A and B mobile phases were as follows: starting from 3% B solution, increasing to 10% B at 10 min, then 1.375%/min, increasing up to 65% B at 44 min. MS analyses were performed in overlapping window DIA modes with successive analyses with the following four events: (1) full scanning (MS1), (2) isolation window 1 DIA (MS2), (3) full scanning (MS1), and (4) isolation window 2 DIA (MS2). The mass data were analyzed with scaffold DIA (Proteome software) with the protein database of A. digitifera from NCBI (34,278 entries). A spectral library was built with Prosit (https://www.proteomicsdb.org/prosit/) using the NCBI database of A. digitifera. The settings were as follows: Fragmentation: HCD, Precursor Tolerance: 10 ppm, Fragment tolerance: 10 ppm, Data acquisition type: Overlapping DIA, Digestion enzyme: Trypsin, Peptide Charge: 2 to 4, Max Missed Cleavages: 1, Fixed modification: Carbamidomethylation, Peptide false discovery rate (FDR): > 1%, and Protein FDR: > 1%.

The proteins were searched with scaffold DIA to identify them from the mass results. The data were shared at jPOSTrepo (announced ID JPST002975 PXD050472).

Searching for PKA Substrate

PKA substrates were isolated from candidate sequences with RRXS/T. The “grep” command in the terminal in Macintosh was used to isolate PKA substrate (grep “RR.S” or grep “RR.T”) candidates containing RRXS, RRXT, or both. A total of 165 proteins with phosphorylation residues were isolated from over 2000 mass spectrometry-identified A. digitifera sperm proteins.

Comparison with the PKA Substrates of the Vertebrates

The identified PKA substrates were compared with the PhosphoSitePlus Substrate of Kinase database (https://maayanlab.cloud/Harmonizome/dataset/PhosphoSitePlus+Substrates+of+Kinases) (Hornbeck et al. 2004, 2015). A total of 327 protein kinase, cAMP-dependent (PRKCA) substrates were identified. The overlap between these substrates and the candidates of PKA substrates was then analyzed.

Searching Orthologs of the PKA Substrate

From the isolated 165 candidates of PKA substrates, the orthologs of these coding sequences (CDs) were screened using ORTHOSCOPE v.1.5.2 (Inoue and Satoh 2019) to confirm that the isolated candidates have orthologs among corals, cnidarians, and vertebrates (Fig. 1c; Supplementary Table 1). This search was performed using the Acropora site (https://orthoscope.jp/orthoscope/Acropora.html) with the default settings applied. Subsequently, one ortholog group was selected from the amino acid sequence FASTA files from ORTHOSCOPE (010_candidates_prot.txt) to search for phosphorylation residues and nucleotides among the FASTA files (010_candidates_nucl.txt) for codon evolution analyses (Fig. 1c). The file was named “**_orthologs.” Among the 165 proteins, 17 proteins orthologs were not successfully isolated with ORTHOSCOPE. Subsequently, orthologs for the remaining 148 proteins were successfully isolated. The script was then executed to determine the number of operational taxonomic units (OTU). Of the 165 candidates, 148 protein orthologs were identified. Subsequently, the script named “pick_phosphorylation_sites.sh” was used to identify and count the number of genes with phosphorylation residues among orthologs and Acropora in this ortholog (i.e., “RRXT = 8/120 RRXS = 30/120, Acropora RRXT = 0/45, Acropora RRXS = 13/45”). In this output, RRXS = 30/120 indicates that 8 orthologs from corals to vertebrates have RRXT from 120 orthologs, and Acropora RRXS = 13/45 indicates that 13 orthologs in the coral Acropora have RRXS from 45 orthologs in the Acropora. The scripts are available at https://github.com/Masaya0606/Isolation-of-PKA-subrtate-in-the-coral-Acropora.git.

Molecular Evolutionary Analyses and Conservation of Phosphorylation Residues in the Orthologs

The molecular sequence evolution of the isolated orthologs was analyzed in terms of non-synonymous and synonymous mutations, and the conservation of phosphorylation residues among the orthologs was checked. As described earlier, the coding sequences of the orthologs were isolated from the nucleotide FASTA files from ORTHOSCOPE (010_candidates_nucl.txt). The isolated coding sequences were aligned using MAFFT, and the aligned files were converted to NEXUS format. The converted files were re-examined using Mesquite, and files containing low-alignment sequences were removed. Subsequently, the files were cleaned and transformed into PHYLIP format to build a maximum likelihood phylogenetic tree using RaxML-NG for subsequent codon evolution analyses. A phylogenetic tree was built with a maximum likelihood criterion using the GTR-gamma model, the bootstrap values of the tree file were removed for the following molecular evolutionary analyses of codons with CodeML analyses.

CodeML, from PAML, was used for molecular sequence evolution. A codon residue model was used to examine molecular evolution. Before the analyses, a phylogenetic tree was built with RaxML-NG using the GTR-gamma model from the aligned PHYLIP file, which was removed from the tree; the edited tree and PHYLIP file for building the tree were used for the CodeML analyses. The conserved sequences of phosphorylation residues (RRXT/RRXS) among the orthologs were examined using the grep command described earlier.

In the CodeML analyses, to examine the codons in the orthologs under positive selection, two models were set: neutral evolution and positive selection. In the analyses, model8a was used for the null hypothesis and model8 for the positive selection. The two models were run separately, using ML phylogenetic trees as tree files and PHYLIP files as nucleotide files. Subsequently, the likelihood ratio between model8 and model8a was calculated (ΔlnL = 2 * (model8a − model8)), and the positive selection was supported when ΔlnL > 2.68 (Yang 1997). The branch non-synonymous/synonymous mutation rate of codons (dN/dS ratio) was obtained from the output file of model8.

Statistical Analyses

A Wilcoxon rank-sum test was performed for the conservation of phosphorylation sites and the dN/dS ratio obtained from code analyses. For the motility analyses, Tukey’s HSD test was applied after the homogeneity of variance was examined via Levene’s test. Kruskal–Wallis test was conducted for swimming velocity in 8-bromo-cAMP. These analyses were performed using R (R Core Team 2021), and the level of statistical significance was P < 0.05.

Results

Involvement of Membrane Potential and cAMP

We examined the effect of potassium ionophore valinomycin and membrane-permeable cAMP, 8bro-cAMP on the sperm motility activation in A. tenuis or A. florida. In the presence of valinomycin, sperm started their sperm motility (Fig. 2a, Tukey HSD, ASW as a negative control vs. valinomycin 10 µM 100 µM, NH4Cl treatment P < 0.001) and swimming velocity was also significantly different from the NH4Cl-induced motility activation sperm (Fig. 2b, Tukey HSD P < 0.05). 8bro-cAMP induced sperm motility activation in artificial sweater (ASW, Fig. 2c, Tukey HSD, NH4Cl-induced motility activation as a control vs. 8-bromo-cAMP 10 µM or 100 µM, P < 0.001); however, the swimming velocity was not significantly different (Fig. 2d, Kruskal’s test, χ2 = 8, df = 8, P = 0.43). This result indicates cAMP-related sperm motility activation in the coral Acropora spp.

Fig. 2
figure 2

Effect of potassium ionophore, valinomycin, and membrane-permeable cAMP, 8bro-cAMP, on coral sperm motility activation. a The effect of valinomycin, the potassium ionophore, on sperm motility activation, and b swimming velocity was examined. Acropora tenuis sperm were used for the analyses (N = 3); c the effect of the membrane-permeable cAMP analog, 8-bromo-cAMP on motility and d swimming velocity was examined in A. florida sperm (N = 4). Tukey’s HSD test for motility and Kruskal–Wallis test were performed for swimming velocity. *Significant difference (P < 0.001)

Full size image

Candidates of the PKA Substrates

We identified 165 sperm-constituent proteins containing RRXT or RRXS sequences among the deduced amino acid sequences (Table 1). These candidates included motor proteins, such as dynein, adenylate kinase, mitochondrial and axonemal proteins, and several uncharacterized proteins. Subsequently, we searched for orthologs of these candidates, successfully isolating 148 orthologs out of the initial 165. Ten candidates were specific to invertebrates, cnidarians, or corals, whereas orthologs for approximately 20 candidates could not be identified. The copy number of several candidates increased in the corals; we included these paralogs in one ortholog group for subsequent analyses.

Table 1 Isolated PKA substrates in sperm proteins
Full size table

Comparison with the PKA Substrates of the Vertebrates

We compared the overlap of the 327 PKA substrates in the PhosphoSitePlus Substrate of Kinase database (Supplementary Table 2). Among the 165 candidates, only 7 substrates coincided with the record (Table 2).

Table 2 Isolated PKA substrates in sperm proteins matched with PhosphoSitePlus Substrates of Kinases
Full size table

Identification of Phosphorylation Residues Among the Orthologs

Most Acropora species contained conserved phosphorylation residues, with threonine residues (RRXT) being more abundant than serine residues (RRXS). These residues are conserved in most dyneins, excluding T-complex protein 1 subunit epsilon-like (Protein no. 76-689 in Table 1). Additionally, phosphorylation residues were preserved in essential sperm components such as mitochondrial proteins (1081, 1583) and tubulin (347).

PKA substrate candidates displayed conservation across animals, with the identification of cnidarian-specific proteins. Notably, several uncharacterized proteins (Protein no. 52-346 in Table 1) were conserved among corals to Homo sapiens. Four E3 ubiquitin ligase candidates for PKA phosphorylation were identified, suggesting conserved ubiquitination from cnidarians to vertebrates and potential regulation by PKA phosphorylation. Eight cnidarian- or coral-specific, uncharacterized proteins were also identified. Seven proteins matched the database, with conserved serine or threonine of PKA phosphorylation residues.

Molecular Evolution of Candidate Proteins

We conducted molecular evolutionary analyses to compare the molecular evolutionary rates of codons and the conservation of phosphorylation residues. However, owing to the time-consuming nature of aligning orthologs across the animal kingdom, analysis was limited to cnidarian- or coral-specific orthologs. The codon site analysis with CodeML revealed that 48 out of the 148 orthologs were under positive selection in the coral Acropora (Table 1). This analysis provided information on whether specific codons were subject to positive selection, as indicated by a low dN/dS ratio in the open reading frames of many orthologs. Consistent with this prediction, the dN/dS ratio for the open reading frames of all isolated orthologs was low (Fig. 3a–d). Notably, the dN/dS ratio and the conservation of PKA phosphorylation residues in threonine among the orthologs were not negatively correlated (Fig. 3c, d).

Fig. 3
figure 3

Conservation rates of serine or threonine residues (RRXS/T) and dN/dS ratio among Acropora or all isolated orthologs. The relationship between serine or threonine residue conservation rates and the dN/dS ratio output in the CodeML analyses in model8 (positive selection) was utilized for plotting. Conservation rates of a RRXS in all orthologs, b Acropora, or d RRXT in all orthologs, and c Acropora were plotted

Full size image

Conversely, the conservation of serine residues in the orthologs under positive selection was significantly higher than that in those under purifying selection (Fig. 4a, b). However, the conservation of threonine residues did not exhibit the same trend (Fig. 4c, d). This trend was consistent across all orthologs. Therefore, 30% of PKA substrate candidates were under positive selection, whereas threonine residues remained conserved. Consequently, codon evolution did not impact the conservation of threonine phosphorylation, and serine residues were more conserved than those under purifying selection. The coding sequences of the seven proteins that matched with the database were subjected to strong purifying selection, whereas only one of them was subjected to positive selection (Table 2).

Fig. 4
figure 4

Conservation rates of serine or threonine residues (RRXS/T) between under positive selection and purifying selection. Violin plots illustrating RRXT or TTXS conservation rates between orthologs under positive and purifying selection conditions. a RRXT among Acropora, b all orthologs, c RRXS among Acropora, and d all orthologs. Differences were compared using the Wilcoxon rank-sum test with continuity correction (W = 1425, P < 0.0005 in RRXS Acropora, and W = 1361.5, P < 0.0005 in RRTS in all orthologs). *Significant difference between positive and purifying selections (Wilcoxon rank-sum test)

Full size image

Discussion

In this study, we identified candidate PKA substrates in coral sperm and explored their diversity. Some candidates are specific to cnidarians or corals. Approximately one-third of these candidates underwent positive selection, indicating adaptive evolution. Proteins crucial for flagellar motility, including tubulin and dyneins, are conserved across animals. Notably, proteins essential for flagellar motility, such as dyneins, are also under positive selection. PKA is conserved across animals, and many candidate substrates exhibit conservation between mammals and phylogenetically early branching animal cnidarians, including corals (Barott et al. 2013; Glass et al. 2023; Speer et al. 2021). Proteins essential for the architecture of motility apparatus axonemes, such as dyneins and β-tubulin, feature PKA phosphorylation sequences. These findings align with those of prior studies (Glass et al. 2023; Speer et al. 2021), and we identified downstream candidate PKA substrates. While PKA and adenylyl cyclase are conserved, signal transduction pathways have evolved through speciation (Bradley and Beltrao 2019). For example, in mammals, mitochondrial PKA controls sperm motility via ATP (Mizrahi and Breitbart 2014). In contrast, valinomycin experiments indicate membrane hyperpolarization commonly contributes to activating soluble adenylyl cyclase (sAC) in the coral Acropora spp. and ascidians (Izumi et al. 1999). Although several mitochondrial proteins, including adenylate kinase 2 and other candidate PKA substrates, were identified in the coral A. digitifera, we did not observe PKA phosphorylation in glucose-6-phosphate isomerase, which is involved in ATP synthesis in mammals (Mizrahi and Breitbart 2014). Although we identified the sperm proteins from only one colony of A. digitifera and the amount of the sperm proteins potentially changes according to the genotype of the colonies, this is the first step to understanding the sperm flagellar motility activation in the coral Acropora spp.

Although the overlap between the substrates and the database was low, we identified the dynein light chain, essential for the flagellar motility initiation among vertebrates. Additionally, the flagellar-specific proteins, such as radial spoke head 4 proteins (RSP4), were not included in the database; however, RSP3, a paralog of RSP4, was downstream of PKA phosphorylation signaling and regulates dynein activity (Bicka et al. 2023; Gaillard et al. 2001). There are many flagellar-specific proteins involved in flagellar motility activation, with many uncharacterized proteins presumed to be involved as well. Nevertheless, identifying these proteins involved in flagellar motility initiation, using techniques such as an anti-phospho PKA substrate antibody, is necessary.

Ubiquitin ligases are present in sperm proteins and are mostly involved in flagellar formation during spermatogenesis (Long et al. 2015). Ubiquitination during cilia and flagella formation, as summarized previously (Long et al. 2015), involves ubiquitination of motor proteins during cilia and flagella formation and disassembly (Huang et al. 2009). The activation of flagellar motility could be influenced by ubiquitination, as proteasomes cause proteolysis of ubiquitinated proteins. In salmonid fish sperm flagella, proteasomes form a complex with the motor protein dynein, as predicted by PKA phosphorylation (Inaba et al. 1998). PKA and proteasomes are coupled to function together for capacitation in mammalian sperm via phosphorylation (Zapata-Carmona et al. 2019). Moreover, proteasomes are present in the sperm of the coral A. digitifera. Thus, the PKA phosphorylation-induced, proteasome-dependent motility regulatory mechanism potentially contributes to the activation of flagellar motility.

In the present study, our findings revealed that one-third of the PKA substrate candidates underwent positive selection, indicating adaptive evolution and functional modification. Orthologs of many PKA substrate candidates in Acropora have been conserved through speciation and evolution in the animal kingdom. Positive selection is favored when functional modifications benefit sperm function. For example, faster-swimming sperm may have a competitive advantage in synchronous spawning (Fitzpatrick 2020; Locatello et al. 2007). Additionally, sneaking males can release sperm to potentially enhance fertilization success (Birkhead and Pizzari 2002; Burness et al. 2004). In other words, proteins associated with sperm function, including flagellar motility, likely underwent adaptive evolution. Despite the validity of these speculations, the identification and characterization of the phosphorylation status of flagellar proteins remain crucial.

In the coral Acropora, PKA substrates play a role in sperm motility activation and chemotaxis. Acropora is a multispecies spawning cnidarian (Baird et al. 2021b; Hayashibara et al. 1993), one of a species-abundant genus of over 110 species identified in the Indo-Pacific (Wallace 1999), and many sympatric species spawn synchronously (Baird et al. 2021b). The initiation of sperm flagellar motility after spawning occurs after the segregation of gamete bundles into sperm and eggs (Morita et al. 2006). The timing of sperm flagellar motility initiation and regulation of flagellar beating, such as chemotaxis, may facilitate contact with conspecifics in a mixture of heterospecifics. The multispecies spawning patterns may have influenced the evolution of flagellar motility, enhancing fertilization success and fitness. Additionally, PKA signaling pathways are conserved among cnidarians, and the conservation of phosphorylation residues in orthologs under positive selection is higher than that under purifying selection. These features are presumably associated with PKA-dependent sperm function to adapt to reproduction.

The higher conservation rates of the serine residues may be attributed to the strict positioning of the threonine residues in terms of the substrate (Pandey et al. 2023) and the structure of the kinase (Durek et al. 2009). PKA substrates under positive selection presumably underwent changes in the 3D structure of the proteins owing to amino acid substitutions. In the present study, the conservation rates of the threonine residues were not significantly different between positive selection and purifying selection. This could be because the threonine residues are structurally highly ordered for phosphorylation. Thus, conformational changes of the proteins induced by positive selection presumably influence phosphorylation by PKA. Additionally, the phosphorylation preferences of serine residues by PKA are reported higher than those of the threonine owing to their phosphoacceptor preferences (Chen et al. 2014). The number of threonine residues available for phosphorylation has been reported to be considerably lower, approximately one-third, than that of serine residues (Kreegipuu et al. 1998), which is consistent with our analyses.

In the present study, seven candidates of PKA substrates that matched with the PhosphoSitePlus Substrate of Kinase database were under purifying selection. Functional modifications are presumably not favored in common substrates that function across various tissues. Nevertheless, distinctive phosphorylation sites contribute to selective phosphorylation in the progression of the cell cycle, an essential process in all types of cells (Alexander et al. 2011). Although PKA signal transduction functions from prokaryotes to eukaryotes (Portela and Rossi 2020), and phosphorylation motifs are conserved among eukaryotes (Bradley and Beltrao 2019), the substrates in the flagellar motility did not overlap with the PhosphoSitePlus Substrate of Kinase database. Moreover, the PKA substrate for motility activation is specific to physiological reactions; thus, the substrates are potentially different. However, further investigations are required to confirm this.

Cnidarian- or coral-specific proteins are likely specialized in corals. Most of these cnidarian or coral-specific proteins were uncharacterized proteins. Therefore, identifying the functions of these proteins and their localization within coral polyps is imperative. As previously mentioned, the predictability of multispecies spawning and flagellar motility contributing to fertilization success with conspecifics has yet to be examined in terms of the functions of these proteins.

The applied membrane-permeable cAMP facilitated motility initiation, highlighting the need for further detailed analyses to confirm the isolated PKA substrate in this study. Although only in silico analyses were used to identify the PKA substrate, immunoprecipitation and examination of the phosphoprotein state before and after motility activation with an anti-PKA phosphorylation antibody are crucial. Additionally, new techniques are available for identifying novel PKA substrates; the PKA and substrate complex can be analyzed using proximity-dependent biotin identification (BioID) methods (Niinae et al. 2021). Preliminary experiments using membrane-permeable cAMP analogs demonstrated flagellar motility activation (Fig. 2), consistent with previous detailed biochemical studies in corals (Glass et al. 2023; Speer et al. 2021). While inhibitor analyses will be convenient for identifying axonemal protein phosphorylation pathways via immunoblot after 2D electrophoresis, the separation of membrane proteins with 2D electrophoresis may pose challenges. As discussed earlier, only a limited number of candidates overlapped with the PKA substrate registered in the PhosphoSitePlus Substrate of Kinase database. Thus, comprehensive analyses are essential to provide more reliable information.

In conclusion, we used in silico analyses to isolate PKA substrate candidates, exploring how orthologs of the candidate proteins are conserved and how orthologs are under positive selection, leading to adaptive evolution. However, molecular evolutionary analyses, such as hypotheses related to internal fertilization-specific positive selection or high intensity of sperm competition via branch residue analyses, could not be conducted owing to sequence diversity in aligning nucleotide files. Nevertheless, the higher conservation of phosphorylation residues in orthologs under positive selection suggests the evolution of specific regions in the PKA substrate protein to acquire desirable functions, accompanied by the conservation of serine and threonine residues for PKA phosphorylation. Furthermore, this study implies the potential conservation of PKA signal transduction, with its functions evolving under various selective pressures, such as sperm competition or environmental changes from external to internal fertilization. This study provides insights into the evolutionary dynamics of reproductive mechanisms in corals, the potential adaptive evolution of sperm-related proteins, and the conservation of PKA signal transduction across diverse reproductive conditions. Additionally, the study may contribute to a better understanding of the molecular mechanisms underlying fertilization success, which could have implications for coral reproductive ecology, evolution, and conservation. However, we did not investigate these mechanisms in flagellar proteins or examine their phosphorylation during motility activation. Therefore, future efforts should focus on identifying flagellar proteins among the candidates and elucidating their phosphorylation during flagellar motility activation in spawning coral Acropora.