Transcriptomic responses of Acropora tenuis planula larvae to three Durusdinium strains.
Successful infection with each symbiont culture (two Durusdinium sp. (Urasan5-A10 and TsIS-G10) and D. trenchii (CCMP2556)) was confirmed by fluorescence microscopy in all treatment groups at 3 dpi. Uptake efficiencies (mean ± SD of three replicates) of Urasan5-A10, CCMP2556, and TsIS-G10 in A. tenuis larvae were 73.3 ± 11.5%, 66.7 ± 25.2%, and 56.7 ± 20.8%, respectively (Supplementary Table S1). Algal cell densities (mean ± SE cells per larva) inoculated with Urasan5-A10, CCMP2556, and TsIS-G10 were 8.0 ± 2.2, 2.6 ± 0.4, and 3.3 ± 0.5 cells/larva, respectively (Supplementary Table S1). After confirmation of successful infection, we performed 3’mRNA sequencing of A. tenuis larvae inoculated with Durusdinium culture strains and those with no exposure (apo-symbiotic). An average of seven million RNA-seq reads per sample were retained after quality trimming, ~ 73% of which were mapped to A. tenuis gene models (Supplementary Table S2). Non-metric multidimensional scaling (NMDS), based on gene expression levels of 7,697 genes for which expression levels (TMM-normalized CPM) were larger than 10 in all samples, showing agreement among the biological replicates (Supplementary Figure S1). We then compared gene expression levels between Durusdinium-exposed and unexposed groups. Then we identified 14 and 5 DEGs in planula larvae inoculated with Urasan5-A10 and CCMP2556 strains, respectively, both of which were isolated from corals (Fig. 1). Among these, two genes (aten_s0305.g7 and aten_s0305.g9) without similarity to sequences in the Swiss-Prot database were commonly detected from both planula larvae inoculated with Urasan5-A10 and CCMP2556 (Fig. 1). In contrast, no genes were differentially expressed by inoculation with the TsIS-G10 strain, which was isolated from a giant clam (Fig. 1).
Acquisition, but transcriptomic neglect of Durusdinium in Acropora tenuis planula larvae.
In order to reveal transcriptomic responses of A. tenuis planula larvae to two dominant symbionts in early life stages, we compared transcriptomic responses of planula larvae inoculated with Durusdinium and Symbiodinium. For larval transcriptomic responses of A. tenuis to Symbiodinium, we used previously reported RNA-seq data of A. tenuis planula larvae inoculated with each of the three Symbiodinium species, S. microadriaticum (dominant symbiont), S. natans (non-symbiotic, never detected in corals), and S. tridacnidorum (occasional symbiont) (Yoshioka et al. 2021a). Numbers of DEGs detected in planula larvae 4 days post-inoculation with S. microadriaticum, S. natans, and S. tridacnidorum were 2,521, 176, and 1,065, respectively (Supplementary Table S3). Then we compared DEG repertoires between planula larvae inoculated with Durusdinium and Symbiodinium (Fig. 2). Among the 14 DEGs in planula larvae inoculated with Durusdinium sp. (Urasan5-A10), five DEGs were exclusive to Durusdinium-inoculation, while nine DEGs, such as a circadian clock gene CLOCK (aten_s0092.g8) and a Ras-related estrogen-regulated growth inhibitor RERG (aten_s0011.g21), were also shared with at least one of the Symbiodinium inoculations (Fig. 2A). On the other hand, among the 5 DEGs in planula larvae inoculated with D. trenchii (CCMP2556), three DEGs were exclusive to D. trenchii inoculation, while two DEGs (aten_s0010.103 and aten_s0305.g9) were shared with at least one of the Symbiodinium inoculations (Fig. 2B). In both comparisons, one gene (aten_s0305.g9) was commonly detected in all samples (Fig. 2), indicating that this gene responded to the presence of Symbiodiniaceae, regardless of genus or species. However, the number of DEGs in larvae inoculated with Durusdinium was considerably lower than those in larvae inoculated with any Symbiodinium.
Dynamic transcriptomic changes in response to algal symbionts in Acropora tenuis primary polyps.
Although infection with Durusdinium affect only 14 genes in A. tenuis planula larvae (see above section), Durusdinium remained in Acropora coral cells until the juveniles achieved centimeter scale. In order to compare transcriptomic responses of Acropora to Durusdinium in later planula larvae, we analyzed transcriptomic responses of A. tenuis primary polyps (millimeter scale) to D. trenchii (CCMP2556) at 10 and 20 dpi, using publicly available RNA-seq data (Yuyama et al. 2018). We mapped RNA-seq data to A. tenuis gene models (Supplementary Table S5) and identified 2,514 and 7,255 DEGs in A. tenuis polyps at 10 and 20 dpi, respectively (Fig. 3A; Supplementary Table S5). Then, we compared DEG repertories between planula larvae at 3 dpi and polyps at 10 and 20 dpi (Fig. 3A). Two DEGs (aten_s0157.g31 and aten_s0262.g12) were specific to planula larvae at 3 dpi. One DEG (aten_s305.g7) was shared between planula larvae at 3 dpi and polyps at 10 dpi, and two DEGs (aten_s0010.g103 and aten_s0305.g9) were shared among all three samples (Fig. 3A). Interestingly, degrees of gene expression levels of shared DEGs were different among samples (Fig. 3B). For instance, the expression level of aten_s0305.g9 was upregulated in planula larvae at 3 dpi and polyps at 10 dpi, but was downregulated in polyps at 20 dpi (Fig. 3B). In addition, gene expression levels of aten_s0010.g103 were upregulated in planula larvae at 3 dpi, but were downregulated in polyps at 10 and 20 dpi (Fig. 3B). These differences suggest that gene expression levels were flexibly controlled between developmental stages.
Identification of core transcriptomic responses to algal symbionts in Acropora tenuis early life stages.
In order to reveal core transcriptomic responses, groups of genes altering gene expression regardless of symbiont types, in early life stages of A. tenuis during symbiosis with Durusdinium and Symbiodinium, we compared DEG repertoires between planula larvae inoculated with S. microadriaticum (at 4 to 12 dpi) and polyps inoculated with D. trenchii (at 10 and 20 dpi). Among upregulated DEGs commonly detected following D. trenchii and S. microadriaticum inoculation, 377 DEGs were common to both (Fig. 4) and eight functional categories, such as transport, cell adhesion, and immunity, were significantly (p < 0.05) upregulated (Fig. 4). Among downregulated DEGs commonly detected following D. trenchii and S. microadriaticum inoculation, 541 DEGs were common to both (Fig. 4) and 11 functional categories, such as cellular function (cell cycle and transcription), metabolisms (lipid, histidine, purine, and one-carbon (methionine, folic acid, and vitamin B12)), biological rhythms, were significantly downregulated (Fig. 4). Among them, symbiosis-related, tandemly duplicated genes reported by Yoshioka et al. (2021a), such as two sugar transporters (aten_s0094.g50 and aten_s0094.g54) and two Anthozoan Notch ligand-like genes (aten_s0035.g9 and aten_s0035.g10) were included (Supplementary Table S6), indicating that these genes could be essential for maintenance of symbiosis in early life stages of A. tenuis.
On the other hand, the majority of DEGs in inoculations with both D. trenchii and S. microadriaticum were specific to the individual symbionts (Fig. 4). Among D. trenchii inoculation-specific DEGs, 3,751 upregulated DEGs, including 22 significantly upregulated functional categories, such as transport, immunity, and stress responses (Fig. 4), and 3,128 downregulated DEGs, including 22 significantly downregulated functional categories, such as cell cycle and immunity, were detected (Fig. 4). In S. microadriaticum-inoculation specific DEGs, 1,006 upregulated DEGs, including 11 significantly upregulated categories, such as transport and cell adhesion (Fig. 4), and 1,429 downregulated DEGs, including 15 significantly downregulated functional categories, such as transcription and immunity, were detected (Fig. 4). Two functional categories, transport and cell adhesion, were upregulated among symbiont-specific DEGs (Fig. 4; Supplementary Table S8) and the former consisted of many genes (488 DEGs in polyps inoculated with D. trenchii and 82 DEGs in planula larvae inoculated with S. microadriaticum (Supplementary Table S8. These indicate that A. tenuis changes its transcriptome depending on the species of algal symbiont and/or developmental stages.
SLC transporter clusters in Acropora tenuis genome involved in coral-algal symbiosis.
Within the category “transport” defined by UniProt, many genes, such as ribosomal proteins, are included. To focus on membrane transporters, we attempted to extract solute carrier (SLC) groups, which constitute a major fraction of transport-related genes in vertebrates (Hediger et al. 2013). We found at least 95 and 19 possible SLC-like genes among D. trenchii inoculation-specific DEGs and S. microadriaticum inoculation-specific DEGs, respectively (Supplementary Table S7). Among those, we found six SLC transporter clusters in the A. tenuis genome (Fig. 5). Scaffold 0094 contains six tandemly located SLC transporters possibly involved in sugar transport, two of which were upregulated by inoculation with both D. trenchii and S. microadriaticum (Fig. 5A). Scaffold 0139 contains five tandemly located SLC transporters possibly involved in amino acid transport, three of which were upregulated by inoculation with D. trenchii, and another that was upregulated solely by inoculation with S. microadriaticum (Fig. 5B). Scaffold 0058 contains three tandemly located SLC transporters possibly involved in amino acid transport, one of which was upregulated by inoculation with S. microadriaticum (Fig. 5C). Scaffold 0205 contains four tandemly located SLC transporters possible involved in amino acid transport, all of which were upregulated by inoculation with D. trenchii (Fig. 5D). Scaffold 0092 contains four tandemly located SLC transporters possibly involved in amino acid and/or peptide transport, all of which were upregulated by inoculation with D. trenchii (Fig. 5D). Scaffold 0313 contains five tandemly located SLC transporters possibly involved in ion transport, three of which were upregulated by inoculation with D. trenchii (Fig. 5D). These indicate that A. tenuis alters expression of SLC transporters depending on the species of algal symbiont. We performed molecular phylogenetic analyses of these tandemly duplicated SLC transporters. The sugar transporter cluster on scaffold 0094 was specifically duplicated in the Acropora lineage (Fig. 5A; Supplementary Figure S2). The amino acid transporter cluster on scaffold 0139 has also been conserved in Nematostella, indicating that these originated before divergence of corals and sea anemones (Fig. 5B; Supplementary Figure S3). The amino acid transporter cluster on scaffold 0058 was duplicated in the common ancestor of Acropora and Montipora (Fig. 5C; Supplementary Figure S4). The amino acid transporter cluster on scaffold 0205 originated before divergence of corals and has also been conserved in Nematostella (Fig. 5D; Supplementary Figure S5). The amino acid transporter cluster on scaffold 0356 was specifically duplicated in Acropora (Fig. 5E; Supplementary Figure S6). The ion transporter cluster on scaffold 0313 was duplicated in the common ancestor of Acropora and Montipora (Fig. 5F; Supplementary Figure S7).