Genome of Halimeda opuntia reveals differentiation of subgenomes and molecular bases of multinucleation and calcification in algae

Significance

Coral reef ecosystems are undergoing significant degradation and reorganization due to ocean warming and acidification. Calcareous algae, crucial primary producers and reef-builders, exhibit diverse morphologies, lifestyles, and adaptative strategies. A significant gap exists, however, in deciphering the genetic basis of algae positioned at an evolutionary crossroad from unicellular to multicellular, from intracellular calcifying to extracellular calcification, and from acidification-sensitive to acidification-tolerant. Genome analysis of the green alga Halimeda opuntia and other algae shed light on unique genetic features associated with multinucleation, cell fragment regeneration, extracellular calcification, and tolerance of CO₂ increases in seawater. Our findings advance the understanding of how calcareous algae respond to environmental changes and have implications in regenerative biology, plant grafting, and coral reef conservation and restoration.

Abstract

Algae mostly occur either as unicellular (microalgae) or multicellular (macroalgae) species, both being uninucleate. There are important exceptions, however, as some unicellular algae are multinucleate and macroscopic, some of which inhabit tropical seas and contribute to biocalcification and coral reef robustness. The evolutionary mechanisms and ecological significance of multinucleation and associated traits (e.g., rapid wound healing) are poorly understood. Here, we report the genome of Halimeda opuntia, a giant multinucleate unicellular chlorophyte characterized by interutricular calcification. We achieve a high-quality genome assembly that shows segregation into four subgenomes, with evidence for polyploidization concomitant with historical sea level and climate changes. We further find myosin VIII missing in H. opuntia and three other unicellular multinucleate chlorophytes, suggesting a potential mechanism that may underpin multinucleation. Genome analysis provides clues about how the unicellular alga could survive fragmentation and regenerate, as well as potential signatures for extracellular calcification and the coupling of calcification with photosynthesis. In addition, proteomic alkalinity shifts were found to potentially confer plasticity of H. opuntia to ocean acidification (OA). Our study provides crucial genetic information necessary for understanding multinucleation, cell regeneration, plasticity to OA, and different modes of calcification in algae and other organisms, which has important implications in reef conservation and bioengineering.

Algae are the foundation of aquatic ecosystems, contributing ~50% of global carbon fixation (1). They are genetically and ecologically diverse. Green algae are one of the four Archaeplastida lineages, arising from primary endosymbiosis of a eukaryote with a cyanobacterium (2). Chlorophytes are one of the three algal phyla that contain both unicellular and multicellular species (the others being rhodophytes and ochrophytes) (3). Genomics has dramatically advanced the molecular genetics and evolutionary biology of chlorophytes. The study of the Chlamydomonas reinhardtii genome revealed animal-like characteristics and spurred a rapid expansion of chlorophyte genomics (4). The subsequent sequencing of the smallest eukaryote Ostreococcus unveiled genomic imprints of sex differentiation and chromosome recombination (5). The genome of the globally distributed and abundant Micromonas provides valuable insights into ecological differentiation and the dynamics of early plant evolution (6). Genomic analysis of unicellular and multicellular algae within core Chlorophyta uncovered the genetic basis of morphological diversity and multiple transitions to macroscopic growth in Ulvophyceae (7, 8). Notably, the unicellular multinucleate lineage within Ulvophyceae, representing a transition from unicellular to multicellular viridiplantae, is receiving increasing attention in genomic research.

Halimeda belongs to Bryopsidales, one of the two notable orders in Ulvophyceae (the other being Dasycladales) that contain unicellular coenocytic species (9), and is commonly found in tropical and subtropical oceans where it plays crucial roles in biogenic reef ecosystems. Unicellular coenocytic algae (also known as siphonous algae) undergo numerous rounds of nuclear division while the cell grows in size but without division (cytokinesis), producing giant cells with numerous nuclei (10). Only a few other lineages within the phylum Ochrophyta are also unicellular coenocytic (SI Appendix, Table S1). Genome assemblies from siphonous Caulerpa lentillifera, Bryopsis sp., and Ostreobium quekettii have unveiled the genetic foundations behind their specialized phenotypes (11–13). Siphonous algae often exhibit a higher density of chloroplasts within their shared cytoplasm (14). The absence of cell walls between nuclei allows for a more efficient distribution of chloroplasts, maximizing the surface area enhancing photosynthesis. Notably, Halimeda, Caulerpa, and Bryopsis are the largest known single-celled organisms, which usually grow up to 40 cm, with C. taxifolia even reaching 1 m. They commonly reproduce asexually by fragmentation but are also capable of sexual reproduction (15). Fragments of the thalli or fronds can regenerate quickly and give rise to new individuals, allowing for rapid colonization of suitable substrates. Deciphering the genetic basis of their remarkable capabilities of cell regeneration and wound healing is important for understanding the evolutionary process and selection pressure for these phenotypic features and has broad implications in medical and conservation applications.

Halimeda is furthermore an excellent candidate for genomics because it calcifies. Calcareous algae along with corals and other calcifying invertebrates account for about 30% of oceanic calcium carbonate (CaCO₃) production, thereby constituting a substantial reservoir of “carbon sinks” (16). Calcification in Halimeda is unique, involving the accumulation of aragonite in the semienclosed interutricle space (IUS), an extracellular compartment formed by the fusion of the outermost filaments with approximately 75% of the cell wall (17). Halimeda is one of the most abundant calcifiers in ancient and modern biogenic reefs (18). Calcification in algae, an energy-consuming process controlled by the CaCO₃ saturation state (Ω) (19), necessitates local supersaturation and nucleation for CaCO₃ crystallization. The supersaturation condition relies on continuous supply of Ca²⁺ and HCO₃⁻ as well as efficient regional regulation of pH in the calcifying fluid, both influencing Ω_CaCO3 (20). Besides, its crystal nucleation requires organic matrices as templates (21, 22). Depending on the crystalline structure, crystal deposit localization, and cell characteristics, algal calcification is classified into biologically induced or organic matrix-mediated processes. However, the regulatory mechanisms governing algal calcification are still poorly understood and underexplored, with genomic information available only for the haptophyte Emiliania huxleyi (23). In addition, biogenic reefs are increasingly degraded due to ocean acidification (OA), which causes direct dissolution of CaCO₃ crystals (24). Investigating the effects of OA on diverse calcified species is critical for unraveling the underlying mechanisms and informing biogenic reef conservation and policy management.

To address these significant gaps in research, we generated a de novo genome assembly of Halimeda opuntia and conducted a targeted comparative genome analysis to investigate the evolutionary innovations associated with its unusual characteristics, in particular multinucleation, cellular regeneration, and calcification. For comparative analyses to decipher the molecular basis behind different calcification modes, we also sequenced the genome of Amphiroa fragilissima (Rhodophyta, Florideophyceae), a multicellular coralline alga that precipitates high-Mg calcite within its cell wall (25). Transcriptomes from diurnal cycles and proteomes following 6 wk of OA treatment were also generated in the two species to investigate their responses to environmental changes.

Results and Discussion

Genomic Features of H. opuntia.

Illumina paired-end short reads and PacBio long reads (totally 232.7 Gb) were combined to assemble the nuclear genome of H. opuntia (SI Appendix, Table S2). We obtained assemblies of 169.53 Mb (SI Appendix, Figs. S1 and S2 and Table S3), with a relatively high degree of genome completeness compared with other green algae (SI Appendix, Table S4). From the genome, 47,597 protein-coding genes were predicted, of which 45,148 (94.9%) were functionally annotated. This indicates a low level of proteomic innovation in this alga. However, transposable elements (TEs) occupy 44.4% of the genome (SI Appendix, Table S5), suggesting a high level of genomic fluidity. We collected a nuclear dataset consisting of 2,362 single-copy orthologous genes (SCOGs) derived from 26 genomes and two transcriptomes. A phylogenetic tree inferred from these SCOGs using a coalescence-based approach showed a sister relationship of Bryopsidales with Chlorophyceae (SI Appendix, Fig. S3A). It is noteworthy that dating divergence times in the phylogeny of Chlorophyta are challenging due to difficulties in interpreting fossil locations with respect to extant taxa (26–28). Our molecular dating analysis based on five time-calibrated nodes indicated that the four unicellular coenocytic algae diverged from the Ulva in the early Neoproterozoic (SI Appendix, Fig. S3A and Table S6). This result is consistent with the findings obtained using the third fossil calibration strategy (26). Relative to their last common ancestor, the speciation of H. opuntia involved expansion of 8,038 gene families and contraction of 536 gene families (Fig. 1A and Datasets S1 and S2).

Fig. 1.

Phylogenetic position and key features of H. opuntia. (A) A maximum-likelihood phylogenetic tree was constructed based on one-to-one homology among 16 algal genomes. Red numbers with plus signs and blue numbers with minus signs indicate expanded and constricted gene families in each species compared to the last common ancestor, respectively. Gray horizontal lines indicate 95% CI of divergence times, and red dots indicate five time-calibrated nodes which were determined from the Timetree website (http://www.timetree.org/). Information of the species and calibration times for divergence time calculation was supplied in SI Appendix, Fig. S3A and Table S6. A. fragilissima is a multicellular red alga containing high-Mg calcite crystals within the cell wall (Right Upper), and H. opuntia is a single-cell multinucleated green alga containing precipitated aragonite crystals in the semienclosed IUS (Right Lower). (B) Expanded gene families (red) and loss gene (blue) involved in nuclear division in the coenocytic algae of H. opuntia, C. lentillifera, Bryopsis sp., and O. quekettii. (C) Copy numbers of gigantism, nuclear division, and wound healing–related gene families in representative green algae.

Molecular Basis of the Giant Cell Size, Multinucleation, and Regeneration in Coenocytic Algae.

In both unicellular and multicellular organisms, cell size and genome size (amount of nuclear DNA per cell or approximately, cellular DNA content) are determined genetically and influenced to a small extent by environmental conditions (29, 30). It is believed that cells sense their sizes against their target size (or cell size-dependent traits) to gate the timing of nucleus division and cell division and orchestrate cell growth versus division to maintain cell size homeostasis (31). Thallus lengths in Halimeda can reach 30 cm or longer, with the nuclear DNA content of H. macrophysa estimated at approximately 3 Gb (derived from 2C = 3.1 pg, using the conversion factor 1 pg = 980 Mb) (32). Certain Halimeda species may maintain a constant nuclear DNA content by regulating the number and/or ploidy level (ranging from 2C to 16C) of nuclei within the same cytoplasm (32, 33). The large cellular DNA content of Halimeda likely serves essential functions beyond its genic role, such as maintaining nuclear-cytoplasmic volume ratios within coenocytic cells and counteracting the effects of increased extranuclear plastid DNA concentrations (34).

To achieve and maintain the giant cell size, large cellular components are required, including cytoskeleton, RNA, and proteins (29, 35). In H. opuntia, C. lentillifera, Bryopsis sp., and O. quekettii, we found expansion of gene families of actin (Act), alpha tubulin (Tuba), dynein light chain (Dynll), kinesin-like protein KIN-7D (Kin7D, SI Appendix, Fig. S4), and KIN-13A (Kin13A, SI Appendix, Fig. S5), which are involved in the coordinated regulation of cytoskeleton dynamics and may be associated with gigantism (Fig. 1 B and C). In support of this potential association, actin, tubulin, dynein, and kinesin are highly expressed in Bryopsis sp. and Acetabularia acetabulum, another giant unicellular chlorophyte with a huge (uninucleate) genome size (13, 36, 37). Transcription factors (TFs) are essential for gene expression, and their number within an organism correlates positively with genome size (38). Certain zinc-finger TFs, such as those implicated in megakaryocytic development (39), are known to play crucial roles. We found expansion of gene families of GATA TF12 (Gata12, SI Appendix, Fig. S6), UBP1-associated proteins 1C (Uba1C), nuclear TF Y subunit C2 (Nfyc2), and AP2-like ethylene-responsive TF SNZ (Snz) (Fig. 1C), consistent with the giant cell size of the species.

The multinucleate phenotype conceivably results from continued nuclear division and cell growth without cytokinesis. Comparative genome analysis indicated that mitosis-related gene families were expanded in the four coenocytic algae (Fig. 1 B and C). These include the regulator of chromosome condensation 1 (Rcc1) gene family, which serves to ensure the completeness of chromosome condensation at the prophase of mitosis (40). Also markedly expanded are gene families of abnormal spindle-like microcephaly-associated protein (Aspm) and augmin-like complex subunit 6 (Augms6), which are related to spindle organization and assembly (41, 42), Tuba, Dynll, Kin7D, Kin13A, and microtubule-associated protein, RP/EB family (Mapre), which are involved in the regulation of microtubule structures and microtubule-based movement of chromosomes (43–45). The expansion of these mitosis-related genes provides the genetic potential for a high rate of nucleus division.

We further looked for the molecular evidence of alterations in the cytokinesis machinery. In eukaryotes, cytokinesis involves myosins that function as a molecular motor to move organelles. In plants, particularly classes VIII (Myo8), XI (Myo11), and XIII (Myo13) myosins are involved (46). Myo11 proteins in Arabidopsis thaliana serve tissue-specific functions such as cytoplasmic streaming and pollen tube growth (47). Closely related Myo13 proteins have been linked to organelle transport and tip growth in green algae Acetabularia (48). Myo8 proteins are associated with microtubule ends and actin and play a role in tip growth, branching, and cytokinesis in moss (49, 50). Combined with the genome analysis of Bryopsis sp. (11), we observed the presence of Myo11 in 37 species examined, Myo13 only in Dasycladales and Chlorella sorokiniana, and Myo8 in all species except for the four unicellular coenocytic algae (Fig. 1B and SI Appendix, Fig. S7). The absence of Myo8 is distinctive in unicellular coenocytic taxa. Myo8 knockout in moss causes cytokinesis defects and result in a multinucleate phenotype (49). The multinucleate phenotype manifesting in the cytokinesis-defective plant mutants indicates that nuclear division can proceed despite incomplete cytokinesis (51). Thus, the loss of Myo8 gene in H. opuntia and other coenocytic algae is potentially an important factor contributing to multinucleation in these species.

H. opuntia and other giant siphonous green algae have a strong ability to regenerate, as fragments of a cell can survive and regrow into complete cells (15). This indicates that an extremely efficient mechanism exists allowing very quick synthesis of new cell membrane/cell wall materials at the breakpoint to prevent cytoplasm leakage and to heal. We examined the four available coenocytic algal genomes for genes encoding homologs of the wound healing–related proteins in plants and animals. We found the presence and expansion of such gene families in these algae (Fig. 1C). These include wound-inducible factors such as auxin-responsive GH3 (Gh3) and respiratory burst oxidase homolog (Rboh) (52, 53), as well as components involved in tissue healing and regeneration, including Act (54). Furthermore, four types of bryohealin (BPL) in Bryopsis and caulerpenyne in Caulerpa are compounds integral to regeneration processes (13, 55). Our phylogenetic analysis revealed the presence of BPL-3 and BPL-4 in H. opuntia and C. lentillifera but total absence of BPL in O. quekettii (SI Appendix, Fig. S8), indicating divergent evolutionary paths of the wound-healing BPL genes among the unicellular coenocytic lineages. Besides, concanavalin A-based dressings have proven efficacy in wound healing applications (56), and we detected expansion of concanavalin A-like lectin/glucanase (Calp) genes in H. opuntia and C. lentillifera (SI Appendix, Fig. S9). In addition, we found remarkable expansion in the siphonous algae of glycine-rich protein (Grp) genes, which are typically arranged in (Gly)_n-X repetitions and play various roles in plant wound healing. Based on glycine repeat arrangements and conserved motifs, plant Grps in cell wall and other tissues have been classified into five classes (57, 58). We found Grps in cyanobacteria, green algae, and land plants in the genome-derived protein databases. Our phylogenetic analysis revealed that plant Grps originated from cyanobacteria and formed polyphyletic clades, with chlorophyte Grps being the closest to plant Grp2 (SI Appendix, Fig. S10). Grp2 in plants is known to function in hormone signaling, stress acclimation, wound-healing, floral development, and cell growth regulation (59). The notable expansions of Grp2 in H. opuntia and other unicellular coenocytic algae highlight their potential roles in wound healing and regenerative growth.

Polyploid Speciation of H. opuntia.

The complexity of H. opuntia nuclear architecture is not limited to multinucleation. H. opuntia genome appears to be allopolyploid, as evidenced by the presence of haploid, diploid, triploid, and tetraploid peaks in the k-mer analysis (SI Appendix, Fig. S1). Smudgeplot analysis showed potential dominance of the tetraploid AAAB structure (Fig. 2A). Through phylogenetic analysis of gene clusters in each collinear block in H. opuntia and C. lentillifera (SI Appendix, Fig. S11), we identified four subgenomes (S1 to S4) that emerged from 1,127 out of the 1,407 genome contigs, covering approximately 90% of the genome assembly (SI Appendix, Table S7). Additionally, over 98% of the syntenic groups from the four subgenomes are located on different contigs.

Fig. 2.

Genome characteristics and allopolyploidy features in H. opuntia. (A) Smudgeplot shows the dominance of the AAAB genome structure. (B) Distribution of Ks values for gene paralogs and orthologs in H. opuntia and C. lentillifera. (C) Boxplot of the Ka/Ks values of orthologs in the four (S1 to S4) H. opuntia subgenomes; *P < 0.05 and **P < 0.01 according to Wilcoxon test. Four homologous gene sets were identified and their ancestral sequences were reconstructed using the CodeML program in PAML software. Homologous genes of each subgenome were compared with their ancestral sequences to calculate the Ka/Ks values. (D) Subgenome bias of genome analysis parameters. The axes represent the subgenome bias indices (SBI) that are defined as follows: Subgenome bias indices = absolute(max-min)/(max+min), in which max and min are the maximum and minimum objects [e.g., number of gene, single-nucleotide polymorphisms (SNPs), heterozygousity rate, mean 4DTv (fourfold-degenerate transversion substitution rate)] among four subgenomes. The solid and dashed rectangles indicate the maximum and minimum values of SBI, respectively. LD_1200 and LD_2400 represent the midday and midnight transcriptome samples. (E) Subgenome expression dominance and enriched GO functional categories. Testable homoeolog pairs (n) are those that could confidently be identified as homoeologous on the basis of synteny and assigned to a subgenome with phylogenetic support (>80% bootstrap) and that had at least one read in each transcriptome dataset. Homoeolog pairs significantly biased toward subgenome S1 homoeolog are shown in blue, and pairs significantly biased toward “other” homoeolog from one of the other three subgenomes are shown in orange. GO enrichment of the homoeologs biased to S1 (fold change >2) is performed.

Eukaryotic multinuclearity can arise either from fusion events between uninucleate cells or from uninucleate cells that replicate their DNA contents without cytokinesis. Fusion of vegetative cells or hyphae in some fungi results in multinucleate cells with potentially genotypically different nuclei (60). Unlike these fungi, Halimeda reproduces sexually and vegetatively through fragmentation and regeneration (15, 61), suggesting that each thallus originates from a single nucleus that has undergone numerous divisions. Further analysis showed nearly identical sequencing depths for the four subgenomes. Thus, there is no evidence of genotypically different nuclei in H. opuntia and with current data, it is more likely that the four subgenomes are within the same nucleus rather than separate nuclei. Further analyses such as in situ hybridization can help verify this proposition in the future.

To determine whether whole-genome duplication (WGD) occurred, potentially contributing to the emergence of subgenomes, we analyzed Ks distributions of 7,770 paralogous genes in H. opuntia, and 1,256 paralogous genes in C. lentillifera, as well as 9,370 orthologous genes between H. opuntia and C. lentillifera. No evidence of ancient WGDs was found, but one recent polyploidization event was identified in H. opuntia (Fig. 2B). The median nonsynonymous to synonymous substitution (Ka/Ks) ratios for gene orthologs in subgenome S1 (0.233) were significantly lower than those in subgenomes S2 (0.279) and S3 (0.269) (P < 0.01 and P < 0.05, Fig. 2C); moreover, S1 retained more genes and exhibited higher transcription levels under both light and dark conditions (Fig. 2D). Combined with other subgenome bias indices of heterozygosity rate, mean 4DTv, and ratio of nonsynonymous vs. synonymous single-nucleotide polymorphisms (Fig. 2D and SI Appendix, Table S8), these data suggest that the four H. opuntia subgenomes evolved asymmetrically, with S1 being dominant. This result is consistent with the view that a dominant subgenome retaining more genes and experiencing stronger purifying selection might emerge after polyploidization to resolve genetic conflicts in organizing distinct subgenomes in a single nucleus (62). Alternatively, current evidence suggests that the observed subgenome biases may primarily reflect legacy from the progenitors (63). Further analysis revealed that the dominant subgenome S1 encoded more highly expressed homoeologs than the other three subgenomes (Fig. 2E), suggesting that this subgenome indeed plays a dominant role in the species. The homoeologs biased to S1 were enriched for the GO terms ion transport, organic acid biosynthesis, nucleobase-containing compound biosynthesis, and endopeptidase activity (Fig. 2E and Dataset S3). Of these, the biased genes encoding Ca²⁺ transport indicate a selection pressure to reinforce calcification phenotype.

In coenocytic organisms, the presence of multiple nuclei within a single cell poses challenges regarding the coordination and regulation of genetic information. Nuclear heterogeneity apparently emerges as an evolutionary solution, as this allows differentiation of expression and function (64). Notably, siphonous green algae, such as Halimeda and Caulerpa, have evolved specialized structures like anchoring holdfasts, stolons, stipes, and leaf-like fronds. In Caulerpa, distinct functions of nuclear populations in those compartments have been reported, with transcripts related to nuclear functions such as translation, DNA replication, and chromatin enrichment being located in stolons (65). This supports the concept of nuclear subfunctionalization in many coenocytes, as evidenced by the presence of regularly spaced “nucleo-cytoplasmic domains”, where neighboring nuclei repel each other, creating space in between and demarcating their own cytoplasmic territories (66). Furthermore, the cytoplasm in siphonous algae display vigorous streaming, facilitating the transportation of nuclei, RNA transcripts, and nutrients throughout the thallus. In H. opuntia, nuclei localized at different thallus regions may express distinct genes from the subgenomes, thereby performing specialized functions. Intracellular trafficking is crucial for the movement of nuclei and chloroplasts, and accordingly, motor proteins such as dynein and kinesin genes are abundant in Bryopsis sp. (13) and H. opuntia (Fig. 1C).

We further generated a large number of gene trees to infer the topology of H. opuntia subgenomes. We aligned the subgenome sequences to de novo assembled transcripts of H. borneensis and identified the blocks shared between the two intrageneric species with length ≥300 bp for use in phylogenetic tree construction. In total, 45% of the 10,920 exon blocks support the topology [outgroup, (S1, (S2, (S3, S4)))] (SI Appendix, Fig. S12A). Based on coalescent multiple sequence alignments and time-scaled phylogenetic trees for the H. opuntia and C. lentillifera genomes, S1 diverged from the ancestral genome approximately 7.11 Mya, S2 diverged next approximately 6.51 Mya, S3 branched from S4 approximately 4.01 Mya, and polyploidization occurred approximately 3.4 Mya (Fig. 3). Divergence of the four subgenomes predated polyploidization, implicating hybridization as the mechanism that led to polyploidy. All the existing data together enabled us to infer a chronicle to account for the emergence of the octoploid H. opuntia through hybridization (SI Appendix, Fig. S12B). If the hybridization involved two tetraploid species, a dominant AABB structure, instead of AAAB that we found, would be expected of Smudgeplot analysis. The tetraploid and hexaploid states may be the evolutionary intermediates, and several hybridization events must have occurred to form the octoploid H. opuntia. These findings expand upon previous research, indicating that Halimeda underwent at least two separate WGD events, potentially facilitated by hybrid speciation (67–69). Both phylogenetic analysis and geographical distribution patterns indicate that nearly all Halimeda species are found in either the Indo-Pacific or Atlantic basins, implying notable vicariance events (70). Niche models further suggest that their limited geographical ranges result from dispersal limitation, with divergence between the two lineages occurring during the Paleogene (65 to 25 Mya) (71). The timing of several hybridization events inferred from H. opuntia genome analysis postdated its dispersal across the global oceans, indicating possible geographic variations in genome structure and underscoring the need for further genome assemblies of Halimeda species.

Fig. 3.

Emergence of major allopolyploid genomic features in H. opuntia at the backdrop of Earth history events. (A) Phylogenetic and syntenic analyses of the four (S1 to S4) subgenomes. N1 to N3 represent the divergence time for each node. (B) Sea level (72), atmospheric CO₂ (73), and biomass of calcareous algae and corals from a biogenic reef of NK1 (18) over the past 10 Mya. T1/T3 and T2/T4 indicate two periods of high biomass production for calcareous algae and corals, respectively. The color gradient from light to dark corresponds to biomass increase.

Accumulating evidence indicates that WGDs often occur under stress such as during prolonged and/or extreme climatic instability, and the increased genotypic diversity resulting from WGDs confers advantages for niche competition and stress adaptation (74). In H. opuntia, the estimated timing of polyploidization coincided with that of the contemporaneous sea level drop (72) and atmospheric CO₂ increase (73) (Fig. 3B). Lowering of the sea level approximately 3.4 Mya may have caused the immobile H. opuntia to experience stress from high-intensity light and ultraviolet exposure and living space compression (75). Moreover, decreased seawater pH due to absorption of increased atmospheric CO₂ might affect calcification through dissolution of CaCO₃ crystals. To counterbalance the environmental instability, subgenome divergence and polyploidization may have been favored, contributing to the expansion of the H. opuntia population. Coincidentally, investigation of the biogenic reef column of core NK-1 in the South China Sea points to calcareous algae accumulation approximately 6.9 to 6.0 Mya and 3.6 to 1.5 Mya (17), i.e., soon after subgenome divergence and polyploidization in H. opuntia (Fig. 3B). Corals represent other important biogenic reef builders in NK-1; however, the timing of their intense growth differs from that of calcareous algae, which may be attributed to distinct responses to external stresses and competition for space (76).

Expanded Capacity for Calcification.

H. opuntia calcifies extracellularly in the IUS, while most other calcifying algae (e.g., coccolithophores, coralline rhodophytes) deposit CaCO₃ intracellularly. Some algae (e.g., Chara) and nonphotosynthetic calcifying organisms (e.g., Foraminifera) may deposit CaCO₃ extracellularly, but none uses structures like IUS. To find the genetic basis of distinct calcification modes, we also sequenced the red alga A. fragilissima, which lives in the same coral reef as H. opuntia but deposits high-Mg calcite crystals within its cell wall. Belonging to two major branches within Archaeplastida, these two algae have diverged over a long history. As such, large genomic differences are anticipated, making general comparative genomics analysis not particularly useful. Nevertheless, we think a focused comparative analysis in the context of calcification can still provide valuable insights. The combination of Illumina paired-end short reads and PacBio long reads led to a 146.42 Mb genome assembly from which 15,287 protein-coding genes were predicted, with 79.9% (12,212 genes) being functionally annotated, and TEs were shown to account for 38.1% of the genome (SI Appendix, Tables S9 and S10). Relative to the last common ancestor, the speciation of A. fragilissima involved expansion of 2,415 gene families and contraction of 3,174 gene families (Fig. 1A and Datasets S4 and S5). We then focused on the expanded gene families and positively selected genes (PSGs) that may play a role in Ω_CaCO3 regulation and crystal nucleation to interpret the calcification modes of H. opuntia and A. fragilissima (Fig. 4 A and B and Datasets S6–S8).

Fig. 4.

Two distinct calcification modes and related gene families proposed for the two calcareous coralline algae. Expanded gene families and PSGs involved in calcification in H. opuntia (A) and A. fragilissima (B). CSC, cellulose synthase complex; ECM, extracellular matrix; IUS, interutricular space; PCW, primary cell wall; SCW, secondary cell wall. (C) Copy numbers of N-acetylglucosaminyltransferase (Ogt), Ca²⁺-ATPase, and carbonic anhydrase (CA) genes in the representative algae. (D) Synteny analysis of Ca²⁺-ATPase genes in the four H. opuntia subgenomes. Gray shading in the background indicates collinear blocks between the four subgenomes and red lines indicate Ca²⁺-ATPase gene pairs in the corresponding collinear blocks. (E) Maximum-likelihood phylogenetic tree of CA genes in the representative algae.

Calcifying ion supersaturation and local alkalinity are crucial for calcification in corals, foraminifera, and calcareous algae, where elevated pH and higher Ca²⁺ and HCO₃⁻ concentrations prevail at calcification sites (77–79). Both algae-expanded calmodulin (Cam) and Ca²⁺/calmodulin-dependent protein kinases (Camk), which influence Ca²⁺ exchange by regulating intracellular Ca²⁺ homeostasis, and potassium-dependent sodium calcium exchanger (Nckx), which transports Ca²⁺ out of the cell across the plasma membrane (80). In comparison to cell wall calcification in A. fragilissima, the semienclosed IUS in H. opuntia requires greater investment to maintain a favorable environment. H. opuntia exclusively expanded sodium/hydrogen exchanger (Nhx), V-ATPase, and bicarbonate transporter (81) (Slc4, SI Appendix, Fig. S13), which potentially serve to elevate seawater pH and HCO₃⁻ concentration in the IUS through removing H⁺ and accumulating HCO₃⁻. The remarkable expansion of Ca²⁺-ATPase genes in H. opuntia, potentially transporting two intracellular Ca²⁺ into the IUS in exchange for four H⁺, is an evolutionary innovation (Fig. 4C). There are two subtypes of Ca²⁺-ATPases, the plasma membrane type (Pmca) and the sarco/endoplasmic reticulum membrane type (Serca). Pmca protein is known to serve as an exporter of Ca²⁺ (80). H. opuntia contains 10 Pmca and 8 Serca, whereas A. fragilissima contains only three Serca and no Pmca (SI Appendix, Fig. S14), a clear discriminating difference between two calcifying pathways in the two coralline algae. While genes duplicated through polyploidization are normally lost to resolve genetic conflicts caused by redundancy (82), synteny analysis showed almost equal distributions of Ca²⁺-ATPases and other calcification-related genes among the four H. opuntia subgenomes (Fig. 4D and SI Appendix, Table S11). The retention of duplicated genes such as Pmca suggests selection pressure imposed by extracellular calcification. In addition, gene families with the highest 20% Ka/Ks in H. opuntia were mainly enriched for Ca²⁺ and other ion transport (SI Appendix, Fig. S15). This provides evidence of selection for the diversification of gene families related to calcifying ion redistribution, which may provide adequate gene mutations or adaptability to sustain the calcification phenotype.

Organic components, including cell wall polysaccharides and proteins and pectin-like compounds, facilitate crystal nucleation by reducing the activation energy for critical nucleus formation (19, 20, 83). Electron micrographs indicate that crystals align closely with cell wall polysaccharide fibrils, integrating into organic matrices and creating a distinct boundary between the cell interior and the cell wall (84). Even in Halimeda, crystals adopt a perpendicular orientation adjacent to the cell wall rather than a random distribution within the IUS (16). Collagen type I (Col1a) and fibronectin type III (Fn3) provide skeletal organic matrices in stony corals for calcification (85), and their homologs exist and were duplicated in H. opuntia (Fig. 4A). Subcellular localization models predict that Fn3 in H. opuntia are transmembrane proteins (SI Appendix, Table S12), consistent with its extracellular site of calcification. In contrast, neither Col1a nor Fn3 were found in A. fragilissima, consistent with this species’ calcification in the secondary cell wall, where its cellular organic components may serve as nucleation matrices. Furthermore, N-acetylglucosaminyltransferase (Ogt) catalyzes O-GlcNAcylation, modifying cell wall glycoproteins in algae and higher plants (86, 87). The enrichment of short interspersed retrotransposon elements-Alu (SINE-Alu) around Ogt (5 kb up and down) in A. fragilissima (SI Appendix, Fig. S16) supports the hypothesis that Ogt expansion may be correlated with the SINE-Alu burst. The expansion of gene families of Ogt, glycoside hydrolase (GH16, GH19, GH43, and GH47), and glycosyltransferase (GT1, GT2, GT23, and GT34) (Fig. 4 B and C and Dataset S8) suggest a potential increase in polysaccharide and glycoprotein biosynthesis together with cell wall remodeling capabilities to support calcification. In particular, we found a remarkable expansion of β-mannan synthase gene family (four copies) (SI Appendix, Fig. S17 and Datasets S4 and S9). This is of interest because mannan provides a strong foundation for calcification in coralline red algae (88).

Coupling of Calcification and Photosynthesis and Respiration.

In both biologically induced or organic matrix-mediated calcification mechanisms, photosynthesis and organic matrix influence nucleation (20). Photosynthesis is the main driving force of calcification because it consumes CO₂ elevating local seawater pH and provides energy needed for calcification (89). Calcification and photosynthesis can be coupled through the function of carbonic anhydrases (CA). CAs catalyze the reversible conversion of CO₂ and water to HCO₃⁻ and H⁺ ions, serving as the carbon-concentrating mechanism (CCM) for photosynthesis in algae (90), as well as catalyzing calcification in calcifying organisms resulting in CO₂ release (23, 91). Theoretically, calcification enhances photosynthesis by providing needed CO₂ (hence reducing demand for the energy-consuming CCM) and in turn benefits from photosynthesis for ATP supply (92). The CA gene family was largely expanded in both investigated species (Fig. 4 C and E). There are five classes of CA, with different subcellular localizations. H. opuntia contained 11 γCA (cytoplasmic, chloroplast, and mitochondrial), 4 αCA (cytoplasmic), and 2 βCA (plasma membrane) (Datasets S10 and S11). Compared to the other three siphonous algae, the significant expansion of αCA and γCA genes in H. opuntia potentially indicates strengthened capacities in photosynthesis and/or calcification.

The photosynthesis-calcification coupling in H. opuntia may also be strengthened by the expansion of the C₄ pathway, particularly phosphoenolpyruvate carboxylase (Pepc), malate dehydrogenase (Mdh1), and pyruvate, orthophosphate dikinase (Ppdk) (Fig. 4A). The high number of cytoplasmic γCA can abundantly convert CO₂ conversion to HCO₃⁻ to facilitate CO₂ diffusion from the IUS on the one hand and feed the C₄ pathway on the other, which could elevate pH in IUS and promote carbon fixation inside the cell. Consistent with the enhancement of photosynthetic capacity, the Lhca and Lhcb gene families encoding light-harvesting proteins and Psb and Psa gene families involved in light-induced water oxidation and electron transfer across the thylakoid membrane are expanded in H. opuntia (SI Appendix, Fig. S18 and Dataset S12). The PSGs of cytochrome b6-f complex iron–sulfur subunit (petC) and chlorophyll synthase (Chlg) contribute to ATP production and chlorophyll biosynthesis, which may potentially enhance photosynthesis (Dataset S13).

To further evaluate the differences in calcification and other relevant processes between the two algae, we performed comparative transcriptomics of H. opuntia and A. fragilissima under varying light/dark conditions. In both species, genes involved in translation, carbohydrate, and energy metabolism were enriched during daytime, and those participated in the metabolism of amino acids, terpenoids, and polyketides were enriched under constant light exposure (SI Appendix, Fig. S19). Genes involved in energy-generating processes (photosynthesis, oxidative phosphorylation, glycolysis) and calcification in H. opuntia exhibited similar expression patterns during the light/dark cycle (Fig. 5 A, Upper). In contrast, inconsistent expression patterns of genes involved in the four processes were observed in A. fragilissima (Fig. 5 A, Lower). Analysis of a range of reef calcifiers indicates that calcification rates are linked to energy production at the organismal level (93). It is likely therefore that H. opuntia, with its semienclosed IUS, may require more energy to continuously maintain CaCO₃ oversaturation. Such a situation may drive H. opuntia to rely heavily on photosynthesis and respiration for sustained calcification, consistent with the expanded gene families and gene expression patterns observed during the light/dark cycle. These genomic and transcriptomic results further support the notion that calcification in H. opuntia is more likely a biologically induced process (20).

Fig. 5.

Dynamics of photosynthesis and calcification genes in the two calcareous algae. (A) Expression patterns of genes involved in the indicated biological processes in H. opuntia (Upper) and A. fragilissima (Lower) during the light/dark cycle. Black shading in the x-axis indicates nighttime. (B) Distributions of differentially expressed proteins (DEPs) involved in cell growth and calcification in two algae depending on CO₂ treatments. The number of DEPs is labeled on the radar graph. (C) Raw scaled heatmap of calcification-related DEPs in two algae.

Proteomic Features Facilitating Plasticity and Adaptation of H. opuntia to CO₂ Fluctuation.

Understanding the molecular mechanism underlying the responses of calcified macroalgae to OA due to high CO₂ level is important to predict the fate of biogenic reefs under global climate change. We analyzed protein profiles for H. opuntia (5,788 proteins identified) exposed to high, medium, and control CO₂ concentrations (1,600, 1,000, and 400 ppm, respectively) for 6 wk. For comparison, we performed the same treatments on A. fragilissima and identified 3,111 proteins. Strikingly, only 7.8% and 10.2% differentially expressed proteins (DEPs) were observed at medium and high CO₂ concentrations compared with those in the control (SI Appendix, Fig. S20A and Dataset S14). These DEPs are primarily related to protein biosynthesis (ribosomes and spliceosomes) and energy generation (glycolysis and oxidative phosphorylation) (Fig. 5B). In contrast, A. fragilissima demonstrated a strong response to OA: DEPs accounted for 35.2% and 52.2% of the total proteins identified at medium and high CO₂ levels, respectively (SI Appendix, Fig. S20B and Dataset S15). In A. fragilissima, the influence on calcification and other vital processes seemed to increase with CO₂ concentration and was accompanied by an increase in the number of down-regulated DEPs (Fig. 5B). These results are consistent with OA having a significant impact on coralline algae containing high-Mg calcite in their skeleton (94). The lack of significant proteomic changes in H. opuntia under different CO₂ conditions is consistent with previous studies showing that OA had no effects on calcification in six Halimeda species, which can grow even at pH <7 (95). The strong resistance of Halimeda species to OA may in part be attributable to the strengthened calcification–photosynthesis coupling discussed above, as well as other features associated with polyploidization and gene family expansion and proteomic modulation, as discussed below.

Despite the small general proteomic shift, we found significant upregulated expression in elevated CO₂-exposed H. opuntia cultures of proteins V-ATPase, Nhx8, and Camk, which are known to transport H⁺ and Ca²⁺ across plasma membrane for calcifying ion remodeling, as well as Col1a, which may serve as a template for mineral nucleation (Fig. 5C). These patterns are consistent with their expansion in the genome described above and with previous proteomic studies on calcifying organisms, which show OA-induced V-ATPase up-regulation in E. huxleyi and Col1a levels in stony coral skeletons (96, 97). In A. fragilissima, based on proteomic analyses, proteins related to the biosynthesis of key organic matrices (e.g., sterol 3-beta-glucosyltransferase (ugt), Ogt, GH16, GH11, GH43, and GT2) and ion-redistribution proteins (NHX8, Cam, Camk, and CA) were significantly down-regulated at elevated CO₂ levels (Fig. 5C), suggesting the potential of substantial impairment to the calcification process.

Marine and freshwater microorganisms extensively adjust the distributions of their proteomic isoelectric point (pI) profiles for survival (98, 99). To investigate the distinct mechanisms underlying the differential responses to CO₂ increases in the two calcareous algae, we examined amino acid composition of the DEPs expressed under different CO₂ levels for their pI profiles. The results showed that under elevated CO₂ treatments, H. opuntia upregulated proteins with higher pI values and downregulated proteins with lower pI values (SI Appendix, Table S13). In contrast, A. fragilissima under the same elevated CO₂ treatments upregulated proteins with lower pI values and downregulated proteins with higher pI values (SI Appendix, Table S14). The contrasting proteomic pI profile shifts between the two algae can explain why H. opuntia is more OA-resistant. The upregulated high pI proteins can provide a neutralizing effect to counter increases in positive charge caused by lower pH.

Concluding Remarks

Using an in-depth functional genomic analysis of the giant unicellular multinucleate chlorophyte H. opuntia, we find initial evidence that the emergence of multinucleation is associated with the loss of myosin VIII and strengthening of the mitosis machinery. This alga has attained the ability to quickly regenerate cell fragments into whole cells likely due to the expansion of wound healing–related gene families. Furthermore, a comparative genomic analysis reveals distinct calcification mechanisms in H. opuntia in contrast to that in the calcareous coralline rhodophyte A. fragilissima. H. opuntia shows strong coupling of calcification with photosynthesis through CO₂ flow and energy generation. Finally, our data link H. opuntia’s OA-tolerance to an ability of tuning proteomic pI values to counter ambient environmental acidity, potentially a mechanism to maintain optimal biochemical activities under elevated CO₂ conditions. This study dissects the genome of a unicellular multinucleate macroscopic alga and unfolds the genetic basis of its unique and ecologically important traits, which is crucial for further understanding the evolution of multicellularity, regeneration, and resilience in calcareous algae. The differential sensitivities of different calcareous algae to elevated CO₂ found here suggest a potential shift of algal community structure in future more acidic oceans, exacerbating the already frequent replacement of degraded coral communities by calcareous algae due to eutrophication (100, 101). In addition, our findings of cell regeneration–related gene families may have implications in medical treatment and plant grafting.

Materials and Methods

Thalli of H. opuntia were collected from a South China Sea coral reef at depths exceeding 15 m (9°36′ N, 115°28′E) in 2018, while thalli of A. fragilissima were collected from a coastal rocky reef in Sanya Bay, China (18°12′ N, 109°28′ E) in the same year. Subsequent to seawater incubation and the removal of living and attached eukaryotes, each macroalga underwent laboratory cultivation at 28 °C under a light intensity of 150 µmol m⁻² s⁻¹ in a 15 L tank with a continuous circulation system of autoclaved natural seawater. One thallus of each species after antibiotic treatment was collected for genome sequencing. See SI Appendix for the details of sequencing, assembly, annotation, evolutionary, polyploid, transcriptomic, and proteomic analysis.

Supplementary Material

Appendix 01 (PDF)

Dataset S01 (XLSX)

Data, Materials, and Software Availability

Whole-genome assemblies of the green alga Halimeda opuntia and the red alga Amphiroa fragilissima have been deposited at NCBI under the accession numbers of PRJNA811316 (102) and PRJNA811381 (103). The raw reads of the RNA-seq have been deposited at NCBI under the accession number PRJNA812619 (104) and PRJNA813324 (105). The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository with the dataset identifier PXD033672 (106) and PXD033648 (107). All data needed to evaluate the conclusions in the paper are present in the paper and/or supporting information.