Recent allopolyploidization and transcriptomic asymmetry in the mangrove shrub Acanthus tetraploideus

Abstract

Background

Mangrove species are vital to the ecosystems of tropical and subtropical coastlines worldwide. Despite the underexplored role of polyploidization in these species, deciphering its impact on gene expression is essential for understanding the connection between polyploidization and species diversification. Our initial investigation, integrating multiple nuclear loci with morphological and cytological data, indicates that the tetraploid Acanthus tetraploideus likely originated from allopolyploidization events involving the diploid species A. ilicifolius and A. ebracteatus. Expanding on these insights, this study utilizes genome-wide evidence to confirm the divergence patterns among extant Acanthus mangrove diploids and to investigate the origin and transcriptome asymmetry of the tetraploid A. tetraploideus.

Results

Phylogenetic analysis and molecular dating revealed a closer evolutionary relationship between A. ebracteatus and A. volubilis than between A. ebracteatus and A. ilicifolius, diverged approximately 6.92 Mya and 9.59 Mya, respectively. Analysis of individual whole transcriptomes revealed that homeologous sequences in A. tetraploideus were preferentially clustered with A. ilicifolius and A. ebracteatus, rather than A. volubilis, in a roughly 1:1 ratio. The high similarity in nucleotide sequences and homologous polymorphisms between the tetraploid A. tetraploideus and its two parental diploids, A. ebracteatus and A. ilicifolius, supports the hypothesis of a recent allopolyploid origin for A. tetraploideus. Estimation of homeolog expression revealed a general attenuation of homeolog expression divergence in A. tetraploideus compared to the in silico parental mix, with 22.87% and 67.66% of genes exhibiting biased homeolog expression, respectively. Further investigation identified remarkable retention of parental expression dominance in the tetraploid, suggesting that parental genetic legacy substantially influences the reconfiguration of homeolog expression in the derived tetraploid. Meanwhile, the observation of numerous novel expression patterns between the two homeolog sets suggests that the transcriptome shock (i.e., the transcriptomic changes induced by interspecific hybridization) associated with allopolyploidization and subsequent post-polyploid evolutionary processes also significantly impact transcriptome asymmetry in A. tetraploideus. While no strong evidence directly links transcriptomic changes to specific adaptive traits, the patterns in unbiased and novelly biased genes in A. tetraploideus suggest adaptations to stable polyploidy. Unbiased genes involved in fundamental cellular processes and novelly biased genes related to chromosome dynamics and cell cycle regulation may stabilize polyploid genomes, supporting the species’ establishment and long-term success. These findings underscore the role of transcriptomic stability in polyploid adaptation.

Conclusions

Our study sheds light on the evolutionary origins and the intricate transcriptional reconfiguration of the tetraploid A. tetraploideus. These insights significantly enhance our comprehension of the pivotal role that polyploidization plays in speciation and adaptative evolution of mangrove species.

Background

Polyploidization is acknowledged as a pivotal force in the evolution and diversification of plants [1,2,3,4]. Throughout history, the majority of angiosperms have experienced at least one polyploidization event, with more than 15% of existing angiosperm species being of polyploid origins [5,6,7,8]. Polyploids originate through two primary mechanisms: allopolyploidy, which involves hybridization between different species, and autopolyploidy, which results from genome duplication within a single species [1, 9, 10]. Studies indicate that polyploids, especially allopolyploids, may possess enhanced environmental tolerance and adaptability to environmental shifts due to the presence of two or more divergent genomes that provide a robust foundation for genetic variation and manipulation [11]. Specifically, studies on polyploid gene expression have shown that rapid changes in gene expression are a prevalent feature in recently formed allopolyploids [12,13,14,15,16]. Additionally, it has been observed that homeologs—duplicated genes derived from parental species—often exhibit unequal expression patterns, known as homeolog expression bias, within natural allopolyploids [17,18,19]. This biased expression can occur either locally or across the entire genome, with the direction and magnitude varying among different taxa, organs, developmental stages, and environmental conditions [16, 17, 20]. However, the bulk of our understanding of the origin and evolutionary trajectories of polyploidy has been derived from research on the Asteraceae [21, 22], Brassicaceae [23, 24], Phrymaceae [25], Poaceae [26, 27], and a range of other crop species [19, 20, 28,29,30]. To enhance our knowledge of the breadth and impact of transcriptomic alterations associated with polyploidization, further investigation into additional plant systems, particularly those that are ecologically significant, is essential.

Mangroves species, a collection of pantropical plants that inhabit the intertidal zone, are highly adapted to saline and hypoxic environments characterized by high levels of ultra-violet radiation. They provide important biological and ecological services within coastal ecosystems [31, 32]. The genus Acanthus, belonging to the Acanthaceae family, is distinguished by its ecological range, which uniquely encompasses both the true mangroves that are typically situated along the seaward edges of coastlines and a significant number of terrestrial species. Three species, namely A. ebracteatus Vahl., A. ilicifolius L. and A. volubilis Wall. are recognized as true mangroves and are distributed in the Indo-West Pacific region (IWP, hereafter), as reported in previous research [32,33,34]. Our recent work has identified a tetraploid mangrove within the genus Acanthus, specifically A. tetraploideus H. Feng & Y. L. Huang [35], based on comprehensive morphological, cytological and molecular analyses. Ecologically, A. tetraploideus occupies nearly the entire range of A. ebracteatus and A. ilicifolius and expands further, particularly toward higher latitudes, indicating superior ecological adaptability. However, the specific adaptive traits and mechanisms underlying this advantage remain unclear due to minimal morphological differences between these species and the lack of related research. While molecular data from eight nuclear loci suggest an allopolyploid origin for A. tetraploideus, derived from the diploid species A. ebracteatus and A. ilicifolius, a genome-wide comparison of nucleotide sequences is essential for further validating the origin of A. tetraploideus and for evaluating the divergence between this tetraploid and its putative diploid progenitors. Furthermore, although the role of allopolyploidization (interspecific hybridization accompanied by or followed by chromosome doubling) in plant speciation and diversification is increasingly recognized [30], it’s influence on mangrove evolution remains largely unexplored. To date, no other extant polyploid species has been reported among true mangroves. Therefore, the discovery of the tetraploid A. tetraploideus offers an unprecedented opportunity to study genomic and transcriptomic changes associated with polyploidy in mangroves, which will lay the foundation for further elucidating the role of polyploidization in the innovation and evolution of this unique group of plants.

This study was formulated to address several critical questions: (1) Is there genome-wide evidence supporting the origin of tetraploid A. tetraploideus from allopolyploid hybridization involving the diploid species A. ebracteatus and A. ilicifolius? (2) Does transcriptomic asymmetry exist in A. tetraploideus, and if so, can this asymmetry be attributed to parental legacy or novel homeologous expression bias? (3) What are the implications of observed homeolog expression patterns for the evolution and adaptation of A. tetraploideus? RNA-sequencing (RNA-seq) serves as a valuable tool for these investigations, enabling the detection of single nucleotide polymorphisms (SNPs) between homologs and facilitating the estimation of sequence variation and quantification of homologous gene expression across the entire transcriptome [30, 36,37,38]. We performed RNA-seq on the tetraploid A. tetraploideus and its closely related diploid mangrove species within the genus Acanthus that are co-distributed in the IWP region, including A. ebracteatus, A. ilicifolius, and A. volubilis. This analysis aimed to confirm the allopolyploid origin of A. tetraploideus at the genome level and to uncover global expression changes associated with polyploidization, as well as their potential functional implications.

Methods

Sample preparation and transcriptome sequencing

Samples of Acanthus tetraploideus and A. ebracteatus were collected from Wenchang, Hainan, China (N 19°37’12’’ E 116°46’48’’ and N 19°37’48’’ E 110°48’36’’, respectively), while A. ilicifolius samples were collected from Sundarbans, India (N 21°51’0’’ E 89°46’12’’) as detailed in Additional file 1. All collections were conducted in accordance with local regulations, and no special permits were required as none of the sampled species are protected under conservation legislation in the respective countries. The voucher specimen of A. tetraploideus (FH3021), A. ebracteatus (FH3123) and A. ilicifolius (Akb0543) were deposited in the herbarium of SunYat-sen University (SYS). The taxonomic identification of the plant materials was undertaken by Dr. Ying Liu in Sun Yat-sen University.

The samples were selected based on our prior researches on the population genetic diversity of Acanthus species [35, 39], which demonstrated low genetic diversity at the species level across the three Acanthus species. Both A. tetraploideus and A. ilicifolius exhibited a single-lineage structure within each species, while A. ebracteatus comprised two intraspecific lineages (i.e., Ae-1 and Ae-2). Only the Ae-1 lineage contributed to the formation of A. tetraploideus. Therefore, samples of A. ebracteatus samples from Hainan belonging to the Ae-1 lineage were selected, while geographically accessible samples were chosen as representatives for A. tetraploideus and A. ilicifolius (Additional file 1). The selected samples of A. tetraploideus and its two putative parental species are highly representative for addressing the core questions of this study. The distribution areas and sampling sites of four Acanthus species are illustrated in Fig. 1 [33, 35, 39].

Fig. 1

Distribution areas and sampling sites of four Acanthus species. Occurrence data for the distribution map were derived from iNaturalist observations (accessed at https://www.inaturalist.org in January 2025) and previous studies [33, 35, 39]. Sampling sites are marked with stars

Each individual plant was evaluated for ploidy level using flow cytometry prior to any downstream analyses [35]. Seedlings from the three species were cultivated in the greenhouse (23 °C ~ 35 °C, 13.5-h day-length, a light intensity sufficient to allow normal growth) of Sun Yat-sen University, Guangzhou. The seedlings were transferred to sandy soil and grown using 1/2 Hoagland’s solution for 5 weeks to help plants adapt to the environment. Fresh leaf samples were harvested from three seedlings per species, which served as biological replicates. Immediately after collection, the samples were flash-frozen in liquid nitrogen and stored at -80 °C for subsequent RNA extraction. Total RNA was extracted from frozen leaf tissues using a modified CTAB method [40] and the integrity of the extracted RNA was confirmed using an Agilent 2100 Bio-analyzer (Agilent Technologies). Transcriptome libraries were constructed for each sample and sequenced on the Illumina HiSeq platform. Raw sequence data were processed to remove adapter sequences and filter for quality using Trimmomatic [41], yielding 34.58–56.82 million high-quality reads per sample, as summarized in Additional file 1. Voucher specimens from each population were deposited in the Herbarium of Sun Yat-sen University (SYSU).

Phylogeny reconstruction and divergence time estimation for diploid mangrove species in Acanthus

For the phylogenetic analysis and molecular dating of diploid mangroves within Acanthus, the sample collection was extended to include not only A.ebracteatus and A. ilicifolius but also individuals of A. volubilis, A. montanus, and A. mollis (Additional file 2). Each of these samples was subjected to identical collection and RNA sequencing procedures. RNA-seq data for A. leucostachyus, Avicennia marina and Av. officinalis were downloaded from the NCBI Sequence Read Archive (SRA) database. Following transcriptome assembly and redundancy removal using Trinity [42], open reading frames and protein sequences for each species were predicted by TransDecoder, which is integrated within the Trinity pipeline. Additionally, coding sequences and protein sequences for Sesamum indicum and Mimulus guttatus were downloaded from SinBase [43] and Phytozome [44], respectively. Single-copy orthologous genes across the 10 samples were identified using OrthoFinder [45] and subsequently used for phylogeny reconstruction with PhyML [46] under the best-fit model determined by JModeltest2 [47]. Divergence times were estimated using MCMCTree within PAML 4.9 [48]; the root constraint for M. guttatus and the common ancestor of other species was set to 68–75 million years ago (Mya), following Bremer et al. (2004) [49]. For each pair of species investigated within Acanthus and Avicennia, orthologous genes were identified based on bidirectional best hit and used for estimation of interspecific genetic divergence. The rates of synonymous (Ks) substitutions were calculated using the maximum-likelihood method implemented in the codeml module of PAML.

Phylogenetic analysis for homologous sequences among tetraploid and diploid mangrove species in Acanthus

The origin of tetraploid A. tetraploideus was investigated by examining the phylogenetic relationships among homologous sequences from A. tetraploideus and the three diploid Acanthus mangroves—A.ebracteatus, A. ilicifolius, and A. volubilis—with the diploid terrestrial species A. leucostachyus serving as an outgroup. Considering the potential loss of replicate genes and incomplete assembly of homeologs in polyploids, homologous sequences were extracted using Multiparanoid [50] across the taxa, ensuring a ratio of 1:1:1:1:1 among A. ebracteatus, A. ilicifolius, A. volubilis, A. tetraploideus, and A. leucostachyus. Gene trees were constructed for each set of homologous sequences at each locus using PhyML, which employs the maximum likelihood method. Additionally, homologous sequences from A. ebracteatus, A. ilicifolius, A. tetraploideus, and A. leucostachyus, with specific ratios of 1:1:1:1 and 1:1:2:1, were extracted and their phylogenetic relationships were analyzed using the aforementioned methods.

Detecting the polymorphisms within A. tetraploideus in relation to nucleotide differences between A. ebracteatus and A. ilicifolius

To determine the extent to which the genetic polymorphisms within A. tetraploideus are derived from differences between its putative parental diploid species, A. ebracteatus and A. ilicifolius, a comparative analysis of single nucleotide polymorphisms (SNPs) in A. tetraploideus was conducted against the nucleotide sequences of these diploid species. The quality of A. ebracteatus transcriptome assembly was slightly better than that of A. ilicifolius, with N50 values of 1,752 bp and 1,506 bp, respectively (Additional file 3). Although the nature of transcriptome reference used for mapping introduced a slight technical bias, it did not distort the key results, as confirmed in our test trial and supported by findings from other similar studies [20]. Therefore, the assembled transcriptome of A. ebracteatus was finally selected as the reference for this analysis. The clean reads from A. tetraploideus, A. ebracteatus and A. ilicifolius were mapped to the reference using BWA [51], and SNPs were identified using SAMtools [52] at sites covered by all samples. Filters were applied to ensure the reliability of the identified SNPs, excluding sites with a mapping quality of ≤ 30, coverage depth of < 10 and minor allele frequency of < 5%. The polymorphisms in A. tetraploideus that corresponded to differences between A. ebracteatus and A. ilicifolius nucleotides were confirmed as homeologous polymorphisms, termed ‘homeoSNPs’ (Additional file 4).

Detecting the nucleotide difference between the tetraploid A. tetraploideus and the diploid A. ebracteatus and A. ilicifolius

The degree of nucleotide divergence between the tetraploid A. tetraploideus and its two diploid counterparts was assessed according to the methodology described by Chen et al. (2016) [36]. Reads from A. tetraploideus were mapped to the assemblies of both A. ebracteatus and A. ilicifolius using BWA, with non-specific matches being disregarded. For the mappings of A. tetraploideus against A. ebracteatus and A. ilicifolius, SNPs were identified, and their proportion relative to the consensus length was determined for each gene. A low percentage of SNPs suggests a shallow divergence between the tetraploid and the diploids, potentially indicative of a recent tetraploid formation and/or a low nucleotide variation rate of the tetraploid homeologs since their origin. Pairwise comparisons were also performed between A. ebracteatus and A. ilicifolius within the context of the tetraploid A. tetraploideus. Orthologous genes between A. ebracteatus and A. ilicifolius were identified using INPARANOID [53], and only 1:1 orthologs with an average coverage of ≥ 10 were selected for further comparison. Log₂-transformed nucleotide difference ratios between homeologs were subsequently calculated.

Comparison of expression patterns in the tetraploid A. tetraploideus and its parental diploids

In allopolyploids, sequence variations between subgenomes, referred to as homeoSNPs, serve as a tool to differentiate subgenome-specific homeologs and determine their expression patterns [20, 30]. In this study, homeoSNPs were identified in A. tetraploideus as previously described and utilized to estimate relative homeolog expression. For each gene under investigation, the relative transcript abundance of the homeologs originating from A. ebracteatus and A. ilicifolius genomes (hereafter denoted as “E” and “I”, respectively) was determined and expressed as the percentage of the E-homeolog in the total gene expression (E%). To evaluate the homeolog expression variations in A. tetraploideus relative to ortholog expression in the two parental species, an in silico parental mix was constructed by merging the filtered reads from A. ebracteatus and A. ilicifolius at a ratio of 1:1. The E-ortholog relative expression ratio for each gene in the in silico parental mix was calculated in the same manner. The global spectra of homeolog expression were visualized using ggplot2 [54]. To ensure a reliable estimate of homeolog expression, only genes with an average depth of ≥ 10 and containing ≥ 3 homeoSNPs were considered.

To quantify parental expression differences in A. tetraploideus and the in silico parental mix, homeologs (or orthologs) were categorized as equally or differentially expressed for each sample based on the relative transcript abundance of the two homeologs (or orthologs), following the methods of Wang et al. (2016) and Xu et al. (2014) [30, 55]. Briefly, parental expression ratios for all genes were assessed using the exact binomial test, with the null hypotheses that parental homeologs (or orthologs) were equally expressed. The resulting p-values were adjusted for multiple testing using the false discovery rate (FDR) correction. Gene with an adjusted p value < 0.05 was considered to exhibit significant deviation from equal expression of E and I homeologs (or orthologs) and were classified as having biased expression. Specifically, for a given gene, if the expression of the E-homeolog (or E-ortholog) was significantly higher than that of the I-homeolog (or I-ortholog), it was designated as ‘E-biased’, conversely, if the I-homeolog (or I-ortholog) expression was significantly higher, it was designated as “I-biased”; if homeologs (or orthologs) showed equal expression, the gene was classified as “not biased”. To further explore homeolog expression patterns in the tetraploid relative to the ortholog expression patterns of the diploid parents, the genes were classified into nine possible expression patterns, grouped into four distinct categories, based on their biased or unbiased expression states in the in silico parental mix compared to the tetraploid [19, 30, 56]. Notably, the significance thresholds for all tests, including p-values and FDR adjusted p-values in the differential expression analysis, were consistently set at < 0.05 to ensure comparability across groups.

Classification of convergence and divergence regulatory groups

To investigate the regulatory mechanisms underlying changes in homeolog expression, genes in A. tetraploideus were classified into three distinct regulatory groups. This classification was based on the expression patterns of genes in A. tetraploideus relative to their expression in the in silico parental mix, following previously established criteria [30, 55]. Regulatory Group I, labeled “convergent regulation”, includes genes for which the expression differences between parental homeologs in A. tetraploideus have been significantly reduced compared to the in silico parental mix, approaching a 1:1 ratio. Regulatory Group II, termed “divergent regulation”, comprises genes for which the expression differences between parental homeologs in A. tetraploideus are significantly enhanced relative to the in silico parental mix, deviating from the 1:1 ratio. Regulatory group III, termed “conserved regulation”, encompasses genes for which the relative expression of parental homeologs in A. tetraploideus remains statistically unchanged compared to the in silico parental mix. The comparison between the tetraploid and the in silico parental mix was conducted using a X² test, followed by FDR corrections.

Assignment of cis- and trans-regulatory divergence

By comparing subgenome-specific expression in the allopolyploid A. tetraploideus with the genome-specific expression in its diploid progenitors, it is possible to dissect expression variation into cis- and trans-origin taking advantage of the slight sequence divergence between the two constitutive genomes of A. tetraploideus and those of its modern parental species, A. ebracteatus and A. ilicifolius (see Results section) [20]. Assignments of cis- and trans-regulatory divergence were conducted using well-defined criteria [55, 57, 58]. Briefly, the parental (P) and allopolyploid (H) datasets were analyzed for evidence of differential expression using the binomial exact test with FDR correction. Genes that were found to be differentially expressed in either the P or H datasets were further analyzed for trans effects (T) using X² exact tests with Yates correction, followed by FDR correction. Seven categories were derived based on the significance status in P, H, and T, and whether the relative expression between the parental homeologs of these genes in the in silico parental mix and the allopolyploid was found to be the same or opposite.

GO enrichment analysis and significance tests

The potential biological functions and Gene Ontology (GO) terms of the investigated genes were determined through a search of the AgBase-Uniprot database using GOANNA [37], with an E-value threshold of 10^− 10 and a maximum of 10 hits for each query. GO enrichment analyses for the genes selected at different steps of the analysis were conducted using Fisher’s exact test in GOBU [59], with significance (p-value) adjusted using FDR. All statistical analyses above were carried out in R version 4.3.2 [60], unless otherwise specified.

Results

Phylogeny and divergence of the three diploid mangrove species in Acanthus

A total of 671 single-copy orthologous genes were identified among the 10 species and utilized for phylogenetic reconstruction and divergence time estimation. The phylogenetic analysis disclosed that the three diploid Acanthus mangrove species formed a monomorphic clade, with A. ebracteatus exhibiting a closer relationship with A. volubilis than with A. ilicifolius (see Figure S1, Additional file 5). Bayesian estimation of divergence times, employing stringent constraints, indicated that the Acanthus mangrove clade diverged from the non-mangrove clade approximately 21.12 Mya (95% credibility interval, CI = 17.11–24.56 Mya), coinciding with the transition from the Paleogene (66-23.03 Mya) to the Neogene (23.03–2.58 Mya) (see Figure S1, Additional file 5). Within the mangrove clade, A. ilicifolius was inferred to have diverged first, 9.59 Mya (95% CI = 7.47–12.30 Mya), followed by the divergence of A. ebracteatus and A. volubilis, which occurred 6.92 Mya (95% CI = 5.12–8.81 Mya). The diversification timescale of Acanthus mangroves was comparable to that observed in another mangrove genus, Avicennia, with Av. marina and Av. officinalis diverging 7.31 mya (95% CI = 5.41–9.44 Mya). An analysis of the synonymous substitution rate, based on 10,615 ortholog pairs, revealed a median Ks of 0.0528 between A. ebracteatus and A. ilicifolius, which is larger than the Ks of 0.0397 observed between A. ebracteatus and A. volubilis (11,761 ortholog pairs), and similar to the Ks of 0.0456 between the two Avicennia species (12,710 ortholog pairs) (see Figure S1, Additional file 5).

Allopolyploid hybridization of A. tetraploideus from two diploid mangroves in Acanthus

In an exploration of the relationship between tetraploid A. tetraploideus and the three diploids in Acanthus using large-scale data, 4,303 1:1:1:1:1 homologous sequences across A. tetraploideus, A. ebracteatus, A. ilicifolius, A. volubilis and A. leucostachyus (as an outgroup) were identified through transcriptome analysis. Phylogenetic trees were constructed for each set of the homologous sequences, yielding a total of 15 distinct types of gene trees being recovered (see Figure S2, Additional file 6). Among these, the majority of gene trees indicated a preferential clustering of A. tetraploideus with A. ebracteatus (1,534 groups, 35.65%) (see Figure S2 A-C, Additional file 6) or with A. ilicifolius (1,809 groups, 42.04%) (see Figure S2 D-F, Additional file 6). Subsequently, there was a preferential clustering of A. tetraploideus with the ancestral branch of A. ebracteatus and A. volubilis (333 groups, 7.74%) (see Figure S2 G, Additional file 6). A smaller number of gene trees (220 groups, 5.11%) showed a preferential clustering of A. tetraploideus with A. volubilis (see Figure S2 H-J, Additional file 6). These results suggest that A. tetraploideus is more closely related to A. ebracteatus and A. ilicifolius than to A. volubilis.

Further investigation into the relationship between A. tetraploideus homeologs and their putative diploid parents involved the identification of 318 1:1:2:1 and 5,317 1:1:1:1 homologous sequences across A. ebracteatus, A. ilicifolius, A. tetraploideus, and A. leucostachyus. Analysis of the first data set uncovered 6 distinct topologies (Fig. 1A): in one topology (120 groups, 37.74%), both homologous gene sets of A. tetraploideus were found to cluster with A. ebracteatus and A. ilicifolius (Fig. 1A, I); in two topologies, one homologous gene set of A. tetraploideus was observed to cluster with A. ebracteatus (93 groups, 29.25%) or A. ilicifolius (58 groups, 18.24%) (Fig. 1A, II-III); and in three topologies (47 groups in total, 14.78%), the two homologous gene sets of A. tetraploideus clustered together into a monophyletic clade and/or with A. ebracteatus and A. ilicifolius into an independent clade (Fig. 1A, IV-VI). The analysis of the second data set revealed three types of topologies (Fig. 1B): in the first topology, the homologous gene set of A. tetraploideus clustered with A. ebracteatus, distinct from A. ilicifolius (Figs. 1B and I, 2 and 545 groups, 47.87%); in the second topology, the homologous gene set of A. tetraploideus clustered with A. ilicifolius, distinct from A. ebracteatus (Fig. 1B, II; 2,176 groups, 40.93%); and in the third topology, A. ebracteatus and A. ilicifolius formed a monophyletic clade, separate from A. tetraploideus (Fig. 1B, III; 596 groups, 11.21%).

Fig. 2

Phylogenetic relationships among homologous sequences of A. ebracteatus, A. ilicifolius, A. tetraploideus, and A. leucostachyus. Phylogenetic trees based on homologous sequences identified according to ratios of 1:1:2:1 (A) and 1:1:1:1 (B) were showed. The ratio of 1:1:2:1 refers to genes for which both homeologs were retained in tetraploid A. tetraploideus. Gene trees in which each homeolog preferentially clusters with the ortholog from one of its putative diploid parents are expected to be the most predominant. The number and percentage of gene groups for each tree-type were presented

To ascertain whether the tetraploid polymorphism is attributable to differences between the two putative parental diploids at the genomic level, the A. ebracteatus assembly was used as the shared mapping reference for both the tetraploid A.tetraploideus and A. ebracteatus x A. ilicifolius reads. A total of 10,450,704 nucleotides from 16,602 unigenes, each with a coverage of ≥ 10x, were identified across samples of the tetraploid A. tetraploideus and the two putative parental diploids (Additional file 4). The polymorphisms identified within A. ebracteatus and A. ilicifolius were filtered out to ensure that any polymorphism detected in A. ebracteatus x A. ilicifolius mapping was a result of interspecific nucleotide differences, not intraspecific variations. Among the 154,815 unambiguous SNPs identified from the A.tetraploideus mappings, 121,179 (78.27%) SNPs aligned with the A. ebracteatus and A. ilicifolius read mapping, respectively. This suggests that the majority of SNPs within tetraploid genes can be accounted for by sequence differences between A. ebracteatus and A. ilicifolius. These homeoSNPs were present in approximately 74% of the total genes analyzed, indicating that the congruence of the subgenomes of tetraploid A. tetraploideus with the genomes of the two diploids is corroborated by evidence at the whole-genome scale.

Nucleotide differences between the tetraploid A. tetraploideus and the parental diploids

When the tetraploid read sequences were compared with the genes of A. ebracteatus and A. ilicifolius, approximately 80% tetraploid reads were found to strictly map to the merged transcript sets of A. ebracteatus and A. ilicifolius in a 1:1 ratio. In A. tetraploideus, a total of 13,099 bp (0.11%) and 21,744 bp (0.23%) nucleotide differences were identified in the mappings to A. ebracteatus and A. ilicifolius, respectively (Additional file 7). These findings indicate a high degree of similarity between the tetraploid homeologs and the transcriptomes of A. ebracteatus and A. ilicifolius. The differentiated sites were observed in 33.36–49.07% of all genes examined. On the gene level, A. tetraploideus showed a small proportion (1.36-2.80%) of genes with nucleotide differences > 1%, consistent with the overall low nucleotide variation observed between the tetraploid homeologs and the putative diploid parents (Fig. 2A). Notably, the relative contribution of the E- and I-homeologs to nucleotide variation was found to be different. Overall, the A. tetraploideus-A. ebracteatus mapping had a lower frequency than the A. tetraploideus-A. ilicifolius mapping in the ‘0’ and ‘0-0.1’ nucleotide variation categories, while the opposite was observed in the ‘0.1-1’ and ‘1–10’ categories. Pairwise comparison of nucleotide differences across each gene also revealed that a larger proportion of genes exhibited higher nucleotide variation in E-homeologs than in I-homeologs in A. tetraploideus (Fig. 2B).

Global spectra of homeolog expression change in the tetraploid A. tetraploideus

To clarify the expression changes of parental homeologs in the tetraploid A. tetraploideus, a comprehensive genome-wide analysis of homeolog expression was conducted for both the in silico parental mix and tetraploid, utilizing the 112,762 homolog-diagnostic SNPs identified from A. tetraploideus. A total of 8,862 homeologous genes with an average depth ≥ 10 and harboring ≥ 3 unequivocal homeoSNPs per gene across all tetraploid samples were selected for further study. To enhance the reliability of our findings, 2,106 and 4,137 from the initial 8,862 genes were identified for having consistent homeologous expression ratios between replicates in the in silico parental mix and A. tetraploideus, respectively, and these genes were retained for further analysis. The global expression profiles of these genes across each plant sample are presented in a boxplot (Fig. 3A). A comparison with the in silico parental mix revealed that the degree of parental homeologous expression divergence was substantially reduced in A. tetraploideus (Kolmogorov-Smirnov test, p < 0.05), suggesting a reduction in parental expression divergence in the tetraploid.

Fig. 3

Nucleotide sequences differences between A. tetraploideus and the diploid progenitors A. ebracteatus and A. ilicifolius. (A) Frequencies of E and I homeologs in different nucleotide differences levels for A. tetraploideus. (B) Distribution of nucleotide difference ratio between E and I homeologs for A. tetraploideus

Homeolog-biased and unbiased expression in the tetraploid A. tetraploideus

In the tetraploid A. tetraploideus, the general reduction of parental expression divergence is also evidenced by the frequency of genes displaying biased homeolog expression towards one of the diploid progenitor genomes (Table 1). The proportion of genes exhibiting biased homeolog expression in A. tetraploideus (22.87%) was significantly lower compared to the in silico parental mix (67.66%) (Fisher’s exact test, p < 0.05). Upon closer examination of the biased expression towards each diploid genome individually, the in silico parental mix had a higher number of E-biased genes (1,007, 47.82%) than I-biased genes (418, 19.85%), suggesting a significantly higher expression of A. ebracteatus genes over A. ilicifolius genes (Exact binomial test, p < 0.05). In contrast, within the tetraploid plants, there was a greater number of I-biased genes (646, 15.62%) compared to E-biased genes (300, 7.25%) (Exact binomial test, p < 0.05).

Furthermore, when comparing the genes exhibiting biased homeolog expression between the tetraploid and the in silico parental mix, a significant difference was observed (Fig. 3B). The number of genes unique to the in silico parental mix was remarkably higher, with 378 E-biased and 91 I-biased genes, compared to the tetraploid, which had only 13 E-biased and 27 I-biased genes. In contrast, for genes with unbiased homeolog expression, none were found to be unique to the in silico parental mix. Collectively, these findings suggest that polyploidization in A. tetraploideus suppresses the divergence of homeologous gene expression.

GO analysis for the biased and unbiased genes (Additional file 8) revealed distinct enrichment patterns. In A. tetraploideus, unbiased genes were significantly enriched in several biological processes, encompassing growth, development, reproduction, metabolism and biological regulation; In contrast, E-biased and I-biased genes did not reveal statistically significant GO term enrichment, suggesting the large number of affected genes are involved in diverse biological functions. In the in silico parental mix, unbiased genes showed slight over-representation of genes associated with “catabolic process”, while E-biased genes were significantly enriched in cellular components such as cytoplasm, plastid stroma, and envelope. I-biased genes showed slight over-representation of genes linked to “transcription regulator activity”.

Table 1 Homeolog expression bias in in Silico parental mix and the tetraploid A. tetraploideus

Homeolog expression in the tetraploid A. tetraploideus in relation to diploid parental legacy

In A. tetraploideus, a comprehensive analysis identified all nine possible types of homeolog expression patterns, which were systematically categorized into four distinct groups (Table 2; Fig. 3C). Nearly half of the genes under analysis (515, 50.29%) demonstrated vertical inheritance of parental expression patterns (Group I, types 1–3), implying that the tetraploid has preserved a considerable degree of parental expression bias. More than 40% of genes (447, 43.65%) exhibited biased homeolog expression in the in silico parental mix, yet shifted to equal expression in the tetraploid (Group II, types 4–5), indicating a notable reduction in parental expression bias for a substantial number of genes. A modest proportion of genes revealed novel homeolog expression bias (Group III, types 6–7) (35, 3.42%) or an opposing directional bias (Group IV, types 8–9) (27, 2.64%) in the tetraploid. Collectively, these results indicate that while the tetraploid largely maintains parental expression patterns, there are also substantial modifications. These alterations, on one hand, generally dampened the parental expression bias in the tetraploid, and on the other hand, induce novel patterns of homeolog expression in a subset of tetraploid genes.

GO enrichment analysis was performed on genes across the four bias groups in tetraploid A. tetraploideus (Additional file 9). This analysis provided limited insights: Group II genes showed slight overrepresentation of genes linked to “structural molecule activity”. However, Group I, Group II and Group IV genes did not reveal statistically significant enrichment in specific GO terms.

Table 2 Expression pattern of the homeolog in tetraploid A. tetraploideus in relation to the in silico parental mix

Regulation of homeolog expression in the tetraploid A. tetraploideus

By categorizing homeolog expression into seven distinct regulatory divergence types (Fig. 4A), our analysis revealed that the trans-regulatory divergence type was more prevalent, constituting 35.45%, compared to the cis-regulatory type, which accounted for 16.11%. Genes subject to a combination of cis- and trans-regulatory divergences, including those designated as “cis + trans”, “cis x trans”, and “compensatory”, represented 6.93% in tetraploid A. tetraploideus. When comparing homeolog expressions in the tetraploid with the parental ortholog expression states in the in silico parental mix from an alternative perspective (Fig. 4B), 57.62% of the analyzed genes showed no significant difference in expression levels between the tetraploid and the in silico parental mix, classified as group III (conserved-regulation). The remaining 42.38% of the genes exhibited significantly divergence from the in silico parental mix, with over 39% of the genes falling under convergent regulation (group I) and less than 4% under divergent regulation (group II).

Fig. 4

Expression divergence of parental and homoeologous genes in the in silico parental mix and tetraploid A. tetraploideus. (A) Relative abundance of E homeolog transcripts. (B) Quantification of E-biased, I-biased and unbiased genes that were unique or shared between the parent mix and the tetraploid. (C) Heatmap of homeolog expression in the tetraploid relative to parental ortholog expression in the in silico parental mix for genes belonging to each of the four groups (I–IV). Each row represents a single gene

Fig. 5

Regulation of homeolog expression in tetraploid A. tetraploideus. (A) Dissecting cis- and trans- regulatory differences in A. tetraploideus. The scatter plots summarize the relative homeolog-specific expression levels in parental (in silico parental mix) and tetraploid data sets. Each point represents a single gene and the bars show the number of genes in each category. (B) Genes belonging to each of the three regulatory groups. (C) Overrepresented GO terms in each of the three regulatory groups for A. tetraploideus

While GO enrichment analysis of the three regulation types in tetraploid A. tetraploideus provided limited information (Additional file 10), a detailed examination of the top-ranking functional categories revealed notable differences among the regulatory types (Additional file 10; Fig. 4C). In Group III, functional categories associated with primary metabolic process, signal transduction, regulation of cellular process, tissue development, cell morphogenesis and response to external stimuli were noticeable. Group II genes were distinguished by functions such as plant organ senescence, pollen maturation, lateral root formation, and organic substance catabolic process. In Group I, GO categories related to the cytosolic ribosome, chloroplast stroma, peptide metabolic process, response to temperature stimuli, and response to toxic substance were conspicuous.

Discussion

Recent allopolyploid origin of the tetraploid A. tetraploideus

In our endeavor to trace the origin of tetraploid A. tetraploideus, this study initiated by validating the divergence pattern among Acanthus species, including three mangrove species—A. ebracteatus, A. ilicifolius, and A. volubilis—as well as three terrestrial relatives—A. leucostachyus, A. montanus, and A. mollis. Phylogenetic and MCMCTree analyses, based on single-copy orthologs derived from transcriptome data, revealed a monophyletic group encompassing all three diploid Acanthus mangroves. These mangroves diverged from their closest terrestrial relative, A. leucostachyus, approximately 21.12 Mya (95% CI = 17.11–24.56 Mya). Synonymous substitution rates (Ks) further delineated the degree of divergence among these diploid species, aligning with their phylogenetic clade groupings as depicted in Additional file 5.

Tetraploid A. tetraploideus (2n = 96) likely originated through either whole-genome duplication of one of the diploids—A. ilicifolius, A. ebracteatus or A. volubilis—to form an autotetraploid, or through hybridization of two of the three diploids (2n = 48) to form an allotetraploid. Our previous study, based on morphological, cytological and molecular evidence, supports the latter hypothesis [35]. In this study, the allopolyploid origin of A. tetraploideus was further validated using genome-wide evidences from transcriptome data. Among over 15,000 genes analyzed, approximately 80% of the SNPs within the tetraploids could be attributed to sequence differences between A. ebracteatus and A. ilicifolius. Phylogenetic analyses of homologous sequences indicated that tetraploid A. tetraploideus harbored the genomic components from both A. ebracteatus and A. ilicifolius (Fig. 1). In phylogenies constructed from homologous sequences occurring in a 1:1:2 ratio for A. ebracteatus vs. A. ilicifolius vs. A. tetraploideus, one set present in A. tetraploideus clustered with the set in A. ebracteatus, while the other set clustered with the set in A. ilicifolius (Fig. 1A). Despite the potential absence of one of the two homeologous gene sets in A. tetraploideus, possibly due to evolutionary processes or assembly method limitations [2, 36, 61, 62], phylogenetic relationships between homologous sequences occurring in a 1:1:1 ratio for the three species also strongly supported for its allopolyploid origin, with one homologous gene set in A. tetraploideus clustered with the set present in either A. ebracteatus or A. ilicifolius with similar probabilities (Fig. 1B). Similar observations were made when A. volubilis was included in the phylogenetic analysis, with the homologous gene set in A. tetraploideus preferentially clustering with that in A. ebracteatus or A. ilicifolius rather than A. volubilis in most gene trees (Additional file 6). These results provide compelling evidence for the allopolyploid origin of tetraploid A. tetraploideus from diploid A. ilicifolius and A. ebracteatus.

Although the divergence between A. ilicifolius and A. ebracteatus is estimated to have occurred around 10 Mya, our nucleotide variation analysis suggests that the origin of A. tetraploideus is likely more recent. This analysis compared A. tetraploideus read sequences with those of A. ilicifolius and A. ebracteatus, showing that the majority of nucleotide differences in alignments were less than 1% (Fig. 2A). Moreover, when classifying nucleotide differences, we observed that the majority of A. tetraploideus sequences compared to A. ebracteatus exhibited no divergence, indicating a recent divergence from its diploid ancestors. Substantial research has demonstrated that the Strait of Malacca, serving as a nexus between the Pacific and Indian Oceans, fluctuates between acting as a geographical barrier and a facilitator of genetic exchange for coastal and marine species, contingent upon variations in sea levels [63,64,65,66]. These fluctuations could have alternately isolated and then brought into contact previously allopatric populations, potentially promoting allopolyploidization by creating opportunity for interaction between distinct populations or species. For example, A. ebracteatus and A. ilicifolius, which are primarily distributed along the Pacific and Indian Ocean coasts respectively, may have been brought into contact under such conditions, contributing to the formation of A. tetraploideus [67]. However, detailed analyses of genomic and population data will be essential to elucidate the evolutionary history of A. tetraploideus.

Pattern of homeologous gene expression bias in A. tetraploideus

Previous research across diverse plant taxa have demonstrated that changes in genome structure and gene expression frequently lead to subgenome specificity in allopolyploids [30, 68, 69]. However, no study has systemically investigated transcriptomic asymmetry in polyploid mangroves until now. In this study, we found that 77.13% of genes in the tetraploid A. tetraploideus exhibited equivalent level of homeologous gene expression, while 22.87% showed biased homeologous expression (Table 1). Taking into consideration potential variations due to different methodological parameters across studies, the proportion of genes exhibiting biased homeologous gene expression is comparable to that reported for several natural allopolyploids: 35% in Coffea arabica [20], 25% in Gossypium hirsutum [70], 27% in Nicotiana tabacum [71], and 22% in Tragopogon miscellus [72]. However, our estimates are higher than the 15% observed in synthetic Brassica napus allopolyploids [28] and lower than the 56.2–72.1% observed in natural tetraploid wheat and the 60-61.1% in synthetic tetraploid wheat [30]. Thus, this variation cannot be solely attributed to an increase in homeologous expression bias over evolutionary time, as previously suggested [20, 69]. Instead, we propose that additional factors, such as the species-specific traits and the genetic relationships among the polyploid and its progenitors, could substantially influence the extent of transcriptomic asymmetry observed in polyploids.

Our study revealed that differential expression between the two putative diploid progenitors, A. ebracteatus and A. ilicifolius, has profound impacts on the homeolog expression patterns in the resulting allotetraploid A. tetraploideus. Upon assessing the relative global ortholog expression levels between the two diploid species, A. ebracteatus and A. ilicifolius, we found that > 60% of the orthologs exhibited differential expression, with A. ebracteatus exhibiting higher overall ortholog expression levels than A. ilicifolius (Table 1). Furthermore, in the context of relative homeologous gene expression within A. tetraploideus, we observed that the global expression divergence between the homeologs from the A. ebracteatus and A. ilicifolius subgenomes was significantly reduced compared to the in silico parental mix (Fig. 3). This suppression of parental expression divergence in A. tetraploideus aligns with findings from a study on allopolyploid wheat [30], but contrasts with results from an allopolyploid rice [55]. Despite the reduction, the relative homeolog expression levels in the parents were partially preserved, and the expression patterns of approximately half of the analyzed genes reflected a vertical transmission of the intrinsic parental ortholog expression states (i.e., the ‘Inheritance’ group in Table 2). A similar pattern was observed in synthetic tetraploid wheat, where the parental inheritance accounted for 56.3–63.1% of the genes analyzed [30]. This finding suggests that differential expression between the diploid parental species may substantially influence the homeolog expression in the resulting tetraploid.

Despite the substantial inheritance of parental expression states, we observed an enhancement of biased expression towards the I-subgenome and the emergence of several novel expression patterns in A. tetraploideus (Tables 1 and 2). Studies indicate that ‘transcriptome shock’ from tetraploid formation disrupts parental gene expression patterns [30]. This genomic asymmetry is expected to evolve under selection pressures, especially in mangroves adapting to stressful environments [34, 68, 73]. Despite the recent allopolyploidization in A. tetraploideus, the species likely displays evolved expression patterns. Our results support the notion that the parental genetic legacy substantially influences homeolog expression in allotetraploids, with new patterns also contributing significantly to differential expression [30, 68, 72, 74, 75].

Regulation of homeologous expression in A. tetraploideus

In allotetraploids, expression variation between homeologs can result from changes in cis- and/or trans-regulation. When diverged cis- and trans-regulatory factors from diploid progenitors merge in a single nucleus, transcription of two divergent alleles is controlled by both native cis-regulatory elements and novel trans-acting factors. This often results in reduced or eliminated parental expression differences in the tetraploids for genes subjected to strong trans-regulation [55]. Our study detected a higher proportion of trans-regulatory genes compared to cis-regulatory ones, which plausibly accounts for the attenuated range of parental expression difference in A. tetraploideus (Fig. 4A). This regulatory pattern aligns with findings in rice [55] and Drosophila [57], but contrast with studies in Arabidopsis [76] and coffee [20], where cis-regulation predominated over trans-regulation.It also differs from cotton [77], in which a predominant “cis + trans” regulatory effect was observed. While differences in the sensitivity of methods used across studies may contribute to this discrepancy, it could also reflect variations in the divergence level between subgenomes, historical selection pressures, or rates and mechanisms of genome plasticity [20, 57, 78].

The significantly higher proportion of genes belonging to regulatory Group III (genes under conserved-regulation) aligns with the observation of the strong effect of trans-regulation in the tetraploid A. tetraploideus (Fig. 4B). While the proportions of genes in the three regulatory groups may vary across tissues, lineages and species, the predominance of Group III genes seems to be a common feature in polyploids, particularly in newly-formed polyploids [30, 55]. However, the predominance of Group III genes seems to diminish over evolutionary time. For example, in tetraploid wheats, the strong convergent regulation of subgenome expression observed during the initial stages of allotetraploidization has been largely replaced by divergent regulation over time, decreasing from 56.8% to 44.7–45.7% in leaf tissue [30]. The high proportion of genes under convergent regulation (57.62%) in the allotetraploid A. tetraploideus may reflect its relatively recent formation [35].

Putative functional implications of transcriptome alterations in allotetraploid A. tetraploideus

Investigations into the evolution of allopolyploid genomes have demonstrated that subgenomes within these species often exhibit structural and functional asymmetry, even in recently emerged allopolyploids. Subgenome dominance and transcriptome changes associated with allopolyploidization have been reported in polyploids of wheat [30], Brassica napus [79], cotton [19], Brassica rapa [80], strawberry [81], and maize [82]. Such asymmetry has been shown or hypothesized to play an important role in the rapid establishment and evolutionary success of allopolyploids compared to their diploid parents. For example, polyploidy wheat exhibits subgenomic asymmetry, which enhances its growth vigor and adaptive potential over its lower-ploidy ancestors [14, 30, 83]. Similarly, homeolog-biased expression in tetraploid cotton is associated with an enhanced ability to cope with various abiotic stresses compared to diploid species [84]. However, exceptions to this pattern exist. For example, in both allotetraploid Coffea arabica [20] and Arabidopsis suecica [85], no evidence of subgenome dominance in gene expression was found, and genes showing large expression difference between homeologs were not biased towards any particular GO category. In our analysis of homeolog expression rewiring in allotetraploid A. tetraploideus, expression differences between homeologs were evident in a substantial number of genes. These biased genes, including both E-biased genes and I-biased genes, are involved in diverse biological process such as metabolism, homeostatic process, biological regulation, growth, development, reproductive and stress response (Additional file 8). However, strict significance testing of GO enrichment analysis did not identify genes with specific functions preferentially expressed by one subgenome over the other. This finding aligns with observations in C. arabica and Arabidopsis suecica, which, like A. tetraploideus, are evolutionarily recent allotetraploids with relatively limited divergence between their ancestral species [20, 35, 85].

It should be noted that both A. tetraploideus and its diploid progenitor species (A. ebracteatus and A. ilicifolius) are true mangrove species, well adapted to extreme intertidal environments characterized by high salinity, drought, hypoxia and high UV radiation. Our preliminary study on environmental niche modeling revealed that the tetraploid A. tetraploideus has expanded its niche space compared to the two diploid species, rather than only exhibiting niche differentiation between the tetraploid and diploids [35]. Furthermore, A. tetraploideus and its diploid progenitor species exhibit only minor morphological differences (i.e., corolla characters), with the tetraploid showing an intermediate or mixed phonotype in corolla color and size relative to its diploid parents [35]. Likely reflecting these characteristics, the vast majority of genes in A. tetraploideus showed equivalent expression between homeologs. These genes were significant enriched in several cellular components, such as organelle, cytosol, nucleoplasm and intracellular protein-containing complex, and were involved in metabolic process, multicellular process, developmental growth, and reproductive process (Additional file 8). The overrepresentation of unbiased gene expression related to diverse fundamental cellular processes and functions may provide a genomic basis for the stable viability of A. tetraploideus as a polyploid.

Additionally, upon closer examination of genes showing novel patterns of homeolog expression bias in A. tetraploideus, it was observed that several genes showing novel I-biased pattern relative to the in silico parental mix (type 6 in Table 2) are associated with chromosome dynamics and cell cycle. Specifically, the regulator of chromosome condensation (RCC1) is recognized as a critical cell cycle regulator, playing roles in the assembly of mitotic spindles, regulation of the G1/S transition, formation of nuclear envelope, and transport of nuclear material [86]. RHO GTPase in plants are essential for modulating cytoskeletal dynamics and have been demonstrated to regulate cell polarity, morphology, migration and cell cycle progression [87]. WPP domain proteins are developmentally associated with the nuclear envelope, a barrier separating the nucleus from the cytoplasm that undergoes a complex reorganization during mitosis, and are known to promote cell division [88]. Furthermore, the structural maintenance of chromosomes (Smc) complex is a key organizer of prokaryotic and eukaryotic genomes, with the Smc5/6 complex playing an important role in DNA damage repair via homologous recombination, DNA replication, and chromosome stability. Instability of the SMC5/6 complex can lead to the chromosome breakage [89]. The non-structural maintenance of chromosome element 4 (Nse4), an essential subunit of the Smc5/6 complex, is vital for maintaining the stability of the entire complex and functions in both somatic nuclei and meiosis to ensure plant viability and fertility [90, 91]. The shift in homeolog expression from unbias in the parental mix to I-bias in the tetraploid for genes encoding RCC1, Rho GTPase, Nse4 and WPP domain-interacting proteins may help prevent aneuploidy and undesirable homeologous exchanges in the allopolyploid. This suggests a potential adaption to polyploidy at the cellular level. Similar evidence for adaption to polyploidy through modifications of the meiotic machinery has been observed in other plants [85].

Since all currently known true mangroves, except A. tetraploideus are diploid, research on polyploidization in mangroves is extremely limited. Feng et al. (2024) [92] investigated the impacts of ancient polyploidization events on the evolution of a modern diploid mangrove tree Sonneratia alba. Wang et al. (2024) [93] studied the evolutionary history of an allotetraploid mangrove associate Barringtonia racemosa, though transcriptome expression was not examined in their study. As a result, it is currently unachievable to compare the transcriptome alteration patterns in A. tetraploideus with those of other mangroves or closely related species, or to determine whether or to what extent polyploid mangroves share a universal pattern of transcriptome asymmetry. In addition, due to the recent discovery of A. tetraploideus, research on the physiological, phenotypic and ecological differences between A. tetraploideus and its diploid progenitor species remains very limited [35]. This lack of data also restricts the ability to link transcriptome changes accompanying polyploidization with specific adaptive traits of A. tetraploideus. Consequently, it is challenging to draw definitive conclusions about whether and how transcriptome changes are related to the adaptive evolution of A. tetraploideus based on the currently available data. We do know that A. tetraploideus exhibits certain advantages over its diploid progenitors, such as occupying a wider ecological niche. However, it is undeniable that polyploids acquire additional genes through hybridization and whole genome duplication, and these duplicated genes provide raw materials for adaptive evolution. This genetic redundancy may enable the tetraploid A. tetraploideus to tolerate environments at higher latitudes more effectively than its diploid progenitors, potentially contributing to its broader distribution range. Nevertheless, a deeper understanding of the underlying connections and mechanisms by which polyploidization affects the evolution and adaptation of A. tetraploideus will require comprehensive exploration incorporating genomic, transcriptomic, and epigenetic studies.

Conclusion

In this study, we conducted a comprehensive analysis to uncover the divergence among diploid Acanthus mangrove species and to decipher the polyploid origin and transcriptome reconfiguration in the tetraploid A. tetraploideus, utilizing RNA-sequencing. Our phylogenetic analysis and divergence time estimation revealed a closer relationship between A. ebracteatus and A. volubilis than with A. ilicifolius. The thorough examination of homologous gene relationships and nucleotide sequences alignments across extensive datasets has provided robust genome-wide evidence supporting the recent allopolyploid origin of tetraploid A. tetraploideus, with its diploid progenitors being A. ebracteatus and A. ilicifolius. Transcriptome profiling for gene expression in tetraploid A. tetraploideus revealed a widespread reduction in the divergence of parental gene expression, consistent with the dominance of convergent-regulatory genes over divergent-regulatory ones, and a significantly higher proportion of trans-regulatory gens compared to cis-regulatory genes within the species. A meticulous analysis of homeolog expression in A. tetraploideus suggests that the species’ tetraploid status is profoundly influenced by its parental genetic heritage. However, the transcriptome shock associated with polyploidization and the subsequent evolutionary processes also significantly impact the transcriptome asymmetry within the new polyploid. Although no strong evidence links transcriptomic changes to specific adaptive traits, the patterns in unbiased and novelly biased genes suggest genetic adaptations to stable polyploidy in A. tetraploideus, potentially supporting its initial establishment and long-term evolutionary success. These findings enrich our knowledge of polyploidy in Acanthus and pave the way for further research into the molecular mechanisms and evolutionary forces that underlie the speciation and adaptation processes of mangrove species.