Strong phylogenetic signal from chloroplast genomes of three Barringtonia species provides the first genomic resources for their conservation

Abstract

Background

The genus Barringtonia (Lecythidaceae) is a vital component of tropical coastal forests and mangrove ecosystems. Among its members, B. racemosa and B. fusicarpa are classified as Endangered and Vulnerable, respectively, due to habitat degradation and anthropogenic pressures, underscoring the urgent need for genetic studies to guide conservation. Chloroplast (cp.) genomes serve as essential resources for phylogenetic reconstruction and conservation genetics. However, the scarcity of cp. genome data for Barringtonia has limited comprehensive evolutionary and conservation-oriented investigations.

Results

We assembled and annotated the first complete cp. genomes of B. racemosa, B. fusicarpa, and B. acutangula. All three genomes exhibit the typical quadripartite structure, ranging from 158,959 bp (B. racemosa) to 159,837 bp (B. acutangula), and contain 132 genes (87 protein-coding, 37 tRNA, 8 rRNA) with a GC content of 36.68%-36.86%. Collinearity and IR boundary analyses revealed high structural conservation without large-scale rearrangements. Interspecific sequence-level variations were detected in simple sequence repeats (SSRs) and long repeats. Nucleotide diversity (π) analysis identified highly polymorphic regions, including rpl20 (π = 0.080), rpoA (π = 0.064), rps3 (π = 0.063), and ndhF (π = 0.060), which represent promising molecular markers for population genetics within the genus. Codon-based selection analyses (Ka/Ks) showed that all protein-coding genes are under strong purifying selection (mean Ka/Ks 0.32–0.37), with no evidence of positive selection. Pairwise genetic distances (p-distances) among Barringtonia species are extremely low (mean 0.0046), while distances to the related genus Bertholletia are ~ 6-fold higher, supporting their generic distinction.

Conclusions

Phylogenetic analysis robustly supports Barringtonia as a monophyletic clade (bootstrap = 100%), with B. racemosa and B. fusicarpa forming a sister lineage to B. acutangula. This study provides the first high-quality cp. genome resources for the two threatened Barringtonia species, revealing strong structural and sequence conservation but no direct chloroplast genomic correlates of endangerment. The identified polymorphic regions and repeat markers lay a foundation for future population genetics, phylogeographic studies, and conservation-oriented genetic management of these ecologically important coastal plants.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12870-026-09322-y.

Background

Global climate change and anthropogenic activities are exerting combined pressures on coastal ecosystems. As critical transitional zones, mangroves serve dual roles as “coastal guardians” [1] and highly efficient blue carbon sinks [2], yet face severe threats from habitat loss and climate shifts [3]. Understanding the genetic diversity, evolutionary history, and adaptive mechanisms of key mangrove-associated taxa is an urgent prerequisite for formulating conservation strategies and predicting ecological resilience under future environmental changes [4]. The genus Barringtonia (Lecythidaceae), widely distributed in tropical coastal forests and mangrove ecosystems of the world [5], often dominates or co-occurs in the landward fringe of mangroves, forming a pivotal ecological gradient between terrestrial forests and typical mangrove communities [6]. These plants stabilize coastal embankments through well-developed root systems [7], provide habitats and food for diverse wildlife [8], and exhibit significant potential in traditional medicine [9] and ornamental horticulture. However, due to their specialized habitat requirements in intertidal zones and typically narrow geographic distributions, multiple Barringtonia species face acute survival threats. According to the IUCN Red List, B. conoidea is already extinct (EX), while B. asiatica, B. macrostachya, B. reticulata, and B. racemosa are classified as Critically Endangered (CR). B. fusicarpa is Vulnerable (VU), and B. acutangula persists primarily in cultivation [10]. The three focal species of this study, B. racemosa (CR), B. fusicarpa (VU), and B. acutangula (cultivation-dependent), were selected to represent contrasting conservation statuses and potential differences in habitat specialization, thereby enabling a comparative assessment of chloroplast genomic features in relation to endangerment levels. Addressing this conservation crisis calls for urgent genetic investigations into population history, genetic diversity, and adaptive potential, which can inform conservation frameworks and sustainable germplasm utilization. However, such assessments ideally require integration of nuclear genomic markers and population‑level data, as chloroplast genomes alone provide only a partial perspective.

Mangroves represent a specialized group of woody plants strictly confined to tropical and subtropical intertidal zones, which have evolved unique physiological and morphological adaptations, such as aerial roots, salt glands, and vivipary—to cope with extreme stresses including high salinity, flooding, hypoxia, and intense UV radiation [11]. Recent genomic studies have further revealed the molecular basis underlying these phenotypic adaptations. Notably, chloroplast genomes, as central hubs for photosynthesis and maternally inherited molecular markers, offer critical insights into how mangroves optimize light capture, counteract photoinhibition and oxidative stress, and coordinate energy metabolism with ion homeostasis through structural variations, gene content shifts, and altered evolutionary rates [12, 13]. Comparative analyses of chloroplast genomes in representative mangrove genera, such as Sonneratia, Bruguiera, and Aegiceras—have demonstrated that photosynthesis-related genes (e.g., psbA, rbcL) and genes involved in transcriptional-translational regulation (e.g., accD, ycf) may undergo positive selection or harbor unique mutations [14]. These molecular variations are hypothesized to correlate with physiological adaptations, including enhanced photosynthetic efficiency, photoprotection against excess light, and coordination of energy metabolism with ion balance. By resolving complete chloroplast genomes, researchers can clarify phylogenetic relationships and identify hypervariable regions (e.g., SSRs, nucleotide hotspots) that serve as candidate molecular markers. Nevertheless, because chloroplast genomes evolve under strong purifying selection, their ability to directly reveal adaptive genomic signatures or assess genetic vulnerability is limited; such inferences require complementary data from nuclear genomes and population genetics [15, 16]. However, it is important to note that chloroplast genomes represent an initial genomic resource, but not necessarily the primary determinant of genetic vulnerability, as adaptive processes may be predominantly mediated by the nuclear genome or gene expression regulation. Therefore, the present study uses chloroplast genomic data as a starting point for exploring evolutionary patterns, while acknowledging that comprehensive assessments of genetic vulnerability require integration of nuclear markers and population-level data.

Although chloroplast genomics has been widely applied in plant systematics and adaptive studies [17], genomic resources for Barringtonia remain critically limited. To date, only the complete chloroplast genome of B. asiatica has been publicly released (NC_070213.1) [18]. The present study therefore provides the first complete chloroplast genomes for the three focal species, B. racemosa, B. fusicarpa, and B. acutangula, which have not been previously reported. Furthermore, no systematic, fine-scale comparative chloroplast genomic studies have been conducted on these closely related Barringtonia species with divergent conservation statuses. This data gap has hindered our ability to address three key questions at the genomic scale: (1) Do chloroplast genome structures of Barringtonia species (particularly endangered vs. non-endangered taxa) exhibit specific variations linked to adaptation or genetic bottlenecks? (2) What is the precise phylogenetic position of Barringtonia within the Lecythidaceae and broader Ericales clade? (3) Do Barringtonia chloroplast genomes bear molecular signatures associated with adaptation to mangrove-specific stresses (e.g., salinity)? It is acknowledged that adaptive signals, if present, may be subtle in the highly conserved plastome, and their absence does not preclude the possibility of adaptive evolution occurring in the nuclear genome or via regulatory mechanisms.

To fill these knowledge gaps and support science-based conservation of endangered Barringtonia species, this study presents the first complete chloroplast genome sequencing, de novo assembly, and functional annotation of B. racemosa, B. fusicarpa, and B. acutangula. Through systematic comparative genomics and phylogenetic analysis, we aim to resolve the following critical questions: (1) What are the similarities and differences in chloroplast genome size, structure, gene composition, and repeat sequences among the three species? How conserved or variable are these genomes? (2) What phylogenetic relationships exist among the three species based on complete chloroplast genome sequences? Is Barringtonia monophyletic? (3) Do genes involved in photosynthesis, transcription-translation, and stress responses exhibit differential selective pressures? Are unique genomic variations observed in endangered species (e.g., B. racemosa, B. fusicarpa) linked to population history or environmental adaptation, thereby revealing genetic vulnerability?

Results

Comparative analysis of the fundamental features of chloroplast genomes

The complete chloroplast genome maps for the three Barringtonia species are presented in Fig. 1. All genomes exhibit the canonical quadripartite structure characteristic of angiosperms, comprising a large single-copy region (LSC), a small single-copy region (SSC), and a pair of inverted repeat regions (IRa and IRb). Genes are distributed across both strands: genes located on the outer circle are transcribed clockwise, while those on the inner circle are transcribed counterclockwise, reflecting bidirectional transcription and functional differentiation within the plastome. Gene categories are color-coded according to their functional roles, highlighting both conserved core functions and lineage-specific features within the chloroplast genome.

Fig. 1.

Genome maps of the Barringtonia cp. genomes. Genes placed outside the circle are transcribed clockwise, whereas genes inside the circle are transcribed counterclockwise. Gene colors differentiate protein-coding genes based on their respective functions. LSC, large single-copy region; SSC, small single-copy region; IRA and IRB, two inverted repeats; GC content, dark grey area in inner circle; AT content, light grey area in inner circle

As shown in Table 1, the total genome lengths vary among the three species: B. racemosa (158,959 bp) < B. fusicarpa (158,977 bp) < B. acutangula (159,837 bp). Total GC content also differs: B. acutangula (36.68%) < B. racemosa (36.83%) < B. fusicarpa (36.86%). Region-specific metrics are as follows: B. racemosa: LSC = 88,225 bp (GC = 34.64%), SSC = 18,556 bp (GC = 30.15%), IRa/IRb = 26,089 bp each (GC = 42.90%). B. fusicarpa: LSC = 88,255 bp (GC = 34.67%), SSC = 18,596 bp (GC = 30.25%), IRa/IRb = 26,063 bp each (GC = 42.92%). B. acutangula: LSC = 88,930 bp (GC = 36.68%), SSC = 18,699 bp (GC = 29.97%), IRa/IRb = 26,104 bp each (GC = 42.87%). All three chloroplast genomes encode 132 genes, primarily involved in photosynthesis and self-replication. These include 37 transfer RNA (tRNA) genes, 8 ribosomal RNA (rRNA) genes, and 87 protein-coding genes (PCGs). Among these, six tRNA genes (trnA-UGC, trnG-GCC, trnI-GAU, trnK-UUU, trnL-UAA, trnV-UAC) and nine protein-coding genes (ndhA, ndhB, petB, petD, atpF, rpl16, rpl2, rps16, rpoC1) each contain a single intron. Three additional protein-coding genes—rps12, clpP, and ycf3, each contain two introns (Table 2).

Table 1.

General characteristics of the chloroplast genomes of the three Barringtonia species

Species	Barringtonia racemosa	Barringtonia fusicarpa	Barringtonia acutangula
Total length (bp)	158,959	158,977	159,837
Length of LSC (bp)	88,225	88,255	88,930
Length of IRa (bp)	26,089	26,063	26,104
Length of IRb (bp)	26,089	26,063	26,104
Length of SSC (bp)	18,556	18,596	18,699
Total GC content (%)	36.83	36.86	36.68
LSC GC content (%)	34.64	34.67	34.46
IRa GC content (%)	42.90	42.92	42.87
IRb GC content (%)	42.90	42.92	42.87
SSC GC content (%)	30.15	30.25	29.97
Number of genes	132	132	132
Number of CDs genes	87	87	87
Number of tRNA genes	37	37	37
Number of rRNA genes	8	8	8

Table 2.

Annotated genes and their classification in the cp. genomes of three Barringtonia species

Category	Gene group	Gene name	Count
Photosynthesis	Subunits of photosystem I	psaA, psaB, psaC, psaI, psaJ	5
	Subunits of photosystem II	psbA, psbB, psbC, psbD, psbE, psbF, psbH, psbI, psbJ, psbK, psbL, psbM, psbN, psbT, psbZ	15
	Subunits of NADH dehydrogenase	ndhA*, ndhB*(2), ndhC, ndhD, ndhE, ndhF, ndhG, ndhH, ndhI, ndhJ, ndhK	12
	Subunits of cytochrome b/f complex	petA, petB*, petD*, petG, petL, petN	6
	Subunits of ATP synthase	atpA, atpB, atpE, atpF*, atpH, atpI	6
	Large subunit of rubisco	rbcL	1
	Subunits photochlorophyllide reductase	-	0
Self-replication	Proteins of large ribosomal subunit	rpl14, rpl16*, rpl2*(2), rpl20, rpl22, rpl23(2), rpl32, rpl33, rpl36	11
	Proteins of small ribosomal subunit	rps11, rps12**(2), rps14, rps15, rps16*, rps18, rps19, rps2, rps3, rps4, rps7(2), rps8	14
	Subunits of RNA polymerase	rpoA, rpoB, rpoC1*, rpoC2	4
	Ribosomal RNAs	rrn16(2), rrn23(2), rrn4.5(2), rrn5(2)	8
	Transfer RNAs	trnA-UGC*(2),trnC-GCA,trnD-GUC,trnE-UUC,trnF-GAA,trnG-GCC*,trnG-UCC,trnH-GUG,trnI-CAU(2),trnI-GAU*(2),trnK-UUU*,trnL-CAA(2),trnL-UAA*,trnL-UAG,trnM-CAU, trnN-GUU(2),trnP-UGG,trnQ-UUG,trnR-ACG(2),trnR-UCU,trnS-GCU,trnS-GGA,trnS-UGA,trnT-GGU,trnT-UGU,trnV-GAC(2),trnV-UAC*,trnW-CCA,trnY-GUA,trnfM-CAU	37
Other genes	Maturase	matK	1
	Protease	clpP**	1
	Envelope membrane protein	cemA	1
	Acetyl-CoA carboxylase	accD	1
	c-type cytochrome synthesis gene	ccsA	1
	Translation initiation factor	infA	1
Genes of unknown function	Conserved hypothetical chloroplast ORF	lhbA, ycf1(2), ycf2(2), ycf3**, ycf4	7
Total			132

The symbol “*” indicates a gene containing a single intron “**” represents a gene with two introns, and “(2)” Indicates two gene copies located within inverted repeats

Comparative analysis of chloroplast genomes and IR boundary changes

Figure 2 presents a comparative analysis of sequence homology and collinearity across the chloroplast genomes of six species: four Barringtonia species, one closely related in-group taxon, and one out-group species (Rhododendron simsii). Within the genus Barringtonia (including B. acutangula, B. fusicarpa, and B. racemosa), chloroplast genomes exhibit exceptionally high collinearity. The order and spatial arrangement of local collinear blocks (LCBs) are nearly identical among species, with the majority of LCBs aligned above the central axis of the plot — indicating that homologous regions are oriented in the same direction and have not undergone major genomic rearrangements.

Fig. 2.

Homology analysis of chloroplast genomes across six species. Legend: Short blocks denote gene locations within the genome, with white for CDS, green for tRNA, and red for rRNA. Colored blocks (Locally Collinear Blocks, LCBs) highlight regions that align between genomes, indicating potential homology without rearrangements. Blocks above the center line show co-linear regions in the same orientation; those below indicate inverted alignments relative to the reference genome. Areas outside of blocks lack homology among the input genomes. Within each block, Mauve plots the similarity score, where the height reflects the average sequence similarity. Lines connecting colored blocks represent collinear relationships

Furthermore, the distribution patterns of functional elements including protein-coding genes (CDS; white blocks), tRNA genes (green blocks), and rRNA genes (red blocks), are highly conserved across Barringtonia species, reflecting strong structural and functional conservation of gene order within the plastome. The similarity profiles within LCBs are also highly consistent and show elevated average identity scores, further confirming the high degree of sequence homology among Barringtonia species. All four Barringtonia species, including B. asiatica, exhibit highly conserved IR boundary configurations and gene order, further confirming the structural stability of the chloroplast genome within the genus. Only minor non-homologous regions lie outside the LCBs.

In contrast, the chloroplast genomes of Barringtonia species and the out-group R. simsii display pronounced divergence. The arrangement and positioning of LCBs differ significantly between these groups. Some LCBs align below the central axis, indicating inverted or complementary orientations, and a substantially higher proportion of the genome lies outside the LCBs as non-homologous regions. This reflects substantial structural and sequence divergence between Barringtonia and the out-group.

Collectively, these results underscore the remarkable structural conservation of chloroplast genomes within the genus Barringtonia, while simultaneously revealing significant genomic differentiation relative to distantly related lineages.

Figure 3 illustrates the structural organization and gene distribution patterns at the boundaries between the inverted repeat regions (IRa and IRb) and the large single-copy (LSC) and small single-copy (SSC) regions, specifically, the junctions JLB (LSC-IRb), JSB (SSC-IRb), JSA (SSC-IRa), and JLA (LSC-IRa), across four Barringtonia species, the closely related taxon Bertholletia excelsa, and the out-group Rhododendron simsii. This comparison clearly reveals the high degree of conservation within Barringtonia and the progressive divergence among taxonomic groups.

Fig. 3.

Comparison of IR boundaries in the cp. genomes of six species. JLA, junction between LSC and IRa; JLB, junction between LSC and IRb; JSA, junction between SSC and IRa; JSB, junction between SSC and IRb

Within Barringtonia, chloroplast genome size, regional lengths, and IR boundary configurations exhibit exceptional conservation: Genome sizes range narrowly from 158,794 to 159,837 bp. IRb/IRa lengths are stable between 26,063 and 26,104 bp. Gene spanning patterns at boundary junctions are highly uniform: At JLB (LSC-IRb), the rpl2 gene extends into IRb by a fixed length of 1,496 bp; its LSC-side segment varies only slightly (73–78 bp). The adjacent rps19 gene has a constant LSC-side length of 261 bp and extends 18 bp into IRb. At JSB (SSC-IRb), the ndhF gene extends into IRb by exactly 2 bp, with its SSC-side segment ranging narrowly from 2,260 to 2,262 bp. At JLA (LSC-IRa), the trnH gene extends 75 bp into IRa, while its LSC-side segment is consistently 2 bp. These precisely conserved numerical features reflect a lineage-specific, structurally stable configuration of IR boundaries within the genus Barringtonia.

In contrast, compared to the closely related Bertholletia excelsa, Barringtonia species display clear divergence in IR boundary architecture: B. excelsa exhibits an expanded IR region (27,841 bp). At JLB, the LSC-side segment of rpl2 is 60 bp (vs. 73–78 bp in Barringtonia). At JSB, ndhF extends 19 bp into IRb (vs. 2 bp in Barringtonia). These differences indicate substantial IR expansion and structural reorganization even among closely related lineages.

The out-group R. simsii exhibits fundamental divergence from Barringtonia: Its total chloroplast genome length is substantially larger (206,912 bp), with an IR region of 47,036 bp- more than double that of Barringtonia. The boundary-associated genes differ entirely: for example, rps12 occupies the JLB position in R. simsii, whereas rpl2 defines this junction in Barringtonia. Gene-spanning patterns at all junctions lack homology with those in Barringtonia, confirming that phylogenetic distance exerts a decisive influence on IR boundary structure.

Analysis of simple sequence repeats and long repeat sequences

The total number of simple sequence repeats (SSRs) identified across the three species ranged from 254 to 267: Barringtonia racemosa had 254 SSRs, B. fusicarpa had 255 SSRs, and B. acutangula had 267 SSRs. Mononucleotide repeats were overwhelmingly dominant in all three species, comprising 62.20% (158 SSRs) in B. racemosa, 62.35% (159 SSRs) in B. fusicarpa, and 63.30% (169 SSRs) in B. acutangula. These mononucleotide repeats exhibited a strong preference for A/T bases, with A/T-type mononucleotide repeats accounting for 94.94%-95.86% of all mononucleotide repeats, while C/G-type repeats comprised only 4.14%–5.06%.

Additionally, B. racemosa contained three pentanucleotide repeats ((ATTAC)₃, (TAATA)₃, (AATAT)₃), B. fusicarpa had one pentanucleotide repeat ((ATTAC)₃), and B. acutangula exclusively contained mono-, di-, tri-, and tetranucleotide repeats. In terms of repeat length distribution, all three species predominantly featured short repeats of 8–10 bp, followed by medium repeats of 11–15 bp. Long repeats (> 16 bp) and very long repeats were rare; notably, B. acutangula contained one ultra-long repeat ≥ 21 bp ((A)₂₁), whereas the longest repeats in B. fusicarpa and B. racemosa were 17 bp and 15 bp, respectively.

Regarding gene association characteristics, SSRs were primarily distributed in intergenic regions across all three species, with a notable concentration around key functional genes such as clpP (14 SSRs in B. racemosa [5.2%], 13 SSRs in B. fusicarpa [4.8%], 13 SSRs in B. acutangula [4.8%]), ycf1 (17 SSRs each in B. racemosa, B. fusicarpa, and B. acutangula [6.3%]), ycf2 (16 SSRs each in B. racemosa, B. fusicarpa, and B. acutangula [5.9%]), and rpl16 (5 SSRs in B. racemosa [1.8%], 6 SSRs in B. fusicarpa [2.2%], 5 SSRs in B. acutangula [1.8%]) (Tables S1, S2, S3). The distribution pattern of intragenic SSRs was conserved within the genus. Notably, genes like clpP and ycf1 harbored multiple SSRs either within or near their sequences, suggesting potential for high variability or relevance to adaptive evolution.

From the pie charts in Figs. 4A-C, it is evident that the proportions of chloroplast genomic regions in B. racemosa, B. fusicarpa, and B. acutangula exhibit highly conserved characteristics: The LSC region accounts for the largest proportion in all three species, reaching 66.1%, 65.1%, and 66.3%, respectively. The IR regions are maintained within a narrow range of 15.0% to 15.7%. The SSC region constitutes between 18.7% and 19.2% of the genome. This distribution pattern reflects the strong structural stability and conservation of chloroplast genomes within the genus Barringtonia. The bar charts in Figs. 4D-F reveal distinct regional and sequence element preferences in the distribution of SSRs across the chloroplast genomes of the three Barringtonia species: In each species, the number of SSRs in the LSC region is significantly higher than in the SSC and IR regions. Within the LSC region, the number of SSRs found in intergenic spacer regions far exceeds those located in protein-coding regions and introns. The SSR counts in both the SSC and IR regions are generally lower, with relatively minor differences observed among various sequence elements. This distribution pattern is consistent across the three Barringtonia species, indicating that SSRs are predominantly enriched in the intergenic spacer regions of the LSC area, whereas functional regions (such as protein-coding sequences) show relatively limited SSR accumulation.

Fig. 4.

Comparison of types and distribution of SSRs in the chloroplast genomes of three Barringtonia species. (A-C) Composition (percentage) of different SSR types in B. racemosa, B. fusicarpa, and B. acutangula. (D-F) Number of SSRs located in protein-coding regions, intergenic spacers, and introns within the large single-copy (LSC), small single-copy (SSC), and inverted repeat (IR) regions of the chloroplast genomes of the corresponding three species

Figure 5 illustrates that the SSR repeat unit distributions in B. racemosa, B. fusicarpa, and B. acutangula share common features alongside subtle differences. All three chloroplast genomes were dominated by mononucleotide repeat units, showing a pronounced A/T base preference. Specifically, the peak counts for T-repeat units were 27 in B. racemosa, 28 in B. fusicarpa, and 30 in B. acutangula, while A-repeat units peaked at 25 in both B. racemosa and B. acutangula, and 26 in B. fusicarpa. In contrast, C/G-type mononucleotide repeats were rare, with most types having fewer than three occurrences. Di-, tri-, and longer sequence repeats were generally present at low levels, typically fewer than 10 instances, constituting minor components of the SSR pool.

Fig. 5.

Distribution of SSR Types in Barringtonia Species (A) B. racemosa, (B) B. fusicarpa, (C) B. acutangula. X-axis: SSR repeat unit; Y-axis: Number of occurrences

Figure 6 illustrates the length distribution and quantity characteristics of different types of long repetitive sequences in the chloroplast genomes of three Barringtonia species. In all three species, repeat sequences are predominantly enriched at a core length of 30 bp, with forward (F) and palindromic (P) repeats being the dominant types. B. racemosa (Fig. 6A): At 30 bp, the number of palindromic repeats (P) reaches 7, and forward repeats (F) reach 6, forming the absolute peak for this species. Reverse (R) and complementary (C) repeats at this length are significantly fewer, each occurring only once (Table S4). B. fusicarpa (Fig. 6B): At 30 bp, both forward (F) and palindromic (P) repeats number 6 each; reverse (R) repeats total 2, and complementary (C) repeats total 1 (Table S5). B. acutangula (Fig. 6C): Shows the highest repeat abundance at 30 bp among the three species: palindromic (P) and forward (F) repeats each reach 10, while reverse (R) and complementary (C) repeats also increase to 5 and 3, respectively, forming a pronounced enrichment peak (Table S6). From an interspecies perspective, the repeat sequence distribution patterns exhibit genus-level conservation: All three species show dominance of forward and palindromic repeats at 30 bp, with reverse and complementary repeats being extremely rare. The numbers of repeat sequences at lengths other than 30 bp remain generally low across all species. However, there are also species-specific differences: B. acutangula exhibits notably higher numbers of all repeat types at 30 bp compared to B. racemosa and B. fusicarpa, indicating a stronger degree of repeat sequence enrichment in its chloroplast genome.

Fig. 6.

Numbers of different repeats in the cp. genomes of (A) Barringtonia racemosa and (B) B. fusicarpa and (C)B. acutangula; F, forward repeat; P, palindromic repeat; R, reverse repeat; C, complement repeat

Overall, these findings indicate that chloroplast genomes within Barringtonia have repeat sequences characterized by length specificity (centered around 30 bp) and type preference (favoring F and P repeats), along with moderate interspecific variation in repeat abundance.

Nucleotide diversity and selection pressure analysis

To assess sequence variability across the chloroplast genomes of the three Barringtonia species (B. racemosa, B. fusicarpa, and B. acutangula), we performed a genome-wide sliding window analysis to calculate nucleotide diversity (π). As shown in Fig. 7, π values vary markedly among different genomic regions. The large single-copy (LSC) and small single-copy (SSC) regions exhibit significantly higher π values than the inverted repeat (IR) regions, with the latter consistently showing low π values (mostly ≤ 0.03). This pattern is consistent with the known functional constraints and slower mutation rates of IR regions.

Fig. 7.

Nucleotide polymorphism analysis of the cp. genomes of Barringtonia racemosa, Barringtonia fusicarpa and Barringtonia acutangula. Names of protein-coding genes and genes of the intergenic region are along the X-axis, and the nucleotide diversity (Pi) value in each window is along the Y-axis

At the individual gene level, high π values are predominantly located in the LSC region. Among all protein-coding genes, rpl20 displays the highest π value (0.080), followed by rpoA (0.064), rps3 (0.063), rpl33 (0.058), matK (0.050), and rps4 (0.048). Within the SSC region, ndhF (0.060) and rps15 (0.045) also show elevated polymorphism. In contrast, genes situated in the IR regions, such as rrn5, rps7, and rpl23, have π values below 0.03, reflecting their high sequence conservation.

To further evaluate the selective pressures acting on protein-coding genes, we performed codon-based selection analyses by calculating the ratio of nonsynonymous (Ka) to synonymous (Ks) substitution rates for pairwise comparisons among the three Barringtonia species, as well as between Barringtonia and closely related taxa (Bertholletia excelsa, Barringtonia asiatica, and the out-group Rhododendron simsii). The results are summarized in Table S7.

Within the genus Barringtonia, all pairwise comparisons (e.g., B. acutangula vs. B. fusicarpa, B. acutangula vs. B. racemosa, B. fusicarpa vs. B. racemosa) yielded mean Ka/Ks ratios consistently below 1 (ranging from 0.32 to 0.37). The majority of individual genes in these comparisons also exhibited Ka/Ks < 1, with only a small number of genes (0–4 out of 19–39 analyzed genes) showing Ka/Ks ≥ 1. Furthermore, a substantial proportion of genes (12–16 genes per comparison) displayed statistically significant purifying selection (P < 0.05). Notably, the comparison between B. fusicarpa and B. racemosa, the two most closely related species, showed the lowest mean Ka/Ks (0.32), consistent with their recent divergence and strong functional constraints.

In comparisons between Barringtonia species and the congeneric B. asiatica, mean Ka/Ks values were slightly higher (0.55–0.59) but still well below 1, indicating continued purifying selection. When comparing Barringtonia species with the more distantly related Bertholletia excelsa (same family, Lecythidaceae) or the out-group Rhododendron simsii (Ericaceae), mean Ka/Ks values remained below 0.4, and the vast majority of genes (> 90%) showed Ka/Ks < 1. These results collectively demonstrate that all protein-coding genes in the chloroplast genomes of Barringtonia species are evolving under strong purifying selection, with no evidence of positive selection at the codon level.

It is important to note that nucleotide diversity (π) quantifies sequence variability but does not directly distinguish between neutral evolution and selection. The Ka/Ks analyses presented here provide a more direct assessment of selective pressures. The high π values observed in genes such as rpl20, rpoA, rps3, and ndhF highlight candidate regions that, despite evolving under purifying selection, exhibit elevated polymorphism and may be useful as molecular markers for population genetics within Barringtonia.

Phylogenetic relationships

Figure 8 focuses on the three Barringtonia species (B. acutangula, B. fusicarpa, and B. racemosa, marked with black asterisks) alongside the congeneric species B. asiatica (ON674119.1). These four taxa form a monophyletic clade representing the genus Barringtonia. All internal nodes within this clade exhibit bootstrap support values of 100%, indicating strong statistical support from chloroplast genomic data for both the monophyly of Barringtonia and the close phylogenetic relationships among its constituent species.

Fig. 8.

Maximum likelihood phylogenetic tree constructed based on the complete chloroplast genome sequences of 18 species. The values at the nodes of the tree represent bootstrap support values (unit: %), which range from 0 to 100 and serve to indicate the phylogenetic credibility of the corresponding branches. The black stars (★) denote the positions of the study species (Barringtonia acutangula, B. fusicarpa, and B. racemosa) in this phylogenetic tree. The “0.01” scale bar at the bottom of the figure is the unit of branch length, corresponding to 0.01 nucleotide substitutions per site, which is equivalent to 1% sequence divergence between taxa

Genetic distance analysis further quantifies the divergence among these species. The pairwise p-distances (based on the complete chloroplast genome alignment) are summarized in Fig. 9 (heatmap) and Table S8. Within the Barringtonia clade, the smallest genetic distance is observed between B. racemosa and B. fusicarpa (0.00279), consistent with their sister relationship. Distances between B. acutangula and the other two species are slightly higher: 0.00382 (B. acutangula vs. B. racemosa) and 0.00462 (B. acutangula vs. B. fusicarpa). The congeneric B. asiatica shows distances ranging from 0.00523 (B. asiatica vs. B. racemosa) to 0.00575 (B. asiatica vs. B. fusicarpa), indicating a low but consistent level of divergence across the genus. The mean pairwise distance within Barringtonia is 0.0046 (range: 0.00279–0.00575), confirming very low sequence divergence.

Fig. 9.

Heatmap of pairwise p-distances among the 18 chloroplast genomes. Colors indicate p-distance values from blue (low distance) to orange-red (high distance). Distances were calculated using MEGA12 with the p-distance model

In contrast, the distance between the Barringtonia clade and the closely related Bertholletia excelsa (MF359948.1) is substantially higher, with values ranging from 0.03255 to 0.03356 (mean ~ 0.0329). This is approximately 6-fold higher than the maximum intrageneric distance (0.00575), strongly supporting their placement as distinct genera within the same family. The Barringtonia clade further groups with Bertholletia excelsa to form a higher‑level branch; however, the node supporting this relationship has only 51% bootstrap support, which is statistically weak. Therefore, the sister relationship between Barringtonia and Bertholletia cannot be robustly concluded from the current plastome dataset. The limited taxonomic sampling (only one additional Lecythidaceae representative) and the potentially rapid radiation within Ericales may both contribute to this low support value. Consequently, the phylogenetic position of Barringtonia within Lecythidaceae remains to be further resolved with expanded taxon sampling and additional genomic markers.

The majority of other clades in the tree exhibit high bootstrap support: The clade uniting Pentaphylax euryoides (MW801315.1) and Planchonella grandifolia (ON881598.1) shows 100% support at the primary node and 93% at the secondary node. The clade containing Camellia japonica (KU951523.1), Symplocos chinensis (MW307951.1), and Styrax chinensis (NC057967.1) exhibits 100% bootstrap support at all relevant nodes, confirming the reliability of these inferred phylogenetic relationships.

The out-group taxon, Rhododendron simsii (MT239364.1), occupies the basal position in the tree. Its corresponding branch length is notably longer than those of other taxa (as indicated by the “0.01” scale bar at the base of the tree), reflecting a higher rate of nucleotide substitution relative to other lineages. This placement is consistent with its designated role as an out-group anchoring the tree’s root and underscores the substantial genetic divergence between Barringtonia and distantly related lineages. The average genetic distance between the Barringtonia clade and Rhododendron simsii is 0.173, which is approximately 30-fold larger than the intrageneric distances, confirming the clear taxonomic separation at the family level and above.

Discussion

Conserved architecture and lineage-specific variation in Barringtonia chloroplast genome

This study reports, for the first time, the complete chloroplast genomes of Barringtonia racemosa, B. fusicarpa, and B. acutangula. Analysis reveals that all three species possess the typical quadripartite structure characteristic of angiosperm chloroplast genomes (Fig. 1). Their genome size, GC content, gene number (132 genes), and gene order are highly conserved (Tables 1 and 2), consistent with the structural stability demonstrated by collinearity analysis (Fig. 2). This high degree of macrostructural conservation is a common feature across many plant lineages, suggesting strong purifying selection acting to maintain core functional integrity within the Barringtonia chloroplast genome. For example, comparative studies of multiple mangrove species have similarly found that their chloroplast genomes typically retain an extremely similar quadripartite structure and a highly conserved gene set [14].

Within this overall conserved framework, our study also identifies subtle, lineage-specific variations among the species. Fine-scale comparison of IR boundaries demonstrates that while the gene-spanning patterns at these junctions (e.g., rpl2, ndhF) are highly consistent across the four Barringtonia species, significant expansions or contractions are evident when compared to the closely related Bertholletia excelsa and the out-group (Fig. 3). This indicates that IR boundary architecture exhibits greater evolutionary plasticity between closely related genera than within the genus Barringtonia itself. Similar IR boundary shifts have been documented in other mangrove lineages (e.g., Acanthus, Ceriops) and are often interpreted as lineage-specific evolutionary events or potential phylogenetic markers [19, 20].

Simple sequence repeats (SSRs) and long repetitive sequences are indispensable components of genomes. They serve not only as highly variable molecular markers but also profoundly influence micro-evolutionary processes through complex mechanisms affecting gene expression and mediating genomic structural rearrangements [21, 22]. For instance, Jiang et al. [23] identified SSRs and long repeats primarily enriched in intergenic spacer (IGS) regions in Rosa, and pinpointed several highly variable IGS regions suitable as effective DNA barcodes. In this study, SSRs in the chloroplast genomes of all three Barringtonia species exhibit a pronounced A/T bias and are predominantly concentrated in the intergenic spacer regions of the LSC (Fig. 4). This distribution pattern may reflect weaker selective constraints in non-coding regions, thereby minimizing the potentially deleterious effects of SSR instability on essential gene function. Notably, a substantial number of SSRs are clustered within or near certain functional genes (e.g., clpP, ycf1), suggesting these regions may harbor elevated mutational potential. Furthermore, B. acutangula exhibits slightly higher total numbers of both SSRs and long repeat sequences compared to the other two species (Figs. 5 and 6), implying its chloroplast genome may have experienced more active insertion events of repetitive elements. Such species-specific differences in the repetitive sequence repertoire may reflect divergent population histories or distinct genomic evolutionary dynamics.

In summary, the chloroplast genomes of Barringtonia exhibit high conservation at the macrostructural level, while displaying predictable variation in specific architectural features such as IR boundaries and repetitive sequence distributions. This pattern of “local dynamism within a stable framework” reflects the evolutionary balance between functional constraint and sequence plasticity.

Phylogenetic relationships and monophyly of Barringtonia based on complete chloroplast genomes

The phylogenetic tree constructed in this study using complete chloroplast genome sequences provides a highly supported framework for the phylogeny of Barringtonia (Fig. 8) and offers initial resolution of relationships within the genus. This underscores the powerful efficacy of chloroplast genomes in resolving phylogenetic questions at both the generic and interspecific levels.

This clear phylogenetic topology provides solid molecular evidence to address long-standing taxonomic uncertainties within Barringtonia and lays the foundation for understanding the sequence of speciation events within the genus. Notably, the sister relationship between the two threatened species (B. racemosa and B. fusicarpa) and the basal position of B. acutangula suggest that the lineage leading to the threatened taxa diverged more recently, potentially reflecting different evolutionary responses to coastal habitat pressures. The high bootstrap support (100%) for the Barringtonia clade, combined with the very low genetic distances (mean p-distance = 0.0046, Fig. 9 and Table S8) and uniformly strong purifying selection (Ka/Ks < 1) among species, indicates that the chloroplast genome has remained highly conserved despite species diversification. This pattern implies that adaptive differentiation in these mangroves may be driven more by nuclear genome evolution or by ecological factors rather than by structural changes in the plastome. Thus, placing Barringtonia within the broader phylogenetic context of Ericales not only clarifies its systematic position but also highlights the conservation of chloroplast genomes in this ecologically transitional lineage, providing a baseline for future studies on the molecular basis of adaptation to intertidal environments.

The chloroplast genome data generated in this study, particularly the identification of species-specific simple sequence repeats and the delineation of lineage-specific patterns in IR boundary dynamics, offer valuable molecular resources for future investigations into whether differences in adaptive capacity or demographic history contribute to their disparate extinction risks (Fig. 3). Furthermore, the basal position of B. acutangula within the Barringtonia clade, combined with its relatively widespread distribution, suggests that this lineage may retain ancestral genomic features that could be informative for reconstructing the evolutionary transition from terrestrial to coastal habitats.

At the family level, the sister-group relationship between Barringtonia and Bertholletia excelsa is recovered, but the supporting node exhibits low bootstrap support (51%). It should be noted that complete chloroplast genome sequences are currently available for only two species of Lecythidaceae worldwide: Barringtonia asiatica and Bertholletia excelsa. Despite the family comprising approximately 20 genera and 450 species (Description from Flora of China, http://www.efloras.org/florataxon.aspx?flora_id=3&taxon_id=10482), no other Lecythidaceae plastomes have been publicly released prior to this study. This scarcity of genomic resources reflects the severely understudied status of this tropical family and unavoidably limited our taxonomic sampling. Therefore, the low bootstrap support observed at this node may be partly attributed to this limited sampling. Additionally, rapid evolutionary radiations within Ericales have been reported to challenge phylogenetic resolution even with plastome data [24, 25], which may also contribute to the weak support. This suggests that deeper phylogenetic relationships within Lecythidaceae remain unresolved and require additional taxon sampling and genomic data for clarification. In contrast, the pronounced genetic divergence from the distantly related out-group Rhododendron simsii confirms the reliability of chloroplast genomes for inferring deep-level phylogenetic relationships.

Adaptive evolution and genetic vulnerability: selective pressures and lineage-specific genomic variation in endangered Barringtonia species

Nucleotide diversity (Pi) serves as a core metric of genomic genetic variation [26], while structural variants (SVs) influence adaptation and speciation; their combined analysis enables a comprehensive understanding of adaptive evolution [21, 27]. Our results reveal significant heterogeneity in selective pressures acting on different regions and genes within the Barringtonia chloroplast genome (Fig. 7). As expected, the IR region exhibits the highest conservation due to its sequence correction mechanisms, whereas the LSC and SSC regions display elevated levels of variation. Notably, several genes encoding ribosomal proteins (e.g., rpl20, rpl33, rps3) as well as genes involved in transcription (rpoA) and photosynthesis (ndhF) exhibit high Pi values (> 0.04).

Beyond their utility as DNA barcodes, these highly polymorphic genes may have biological functions relevant to environmental adaptation. For instance, ndhF encodes a subunit of the NADH dehydrogenase complex, which participates in chloro-respiration and cyclic electron transport around photosystem I, processes that help mitigate oxidative stress under high light and salinity [14, 20]. The high nucleotide diversity observed in ndhF among Barringtonia species could thus reflect local adaptation to varying light and salinity conditions across their coastal habitats. Similarly, rpl20 and rps3 are components of the plastid ribosome; their elevated polymorphism might influence translation efficiency under stress, while rpoA (RNA polymerase subunit) could affect the transcriptional regulation of photosynthetic genes in response to environmental cues. Although direct evidence linking these polymorphisms to adaptive traits is lacking, these candidate genes warrant further functional investigation.

These highly variable genes or intergenic regions represent ideal candidate loci for developing high-resolution DNA barcodes applicable primarily within the genus Barringtonia, as their broader applicability across the family Lecythidaceae or beyond requires further validation. In particular, it should be noted that some of these genes (e.g., ndhF) have been reported to be lost or pseudogenized in certain other plant lineages, which may limit their universal utility [14, 20]. This finding is consistent with conclusions from multiple studies on mangrove chloroplast genomes [14]. For example, numerous studies report that most protein-coding genes in mangrove chloroplast genomes are under purifying selection, with Ka/Ks ratios generally below 1. However, some studies have detected signatures of positive or relaxed selection on individual genes (e.g., ycf2, rbcL, accD) in certain mangrove species, suggesting potential roles in adaptation to the intertidal environment, energy metabolism, or stress response [20]. We did not detect clear, strong signals of positive selection in the endangered Barringtonia species analyzed here. This may indicate that adaptive differentiation at the chloroplast genome level is not pronounced, or that key adaptations are more likely mediated through the nuclear genome or gene expression regulatory mechanisms.

Our Ka/Ks analyses further substantiate the overall pattern of strong purifying selection across the Barringtonia chloroplast genome (Table S7). Within the genus, all pairwise comparisons yielded mean Ka/Ks ratios well below 1 (0.32–0.37), with the closest relatives (B. fusicarpa and B. racemosa) showing the lowest value (0.32). The vast majority of individual genes were under purifying selection, and a substantial proportion (12–16 per comparison) showed statistically significant Ka/Ks < 1 (P < 0.05). Notably, even comparisons between Barringtonia species and the more distantly related Bertholletia excelsa (same family) or Rhododendron simsii (out-group) yielded mean Ka/Ks < 0.4, with > 90% of genes under purifying selection. These results collectively demonstrate that no positive selection was detected at the codon level, which is consistent with the high functional conservation expected for chloroplast genomes.

The absence of positive selection in the Barringtonia plastome may be explained by several non-mutually exclusive factors. First, chloroplast genomes are generally highly conserved and evolve under strong purifying selection due to their essential roles in photosynthesis and energy production [12, 13]. Second, adaptive processes in response to salinity and flooding may primarily involve nuclear-encoded genes rather than plastid genes, as seen in other mangroves [28]. Third, our limited taxon sampling within the genus (only three species) reduces the statistical power to detect positive selection, and the relatively recent divergence of these species may not have allowed sufficient time for adaptive substitutions to accumulate in the plastome. Thus, the lack of positive selection signals does not rule out adaptive evolution; it may simply be occurring in the nuclear genome or through regulatory mechanisms.

Critically, we examined the chloroplast genomic characteristics of the endangered B. racemosa and B. fusicarpa relative to the near-threatened B. acutangula from a conservation genetics perspective. The three genomes are highly conserved, with no gene loss or major structural rearrangements, and no specific chloroplast variants were found that directly correlate with endangered status. Therefore, the endangered status of B. racemosa and B. fusicarpa cannot be directly attributed to their chloroplast genome features. Although subtle differences were observed (e.g., a greater abundance of repetitive sequences in B. acutangula), these do not explain the disparity in conservation status.

Chloroplast genomes are maternally inherited and evolve under strong purifying selection; thus, they may not capture the nuclear genetic diversity or inbreeding effects that contribute to endangerment. Because the nuclear genome was not analyzed here, we refrain from concluding that genetic vulnerability is linked to the nuclear genome, such a claim would exceed the evidence. It should be noted that the plant materials used in this study were collected from cultivated individuals; therefore, inferences about natural population history and genetic vulnerability should be made with caution and require validation using wild populations and nuclear genomic data. Instead, future studies should integrate nuclear genomic diversity, demographic history, mating systems, and ecological data to fully elucidate the mechanisms of endangerment in these species. This aligns with the principle that maternally inherited chloroplast markers alone cannot fully assess genetic vulnerability, as emphasized by Asaf et al. [28] for Avicennia marina.

Conclusions

This study reports the first complete chloroplast genomes of three Barringtonia species: the endangered B. racemosa and B. fusicarpa, and the near-threatened B. acutangula. All three genomes exhibit the canonical quadripartite structure, high collinearity, and conserved gene content (132 genes), with no large-scale rearrangements or gene loss. Species-specific variations are evident in fine-scale features, including the distribution of simple sequence repeats (SSRs), long repeats, and inverted repeat (IR) boundary configurations.

Phylogenetic analysis robustly supports the monophyly of Barringtonia (100% bootstrap support), with B. racemosa and B. fusicarpa forming a sister lineage to B. acutangula. Genetic distances among the four Barringtonia species are extremely low (mean p-distance = 0.0046), and all protein-coding genes evolve under strong purifying selection (mean Ka/Ks = 0.32–0.37), with no evidence of positive selection.

Importantly, no chloroplast genomic variants were found that directly correlate with the endangered status of B. racemosa and B. fusicarpa. Their endangerment is therefore unlikely to stem from functional impairment of the plastome, but rather may be driven by erosion of nuclear genomic diversity or population demographic factors. The highly polymorphic regions identified (e.g., rpl20, rpoA, rps3, ndhF) and the SSR markers developed here provide valuable tools for future population genetics and conservation studies.

This study establishes a critical genomic foundation for systematics, species identification, and conservation-oriented genetic management of Barringtonia. Future research should integrate nuclear genomic data, population resequencing, and ecological information to fully elucidate the mechanisms underlying genetic vulnerability in these threatened coastal species.

Materials and methods

Plant materials

This study focused on three Barringtonia species: Barringtonia racemosa (L.) Spreng., Barringtonia fusicarpa H. H. Hu, and Barringtonia acutangula (L.) Gaertn. Healthy, mature leaves were collected from introduced plants cultivated in the green space of the living area at the Subtropical Crops Research Institute of Guangxi Zhuang Autonomous Region, China. As all plant materials were obtained from artificially cultivated individuals rather than directly from the wild, no specific collection permits were required. This study complied fully with all relevant institutional and national regulations.

Species identification was carried out by Professor Taiping He, a plant taxonomist at Guangxi University. Voucher specimens of the three species have been deposited in the Southeast Guangxi Distinctive Medicinal Plant Herbarium, Yulin Normal University (coordinates: 22°41′N, 110°12′E), under the following accession numbers: B. racemosa (YLSY-BR004), B. fusicarpa (YLSY-BF001), and B. acutangula (YLSY-BA001). Collected leaf tissues were immediately flash-frozen in liquid nitrogen and stored at − 80 °C until DNA extraction.

DNA extraction, and sequencing

Total genomic DNA was extracted from leaf tissue using a modified cetyltrimethylammonium bromide (CTAB) protocol [29]. DNA quality and concentration were assessed via agarose gel electrophoresis and NanoDrop spectrophotometry. The final DNA concentrations for the three species were 56.62 ng/µL (B. racemosa), 35.36 ng/µL (B. fusicarpa), and 49.26 ng/µL (B. acutangula). Qualified DNA samples were fragmented to a target size using a Covaris ultrasonicator. Subsequent steps included size selection and purification, end repair, A-tailing, and adapter ligation to construct Illumina sequencing libraries [30]. Finally, paired-end 150 bp (PE150) sequencing was performed on an Illumina NovaSeq platform at BMK Biotechnology Co., Ltd., Beijing, China.

Chloroplast genome assembly and annotation

Raw sequencing reads were initially subjected to quality control using fastp v0.20.0 [31]. This step removed adapter sequences, primer contaminants, and low-quality reads defined as those with an average Phred quality score below Q5 or containing more than five ambiguous nucleotides (N’s). After quality filtering, clean reads for each sample were mapped to the reference chloroplast genome of a closely related species using BWA v0.7.17 to assess sequencing depth and coverage uniformity. The sequencing depth and genome coverage statistics were summarized in a table, including the number of cleaned reads (×10⁶), total clean data (Gb), number of mapped reads (×10⁶), mapping rate (%), average coverage (×), proportion of genome positions covered by at least 1× (%), assembled genome length (bp), GC content (%), and mean insert size (bp) with standard deviation (Table S9). These metrics provide a comprehensive evaluation of assembly reliability.

High-quality reads were then assembled de novo into chloroplast contigs using SPAdes v3.10.1 [32], with k-mer sizes set to 55, 87, and 121 to optimize assembly across varying sequence complexities. The resulting contigs were scaffolded using SSPACE v2.0 [33], and potential gaps within scaffolds were closed using GapFiller v2.1.1 [34], yielding complete circularized chloroplast genome sequences. Chloroplast genome annotation was performed by integrating ab initio gene prediction with homology-based alignment. Protein-coding genes were predicted using Prodigal v2.6.3, ribosomal RNA (rRNA) genes were identified via HMMER v3.1b2 [35], and transfer RNA (tRNA) genes were detected using ARAGORN v1.2.38 [36]. Concurrently, chloroplast genome sequences of closely related species were downloaded from the NCBI database and aligned against our assemblies using BLAST v2.6 [37] to validate gene content and positions. The preliminary annotations generated by both methods were manually curated to resolve discrepancies, refine gene boundaries (particularly for multi-exon genes), and eliminate redundant or spurious annotations. Final annotated genomes were visualized using the online tool OrganellarGenomeDRAW (OGDRAW) version 1.3.1 [38]. The three newly assembled chloroplast genome sequences have been submitted to the GenBank database (https://www.ncbi.nlm.nih.gov/genbank/) (Accession numbers PX987919.1, PX987920.1, and PX987918.1). To assess structural conservation and detect potential rearrangements, global alignments and synteny analyses among Barringtonia species and related taxa were conducted using Mauve [39].

Analysis of SSRs, nucleotide polymorphisms, selection pressure and ir boundary changes

Simple sequence repeats (SSRs) were identified using MISA v1.0 [40] with the following minimum repeat thresholds: mononucleotide repeats ≥ 8 units, dinucleotide repeats ≥ 5 units, and tri-, tetra-, penta-, and hexanucleotide repeats ≥ 3 units. Longer repetitive sequences-including forward, palindromic, reverse, and complement repeats-were detected using Vmatch v2.3.0 (in conjunction with custom Perl scripts) [41]. The search parameters were set to a minimum repeat length of 30 bp and a Hamming distance of 3 to allow for limited mismatches within repeated motifs. To evaluate sequence variation within the genus Barringtonia, nucleotide diversity (π) was calculated across the entire chloroplast genome and in specific genomic regions using DnaSP v5.10 [42]. To assess selective pressures acting on protein‑coding genes, we performed codon‑based selection analyses by calculating the ratio of nonsynonymous (Ka) to synonymous (Ks) substitution rates. Pairwise comparisons were conducted among the three Barringtonia species, as well as between each Barringtonia species and closely related taxa, including Bertholletia excelsa (MF359948.1), Barringtonia asiatica (ON674119.1), and the out-group Rhododendron simsii (MT239364.1). Protein-coding gene sequences were aligned using MAFFT v7.427 with the “auto” strategy, and poorly aligned regions were removed using trimAl v1.4.rev15 with the 15.rev15ligned using MKa and Ks values were calculated using the Nei-Gojobori method (with Jukes-Cantor correction) implemented in KaKs_Calculator v3.0. Only genes with valid Ka/Ks ratios were included in the analysis. Statistical significance of purifying selection (Ka/Ks < 1) was assessed using Fisher’s exact test, with P < 0.05 considered significant. The resulting Ka/Ks ratios and related statistics are summarized in Supplementary Table S7. The expansion and contraction dynamics of the IR boundaries relative to the large LSC and small SSC regions were visualized and compared among Barringtonia species using the online tool IRscope [43].

Phylogenetic and genetic distance analysis

To determine the phylogenetic placement of the three Barringtonia species within the genus and the order Ericales, we assembled a dataset of 18 complete chloroplast genomes. This dataset includes the three newly sequenced genomes, the congeneric Barringtonia asiatica (ON674119.1), and 14 additional plastomes from representatives of Ericales retrieved from NCBI GenBank. Due to the current scarcity of publicly available Lecythidaceae chloroplast genomes-only Barringtonia asiatica and Bertholletia excelsa (MF359948.1) were accessible at the time of this study, we expanded our sampling to include multiple families within Ericales to provide a broader phylogenetic context. The ingroup thus comprises the four Barringtonia species, Bertholletia excelsa (Lecythidaceae), and representatives of the following Ericalean families: Styracaceae (Styrax chinensis), Actinidiaceae (Actinidia chinensis), Clethraceae (Clethra fargesii), Theaceae (Camellia japonica), Symplocaceae (Symplocos chinensis), Sapotaceae (Planchonella grandifolia), Ebenaceae (Diospyros kaki), Pentaphylacaceae (Pentaphylax euryoides), Primulaceae (Primula sinensis), Balsaminaceae (Impatiens balsamina), Fouquieriaceae (Fouquieria diguetii), and Polemoniaceae (Polemonium chinense). As the out-group, we selected Rhododendron simsii (Ericaceae, Ericales) to root the tree.

All 18 chloroplast genome sequences were aligned using MAFFT v7.505 under the“auto” strategy [44]. The alignment was subsequently trimmed with trimAl v1.4.rev15 using the “automated1” setting to remove poorly aligned regions. A maximum likelihood (ML) phylogenetic tree was inferred using RAxML v8.2.10 under the GTRGAMMA substitution model, with nodal support assessed by 1,000 bootstrap replicates [45]. To quantify the genetic divergence among species, pairwise genetic distances (p-distances) were calculated based on the trimmed alignment using MEGA12 under the p-distance model, with all ambiguous positions removed for each sequence pair.

Supplementary Information

Supplementary Material 1. ^{(185.8KB, docx)}

Authors’ contributions

Conceptualization, YP.Y., XH.T. and F.L.; methodology, XH.T. and F.L.; validation, YP.Y., XH.T. and F.L.; formal analysis, XH.T. and F.L.; investigation, YJ.W., XY.H., LY.M. and Q.Q.; resources, YQ.T., YJ.Q. and L.L.; data curation, XH.T. and F.L.; writing-original draft preparation, F.L.; writing—review and editing, XH.T. and YJ.W.; visualization, YJ.W. and XY.H., supervision, YP.Y.; project administration, YP.Y.; funding acquisition, YP.Y. and F.L. All authors have read and agreed to the submitted version of the manuscript.

Funding

This work was supported by the following projects: The Natural Science Foundation Joint Special Project of Guangxi Province, China (Grant No. 2026GXNSFHA00640042); The Special Project for Basic Scientific Research of Guangxi Academy of Agricultural Sciences (Grant No. Gui Nongke 2024YP134); Yulin Normal University High-Level Talents Scientific Research Start-up Fund Project (grant No. G2025ZK13); The Stable Funding Research Team Project of Guangxi Academy of Agricultural Sciences (Grant No. Gui Nongke 2026YT015); The Project for Enhancing Young and Middle-aged Teachers’ Research Basic Ability in Colleges of Guangxi (Grant No. 2024KY0591).

Data availability

The complete chloroplast genome sequences of Barringtonia racemosa (L.) Spreng., Barringtonia fusicarpa H. H. Hu, and Barringtonia acutangula (L.) Gaertn. have been deposited in GenBank (NCBI) under the accession numbers PX987919.1, PX987920.1, and PX987918.1, respectively.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.