Comparative plastome analyses of Lewinskya (Orthotrichaceae): insights into genome structure, molecular evolution, and phylogenetic relationships

Abstract

Background

Mosses are key components of terrestrial ecosystems and provide important systems for studying plant diversity, adaptation, and genome evolution. Lewinskya is a species-rich moss genus in Orthotrichaceae, but species delimitation and phylogenetic reconstruction within the genus remain difficult because diagnostic characters are often subtle or convergent. Chloroplast genomes can provide useful genomic resources and complementary evidence for comparative and systematic studies. This study aimed to generate new Lewinskya plastome resources and evaluate plastome structure, sequence variation, codon usage, and plastid-based phylogenetic relationships in the genus.

Results

Five newly sampled Lewinskya chloroplast genomes were assembled from genome-skimming data, including three circular plastome assemblies and two high-quality single-scaffold assemblies. Together with the published plastome of L. incana, the six Lewinskya plastomes ranged from 122,258 to 123,526 bp and showed conserved genome organization, gene content, GC composition, and inverted repeat boundaries. Each plastome encoded 128 genes, including 83 protein-coding genes, 37 transfer RNA genes, and eight ribosomal RNA genes. A total of 520–542 simple sequence repeats were detected per plastome, with mononucleotide repeats being dominant and most repeats located in the large single-copy region. Comparative analyses revealed no large-scale rearrangements, but several localized divergence regions were detected. Nucleotide diversity analysis identified 11 highly variable regions, including five genic regions (rps18, rpl22, infA, rpl32 and rps3) and six intergenic spacers, most of which were located in the large single-copy region. Codon usage patterns were highly similar among species and showed a preference for A/T-ending codons. Phylogenetic analyses based on 78 plastid protein-coding genes from 26 Orthotrichaceae plastomes strongly supported the sampled Lewinskya species as a clade, although some deeper relationships within the genus remained weakly resolved.

Conclusions

The newly assembled Lewinskya plastomes expand genomic resources for Orthotrichaceae and show that chloroplast genome evolution in the sampled species is structurally conservative but contains informative localized variation. The identified repeat loci and highly variable regions provide candidate markers for future species identification and population-level studies. Plastome-scale data offer useful evidence for Lewinskya systematics, but broader taxon sampling and integration with nuclear genomic and morphological evidence will be needed to resolve difficult interspecific relationships.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12870-026-09510-w.

Background

Mosses (Bryophyta) are key components of terrestrial ecosystems, contributing substantially to primary production, water retention, nutrient cycling, and microhabitat formation, and many lineages are widely used as sensitive biomonitors of atmospheric contamination and environmental change [1–5]. Recent bryophyte genomic studies have further expanded our understanding of moss diversity, phylogeny, and adaptation, especially through genome-, transcriptome-, and super-pangenome-based analyses [6, 7]. Chromosome-level moss genome assemblies and organelle genome studies have also provided new insights into stress-response mechanisms, mitochondrial genome evolution, and plastid phylogenomic relationships in bryophytes [8–11]. Lewinskyinae F. Lara, Garilleti & Draper is a subtribe of mosses in the moss family Orthotrichaceae Arn., representing one of two major lineages (with Orthotrichinae F. Lara, Garilleti & Draper) in tribe Orthotricheae Engler [12, 13]. Following recent generic realignments, Lewinskya F. Lara, Garilleti & Goffinet is recognized as a distinct epiphytic genus within Lewinskyinae [12–14]. It is globally distributed and comprises approximately 75 recognized taxa [15], with notable diversity in East Asia, including 14 species and two varieties currently reported from China [16]. Ecologically, Lewinskya species are primarily corticolous epiphytes (occasionally saxicolous), and their broad climatic amplitude makes the genus relevant for studies of niche adaptation, dispersal, and environmental sensitivity [14, 17–22].

Despite the clarified generic framework, Lewinskya remains taxonomically challenging because many diagnostic characters are subtle, developmentally plastic, or convergent across lineages. Orthotrichum Hedw. was historically treated as a broadly circumscribed, morphologically heterogeneous assemblage [23–25]. Molecular phylogenetic analyses subsequently revealed Orthotrichum sensu lato (s.l.) to be polyphyletic [26–29]. This revision resulted in the delineation of four distinct genera: Lewinskya, Nyholmiella Holmen & E. Warncke, Orthotrichum sensu stricto (s.str.) and Pulvigera Plášek, Sawicki & Ochyra [14, 29–32]. Although Lewinskya is morphologically diagnosable (e.g., superficial/phaneroporous stomata versus cryptoporous stomata in Orthotrichum s.str.) [14, 32], extensive character overlap among closely related species persists. Recent integrative studies further suggest that pseudo-cryptic diversity and complex biogeographic histories can obscure species limits in Lewinskya, even when combined morphological and genomic evidence is applied [22], underscoring how morphological conservatism can mask true diversity in Orthotrichaceae [33].

Prior molecular systematics have improved higher-level delimitation but often fall short of resolving intrageneric relationships in Lewinskya. Traditional Sanger-based phylogenies that rely on a small set of loci—typically nuclear ribosomal regions (e.g., ITS) and a few plastid fragments (e.g., rps4, trnG, trnL-F)—frequently yield low support, unstable topologies, or insufficient signal for recent radiations and closely allied taxa [12, 27, 34]. Even multilocus frameworks within Orthotricheae have historically sampled only a handful of nuclear and plastid markers, limiting their ability to robustly infer species-level branching patterns [12, 35]. More recent genomic approaches, including target enrichment-based datasets [13, 35] and Genotyping-by-Sequencing (GBS) methods [22], have expanded the number of nuclear loci dramatically, yet challenges remain in consistently recovering well-supported relationships among Lewinskya species [13, 22]. These unresolved relationships constrain downstream inference on character evolution, species boundaries, and historical biogeography within the genus.

Complete chloroplast genomes (plastomes) provide an additional, information-rich system that can complement nuclear datasets in taxonomically difficult groups. Plastomes are typically organized as a quadripartite structure comprising large and small single-copy regions separated by inverted repeats, and they encode a largely conserved gene set, providing thousands of homologous characters for phylogenomic inference [36–38]. In mosses, plastomes generally retain a broadly conserved architecture while still exhibiting informative variation in gene content, repeats, and substitution patterns across lineages [38–40]. Beyond phylogenetic reconstruction, plastome-scale datasets facilitate comparative analyses of simple sequence repeats (SSRs), dispersed repeats, IR boundary dynamics, and mutation “hotspots”, all of which are valuable for developing lineage-specific markers [41, 42]. The use of whole plastomes as “super-barcodes” has shown strong discriminatory power in multiple plant systems [43], and a similar strategy may help overcome the limited signal of single-locus barcodes in Lewinskya. At the same time, plastid inheritance is typically uniparental and can be decoupled from species history via introgression (“chloroplast capture”) or cytonuclear discordance, so plastome-based inferences should be interpreted as a complementary genomic perspective rather than a substitute for nuclear evidence [44, 45].

Against this backdrop, plastome resources for Lewinskya remain limited and unevenly annotated, limiting both comparative genomics and plastid-based marker development. The present study aims to (1) generate and annotate new chloroplast genome assemblies for five Lewinskya species through genome-skimming sequencing; (2) conduct comprehensive comparative plastome analyses, including SSR profiling, repeat characterization, codon usage analysis, and IR boundary assessment; (3) identify mutation hotspots (highly variable regions) as candidate molecular markers for species identification and population-level applications; and (4) reconstruct a plastome-informed phylogenetic framework and evaluate interspecific relationships within Lewinskya from a plastid genomic perspective. Collectively, these objectives will expand foundational genomic resources for Orthotrichaceae and provide a plastid genomic framework and candidate molecular markers that may facilitate future efforts to resolve systematic uncertainties in the species-rich genus Lewinskya.

Materials and methods

Plant sampling and DNA extraction

Leaf samples from five Lewinskya species (L. laevigata (J.E. Zetterst.) F. Lara, Garilleti & Goffinet, L. pylaisii (Brid.) F. Lara, Garilleti & Goffinet, L. rupestris (Schleich. ex Schwägr.) F. Lara, Garilleti & Goffinet, L. sordida (Sull. & Lesq.) F. Lara, Garilleti & Goffinet, and L. speciosa (Nees) F. Lara, Garilleti & Goffinet) were collected for chloroplast genome sequencing. Voucher specimens were identified by Dr. Qinghua Wang from the Institute of Botany, Chinese Academy of Sciences (PE) and deposited in the Herbarium of the Institute of Botany, Chinese Academy of Sciences (PE) and the Herbarium of the Kunming Institute of Botany, Chinese Academy of Sciences (KUN) (Table S1). Total genomic DNA was extracted from dried gametophytic or sporophytic tissue using a modified cetyltrimethylammonium bromide (CTAB) protocol [46]. DNA quality and quantity were assessed by agarose gel electrophoresis and fluorometric quantification.

The five newly assembled Lewinskya plastomes were used together with the publicly available circularly assembled and annotated plastome of L. incana NC042174 [47] for comparative genomic analyses, repeat analyses, nucleotide diversity analysis, and codon usage analysis. This accession was selected as the reference comparative plastome because it provided a complete annotated circular plastome from the same genus.

Sequencing, assembly and annotation

Genomic DNA was fragmented by sonication to an average insert size of ~ 350 bp, and paired-end libraries were constructed following Illumina and DNBSEQ-T7 protocols. Sequencing was performed on the NovaSeq and DNBSEQ-T7 platforms at Novogene (Beijing, China), generating 150-bp paired-end reads. Raw sequencing reads were first quality-checked using fastp [48]. Clean reads were then assembled into chloroplast genomes with GetOrganelle [49], and the assembly was verified by visualizing contigs with Bandage [50]. All sequences were annotated using CPGAVAS2 [51] and cross-validated with GeSeq [52]. Manual adjustments were conducted in UGENE [53] where necessary. Circular genome maps were produced using OGDRAW [54]. Finally, the annotated sequences of the newly assembled chloroplast genomes were submitted to NCBI GenBank under accession numbers PV998914-PV998918.

SSR and repeat sequence analysis

Three types of repeat sequences were identified in the Lewinskya chloroplast genomes: simple sequence repeats (SSRs), tandem repeats and long repeats. SSRs were detected using MISA v2.1 [55]. The minimum repeat thresholds were set to 8 for mononucleotides, 5 for dinucleotides, and 3 for tri-, tetra-, penta-, and hexanucleotides. Tandem repeats were identified using Tandem Repeats Finder v4.09 [56] with default parameters. Long repeats, including forward (F), reverse (R), complement (C), and palindromic (P) types, were detected using CPStools [57] with a minimum repeat size of 30 bp and a Hamming distance of 3. SSR loci were further mapped to annotated plastome coordinates and classified as located in coding sequences (CDS), intergenic spacer regions (IGS), genic non-CDS regions, or RNA genes, and CDS-located SSRs were additionally summarized by functional gene group.

Comparative genomic analysis

To investigate expansion and contraction of the inverted repeat (IR) boundaries, genes at the junctions of the large single-copy (LSC), small single-copy (SSC), and IR regions were analyzed and visualized using IRscope [58]. Junction plots were further cross-checked using CPJSdraw [59]. Gene arrangement and collinearity across Lewinskya chloroplast genomes were assessed using Mauve [60]. Sequence divergence among the chloroplast genomes of Lewinskya species was analyzed using mVISTA [61] in LAGAN mode. Nucleotide diversity (π) across the Lewinskya chloroplast genomes was calculated using DnaSP v6.10 [62]. A sliding window analysis with a window length of 600 bp and a step size of 200 bp was conducted to identify regions of high genetic variability. Regions with π > 0.010 in genic regions and π ≥ 0.050 in intergenic spacer regions were treated as highly variable regions. These thresholds were chosen to highlight the upper tail of variability in the present dataset rather than to define universal cut-off values.

Codon usage bias

Codon usage bias was examined using CPStools [57]. Coding sequences (CDS) of at least 300 base pairs (bp) were extracted from the plastome sequences. For genes duplicated in the IR regions, only one copy was retained to avoid overrepresentation. Relative synonymous codon usage (RSCU) values were calculated for the remaining CDS. To facilitate direct comparison among species, RSCU values were visualized using the online RSCU-Plot platform (https://pcg-lab.shinyapps.io/RSCU-Plot/) [63], following its recent use in chloroplast genome comparative analyses [64]. To further assess factors influencing codon usage bias, PR2-plot, ENC-plot, and neutrality plot analyses were performed for CDSs ≥ 300 bp. In the PR2-plot, G3/(G3 + C3) was plotted against A3/(A3 + T3). In the ENC-plot, ENC values were plotted against GC3s with the expected ENC curve. In the neutrality plot, GC12 was regressed against GC3.

Phylogenetic analysis

A total of 26 chloroplast genome sequences of Orthotrichaceae were sampled for phylogenetic reconstruction (Table 1, Table S2). Macrocoma tenuis subsp. sullivantii (Müll. Hal.) Vitt was selected as the outgroup based on previous phylogenetic studies [12, 13], and taxon details are listed in Tables S1 and S2. For chloroplast genomes lacking complete annotation information, protein-coding genes were re-annotated using the chloroplast genome of L. incana (NC042174) as the reference. Protein-coding sequences (CDSs) were then extracted from all 26 plastomes and grouped by gene name. Each CDS was aligned separately using MAFFT v7 [65] with the auto strategy. Ambiguously aligned and gap-rich regions were removed from each gene alignment using trimAl v1.2 [66] with the -gappyout option. The trimmed alignments were concatenated into a plastid CDS supermatrix. The final concatenated matrix included 26 taxa, 78 protein-coding genes, with a total aligned length of 58,023 bp.

Table 1.

Plastome features of the five newly sequenced Lewinskya accessions and the published annotated plastome of L. incana NC042174. IRs, the combined length of IRa and IRb. Assembly topology: C, circular plastome assembly; S, single-scaffold plastome assembly

Taxa	GenBank accession No.	Length (bp)				GC content (%)				Number of genes				Assembly topology
		Total	LSC	SSC	IRs	Total	LSC	SSC	IRs	Total	tRNA	rRNA	CDS
L. incana	NC042174	123,464	85,222	18,514	19,728	28.4	25.6	25.0	43.7	128	37	8	83	C
L. laevigata	PV998915	123,017	86,005	18,458	18,554	28.4	25.6	25.1	44.7	128	37	8	83	C
L. pylaisii	PV998916	123,077	86,092	18,443	18,542	28.3	25.6	25.1	44.8	128	37	8	83	C
L. rupestris	PV998914	122,258	85,280	18,440	18,538	28.5	25.6	25.5	44.1	128	37	8	83	S
L. sordida	PV998917	123,273	85,348	18,415	19,510	28.4	25.6	25.0	43.9	128	37	8	83	C
L. speciosa	PV998918	123,526	85,262	18,438	19,826	28.4	25.6	25.0	43.5	128	37	8	83	S

Maximum likelihood (ML) phylogenetic analysis was performed using IQ-TREE [67]. The best-fit substitution models were selected using ModelFinder with the MFP+MERGE option under the Bayesian information criterion, and the GTR + F+I+R3 model was selected for the ML analysis. Branch support was assessed using 1,000 SH-aLRT replicates and 1,000 ultrafast bootstrap replicates [68]. Bayesian inference (BI) was conducted using MrBayes v3.2 [69] under the GTR + F+I+G4 model. Two independent runs were performed for 2,000,000 generations with four Markov chains, sampling every 1,000 generations. The first 25% of trees were discarded as burn-in. Convergence was assessed by confirming that the average standard deviation of split frequencies was below 0.01. The resulting phylogenetic trees were visualized and edited using FigTree v1.4.4 [70].

Results

Sequencing output and plastome assembly quality

Genome-skimming sequencing of the five newly sampled Lewinskya accessions generated 115,604,686–205,638,244 raw reads per accession, corresponding to 17.34–30.85 Gb of raw data (Table S3). After quality filtering, 115,380,716–204,939,600 clean reads were retained, yielding 17.31–30.74 Gb of clean data. The effective rate ranged from 99.44% to 99.84%, with low error rates of 0.02–0.05%. The Q20 and Q30 values were 97.27–99.13% and 90.40–97.20%, respectively, indicating high sequencing quality.

Plastome assembly evaluation showed that the five newly generated plastome assemblies had high read support and nearly complete plastome coverage (Table S4). Three accessions, L. laevigata PV998915, L. pylaisii PV998916, and L. sordida PV998917, were recovered as circular plastome assemblies, whereas L. rupestris PV998914 and L. speciosa PV998918 were recovered as single-scaffold plastome assemblies. The average plastome k-mer coverage ranged from 63.5 to 196.7, and the average plastome base coverage ranged from 271.2× to 641.3×. Plastome coverage was 99.99–100%, with zero or one zero-depth base and no ambiguous bases detected in the final assemblies. These results support the reliability of the newly assembled plastomes for subsequent comparative genomic and phylogenetic analyses.

Plastome structure and features

The five newly sequenced Lewinskya plastome assemblies ranged from 122,258 bp in L. rupestris PV998914 to 123,526 bp in L. speciosa PV998918 (Table 1). The published annotated circular plastome of L. incana NC042174 was also included for comparison. Across the six accessions, plastome lengths ranged from 122,258 to 123,526 bp. The large single-copy (LSC) regions were 85,222–86,092 bp, the small single-copy (SSC) regions were 18,415–18,514 bp, and the combined lengths of IRa and IRb were 18,538–19,826 bp. Overall GC content ranged from 28.3% to 28.5%, with higher GC content in the IR regions (43.5–44.8%) than in the LSC (25.6%) and SSC regions (25.0–25.5%).

Plastome maps were generated for the five newly sequenced accessions and L. incana NC042174 (Fig. 1). Circularity was confirmed for L. laevigata PV998915, L. pylaisii PV998916, and L. sordida PV998917, whereas L. rupestris PV998914 and L. speciosa PV998918 were retained as high-quality single-scaffold assemblies with no ambiguous bases, complete gene sets, and clearly defined LSC/SSC/IR boundaries (Table 1 and S4).

Fig. 1.

Plastome maps of the five newly sequenced Lewinskya accessions and the published circular plastome of L. incana NC042174. Circularity was confirmed for L. laevigata PV998915, L. pylaisii PV998916, and L. sordida PV998917; L. rupestris PV998914 and L. speciosa PV998918 were recovered as high-quality single-scaffold plastome assemblies. For L. rupestris and L. speciosa, circular maps are shown for visualization and do not indicate confirmed circularization. Within the circle, darker grey shades indicate GC content, whereas lighter grey shades indicate AT content. The plastid genome boundaries are divided into the LSC, SSC, IRa, and IRb regions. Genes from various functional groups are color-coded

All five newly sequenced plastomes encoded 128 genes, including 83 protein-coding genes, 37 tRNA genes, and eight rRNA genes. The same gene complement was present in L. incana NC042174 (Tables 1 and 2). Eighteen genes contained introns: 15 genes had one intron, whereas clpP, rps12, and ycf3 had two introns. Overall, the sampled Lewinskya plastomes showed conserved genome size, GC content, structure, and gene content.

Table 2.

Gene types and gene numbers of Lewinskya species

	Gene group	Gene name	Number
Photosynthesis	Photosystem I	psaJ, psaI, psaA, psaB, psaM, psaC	6
	Photosystem II	psbH, psbN, psbT, psbB, psbE, psbF, psbL, psbJ, psbZ, psbC, psbD, psbA, psbK, psbI, psb30, psbM	16
	NADH dehydrogenase	ndhC, ndhK, ndhJ, ndhBa, ndhH, ndhAa, ndhI, ndhG, ndhE, ndhD, ndhF	11
	Cytochrome b/f complex	petDa, petBa, petG, petL, petA	5
	Large subunit of rubisco	rbcL	1
	ATP synthase	atpB, atpE, atpA, atpFa, atpH, atpI	6
	Subunits photochlorophyllide reductase	chlB, chlL, chlN	3
Self-replication	Proteins of large ribosomal subunit	rpl23, rpl2a, rpl22, rpl16a, rpl14, rpl36, rpl20, rpl33, rpl32, rpl21	10
	Proteins of small ribosomal subunit	rps19, rps3, rps8, rps11, rps12b, rps18, rps4, rps14, rps2, rps7, rps15	11
	Subunits of RNA polymerase	rpoC2, rpoC1a, rpoB	3
	Ribosomal RNAs	rrn16(2), rrn23(2), rrn4.5(2), rrn5(2)	8
	Transfer RNAs	trnI-CAU, trnP-UGG, trnW-CCA, trnR-CCG, trnM-CAU, trnV-UACa, trnF-GAA, trnL-UAAa, trnT-UGU, trnS-GGA, trnfM-CAU, trnG-GCC, trnS-UGA, trnT-GGU, trnE-UUC, trnY-GUA, trnD-GUC, trnH-GUG, trnK-UUUa, trnQ-UUG, trnS-GCU, trnG-UCCa, trnR-UCU, trnC-GCA, trnL-CAA, trnV-GAC(2), trnI-GAU(2)a, trnA-UGC(2)a, trnR-ACG(2), trnN-GUU(2), trnL-UAG, trnP-GGG	37
Other genes	Maturase	matK	1
	Protease	clpP b	1
	Envelope membrane protein	cemA	1
	Acetyl-CoA carboxylase	accD	1
	c-type cytochrome synthesis gene	ccsA	1
	Translation initiation factor	infA	1
	Conserved open reading frames	ycf4, ycf3b, ycf2, ycf66a, ycf1	5
Total			128

^a indicates one intron

^b indicates two introns, and (n) indicates gene copy number

Distribution of SSRs and long repeats

A total of 520–542 SSRs were identified in the six Lewinskya plastomes (Fig. 2, Table S5). The highest number of SSRs was detected in L. speciosa PV998918 (542), followed by L. laevigata PV998915 (541), whereas L. rupestris PV998914 contained the fewest SSRs (520). Across all species, most SSRs were located in the LSC region, which contained 416–431 SSRs, accounting for 77.9–80.0% of the total SSRs. The SSC region contained 70–77 SSRs, while the combined IR regions contained 34–42 SSRs.

Fig. 2.

3D bar charts showing the SSR distribution of SSR types across genomic compartments in six Lewinskya species (X-axis: SSR Type; Y-axis: Species; Z-axis: Count). (A) SSR counts in whole chloroplast genomes. (B) SSR counts in the LSC region. (C) SSR counts in the SSC region. (D) SSR counts in IR regions

Mononucleotide repeats were the predominant SSR type in all six plastomes, with 390–413 repeats per plastome, representing 75.6% of all SSRs. Trinucleotide repeats were the second most abundant type, ranging from 89 to 96, followed by dinucleotide repeats (19–21) and tetranucleotide repeats (17–20). No penta- or hexanucleotide repeats were detected (Fig. 2, Table S5). In the IR regions, dinucleotide repeats were absent, whereas trinucleotide and tetranucleotide repeats were conserved, with six and four repeats, respectively, in each plastome. Mapping the 3,207 classified SSRs to annotated genomic regions showed that 1,248 SSRs were located in CDS regions, 1,418 in IGS regions, 517 in genic non-CDS regions, and 24 in RNA genes (Fig. S1A, Table S6). Non-CDS regions collectively contained 1,959 SSRs, accounting for 61.09% of all classified SSRs. Among CDS regions, the highest length-normalized SSR densities were observed in other protein-coding genes, conserved open reading frames (ycf genes), and RNA polymerase subunit genes (Fig. S1B; Table S7).

The numbers of interspersed repeats ranged from 84 in L. rupestris PV998914 to 102 in L. incana NC042174 (Fig. 3). Palindromic repeats were the most abundant interspersed repeat type (43–46), followed by forward repeats (20–27), reverse repeats (14–17), and complementary repeats (5–18). Tandem repeats ranged from 39 to 45, with the highest number detected in L. pylaisii PV998916.

Fig. 3.

Distribution of complementary, reverse, forward, palindromic and tandem repeats in six Lewinskya plastomes. The x-axis indicates species and accession numbers, and the y-axis indicates the number of repeats identified in each plastome

Comparative genomic analysis and IR boundary dynamics

Comparative analysis of the six Lewinskya plastomes revealed a highly conserved genome structure. The Mauve alignment resolved all plastomes into a single locally collinear block (LCB) spanning the aligned plastome region, indicating that gene order was conserved across the sampled species (Fig. 4). No large-scale inversions, translocations, or rearrangements of gene blocks were detected. The IR/SC junction comparison also showed a generally conserved boundary structure, although minor differences in regional length and junction position were observed among species (Fig. 5). The LSC region ranged from 85,222 bp in L. incana to 86,092 bp in L. pylaisii, the SSC region ranged from 18,415 bp in L. sordida to 18,514 bp in L. incana, and each IR ranged from 9,269 bp in L. rupestris to 9,913 bp in L. speciosa. Based on the sum of LSC, SSC and two IR regions, plastome sizes ranged from 122,258 bp in L. rupestris to 123,526 bp in L. speciosa. The genes flanking the four junctions were largely conserved among the six plastomes (Fig. 5). At the LSC/IRb boundary, rps12 was located immediately upstream of IRb, whereas trnV was positioned within IRb. The ccsA gene was located near the IRb/SSC boundary, and chlL occurred at the beginning of the SSC region. At the SSC/IRa boundary, ndhF was positioned at or very close to the JSA junction in all six plastomes, while trnN was located within IRa.

Fig. 4.

Mauve alignment of the six Lewinskya plastomes showing conserved gene arrangement and collinearity (white blocks represent protein-coding genes, black blocks represent tRNA genes, green blocks indicate intron-containing tRNA genes, and red blocks represent rRNA genes). Features of locally collinear blocks (LCB) using MAUVE. The coloured LCB boxes indicate the most conserved collinear regions, and block histograms show the sequence similarity

Fig. 5.

Comparisons of the borders of a large single-copy (LSC), inverted repeat (IR), and small single-copy (SSC) regions among Lewinskya plastomes. JLB (LSC/IRb), JSB (IRb/SSC), JSA (SSC/IRa), and JLA (IRa/LSC) denote the junction sites between the two corresponding regions in the six plastomes

The mVISTA comparison, using L. incana NC042174 as the reference, further supported high overall sequence similarity among the six plastomes (Fig. 6). Most aligned regions showed high conservation, with coding regions generally more conserved than non-coding regions. The IR regions were also more conserved than the LSC and SSC regions. Several localized divergence peaks were detected, mainly in non-coding regions, including rps3-rpl16, petL-psbE, trnT-rps4, rps2-rpoC2, and rps12-trnV. Among protein-coding genes, ycf2 showed comparatively elevated sequence divergence.

Fig. 6.

mVISTA comparison of six Lewinskya plastomes (L. incana NC042174 as reference). The percentage identity ranging from 50 to 100% is represented by the vertical scale

Sequence divergence and nucleotide diversity

To quantify sequence divergence across the chloroplast genomes, we calculated nucleotide diversity (π) for genic regions and intergenic spacer (IGS) regions based on six aligned Lewinskya plastomes. The average π value of IGS regions was 0.01612, approximately five times higher than the mean π value of genic regions (0.00327) (Fig. 7, Table S8). The sliding analysis showed that most regions had relatively low nucleotide diversity, but several regions displayed elevated π values. In total, 11 highly variable regions were identified based on π thresholds (Fig. 7). Five of these regions corresponded to genic regions, namely rps18, rpl22, infA, rpl32, and rps3, each displaying π values greater than 0.010. The remaining six were IGS regions: rps3-rpl16_1, rps2-rpoC2, rpl16_2-rpl14, psbN-psbT, psbE-psbF, and trnI-CAU-rpl23, with π values greater than or equal to 0.050. Four of the five highly variable genic regions were located in the LSC region, whereas rpl32 was located in the SSC region. All six highly variable IGS regions were located in the LSC region.

Fig. 7.

Sliding window analysis of Lewinskya species. (A) Nucleotide diversity (π) values of common genes within the chloroplast genomes; (B) Nucleotide diversity (π) values of IGS within the chloroplast genomes. Red dashed lines indicate the thresholds used for highly variable regions: π = 0.010 for genic regions and π = 0.050 for IGS regions. Numbered points indicate regions exceeding the corresponding thresholds. Plastome regions are shown above each panel. IR genes are represented once in panel A. LSC, large single-copy region; SSC, small single-copy region; IRa and IRb, inverted repeat regions

Codon usage bias

Comparative RSCU visualization showed highly similar codon usage patterns among the six Lewinskya plastomes (Fig. 8, Table S9). Most preferred synonymous codons ended with A or T, indicating a clear bias toward A/T-ending codons in plastid protein-coding genes. High RSCU values were consistently observed for codons such as TTA (Leu), TCT (Ser), GCT (Ala), CGT (Arg), CCT (Pro), GTT (Val), AAA (Lys), and ATT (Ile), whereas several G/C-ending codons showed lower usage. Met (ATG) and Trp (TGG) had RSCU values of 1 because each is encoded by a single codon. Overall, codon usage patterns were conserved among the sampled species.

Fig. 8.

Comparative analysis of relative synonymous codon usage (RSCU) in six Lewinskya plastomes. RSCU values were visualized using RSCU-Plot. Codons with RSCU > 1 indicate preferential usage, whereas codons with RSCU < 1 indicate lower-than-expected usage among synonymous codons

PR2-plot, ENC-plot, and neutrality plot analyses indicated that codon usage bias in Lewinskya plastid protein-coding genes was influenced by both nucleotide composition and selective constraints (Fig. 9, Tables S10 and S11). Most genes deviated from the PR2 central position and 75.5% were located below the expected ENC curve. The neutrality plot showed a significant but weak GC12–GC3 association (slope = 0.558, R² = 0.106, P = 7.099 × 10⁻⁹), supporting a combined influence of mutation pressure and selective constraints.

Fig. 9.

PR2, ENC, and neutrality plot analyses of codon usage bias in Lewinskya plastid protein-coding genes. (A) PR2-plot showing the relationship between G3/(G3 + C3) and A3/(A3 + T3). Dashed lines indicate the central position of 0.5. (B) ENC-plot showing the relationship between ENC and GC3s. The dashed curve represents the expected ENC values under GC3s compositional constraints. (C) Neutrality plot showing the regression of GC12 against GC3. GC12 was calculated as the mean of GC1 and GC2

Phylogenetic analysis

Phylogenetic analyses were conducted using a concatenated dataset of 78 plastid protein-coding genes from 26 Orthotrichaceae chloroplast genomes. The ML and BI analyses recovered largely congruent topologies (Fig. 10). With Macrocoma tenuis subsp. sullivantii designated as the outgroup, the sampled Orthotrichoideae taxa were resolved into two main lineages, corresponding to Lewinskyinae and Orthotrichinae.

Fig. 10.

Phylogenetic relationships of Orthotrichoideae plastomes inferred from the concatenated plastid protein-coding gene dataset, with Macrocoma tenuis subsp. sullivantii (MT591413) designated as the outgroup. Values above branches indicate SH-aLRT support, maximum-likelihood bootstrap support (ML_BS), and Bayesian posterior probability (PP_BI), respectively, in the order SH-aLRT/ML_BS/PP_BI. “*” denotes SH-aLRT = 100, ML_BS = 100, or PP_BI = 1.00

Within Lewinskyinae, the sampled Lewinskya species formed a strongly supported clade (SH-aLRT/ML_BS/PP_BI = 100/100/1.00). Several terminal relationships within Lewinskya were well supported, including the two L. acuminata accessions (SH-aLRT/ML_BS/PP_BI = 100/100/1.00), L. affinis + L. incana (SH-aLRT/ML_BS/PP_BI = 99/99/1.00), and L. speciosa OZ336986 + L. speciosa PV998918 (SH-aLRT/ML_BS/PP_BI = 100/100/1.00). The three L. rupestris accessions formed a clade (SH-aLRT/ML_BS/PP_BI = 100/100/1.00), with L. rupestris MT591405 and L. rupestris PV998914 recovered as sister taxa (SH-aLRT/ML_BS/PP_BI = 67.1/72/1.00). In contrast, several deeper nodes within Lewinskya were weakly supported, including L. pylaisii + L. speciosa MT591406 (SH-aLRT/ML_BS/PP_BI = 56.2/50/0.54) and the adjacent backbone node uniting this clade with L. laevigata and the L. acuminata clade (SH-aLRT/ML_BS/PP_BI = 57/42/0.59).

Pulvigera lyellii was placed outside the sampled Lewinskya clade. The three sampled Ulota species formed a distinct clade within Lewinskyinae (SH-aLRT/ML_BS/PP_BI = 100/100/1.00), with U. bruchii and U. curvifolia recovered as sister taxa (SH-aLRT/ML_BS/PP_BI = 99.9/100/1.00). In Orthotrichinae, the sampled Orthotrichum species formed a strongly supported clade (SH-aLRT/ML_BS/PP_BI = 100/100/1.00). Within this clade, O. crenulatum and O. rogeri were recovered as sister taxa (88.7/88/1.00), and this pair was sister to O. camanchacanum (SH-aLRT/ML_BS/PP_BI = 98.5/100/1.00), with O. stellatum placed as the earliest-diverging sampled species of the clade. Nyholmiella obtusifolia and Stoneobryum bunyaense were placed outside the sampled Orthotrichum clade.

Discussion

Lewinskya is a species-rich moss genus in Orthotrichaceae that has been segregated from the former broad concept of Orthotrichum s.l. and is now recognized as a distinct lineage within Lewinskyinae [12–16]. Species of Lewinskya are mainly corticolous epiphytes, although some taxa also occur on rock substrates, and the genus shows broad ecological and geographical distributions [14, 17–22]. Despite this diversity, species delimitation and phylogenetic reconstruction in Lewinskya remain challenging because many diagnostic characters are subtle, developmentally variable, or convergent among closely related taxa [14, 22, 33]. Previous molecular studies have greatly improved the generic and suprageneric framework of Orthotrichaceae, but relationships among some Lewinskya species remain difficult to resolve using limited molecular markers or even some genomic datasets [12, 13, 22, 27, 35]. In this study, we newly sequenced and assembled five Lewinskya chloroplast genomes and compared them with the published plastome of L. incana [47]. These six plastomes provide additional genomic resources for Lewinskya and allow evaluation of plastome structure, repeat distribution, sequence divergence, codon usage bias, and plastid-based phylogenetic relationships within an evolutionary context.

Plastome conservation and structural features in Lewinskya

The six sampled Lewinskya plastomes showed a highly conserved chloroplast genome organization. Their total lengths ranged from 122,258 bp to 123,526 bp, and the typical LSC, SSC, and two IR regions could be identified, consistent with the general organization of land plant plastomes [36–38, 71]. Although L. rupestris PV998914 and L. speciosa PV998918 were retained as high-quality single-scaffold assemblies rather than confirmed circular assemblies, they had no ambiguous bases, contained complete gene sets, and showed clearly defined LSC/SSC/IR boundaries. Therefore, these assemblies are suitable for comparative plastome analyses, although their assembly topology should still be described cautiously.

The sampled Lewinskya plastomes were also highly conserved in gene content. Each plastome encoded 128 genes, including 83 protein-coding genes, 37 tRNA genes, and eight rRNA genes. No gene loss or large-scale structural rearrangement was detected among the sampled species. Such conservation is consistent with the generally stable plastome architecture reported in mosses and other bryophyte lineages [38, 40, 72–74]. The Mauve alignment further supported this conclusion by recovering a single locally collinear block across the six plastomes, indicating conserved gene order and the absence of major inversions, translocations, or rearrangements. These results suggest that Lewinskya plastomes have experienced limited structural change during diversification, probably reflecting strong functional constraints on chloroplast genome organization.

The overall GC content of the Lewinskya plastomes was low and stable, ranging from 28.3% to 28.5%. This AT-rich pattern is consistent with many bryophyte plastomes [40, 72, 74]. As expected, the GC content was higher in the IR regions than in the LSC and SSC regions. This pattern is commonly associated with the presence of GC-rich rRNA genes in the IRs and has been widely observed in land plant plastomes [36, 71]. The higher conservation of IR regions may also contribute to the lower sequence divergence observed in these regions. Overall, the conserved genome size, structure, gene content, and GC composition indicate that chloroplast genome evolution in Lewinskya is structurally conservative.

SSRs and long repeats as sources of plastome variation

Chloroplast simple sequence repeats (SSRs) are useful for evaluating plastome variation and developing molecular markers for population genetics, phylogeography, and species identification [75, 76]. In the six Lewinskya plastomes, SSRs were abundant, with 520–542 SSRs detected per plastome. Mononucleotide repeats were the dominant SSR type, accounting for approximately three quarters of all SSRs, whereas penta- and hexanucleotide repeats were absent. This pattern is consistent with the AT-rich composition of chloroplast genomes, because mononucleotide SSRs in plastomes are often dominated by A/T motifs [75, 76]. Similar SSR patterns have also been reported in other moss plastomes, such as Rhodobryum giganteum [72].

Most SSRs in Lewinskya were located in the LSC region, whereas fewer SSRs occurred in the SSC and IR regions. This uneven distribution is consistent with the higher conservation of IR regions and the greater sequence variability of single-copy regions. In addition, most classified SSRs were located in non-coding regions, including intergenic spacers and genic non-CDS regions. Because non-coding SSRs are less constrained by protein-coding function, they may be more likely to accumulate length variation and therefore have potential value for population-level studies. These SSR loci may provide useful candidate markers for future analyses of genetic diversity, population structure, and phylogeographic patterns in Lewinskya. However, their practical utility should be further validated using broader population sampling.

Long repeats were also detected in all sampled Lewinskya plastomes, with palindromic repeats being the most abundant interspersed repeat type. In some plant plastomes, SSR accumulation, IR expansion, inversions, and other structural changes have been reported [77–79]. However, no major rearrangements were detected among the six Lewinskya plastomes, even though interspersed and tandem repeats were present. This suggests that the repeats detected here have not produced obvious structural instability at the plastome scale. The coexistence of abundant repeats and conserved gene order indicates that repetitive elements in Lewinskya may contribute mainly to local sequence variation rather than large-scale genome restructuring.

Comparative plastome divergence and IR boundary stability

Comparative genomic analyses further revealed high plastome conservation among the sampled Lewinskya species. The mVISTA comparison showed high overall similarity across the six plastomes, with coding regions generally more conserved than non-coding regions and IR regions more conserved than the LSC and SSC regions (Fig. 6). This pattern is consistent with the expectation that protein-coding genes are subject to stronger functional constraints, whereas intergenic spacers and other non-coding regions can accumulate substitutions and indels more readily [36, 40, 73]. Localized divergence peaks were mainly detected in non-coding regions, including rps3-rpl16, petL-psbE, trnT-rps4, rps2-rpoC2, and rps12-trnV. Among protein-coding genes, ycf2 showed comparatively elevated sequence divergence. These regions may provide useful candidates for future marker development.

The IR/SC junctions were generally conserved among the six Lewinskya plastomes. The genes flanking the LSC/IRb, IRb/SSC, SSC/IRa, and IRa/LSC junctions occupied similar positions across the sampled species, although minor differences in regional length and junction position were observed. Such small IR boundary shifts likely contributed to the limited plastome size variation among Lewinskya species. No drastic IR expansion or contraction, gene loss, or junction-associated pseudogene formation was detected. This pattern agrees with the generally conserved plastome structures reported in other moss lineages [40, 72, 74]. Therefore, plastome size variation in Lewinskya appears to be driven mainly by small-scale length changes rather than major structural rearrangements.

Highly variable regions as candidate molecular markers

Nucleotide diversity analysis provided a more detailed view of variable regions in Lewinskya plastomes. The average π value of intergenic spacer regions was approximately five times higher than that of genic regions, indicating that non-coding regions are the major contributors to plastome sequence variation in the sampled taxa. This result is consistent with the general expectation that coding sequences are more constrained by functional requirements, whereas intergenic regions can tolerate more mutations [36, 73].

A total of 11 highly variable regions were identified in the six Lewinskya plastomes. Five of these were genic regions, namely rps18, rpl22, infA, rpl32, and rps3, and six were intergenic spacers, namely rps3-rpl16_1, rps2-rpoC2, rpl16_2-rpl14, psbN-psbT, psbE-psbF, and trnI-CAU-rpl23 (Fig. 7). Most of these highly variable regions were located in the LSC region, whereas no comparable hotspot was detected in the IR regions. This distribution is consistent with the stronger conservation of IR regions and the relatively faster evolution of single-copy regions.

These highly variable regions may be useful candidate markers for Lewinskya species identification and population-level studies. Plastome-wide comparisons can identify variable loci with higher discriminatory potential than traditional single-locus barcodes, especially in taxonomically difficult groups [43, 47]. This is particularly relevant for bryophytes, where standard plant barcoding markers may show limited variation among closely related species [80]. Although ycf1 has been proposed as a promising plastid barcode in land plants [81], the most informative loci can differ among lineages. In Lewinskya, the hotspots identified here provide candidate regions for future marker development, but their diagnostic performance should be tested using expanded taxon sampling and multiple populations per species.

Codon usage bias in Lewinskya plastomes

Codon usage patterns were highly similar among the six Lewinskya plastomes, suggesting conserved protein-coding sequence evolution within the sampled taxa. RSCU analysis showed a clear preference for A/T-ending codons, which is consistent with the low overall GC content of Lewinskya plastomes (Fig. 8). Similar A/T-ending codon preferences have been reported in other moss plastomes [72, 74], indicating that nucleotide composition is an important factor shaping plastid codon usage in bryophytes.

The PR2-plot, ENC-plot, and neutrality plot analyses suggest that codon usage bias in Lewinskya plastid genes is not determined by nucleotide composition alone. Most genes deviated from the PR2 central position (Fig. 9A), and a large proportion of genes were located below the expected ENC curve (Fig. 9B). The neutrality plot showed a significant but weak association between GC12 and GC3 (Fig. 9C). Together, these patterns indicate that both mutation pressure and selective constraints may contribute to codon usage bias in Lewinskya plastomes, which is consistent with general interpretations of codon usage evolution [82]. However, because plastid protein-coding genes are functionally conserved and the sampled species are closely related, the observed codon usage bias should be interpreted as a genome-wide tendency rather than strong evidence for lineage-specific adaptive divergence.

Plastome phylogeny and systematic implications

The plastome-based phylogenetic analysis was conducted using 26 Orthotrichaceae chloroplast genomes and recovered the sampled Lewinskya species as a strongly supported clade within Lewinskyinae. This result supports the current recognition of Lewinskya as a distinct genus separated from Orthotrichum s.str [12, 14, 32]. At a broader level, the sampled Orthotrichoideae taxa were resolved into two main lineages corresponding to Lewinskyinae and Orthotrichinae, which is consistent with recent taxonomic and phylogenetic frameworks for Orthotrichaceae [12, 13, 83]. The placement of Pulvigera outside the sampled Lewinskya clade and the recovery of distinct Ulota and Orthotrichum clades further support the usefulness of plastome-scale data for resolving higher-level relationships within Orthotrichaceae.

Within Lewinskya, several terminal relationships were well supported. For example, the two L. acuminata accessions, the L. affinis + L. incana relationship, two of the L. speciosa accessions, L. speciosa OZ336986 and L. speciosa PV998918, and the three L. rupestris accessions were recovered with strong support. These results indicate that plastid protein-coding genes can provide useful phylogenetic signal for some species-level relationships in Lewinskya. However, not all relationships within the genus were strongly resolved. Several deeper nodes within Lewinskya showed weak support, including the node involving L. pylaisii and L. speciosa MT591406 and the adjacent backbone node connecting this lineage with L. laevigata and the L. acuminata clade. Therefore, the plastome phylogeny should be interpreted as a plastid genomic perspective rather than a complete resolution of the species tree.

This limitation is important because plastid genomes are usually inherited as a single linked unit and may not fully reflect organismal evolutionary history, especially when incomplete lineage sorting, hybridization, introgression, or chloroplast capture has occurred [44, 45]. Similar challenges have been reported in phylogenetic studies of Orthotrichaceae, where broader taxon sampling and genomic-scale nuclear evidence have been needed to address difficult systematic questions [12, 13, 22, 35]. The weakly supported backbone nodes observed here suggest that relationships among some Lewinskya species may require denser sampling and complementary nuclear genomic data. Thus, while the plastome data improve available genomic resources and provide useful evidence for Lewinskya systematics, future studies integrating plastid genomes, nuclear markers, morphology, and broader geographic sampling will be necessary to clarify the evolutionary history of this taxonomically challenging genus.

Overall, this study expands plastome resources for Lewinskya by providing five newly assembled chloroplast genomes and a comparative plastome analysis within the genus. The sampled Lewinskya plastomes show conserved genome size, gene content, structure, GC composition, and IR boundaries, indicating strong structural conservation during plastome evolution. At the same time, SSRs, long repeats, and highly variable regions reveal localized sequence variation that may be useful for future marker development. The plastome-based phylogeny supports the monophyly of the sampled Lewinskya taxa and provides additional evidence for relationships within Orthotrichaceae, while also showing that some deeper interspecific relationships remain insufficiently resolved. These plastome data therefore provide a useful genomic foundation for future studies of species identification, comparative genomics, and evolutionary relationships in Lewinskya and related mosses.

Conclusions

The present study provides five newly assembled high-quality chloroplast genomes for Lewinskya and expands available plastome resources for Orthotrichaceae. Comparative analyses of six Lewinskya plastomes revealed highly conserved genome size, structure, gene content, GC composition, and inverted repeat boundaries, indicating strong structural conservation during plastome evolution in the sampled species. At the same time, simple sequence repeats, long repeats, and highly variable regions revealed localized sequence variation, providing candidate molecular markers for future species identification, population-level analyses, and phylogeographic studies. The plastome-based phylogeny strongly supported the sampled Lewinskya species as a clade and provided additional evidence for relationships within Orthotrichaceae. However, several deeper interspecific relationships within Lewinskya remained weakly supported, suggesting that plastome data should be interpreted as a complementary genomic perspective. Broader taxon sampling and integration of plastid genomes, nuclear genomic data, and morphology will be necessary to further clarify the evolutionary history of this taxonomically challenging genus.

Abbreviations

plastomes

complete chloroplast genomes

SSRs

Simple sequence repeats

IR

Inverted repeat

LSC

Large single-copy region

SSC

Small single-copy region

CDS

Coding sequence

IGS

Intergenic spacer regions

LCB

locally collinear block

JLA

Boundary between LSC and IRA

JLB

Boundary between LSC and IRB

JSA

Boundary between SSC and IRA

JSB

Boundary between SSC and IRB

RSCU

Relative synonymous codon usage

ML

Maximum likelihood

BI

Bayesian inference

SH-aLRT

Shimodaira-Hasegawa approximate likelihood-ratio test

BS

Bootstrap

PP

Posterior probability

Authors’ contributions

Conceptualization, W.H.; Methodology, W.H., Y.J.Z., Z.X.S. and K.Z.; Software and analysis, W.H., Y.J.Z., K.Z., Y.Q.L. and H.F.L.; Resources, Y.J.; Data curation, Y.J.; Writing−original draft preparation by W.H., review and editing by Y.L., X.H.L., Z.X.S. and Y.J.; Visualization, W.H., K.Z. and Y.Q.L.; Supervision and project administration, W.H., X.H.L. and Y.J.; Funding acquisition, X.H.L. and W.H. All authors have read and approved the final version of the manuscript.

Funding

This study was supported by the Scientific Research Foundation for High-level Talent of Sanming University (Grant No. 20YG02), the Natural Science Foundation of Fujian Province (Grant No. 2023J011022) and the Fujian Provincial Overseas Study and Training Program for Young and Middle-aged Backbone Teachers in Higher Education.

Data availability

The complete chloroplast genome sequences generated in this study have been deposited in the NCBI GenBank ( https://www.ncbi.nlm.nih.gov/nuccore/) under the following accession numbers: PV998914 to PV998918. All additional data supporting the conclusions of this study are included within the manuscript and supplementary files.

Declarations

Ethics approval and consent to participate

The study was conducted in accordance with ethical guidelines and principles. No human or animal subjects were involved. Plant samples were collected in compliance with local regulations and with the permission of the relevant authorities. Voucher specimens were deposited in the Herbarium of the Institute of Botany, Chinese Academy of Sciences (PE) and the Herbarium of the Kunming Institute of Botany, Chinese Academy of Sciences (KUN).

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.