Plastome characterization and its phylogenetic implications on Lithocarpus (Fagaceae)

Abstract

Background

The genus Lithocarpus is a species-rich dominant woody lineage in East Asian evergreen broad-leaved forests. Despite its ecological and economic significance, the plastome structure and evolutionary history of the genus remain poorly understood. In this study, we comprehensively analyzed the 34 plastomes representing 33 Lithocarpus species. Of which, 21 were newly assembled. The plastome-based phylogenomic tree was reconstructed to reveal the maternal evolutionary patterns of the genus.

Results

The Lithocarpus plastomes exhibit a typical quadripartite structure, ranging in length from 161,010 to 161,476 bp, and containing 131 genes, including 86 protein-coding genes, 8 rRNA genes, and 37 tRNA genes. Remarkably, the infA gene was identified as a pseudogene in 17 species. Significant variability was observed in simple sequence repeats (SSRs) as well as in the boundary regions between the two single-copy regions and the inverted repeat region (SC/IR) across the plastomes. Additionally, four genes (accD, atpF, rpl32, and rps8) were found to be under positive selection. The monophyletic status of Lithocarpus was strongly supported by plastome-based phylogeny; however, the phylogenetic tree topology showed a significant difference from that obtained by the nuclear genome-based phylogeny.

Conclusions

The plastome of Fagaceae is generally conserved. Nevertheless, genes related to metabolism, photosynthesis, and energy were under strong positive selection in Lithocarpus, likely driven by environmental pressures and local adaptation. The plastome-based phylogeny confirmed the monophyletic status of Lithocarpus and revealed a phylogeographic pattern indicating limited seed-mediated gene flow in the ancestral lineage. The prevalence of cytonuclear discordance in Lithocarpus and other Fagaceae genera suggests that ancient introgression, incomplete lineage sorting, and asymmetrical seed- and pollen-mediated geneflow might contribute to this discordance. Future studies are essential to test these hypotheses and further elucidate the divergence patterns of this unique Asian evergreen lineage.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12870-024-05874-z.

Background

The plastome, a crucial uniparentally inherited organelle in plant cells, is responsible for photosynthesis and other metabolic processes essential for plant adaptation to their environment [1, 2]. Although nuclear genome datasets have become increasingly prominent in phylogenetic and plant genome analyses, plastomes remain indispensable for tracing the maternal evolutionary history of angiosperm taxa, offering critical insights that complement nuclear data [3, 4]. One key advantage of plastomes in phylogenetic reconstruction is their conserved gene content and structure, which provide a high degree of homology across diverse plant lineages [3, 5]. Moreover, the low or no recombination rates within plastomes enhance the reliability of phylogenetic inferences by preserving the evolutionary signature of maternal lineages, facilitating the resolution of ambiguous phylogenetic relationships [4]. Plastome-based phylogenetic approaches are thus powerful tools in plant phylogenetics and evolutionary research, enabling the detection of introgression events such as plastome capture [e.g., 6, 7], monitoring seed-mediated gene flow in spatial population genetic studies [e.g., 8], and examining the structure, diversity, and evolution of organellar genomes [e.g., 9].

The genus Lithocarpus Blume, commonly known as tanoak or stone oak, includes approximately 330 to 347 species, making it the second-largest genus in the family Fagaceae [10]. The northern Indo-China and Malaysian regions are two important species diversity centers of the genus [11–14]. These species are usually dominant trees in evergreen broadleaved forests and play a critical role in maintaining regional microclimate and biodiversity [15, 16]. They also provide important ecological services to society as sources of starches [17, 18], timber [18], and important sugar substitutes in the therapy of type two diabetes [18, 19]. However, Asian Fagaceae species have faced severe population reductions and habitat degradation in recent decades. Estimates indicate that Asian evergreen forests have experienced severe population declines, habitat loss, and habitat degradation [20, 21]. According to an International Union for Conservation of Nature (IUCN) primary assessment, about one-third of Asian Fagaceae species may be endangered [22]. These evergreen fagaceous lineages in the Asian tropics and subtropics are particularly vulnerable to intense human disturbance, driven by the high productivity and biodiversity of these regions [20, 21, 23, 24]. Understanding the genomics of Lithocarpus is crucial for characterizing biodiversity and enhancing conservation efforts among these species.

As a species-rich dominant lineage in evergreen broadleaved forests, Lithocarpus has been the focus of several phylogenetic studies in recent years. Sanger-based chloroplast DNA (cp.DNA) barcode sequencing has been widely applied to infer the phylogenetic structure and population genetic structure of fagaceous plants [25–29]. Cannon and Manos [26] resolved two main clades within Southeast Asian Lithocarpus using cp.DNA, corresponding to Borneo versus other widespread East Asian regions. Their research revealed a strong geographical structure, high genetic diversity, and endemism of cp. DNA haplotypes within Lithocarpus, likely owing to limited migration and extinction events in Southeast Asia [26].

Phylogenetic inferences based on the atpB -rbcL spacer (cp.DNA) and internal transcribed spacer (ITS) (nuclear ribosomal DNA; nrDNA) reveal an incongruent tree topology [29]. These findings suggest a complex evolutionary history of Lithocarpus, influenced by East Asian geographic events and ancient introgression [29–31]. However, the tree topology inferred from these Sanger-sequencing-based markers only received medium to low support.

Next-generation sequencing (NGS) technologies offer powerful tools to resolve phylogenetic relationships not only at evolutionarily deep nodes, but also among closely related species characterized by recent interspecific gene flow within Fagaceae [32–34]. Recent phylogenomic reconstructions of Fagaceae using genomic DNA resequencing [30] and hybrid enrichment sequencing (Hyb-Seq) [31] yielded similar tree topologies for the main lineages in Fagaceae and the species phylogeny within Lithocarpus specifically. These studies resolved a sister relationship between Chrysolepis and Lithocarpus [30, 31], whereas the phylogenetic tree inferred from plastome sequences showed Lithocarpus as the sister group to Castanopsis + Castanea [30, 31]. Such notable discordance between the plastome- and nuclear genome-based trees may be the result of incomplete lineage sorting, horizontal transfer, or ancient gene flow among the ancestral lineages [30, 31]. The accumulated findings of phylogenomic studies on Fagaceae offer significant insights into the evolutionary history of Lithocarpus and related taxa. However, these studies have included only a limited number of Lithocarpus species. Meanwhile, phylogenetic studies of Lithocarpus based on NGS data remain scarce, with only sporadic reports of plastomes published to date [e.g., 35–40]. Moreover, the monophyletic status of Asian Lithocarpus has not been consistently resolved. Some studies indicated that the plastome-based phylogeny was polyphyletic in Lithocarpus [41], in contrast to the monophyly inferred by Zhou et al. [30]. These studies provide valuable insights into the evolution of Lithocarpus, but also highlight substantial uncertainties in its plastome evolutionary history. Therefore, the evolutionary history within Lithocarpus warrants further investigation.

In recent years, the gradual sequencing of plastomes from Fagaceae species has substantially enhanced our understanding of the structure and divergence patterns of plastomes within this family [e.g., 6, 42, 43]. These studies have shown that the plastome structure of the family Fagaceae is relatively conserved in structure in terms of size (158,163–161,419 bp), GC content (36.8–37.1%), and gene order [6, 42, 43]. Despite this, molecular signatures of adaptive evolution have been observed in certain protein-coding genes in Fagaceae [42–47], though these previous studies have mainly focused on the genera Quercus [43, 48], Castanea [46], Fagus [44, 47], and Castanopsis [49]. Despite Lithocarpus accounting for approximately one-third of the species diversity within the family Fagaceae, investigations of Lithocarpus plastomes have remained limited. Most studies either report the plastome structure of a single species or plastomes of Lithocarpus were only used as molecular evidence for the classification of new species [e.g., 37, 39, 40]. No comprehensive study has ever been conducted on the plastome structure, gene function, or molecular signatures of the adaptive evolution of Lithocarpus.

In this study, we newly sequenced and assembled the plastomes of 21 species representing the main morphological groups of the genus Lithocarpus. Utilizing these new plastomes in combination with the previously reported Lithocarpus plastomes, we analyzed the plastome structure and sequence divergence pattern and reconstructed the plastome-based phylogenomic tree, aiming to (1) examine the plastome structure and sequence divergence patterns of Lithocarpus and (2) infer the maternal evolutionary history of Lithocarpus. This study also provides important insights into the evolution and adaptation of this unique East Asian fagaceous lineage.

Materials and methods

Plant material and DNA extraction

Twenty-one species of Lithocarpus from East Asia (including Tibet, Yunnan, Hunan, Hainan, and Guangdong provinces of China, among other areas) were sequenced, representing five of the thirteen subgenera (i.e., Synaedrys, Pachybalanus, Gymnobalanus, Pseudosynaedrys, and Pasania) proposed by Camus [11]. Genomic DNA was isolated using an optimized CTAB method as described by Doyle and Doyle [50]. The quality of the genomic DNA was checked by 1% agarose gel electrophoresis, and the DNA concentration was measured using a Qubit^® 2.0 fluorometer (Invitrogen, Carlsbad, CA, USA) and then adjusted to 20 ng/uL for library construction. Detailed collection information and voucher specimens for each species are summarized in Table 1. The herbarium voucher specimens were identified by Dr. Min Deng, a Fagaceae taxonomy expert. The voucher specimens were deposited in the Herbarium of Yunnan University (YUKU), Kunming, China.

Table 1.

Collection information and Genbank accessions of the plant materials used for this study

Taxa	subgenus	Locality	Voucher information	Herbarium information	GenBank Accession
Lithocarpus amygdalifolius (Skan) Hayata	Pachybalanus	Nanning, Guangxi, China	DM26426	Yunnan University	PQ276669
L. areca (Hickel et A. Camus) A. Camus	Pasania	Chongzuo, Guangxi, China	DM26599	Yunnan University	PQ276671
L. brevicaudatus (Skan) Hayata	Pasania	Wuzhishan, Hainan, China	DM24894	Yunnan University	PQ276666
L. cleistocarpus (Seemen) Rehder et E. H. Wilson	Pachybalanus	Zhaotong, Yunnan, China	DM23979	Yunnan University	OM112296
L. dealbatus (J. D. Hooker et Thomson ex Miquel) Rehder	Pasania	Kunming, Yunnan, China	DM25628	Yunnan University	PQ276668
L. elizabethiae (Tutcher) Rehder	Pasania	Kunming, Yunnan, China	DM25627	Yunnan University	PQ276667
L. fenestratus (Roxburgh) Rehd	Pasania	Yuxi, Yunnan, China	DM24080	Yunnan University	OM112300
L. fenzelianus A. Camus	Pachybalanus	Wuzhishan, Hainan, China	DM24905	Yunnan University	OM388302
L. fohaiensis (Hu) A. Camus	Pachybalanus	Xishuangbanna, Yunnan, China	DM22706	Yunnan University	PQ276656
L. glaber (Thunb.) Nakai	Pasania	Zhuzhou, Hunan, China	DM24713	Yunnan University	OM388303
L. grandifolius (D. Don) S. N. Biswas	Pasania	Nanchuan, Chongqing, China	DM24500	Yunnan University	PQ276659
L. gymnocarpus A. Camus	Gymnobalanus	Honghe, Yunnan, China	DM24655	Yunnan University	PQ276663
L. konishii (Hayata) Hayata	Gymnobalanus	Shenzhen, Guangdong, China	DM24770	Yunnan University	PQ276665
L. longzhouicus (C. C. Huang & Y. T. Chang) J. Q. Li & L. Chen		Nanning, Guangxi, China	DM26595	Yunnan University	PQ276670
L. obscurus C. C. Huang et Y. T. Chang		Motuo, Tibet, China	DM23454	Yunnan University	OM112297
L. pachylepis A.Camus	Synaedrys	Honghe, Yunnan, China	DM24588	Yunnan University	PQ276660
L. sp.		Wenshan, Yunnan, China	DM24476	Yunnan University	PQ276657
L. sphaerocarpus (Hickel & A.Camus)		Wenshan, Yunnan, China	DM24498	Yunnan University	PQ276658
L. tabularis Y.C.Hsu & H.Wei Jen		Honghe, Yunnan, China	DM24606	Yunnan University	PQ276662
L. uvariifolius (Hance) Rehd.	Synaedrys	Shenzhen, Guangdong, China	DM24743	Yunnan University	PQ276664
L. xylocarpus (Kurz) Markgraf	Pseudosynaedrys	Honghe, Yunnan, China	DM24591	Yunnan University	PQ276661

Illumina sequencing, assembly, and annotation

High-quality DNA was used to build genomic libraries. The paired-end (PE) read library was constructed using TruSeq DNA sample preparation kits (Illumina, San Diego, CA, USA). Sequencing was performed using 150-bp paired-end reads on the Illumina HiSeq2500 platform with an insert size of approximately 400 bp. Raw reads were filtered and trimmed to remove the low-quality reads using Fastp [51] with default parameters. Approximately 2 GB of clean data were generated per library. All sequencing was conducted by Biomarker Technologies Inc. (Beijing, China). Additionally, twelve sets of raw reads of Fagaceae whole-genome sequencing data, including seven Lithocarpus species and five species from other genera of Fagaceae, reported by Zhou et al. [30], were downloaded and used for subsequent analyses. Detailed GenBank accession information for the data is summarized in Table S1.

High-quality clean data were assembled using GetOrganelle v1.7.2 [52] with the following parameter settings: ‘-R 10 -k 21,45,65,85,105,115,127 -F embplant_pt”, utilizing Lithocarpus balansae (GenBank accession number, KP299291) as the reference genome. Genome annotation was performed using CPGAVAS [53] and confirmed with DOGMA (http://dogma.ccbb.utexas.edu/) [54] and BLAST [55]. Additionally, tRNAs were identified using tRNAscan-SE [56]. The manual correction was made to locate the start and stop codons and exon-intron boundaries using Geneious Prime [57], with L. hancei (MW375417) and L. balansae (KP299291) as reference genomes. Complete plastome maps were generated using OGDRAWv1.2 (Draw Organelle Genome Maps, http://ogdraw.mpimp-golm.mpg.de/) [58]. All annotated plastome sequences have been deposited in GenBank, under the accession numbers listed in Table 1. Complete plastome sequences of seven Lithocarpus species reported in previous studies [30] and six whole plastomes of Lithocarpus species (L. hancei, L. dealbatus, L. balansae, L. polystachyus, L. longinux, and L. litseifolius) downloaded from NCBI (Table S1), along with the newly sequenced and assembled Lithocarpus plastomes. In total 34 plastomes representing 33 species were used for subsequent analyses.

Genomic feature analyses

Relative synonymous codon usage (RSCU) values were calculated by dividing the observed frequency by the expected frequency (RSCU > 1 indicates higher than expected codon use and RSCU < 1 indicates lower use [59]. Using MEGA X software (https://www.megasoftware.net/) [60], we determined the RSCU values for 34 plastomes of Lithocarpus (21 newly obtained and 13 previously published) [60], revealing variations in synonymous codon usage. The RSCU cluster diagram was created using the R package pheatmap (https://CRAN.R-project.org/package=pheatmap) [61].

Simple sequence repeats (SSR) within the 34 complete plastomes of Lithocarpus were detected using MISA (http://pgrc.ipk-gatersleben.de/misa) with motif sizes ranging from one to six nucleotides [62]. Thresholds for the minimum number of repeat units were set as follows: 10 for mono-nucleotide SSRs, 5 for di-nucleotide SSRs, 4 for tri-nucleotide SSRs, and 3 for tetra-nucleotide, penta-nucleotide, and hexa-nucleotide SSRs, respectively.

Comparative genomic analyses and sequence divergence

The boundaries of the large single-copy (LSC), small single-copy (SSC), and inverted repeat (IR) regions of the Lithocarpus plastomes, along with those of other Fagaceae species, were drawn using the IRscope online tool (https://irscope.shinyapps.io/irapp/) [63]. The mVISTA program in Shuffle-LAGAN mode was used to compare the 34 complete plastomes of Lithocarpus with the annotation of L. balansae (KP299291) used as the reference [64]. The synonymous substitution rate (K_s), nonsynonymous substitution rate (K_a), and the K_a/K_s ratio of coding sequences (CDS) in the whole plastome regions were calculated among the 34 Lithocarpus plastomes using KaKs Calculator 2.0 [65]. The Ka/Ks values were visualized with a boxplot generated using the R package ggplot2 (https://cran.r-project.org/package=ggplot2) [66]. A one-sample t-test with µ = 1 was performed for each gene to evaluate statistical significance.

Phylogenetic analyses

To infer the phylogenetic structure of Fagaceae, 61 plastomes, including 34 individuals of Lithocarpus, 2 individuals of Notholithocarpus, 12 species of Quercus, 3 species of Castanopsis, 3 species of Castanea, 2 species of Chrysolepis, 2 species of Trigonobalanus, and 2 species of Fagus, with Betula pendula as an outgroup, were used to reconstruct the plastome-based phylogenomic tree. Of these, 21 plastomes were newly obtained in this study, and 40 were downloaded from NCBI. Betula pendula of the family Betulaceae was used as an outgroup to root the tree [30, 67]. GenBank accession numbers of the plastomes are provided in Supplementary Tables 1 and Table S1.

Maximum-likelihood (ML) and Bayesian inference (BI) phylogenetic analyses were conducted using three data partitions: (1) the whole plastome sequences, (2) protein-coding exons, and (3) introns and intergenic spacers. All sequences were aligned using MAFFT 7.0 [68] with default parameters. The best-fit nucleotide substitution model (TVM + F + R10) was identified by ModelFinder [69]. The ML tree was reconstructed using IQ-tree v2.0 [70] with 1000 ultrafast bootstrap replicates and default settings. The BI tree was reconstructed using MrBayes v3.2.6 [71]. Markov Chain Monte Carlo (MCMC) analysis was performed over 10,000,000 generations, with tree sampling every 1000 generations. The MrBayes output was inspected using Tracer ver.1.7.1 [72] to ensure proper convergence and mixing (effective sample sizes all > 200), and a maximum clade credibility tree was generated after a 20% burn-in. The result was visualized and edited with Figtree v1.4 (https://github.com/rambaut/figtree/releases) [73]. Phylogenetic trees were plotted on the world map using the R package phytools v2.0 (https://cran.r-project.org/web/packages/phytools/index.html) [74]. In addition, we utilized the R package ape v5.8 (10.1093/bioinformatics/bty633) [75] to trim the nuclear phylogenetic tree obtained by Liu et al. [31] and our plastome tree to illustrate the discordance between the plastome and nuclear genomes.

Morphological traits divergence pattern

We gathered the leaf epidermal trichome (Bubble Trichome Group [BBT]; Appressed parallel tufts Group [APT]; Glabrous Group [Glabrous]) [76, 77], and acorn type (Acorn fruit [AR]; Enclosed receptacle fruit [ER]) [78, 79] characteristics of Lithocarpus reported in previous studies, then mapped these traits at the tips of the phylogenetic tree to illustrate the morphological divergence pattern to these key taxonomic significant characteristics.

Results

Plastome features in Lithocarpus species

Illumina sequencing generated 2 to 5 G of raw data for each sampled species library. After sequencing, trimming, and quality control of reads, 21 high-quality plastomes were newly assembled (Fig. S1), and 13 additional Lithocarpus plastomes were reassembled and compared. GenBank accession numbers and the sources of the plastomes used in this study are provided in Supplementary Tables S1 and S2.

The sizes of plastomes of the 34 Lithocarpus individuals ranged from 161,010 (L. pachylepis) to 161,476 bp (L. dealbatus) (Table 2). The plastome of Lithocarpus is a typical single-circular molecule with a four-segment structure, comprising a large single-copy region (LSC) (90,394–90,731 bp), a small single-copy region (SSC) (18,933–19,255 bp), and the two inverted repeat regions (IRA and IRB, respectively) (25,632–25,911 bp) (Fig. 1; Table 2). All of the 34 plastomes showed a similar total GC content, ranging from 36.7 to 36.8%. The GC contents of the LSC and SSC regions were 34.50–34.60% and 30.70–31.00%, respectively, while the IR regions had a higher GC content of 42.70–42.80% (Table 2).

Table 2.

Characteristics of the 34 complete plastomes in Lithocarpus

Species	Size (bp)	LSC (bp)	IR	SSC (bp)	GC content (%)	LSC GC content (%)	IR GC content (%)	SSC GC content (%)	Total genes	CDS	tRNAs	rRNAs
Lithocarpus amygdalifolius	161,283	90,538	25,883	18,979	36.7	34.6	42.7	30.7	131	86	37	8
L. areca	161,144	90,427	25,871	18,979	36.7	34.6	42.7	30.9	131	86	37	8
L. balansae	161,020	90,596	25,632	19,160	36.7	34.5	42.8	30.8	131	86	37	8
L. brevicaudatus	161,208	90,462	25,893	18,960	36.8	34.6	42.7	30.7	131	86	37	8
L. calophyllus	161,288	90,510	25,904	18,970	36.7	34.6	42.7	30.7	131	86	37	8
L. cleistocarpus	161,178	90,558	25,762	19,096	36.7	34.6	42.8	30.8	131	86	37	8
L. corneus	161,112	90,543	25,676	19,217	36.8	34.6	42.8	30.8	131	86	37	8
L. dealbatus	161,118	90,460	25,781	19,096	36.8	34.6	42.8	30.8	131	86	37	8
L. dealbatus	161,476	90,731	25,879	18,987	36.8	34.6	42.7	30.9	131	86	37	8
L. elizabethiae	161,273	90,619	25,779	19,096	36.7	34.6	42.8	30.8	131	86	37	8
L. fenestratus	161,184	90,524	25,804	19,052	36.7	34.5	42.7	30.8	131	86	37	8
L. fenzelianus	161,148	90,421	25,893	18,941	36.7	34.6	42.7	30.7	131	86	37	8
L. fohaiensis	161,370	90,615	25,911	18,933	36.8	34.6	42.7	31	131	86	37	8
L. glaber	161,302	90,556	25,883	18,980	36.7	34.6	42.7	30.7	131	86	37	8
L. grandifolius	161,193	90,510	25,797	19,089	36.7	34.5	42.7	30.7	131	86	37	8
L. gymnocarpus	161,295	90,541	25,875	19,004	36.7	34.6	42.7	30.8	131	86	37	8
L. haipinii	161,289	90,537	25,880	18,992	36.7	34.6	42.7	30.7	131	86	37	8
L. hancei	161,228	90,509	25,880	18,959	36.7	34.6	42.7	30.7	131	86	37	8
L. konishii	161,374	90,614	25,893	18,974	36.7	34.6	42.7	30.8	131	86	37	8
L. litseifolius	161,322	90,551	25,897	18,977	36.7	34.6	42.7	30.7	131	86	37	8
L. longanoides	161,281	90,539	25,883	18,976	36.7	34.5	42.7	30.7	131	86	37	8
L. longinux	161,420	90,409	25,878	19,255	36.8	34.6	42.7	31	131	86	37	8
L. longipedicellatus	161,408	90,615	25,897	18,999	36.8	34.6	42.7	30.9	131	86	37	8
L. longzhouicus	161,143	90,399	25,888	18,978	36.8	34.6	42.7	30.9	131	86	37	8
L. obscurus	161,349	90,616	25,882	18,969	36.8	34.6	42.7	30.9	131	86	37	8
L. pachylepis	161,010	90,563	25,715	19,017	36.8	34.6	42.7	30.8	131	86	37	8
L. polystachyus	161,217	90,491	25,879	18,968	36.7	34.6	42.7	30.7	131	86	37	8
L. sp.	161,291	90,636	25,780	19,095	36.7	34.5	42.8	30.8	131	86	37	8
L. sphaerocarpus	161,283	90,632	25,772	19,107	36.7	34.5	42.8	30.8	131	86	37	8
L. tabularis	161,131	90,512	25,776	19,067	36.7	34.6	42.8	30.8	131	86	37	8
L. tephrocarpus	161,233	90,465	25,899	18,970	36.7	34.6	42.7	30.7	131	86	37	8
L. truncatus	161,368	90,575	25,893	19,007	36.7	34.6	42.7	30.8	131	86	37	8
L. uvariifolius	161,155	90,394	25,902	18,957	36.7	34.6	42.7	30.7	131	86	37	8
L. xylocarpus	161,239	90,613	25,765	19,096	36.8	34.6	42.8	30.8	131	86	37	8

Fig. 1.

Plastomes map of Lithocarpus species. The transcription of genes seen outside the outer layer circle is done clockwise, whereas the transcription of genes inside is done counterclockwise. Different functional groups of genes are color-coded. The GC content of the plastome is indicated by the dark gray area in the inner circle. LSC, large single-copy region; SSC, small single-copy region; IRa, IRb, inverted repeat A and B, respectively

A total of 131 genes were annotated in the plastome of the 21 newly assembled Lithocarpus species, including 86 protein-coding genes (PCGs), 37 tRNA genes, and 8 rRNA genes. These genes were categorized based on their functions as being related to photosynthesis, self-replication, and other functions (Table 3). Among these genes, 18 contained introns, including 12 protein-coding genes and 6 tRNA genes. Most genes had a single intron, while the ycf3 clpP and rps12 genes contained two introns (Table 3). Additionally, the infA gene was identified as a pseudogene in 17 Lithocarpus plastomes (Fig. 2).

Table 3.

Gene composition within the plastomes of Lithocarpus species

Category of Genes	Group of gene	Name of gene
Self‒replication	Ribosomal RNA genes	rrn4.5 × 2, rrn5 × 2, rrn16 × 2, rrn23 × 2
	Transfer RNA genes	trnC-GCA, trnD-GUC, trnE-UUC, trnF-GAA, trnfM-CAU, trnG-GCC, trnG-UCC, trnH-GUG, trnK-UUU*, trnL-UAA, trnL-UAG, trnM-CAU, trnP-UGG, trnQ-UUG, trnR-UCU, trnS-GCU, trnS-GGA, trnS-UGA, trnT-GGU, trnT-UGU, trnV-UAC, trnW-CCA, trnY-GUA, trnA-UGC*×2, trnI-CAU×2, trnL-CAA×2, trnI-GAU*×2, trnV-GAC×2, trnR-ACG×2, trnN-GUU×2
	Ribosomal protein (small subunit)	rps11, rps12*×2, rps14, rps15, rps16*, rps18, rps19, rps2, rps3, rps4, rps7 × 2, rps8
	Ribosomal protein (large subunit)	rpl14, rpl16*, rpl2*×2, rpl20, rpl22, rpl23 × 2, rpl32, rpl33, rpl36
	RNA polymerase	rpoA, rpoB, rpoC1*, rpoC2
	Translational initiation factor	infA#
Genes for photosynthesis	Subunits of photosystem I	psaA, psaB, psaC, psaI, psaJ
	Subunits of photosystem II	psbA, psbB, psbC, psbD, psbE, psbF, psbH, psbI, psbJ, psbK, psbL, psbM, psbN, psbT
	Subunits of ATP synthase	atpA, atpB, atpE, atpF*, atpH, atpI
	Large subunit of Rubisco	rbcL
	Subunits of NADH Dehydrogenase	ndhA*, ndhB*×2, ndhC, ndhD, ndhE, ndhF, ndhG, ndhH, ndhI, ndhJ, ndhK
Other genes	Maturase	matK
	Envelope membrane protein	cemA
	Subunit of acetyl‒CoA	accD
	Synthesis gene	ccsA
	ATP‒dependent protease	clpP**
	Genes of unknown function	ycf1, ycf2 × 2, ycf3**, ycf4

Notes Gene*: Gene with one introns; Gene**: Gene with two introns; #Gene: Pseudo gene; Gene (×2): Number of copies of multi-copy genes

Fig. 2.

Example of infA pseudogenes in the plastomes of Lithocarpus. “*” in the black boxes showing the termination codons

Codon usage

Sixty-four codons encoding 20 amino acids were detected in the 34 Lithocarpus plastomes. Lithocarpus longipedicellatus had the highest number of codons, with a total of 25,091 codons, while L. balansae had the fewest, with 24,886 codons. The two most frequently used amino acids in Lithocarpus species were leucine (Leu) and isoleucine (Ile), whereas cysteine (Cys) was the least common amino acid based on codons. There were six synonymous codons for Leu, serine (Ser), and arginine (Arg), but only one codon each for methionine (Met) and tryptophan (Try) (L. glaber is shown as an example in Fig. 3). Among the 64 codons, RSCU values of 30 codons were greater than 1.00, with 29 codons ending in A/U. Conversely, 31 out of 34 codons with RSCU values less than 1.00 ended with G/C (Table S2). Approximately half of the codons were used more frequently, as indicated by RSCU values exceeding 1 (Fig. 4). Notably, the codon usage of Lithocarpus plastomes showed a clear bias towards A/U at the third codon position. The codon usage and RSCU of the plastomes of 34 Lithocarpus individuals are summarized in Table S2.

Fig. 3.

Comparative analysis of plastome codon usage bias of Lithocarpus glaber. The colors indicate different codes, and the RSCU value frequency is illustrated as height in the upper diagram

Fig. 4.

The heat map of codon usage bias in the plastomes of Lithocarpus. The red color indicates higher RSCU values and the blue color indicates lower RSCU values

Analysis of SSRs

The number of SSRs observed in Lithocarpus species was high within the range of the family Fagaceae, comparable to that found in Quercus (Fig. 5B). The 34 Lithocarpus plastomes SSRs ranged in numbers of repeated from 117 (L. dealbatus) to 133 (L. tephrocarpus) (Fig. 5A). Among these, the mononucleotide repeats were the most abundant (62.02–68.22%), particularly A/T repeats (56.59–65.57%), followed by dinucleotide repeats (12.88–15.87%) (Fig. 5A, Table S3). Most of the SSRs (69%) were distributed in the intergenic spacer (IGS) region across all plastomes (Fig. 5C). However, a few SSRs with specific repeat units were unique to different Lithocarpus species. As shown in Table S3, the SSRs ACT/AGT, AAAC/GTTT, AACTC/AGTTG, AAATAT/ATATTT, and AATATC/ATATTG, among others, were unique to L. fohaiensis, L. grandifolius, L. elizabethiae, L. longipedicellatus, and L. gymnocarpus, respectively.

Fig. 5.

Analyses of simple sequence repeats (SSRs) in Lithocarpus plastomes. (A) Types and numbers of SSRs in the plastomes of Lithocarpus; (B) Types and numbers of SSRs in the plastomes of Fagaceae; (C) Distribution of SSRs in the intergenic spacer (IGS), coding sequences (CDS), intron, and Transfer RNA (tRNA) regions of the plastomes of Lithocarpus

Expansion and contraction of IR region in the Lithocarpus plastome

The LSC/IRb, SSC/IRa, and IRa/LSC boundaries among the 34 Lithocarpus plastomes were generally conserved. The LSC/IRb boundary genes were rps19 and rpl2, the SSC/IRa boundary was located within the ycf1 gene, and the IRa/LSC boundary genes were rpl2 and trnH. In contrast, the IRb/SSC junctions of plastomes within Lithocarpus were significantly variable and could be categorized into two distinct types. The expansion of the IR into the ndhF gene (type I) was observed in 16 Lithocarpus species, while the expansion into the ycf1 gene (type II) occurred in the remaining Lithocarpus species (Fig. S2). Notably, type I was found only in Lithocarpus and Fagus species (Fig. S3).

Plastome comparison and evolution

The gene arrangement of the 34 Lithocarpus plastomes was conserved (Fig. S4). Based on the comparison of Lithocarpus plastomes using mVISTA (Fig. S5), it is also evident that the plastomes of Lithocarpus exhibit a high degree of similarity. Overall, non-coding and single-copy (SC) regions exhibited more nucleotide divergence than coding and inverted repeats (IRs) (Fig. S5).

Our study showed that Ka or Ks values of certain genes of Lithocarpus plastomes are zero, rendering the calculation of Ka/Ks ratios unfeasible. After excluding these anomalous values, we assessed the selection pressure on 56 protein-coding genes by calculating the ratios of Ka to Ks substitutions. Most K_a/K_s ratios were less than 1 or undefined. A high K_a/K_s ratio (> 1) was observed in four genes (accD, atpF, rpl32, and rps8) with a p-value < 0.01, indicating these genes might be under a positive selection (Fig. 6).

Fig. 6.

Boxplots of Ka/Ks values among every shared protein-coding gene of Lithocarpus. Statistics analysis was performed using the one-sample t-test and the resulting probabilities (P-values *<0.05, **< 0.01, ***< 0.001) are shown

Plastome-based phylogeny

The phylogenetic trees reconstructed by ML and BI searching methods based on the three data partitions of the plastome (complete plastome, protein-coding exons, and non-coding regions) show the identical tree topology on the main clades, except for minor difference at the terminal tips (Fig. 7, S6, S7). Among these, the protein-coding tree inferred by the Bayesian inference method received the highest support values for the main clades (Fig. 7). Of which, the monophyly of most Fagaceae genera was strongly supported (bootstrap values ranging [BS] = 100, BI posterior probability [PP] = 1.00), except for Quercus, which exhibited polyphyly. Fagus and Trigonobalanus were resolved as a grade in two successive divergences. Notholithocarpus and Chrysolepis were nested within the Quercus clade, forming sister groups with Quercus subgenus Quercus and subgenus Lobatae, respectively. Castanopsis and Castanea were inferred to be sister taxa (BS = 93, PP = 1.00), while three Quercus species were determined to be sister taxa to Lithocarpus (BS = 98, PP = 1.00). Within Lithocarpus, two clades (Clade I and Clade II) were resolved. Clade I included L. obscurus (mainly distributed in Mêdog, Tibet) and L. dealbatus (from India). Clade II primarily comprised species from the southern Qinling Mountains of China, with two main subclades: Clade II-1, consisting of species from Southwest China, and Clade II-2, comprising species from the central-east China–Japan region (Fig. 8).

Fig. 7.

Phylogenetic relationships of Fagaceae inferred from ML and BI analyses based on the plastome protein-coding regions. The numbers near each node are bootstrap support values and posterior probability, respectively

Fig. 8.

Cladogram and geographic distribution of Lithocarpus. The map was created using the open-source R package phytools v2.0. The dots on the map show the sampling sites and the dot color indicates the main clades in Lithocarpus illustrated in Fig. 7

Leaf epidermal trichome and acorn morphology divergence pattern

The Clade I and Clade II inferred by the plasma-based phylogenetic tree were not supported by either the leaf trichome features or acorn type (Fig. 7). The only species with BBT trichome and ER acorn―L. corneus was inferred as a basal group in Lithocarpus, but ATP and Glabrous leaf trichome types and ER and AR fruit types are paraphyletic on the trimmed nuclear-based phylogenetic tree of Liu et al. [31] (Fig. 9).

Fig. 9.

Conflicts between the plastome (A) and nuclear (B, modified from the nuclear tree of Liu et al. [31]) species trees, visualized using the R package ape v5.8

Discussion

Plastome characteristics and adaptive evolution of Lithocarpus

The plastome structure across the 34 Lithocarpus samples was largely conserved, though variations in gene content, SSRs, and SC/IR boundaries were observed. Such variation is similar to that reported among other Fagaceae genera [42, 43, 46, 47]. The conservation of plastome structure is likely constrained by the need to maintain the stability of plastome functionality [80].

The protein-coding infA gene encoding translation initiation factor 1, which aids in the assembly of the translation initiation complex [81], was found to be a pseudogene in 17 Lithocarpus species. The loss and pseudogenization of infA have been previously documented in various genera of Fagaceae [42, 43, 82] and other flowering plant families [83, 84]. In some seed plant lineages, the infA gene has been identified as translocated from the plastome to the nuclear genome [80], possibly as a consequence of relaxed purifying selection on the plastome.

Typically, IR regions of angiosperm plastomes begin near the rps19 gene and end consistently downstream of either trnN-GUU or the truncated ycf1 gene [85]. While IR expansion has been documented in specific lineages, usually within the LSC region, the IR/SSC junctions are considered relatively stable [85]. However, this study discovered obvious IR region expansion at the IR/SSC boundaries in Lithocarpus (Fig. S2). A similar IR expansion was also detected in the early derived taxa of Fagaceae (i.e., Fagus) (Fig. S3). These findings suggest that IR expansions are independent events in the Fagaceae.

An analysis of different SSR repeat types revealed that mononucleotide repeats, particularly A/T repeats, were the most prevalent, while the remaining SSRs exhibited high A/T content. Such an A/T bias has been widely reported among Fagaceae plastomes [45–47, 86] and many other plant plastomes [87, 88]. These findings align with the proposition that the plastome not only exhibits abundant A/T content but also harbors a considerable number of short polyadenine (polyA) and polythymine (polyT) repeats [89]. Although the numbers of repeat motifs are similar across Lithocarpus, certain motifs are species-specific. These loci may contain crucial information that may be used to untangle the intraspecific genetic structure of Lithocarpus.

Our study identified four genes that underwent positive selection, including those involved in energy storage (accD), photosynthesis (atpF), and protein synthesis (rpl32, and rps8). Previous studies have shown that the accD gene, which encodes the β-carboxyl transferase subunit of acetyl-CoA carboxylase, plays a role in leaf growth [90, 91], leaf longevity [92], and fatty acid biosynthesis [93, 94]. These processes enhance photosynthesis and reserve energy, helping plants cope with seasonal resource constraints and defense responses [95, 96]. The accD gene has also exhibited signatures of positive selection in some evergreen angiosperm lineages in the tropics and subtropics, e.g., Alpinia [97], Pterocarpus [98], and some species of Araceae [99]. The wide distribution of Lithocarpus in heterogeneous environments of East Asia may boost genetic divergence.

The gene atpF encodes a subunit of H+-ATP synthase, which is essential for chloroplast electron transport, photorespiration, and stress resistance in plants [100]. ATP synthase is crucial for providing the energy required for photosynthesis [101]. Notably, the atpF gene is also associated with deciduous versus evergreen habits in oaks, showing stronger signatures of positive selection in the subalpine sclerophyllous oak Quercus aquifolioides compared to deciduous oaks from mesic habitats [45]. Similarly, the positive selection detected in the atpF gene in Lithocarpus may reflect its role in photosynthesis and in maintaining energy for year-round leaf retention in these species.

Additionally, rpl32 and rps8 encode ribosomal proteins L32 and S8, respectively [102]. Environmental stress can cause oxidative damage, affecting the translation system, particularly in systems of prokaryote origin such as the chloroplast [103]. These ribosomal protein-coding genes have undergone strong selection, possibly driven by regional adaptation. These genes likely help Lithocarpus maintain its leaves year-round, enhancing the species’ ability to acclimatize and cope with the hot and humid conditions in the tropics and subtropics.

Phylogenetic analysis

Our plastome-based phylogenetic trees showed similar topologies using three distinctly different data partitions, with only minor differences among the tips. The phylogenetic tree of protein-coding exons received the highest credibility support (Fig. 7, S6, S7). Generally, regions under relatively relaxed selection, such as introns, exhibit higher polymorphism than exons [104]. Consequently, coding regions typically show higher sequence homology than non-coding regions, as they contain fewer conflicting phylogenetic signals [105], often leading to higher resolution.

Compared to previous phylogenetic studies based on single or multiple locus DNA sequences [106, 107], our plastome tree provides a robust phylogenetic framework for the family Fagaceae, with major nodes showing strong support (i.e., PP = 1.00 and BS ≥ 93). The results further confirmed that whole plastome sequencing can enhance the phylogenetic resolution within a given lineage [108, 109]. Most fagaceous genera were resolved as monophyletic based on the plastome sequence data, e.g., Lithocarpus, Castanopsis, and Castanea, consistent with the results of previous studies [30, 31, 38], except for Quercus, which was inferred to be polyphyletic on the plastome tree in our study. This result is consistent with that obtained by Zhou et al. [30], suggesting possible a widespread ancient gene flow in the ancestral lineage [30, 31].

This study is the first to analyze the maternal evolutionary history of the genus Lithocarpus based on a large dataset of plastomes. Notably, our plastome-based phylogenomic tree resolved two main clades in Lithocarpus with high credibility support. One clade, composed of two species from the southern Himalaya lowlands, formed a sister group to the rest of the Lithocarpus species. The latter clade includes two subclades corresponding to the two geographical regions of Southwest China (Clade II-1) versus China–Japan (Clade II-2), suggesting possible phylogeographic structure in the plastome of Lithocarpus (Fig. 8). These findings are consistent with previous biogeographic studies using cp. DNA sequences on Lithocarpus by Cannon & Manos [26] and Yang et al. [29]. Non-recombining genetic units, such as plastomes, can show significant divergence even within continuous populations, owing to limited seed dispersal capabilities [110]. Lithocarpus species show the highest acorn morphological diversity within the Fagaceae family, and their acorns are generally large [106, 111]. Fagaceous acorns are primarily dispersed by gravity and/or by hoarding animals (e.g., rodents and jays) with limited dispersal abilities, meaning they are typically dispersed within 50 m of the maternal tree [112–114]. Additionally, fagaceous seeds are highly sensitive to moisture loss [115, 116], resulting in high mortality at the post-dispersal stage [117, 118]. The phylogeographic structure detected in the plastome-based phylogenetic tree of Lithocarpus in this study may indicate restricted seed-mediated gene flow in ancestral lineages.

Compared to previous phylogenetic reconstruction using nuclear DNA datasets [29–31], there is a noticeable discordance in tree topologies between the nuclear- and plastome-based datasets from Lithocarpus (Fig. 9). Similar discrepancies have also been observed in the deep nodes of other fagaceous lineages, especially Quercus [6, 30, 31]. Four hypotheses may explain this inconsistency: (1) convergent evolution of plastome sequences [119]; (2) introgressive hybridization [7, 120]; (3) incomplete lineage sorting [121]; and (4) different rates of pollen- and seed-mediated gene flow [122, 123].

In the present study system, the probability of sequence convergence across an entire plastome was low, given the large size of the plastome used in this study. Therefore, convergent evolution is unlikely to explain the observed incongruence. Hybridization and introgression commonly occur among oaks with sympatric distributions [124–126]. When interspecific gene flow is asymmetric, one parental species may experience assimilation of its nuclear genome, while its maternal plastome is retained in the populations. This phenomenon is commonly referred to as the introgression-induced chloroplast capture [127, 128]. Introgression-induced chloroplast capture has been identified as a mechanism that can distort phylogenetic relationships, often resulting in geographic clustering of introgressed taxa [129]. In the Fagaceae family, natural hybridization and introgression are commonly observed among the species within the same Sect. [126] and sometimes even among more distantly related species [124, 125]. Recently, phylogenetic work on Fagaceae has revealed a secondary increase in the speciation rate of Lithocarpus during the Oligocene and Miocene [30], suggesting that interspecific hybridization may have occurred during the early stages of its diversification. Incomplete lineage sorting among taxa is often associated with radiations [130, 131]. Previous research has suggested that Lithocarpus might have experienced such radiation during the Neogene [30]. Accordingly, the possibility of incomplete lineage sorting causing cytonuclear discordance cannot be totally discounted. Furthermore, pollen and seed dispersal are critical determinants of gene flow [132]. Gene flow via pollen is significantly greater than that occurring via seeds, leading to broader genetic exchange for the nuclear genome compared to the plastome [133]. Differences between seed- and pollen-mediated gene flow can result in cytonuclear discordance in phylogenetic studies, as previously observed in Quercus [122] and Carya [123]. In nature, Lithocarpus species are primarily pollinated by insects [134], and the long-distance transmission of pollen enhances gene flow among populations. In contrast, the seed dispersal of these species with large recalcitrant seeds is more limited [122, 123]. The contrasting patterns of pollen- and seed-mediated gene flow among the ancestral populations could contribute to the cyto-nuclear discordance observed in Lithocarpus. All these hypotheses should be tested in future studies.

Furthermore, the plastome-based phylogenetic framework does not align with key taxonomical groupings based on acorn morphology (e.g., enclosed receptacle fruit, ER; and acorn fruit, AC) [106], nor with leaf epidermis characteristics (e.g., Bubble Trichome Group, BBT; Appressed parallel tufts Group, APT; and Glabrous Group) [77] in Lithocarpus (Fig. 7). In contrast, the phylogenomic tree inferred from nuclear DNA appears to be more consistent with species groupings based on leaf epidermal features [29, 31], indicating that these fine morphoanatomical traits are phylogenetically informative within lower taxonomical ranks in the genus Lithocarpus. The paraphyletic pattern of the acorn traits on these nuclear trees may indicate that these traits have multiple independent origins, possibly as a consequence of convergent adaptation to cope with animal predation [111].

Conclusions

Lithocarpus plastomes are conserved in terms of their genome structure, size, gene arrangement, and codon bias, but they show variation in gene content, SSRs, and the borders of SC/IR sequences. Most of the plastome genes in Lithocarpus are under purifying selection, while four genes related to metabolizing, photosynthesis, and energy storage (accD, atpF, rpl32, and rps8) underwent positive selection, indicating their roles in adaptation to diverse environments. The monophyletic status of Lithocarpus was strongly supported by the present plastome-based phylogenies. The plastome tree topologies revealed geographically structured relationships that conflict with previous nDNA phylogenies. Cytonuclear discordance, observed in Lithocarpus and other genera within the Fagaceae, may result from ancient introgression, incomplete lineage sorting, and asymmetric seed- and pollen-mediated gene flow. Future investigations, including whole-genome high-throughput sequencing to compare plastome and nuclear evolutionary histories, are essential to test these hypotheses and understand the drivers of such discordance, shedding light on the patterns and processes of diversification among fagaceous species.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Abbreviations

AC

Acorn fruit

APT

Appressed parallel tufts group

BBT

Bubble trichome group

BI

Bayesian inference

bp

Base pair

BS

Bootstrap value

CDS

Coding sequences

cpDNA

Chloroplast DNA

CTAB

Cetyl trimethylammonium bromide

ER

Enclosed receptacle fruit

Hyb

Seq-Hybrid enrichment sequencing

IGS

Intergenic spacer region

IR

Inverted repeat region

ITS

Internal transcribed spacer

IUCN

International Union for Conservation of Nature

LSC

Large single copy region

ML

Maximum Likelihood

NGS

Next-generation sequencing

nrDNA

Nuclear ribosomal DNA

PE

Paired-end

PP

Posterior probability

PCGs

Protein-coding genes

rRNA

Ribosomal RNA

RSCU

Relative synonymous codon usage

SSC

Small single copy region

SSR

Simple sequence repeat

tRNA

Transfer RNA

YUKU

Herbarium of Yunnan University

Author contributions

M. D., XL.J., and SJ. Z. designed the research. CY. W. and LF. Y. collected and analyzed the data; LF.Y., XL. J., and M. D. prepared and revised the manuscript. All authors read and approved the final manuscript.

Funding

This work was supported by the Foundation for National Key Research and Development Program of China [Grant number 2023YFF1305001], National Natural Science Foundation of China [Grant number 32460060 & 31972858] and Yunnan Key Laboratory for Integrative Conservation of Plant Species with Extremely Small Populations [Grant number PSESP2021F01] and Southeast Asia Biodiversity Research Institute, Chinese Academy of Sciences [Grant number Y4ZK111B01]. The funders were not involved in the design of the research, collection, analysis, interpretation of data, and manuscript preparation.

Data availability

The 21 newly sequenced plastomes have been submitted to NCBI with accession numbers: OM112296-OM112297, OM112300, OM388302-OM388303, and PQ276656-PQ276671. The resulting DNA alignments and trees are available on GitHub (github.com/yanglifang116/Lithocarpus_plastomes).

Declarations

Ethics approval and consent to participate

The collection of all samples fully complied with national and local legislation. The plant samples used in this study were not listed as national key protected species and were not collected from national parks or nature reserves. In accordance with national and local laws, no specific permits were required for the collection of these plants.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.