Introduction

The genus Quercus L. (oak) comprises approximately 450 species worldwide, which is widely distributed in the Northern Hemisphere1,2. The innovative traits of Quercus and geoclimatic changes have facilitated its rapid spread and differentiation in the Northern Hemisphere over the last 56 million years3,4,5,6. Quercus is one of the largest woody genera and has important ecological, economic, and cultural functions2,7. Eight sections belonging to two subgenera were identified and documented in 20171,8. The subgenus Quercus contains the Lobatae, Ponticae, Protobalanus, Quercus, and Virentes sections, whereas the subgenus Cerris contains the Cerris, Cyclobalanopsis, and Ilex sections1.

Due to agriculture, biological resource-use, climate change, and other factors, an estimated 31% of oak species are threatened with extinction and 41% are of conservation concern2. Thus, the biological conservation of Quercus needs to be strengthened in terms of ecosystems, species, and genetic diversity. With the rapid development of high-throughput sequencing technology, increasing amounts of genomic information are available for use in conservation efforts9,10,11.

Chloroplast (cp) genomes are much smaller than nuclear genomes, normally ranging from 115 to 165 kilobase pairs (kb), with a pair of inverted repeats (IRs) separating a large single-copy (LSC) region and a small single-copy (SSC) region12. As they are non-recombinantly and uniparentally inherited and have low rate of nucleotide substitutions, cp genomes have been increasingly targeted to resolve the deep phylogeny of plants13,14. The exploration of gene rearrangements, structural changes, and repeat sequences has become an ideal strategy for species identification, population genetics, and genetic engineering15.

The Quercus section Ilex originated along the East Tethys seaway during the middle Eocene, followed by climate change and the Himalayan orogeny, which dominating the formation of the current distribution pattern16. The Quercus section Ilex can adapt to the most diverse climatic ecosystems and has the highest genetic diversity among all sections in Quercus17. Till now, previous studies have sequenced and published the cp genomes of 16% of Quercus species (70 species in Table S1, available until December 2023), including 49 species of the subgenus Cerris (30%) and 21 of the subgenus Quercus (8%). Currently, 22 species for which the cp genome has been sequenced in section Ilex, accounting for the largest proportion of all sections. However, very few cp genome resources exist for species threatened with extinction and species of conservation concern in the genus Quercus (< 10%). Among section Ilex, only three of the eight threatened species have been sequenced (Table S1). Therefore, additional genetic and genomic resources need to be explored for these threatened species.

Quercus marlipoensis (critically endangered; CR) and Q. kingiana (endangered; EN) are threatened species distributed in tropical area of Southeast Asia2. Except for phylogenetic investigations that construct nuclear gene trees, no other studies on all the other aspects of these two threatened species have been conducted. Therefore, we sequenced and assembled the cp genomes of Q. marlipoensis and Q. kingiana, and performed the comparative analyses of all cp genomes of section Ilex. In this study, we aimed to (1) examine abundant simple sequence repeats (SSRs) and highly variable regions in the whole cp genomes of Q. marlipoensis and Q. kingiana to identify markers for phylogenetic and genetic studies; (2) study the consistency and difference of cp genome structure within and among sections of Quercus to provide new insights into the evolution of cp genomes in each section; and (3) perform phylogenetic reconstruction and selection pressure analysis for all cp genomes of section Ilex to understand the development and evolution of this section. Our findings are expected to enrich the molecular data for phylogenetic studies and conservation of endangered species in the Quercus section Ilex.

Results

Structural characteristics of cp genomes

The length of the cp genomes for Q. marlipoensis and Q. kingiana were 160,995 and 161,167 bp, respectively. The two cp genomes had the same circular quadripartite structure, comprising a long single-copy (LSC) region, a small single-copy (SSC) region, and two inverted repeat (IR) regions (Fig. 1). The GC content of the two cp genomes were 36.8% and 36.9%, respectively. Furthermore, the GC content of the IR region (approximately 42.7%) was significantly higher than that of the LSC and SSC regions (34.7% and 31.07%, respectively) (Table 1). In addition, the name, number, and order of annotated genes in the two cp genomes were consistent. The 131 annotated genes included 86 protein-coding genes (PCGs), 37 transfer RNA (tRNA) genes, and 8 ribosomal RNA (rRNA) genes (Tables 1 and 2). Remarkably, 18 genes (seven PCGs, seven tRNAs, and four rRNAs) were duplicated in the IR regions. Eighteen genes contained introns, including 12 PCGs and 6 tRNAs. Nine PCGs (rps16, rpl2, rpl16, rpoC1, atpF, ndhA, ndhB, petB, and petD) and six tRNAs (trnA-UGC, trnG-UCC, trnI-GAU, trnK-UUU, trnL-UAA, and trnV-UAC) contained one intron, whereas the remaining three PCGs (ycf3, clpP1, and rps12) contained two introns (Table 2).

Fig. 1
figure 1

Gene map of the cp genomes of Q. marlipoensis and Q. kingiana. The outermost circle represents the genes annotated in the chloroplast genomes. Genes outside the circle are transcribed in the counter-clockwise direction, whereas those inside the circle are transcribed in the clockwise direction. Genes of different functional groups are identified by color. The length and boundary of the LSC, SSC, and two IR regions are indicated in the inner circle. Inside the inner circle, the dark gray area indicates GC content while the lighter gray corresponds to the AT content of the genome. LSC large single copy, SSC small single copy, IR inverted repeat.

Full size image
Table 1 Basic features of the cp genomes of Q. marlipoensis and Q. kingiana.
Full size table
Table 2 Genetic classification of the cp genomes.
Full size table

Repeated sequences

In the two cp genomes, 113 (Q. marlipoensis) and 122 (Q. kingiana) SSRs were detected (Table S2). The number of mononucleotide repeats was the highest, accounting for approximately 68% of the total, while the number of pentanucleotide repeats was the lowest. There were a slightly more tetranucleotide than trinucleotide repeats. No hexanucleotide repeats were detected in either species (Fig. 2A). Further, the proportion of A/T bases in the SSRs was significantly higher than that of G/C, indicating a strong A/T bias (Fig. 2B). Additionally, from the distribution of SSRs, most of the SSRs were located in the LSC and inter-genic spacer (IGS) regions, while a few were located in the IR and gene regions (Fig. 2C,D).

Fig. 2
figure 2

Type, distribution, and numbers of SSRs in the cp genomes of Q. marlipoensis and Q. kingiana. (A) The number of SSRs with six repeated types (mononucleotides, dinucleotides, trinucleotides, tetranucleotides, pentanucleotides, and hexanucleotides); (B) the number of SSRs with different base types; (C and D) the total number of SSRs in different regions of the two cp genomes. LSC large single copy, SSC small single copy, IRs inverted repeats, IGS inter-genic spacer, CDS coding sequence.

Full size image

Twelve and eight tandem repeats were also detected in the cp genomes of Q. marlipoensis and Q. kingiana, respectively (Table S3). The size of the basic repeat units mainly ranged between 20 and 29 bp (Fig. 3B). Among the four types of scattered repeats detected, the numbers of forward repeats (17 and 20) and palindromic repeats (19 and 24) were the highest, whereas the numbers of reverse repeats (3 and 4) and complementary repeats (0 and 1) were smaller in the Q. marlipoensis and Q. kingiana cp genomes (Fig. 3A). The scattered repeats ranged in size from 30 to 80 bp, mainly concentrated between 30 and 39 bp (Fig. 3C).

Fig. 3
figure 3

The distribution of the number and length of tandem and scattered repeats in the cp genomes of Q. marlipoensis and Q. kingiana. (A) Numbers of scattered repeats in the two cp genomes; (B) The length distribution of tandem repeats; (C) The length distribution of scattered repeats. F: forward repeats; R: reverse repeats; C: complementary repeats; P: palindromic repeats.

Full size image

Codon usage bias

According to the results of CodonW analysis, 17,070 and 16,929 codons were encoded by the PCGs extracted from the Q. marlipoensis and Q. kingiana cp genomes, respectively. The GC content and codon preference indexes of the two cp genomes indicated a weak usage bias in the codons (Table 3). The relative synonymous codon usage (RSCU) analysis of the two cp genomes indicated that 30 of 59 synonymous codons had RSCU values > 1, with only two codons ending in G/C (UUG and UCC) and the remaining 28 codons ending in A/U (Table S4). The results showed that the codon usage pattern of the cp genomes was inclined toward A/U endings. In addition, the RSCU value of the codons encoding leucine (Leu) was the highest (Fig. 4).

Table 3 Codon parameter characterization of the cp genomes of Q. marlipoensis and Q. kingiana.
Full size table
Fig. 4
figure 4

Relative synonymous codon usage (RSCU) analysis of the cp genomes of Q. marlipoensis and Q. kingiana.

Full size image

We also analyzed the factors influencing codon usage bias. In the PR2-bias-plot analysis, the four bases of the codon at the third position were unevenly distributed in the four regions of the PR2-bias-plot (A3 ≠ T3, G3 ≠ C3), indicating that codon usage bias varied with base position (Fig. 5A,B). In the neutrality-plot analysis, the GC12 and GC3 values were positively correlated, but the correlation was not significant. This suggested that the codon usage bias was influenced more by natural selection than other factors (Fig. 5C,D). Additionally, in the ENC-plot analysis, some genes were distributed along or near the standard curve, while others were located farther below the curve, indicating that codon usage bias was affected by natural selection and mutations (Fig. 5E,F). Through the above three analyses, we found that the codon usage bias of cp genomes was affected by base mutations, natural selection, and other factors.

Fig. 5
figure 5

PR2-bias-plot analysis (A and B); Neutrality-plot analysis (C and D); and ENC-plot analysis (E and F) of the cp genomes of Q. marlipoensis and Q. kingiana.

Full size image

Comparative chloroplast genomic analyses of Quercus section Ilex

The online software IRscope (https://irscope.shinyapps.io/irapp/) was used to assess the expansion and contraction of IR regions in 19 species of Quercus section Ilex. The junction of LSC/IRb (JLB) was located in the IGS of the rps19 and rpl2 genes. The rps19 gene in most species was shifted 11 bp away from the JLB, except for Q. semecarpifolia, which had a 12 bp shift. Both SSC/IRb (JSB) and SSC/IRa (JSA) boundaries were located within the ycf1 gene. The ycf1 gene at the JSB boundary expanded (2–93 bp), while the ycf1 gene at the JSA boundary contracted (4567 to 4636 bp). In addition, ndhF gene was distributed at the JSB boundary in Q. acrodonta, Q. bawanglingensis, and Q. franchetii; where it contracted to the IRb region by 22–28 bp (Fig. 6).

Fig. 6
figure 6

Comparison of the junction regions (JLA, JLB, JSB, JSA) among 19 cp genomes of Quercus section Ilex. The boundary genes are denoted with colored boxes. The number above the gene boxes indicates the distance between the end of the gene and the junction regions (JLA, JLB, JSB, JSA).

Full size image

The cp genomes of Quercus section Ilex were highly conserved. The genome structures and gene sequences of cp genomes had strong collinearity without gene rearrangement or inversion (Fig. S1). However, sequences in the non-coding regions were more variable than those in the coding regions. In addition, sequence variability in the SC regions was significantly higher than that in the IR regions. Further, we found high variability in the exon regions of three PCGs (rpoC2, rps19, and ycf1) and in the conserved non-coding segments of four IGS regions (rps16/psbK, psbM/trnD-GUC, psbZ/trnG-UCC, and rpl32/trnL-UAG) (Fig. S2).

Within Quercus section Ilex, polymorphic site analysis detected 2437 variable sites, including 1174 singleton variable sites and 1263 parsimony informative sites. We also calculated nucleotide diversity (Pi) values to determine the diversity level of the cp genomes. The Pi values ranged from 0 to 0.01673 with the mean value of 0.00392. We detected eight highly variable regions with Pi values greater than 0.01, with the highest Pi value of 0.01673 occurring in the IGS region between accD and psaI genes (Fig. 7).

Fig. 7
figure 7

Sliding window analysis of 19 cp genomes of Quercus section Ilex (window length: 800 bp; step size: 200 bp). The X-axis represents the nucleotide position of the window’s mid-point, and the Y-axis represents the nucleotide diversity (Pi) value per window.

Full size image

Comparison of chloroplast genome among different sections of Quercus

To study the consistency and difference of cp genome in each section from the perspective of genome structure and nucleotide diversity, we compared the cp genomes of the four sections of Quercus, including 19 of section Quercus, 37 of section Cyclobalanopsis, three of section Cerris, and 26 of section Ilex.

The expansion and contraction of IR regions were analyzed respectively in the four sections (Fig. S3–6). The results showed that four genes (rps19, ycf1-JSB, ycf1-JSA, and trnH) were located at four junctions and affected the expansion and contraction of IR regions. The boundary positions of the four boundary genes were exactly the same of the three species in section Cerris, which indicated that there was no contraction and expansion of IR regions in section Cerris (Fig. S5). There were differences in the boundary positions of the four boundary genes in the other three sections, among which section Ilex had the highest number of expansion and contraction types (Fig. S7). The rps19 and trnH genes in section Quercus and Ilex were contracted more significantly than section Cyclobalanopsis. While the ycf1_JSB and ycf1_JSA genes in section Cyclobalanopsis were expanded more significantly than the other three sections (Fig. 8).

Fig. 8
figure 8

The box plots of expansion and contraction of the four boundary genes in the four sections of Quercus (section Quercus, Cyclobalanopsis, Cerris, and Ilex).

Full size image

Nucleotide diversity analysis was also performed for these four sections. After polymorphic site detection, it was found that section Ilex had the most singleton variable sites among the four sections, showing the highest level of sequence diversity. However, section Cerris was the most conservative and contained the fewest variation sites (Table 4). The sliding window analysis obtained the nucleotide diversity value (Pi) between the whole genome sequences of each section, and it was found that section Ilex had the highest mean value, followed by section Quercus (Fig. 9).

Table 4 The distribution of singleton variable sites and parsimony informative sites in the four sections of Quercus.
Full size table
Fig. 9
figure 9

Nucleotide diversity values of cp genomes in the four sections of Quercus (section Quercus, Cyclobalanopsis, Cerris, and Ilex).

Full size image

Phylogenetic analysis

Using the whole cp genome data, the phylogenetic relationship of the two newly sequenced cp genomes in this study and their closely related species in the Quercus genus was reconstructed. A total of 33 taxa were used to reconstruct the phylogenetic tree, and the majority of branches had high bootstrap support values (Fig. 10). First, two distinct groups were recognized among all Quercus species, with the subgenera Quercus and Cerris. In the subgenus Cerris, the first clade was composed of section Cerris and part of section Ilex, whereas the second clade comprised section Cyclobalanopsis and another part of section Ilex. Quercus marlipoensis and Q. kingiana were nested within the second clade. Quercus marlipoensis had a close relationship with species from Himalayan subalpine areas. However, Quercus kingiana was more closely related to section Cyclobalanopsis.

Fig. 10
figure 10

Maximum Likelihood (ML) phylogenetic tree among 32 cp genomes of Quercus species and an outgroup Fagus engleriana. Values above the branch represent bootstrap support, where BS less than 50% is represented by a “—”.

Full size image

Selective pressure analysis

The selection pressure of common PCGs in the cp genomes of Quercus section Ilex was detected using the site model of the PAML v4.9j software. Using the ratio of dN (non-synonymous substitution rate) to dS (synonymous substitution rate), the evolution rates of the PCGs were calculated. We detected 27 and 36 PCGs with positive selection sites at M2 and M8 site models, respectively. Then, the likelihood ratio test (LRT) was performed for these 36 genes, where those with P < 0.05 were positively selected genes. A total of 11 PCGs were positively selected: ccsA, ndhB, ndhD, ndhF, rbcL, rpl22, rpl32, rpoC2, rps12, ycf1, and ycf2 (Table 5). Subsequently, based on the Bayes empirical Bayes (BEB) algorithm, the 11 PCGs had 103 positive selection sites (Table S5).

Table 5 Likelihood ratio test (LRT) and positive selection sites under different site models of PCGs in the Quercus section Ilex.
Full size table

Discussion

Chloroplast (Cp) genome characterization of newly sequenced species

High-throughput sequencing technology has accelerated chloroplast genomics research. However, less than 20% of Quercus species have completed chloroplast genome studies, and the data is still insufficient. In the current study, we assembled the complete cp genomes of Q. marlipoensis and Q. kingiana. Similar to other Quercus species, these two cp genomes exhibited a quadripartite circular structure18,19,20. In addition, their cp genome content and composition were similar to those of previously sequenced cp genomes, indicating high conservation across Quercus cp genomes21,22. However, the length of cp genomes differed between the two species. Such differences are predominantly caused by IR contraction and expansion, a common evolutionary phenomenon noticed across the cp genomes of angiosperms23,24,25.

Repeat sequences exist widely in plant genomes and have likely contributed to plant evolution through repeatability of IR regions26,27. We detected 235 SSRs in both species, which were largely composed of A/T repeat units. This observation is consistent with previous reports on the AT richness of cp genomes. Indeed, SSRs serve as valuable molecular markers and have been extensively studied across various research fields, including ecology, evolution, population genetics, and polymorphism studies at both species and population levels28,29,30. Our observation of most SSRs being distributed in the IGS and LSC regions is consistent with similar findings in other angiosperm lineages21,31.

GC content is a major factor in codon usage bias and may play an important role in the evolution of genome structure32. In our study, 30 out of the 59 synonymous codons exhibited RSCU values greater than 1. The third base of the codons was biased towards A/U, a phenomenon that was commonly noticed in angiosperms33,34,35. Codon usage bias is a natural phenomenon that is caused by mutations, selection, genomic composition, and other related factors36.

Sequence differences and evolution of Quercus

IR regions are crucial for stabilizing cp genome structure and are highly conserved across angiosperms. Variation in angiosperm cp genome length is primarily attributed to the expansion and contraction of IRs37,38,39, highlighting the importance of their analysis in evolutionary studies40. By comparing the four boundary regions of the cp genomes across 19 species of Quercus section Ilex, we found a consistent gene distribution and size, indicating strong uniformity in this group. In addition, the cp genomes in the Quercus section Ilex displayed good collinearity in the current study.

Both mVISTA sequence variation and nucleotide diversity analyses showed differences in the degree of variation among different regions of the cp genomes. Specifically, the SC regions exhibited higher variability compared to that in the IR regions. This variability may be caused by the presence of more conserved rRNA genes in the IR regions. Further, the IGS regions showed higher variations than that in the coding regions, a pattern consistent with observations in other Quercus species18,41. Importantly, these highly variable regions can be used to develop DNA barcodes for species identification and systematic classification42. The ycf1 gene and two IGS regions, trnH/psbA and trnK/rps16, have previously been identified as practical barcodes for plants31,43,44.

The cp genome structure of the four sections of Quercus was compared, and it was found that there were conservation and differences within and among the four sections. Within section Cerris, cp genomes exhibited high conservation, most likely due to their close phylogenetic relationships45. However, the differences within other sections may be the result of adaptations between species to the local environment. Of course, a small percentage of the differences may be due to personal errors encountered in assembly and annotation, which require us to proofread manually. There were also significant differences in cp genome structure among different sections. The expansion and contraction of boundary genes of IR regions in section Cyclobalanopsis were mostly significantly different from those in other sections. This may be due to the fact that this section contained more species, so the structural differences between sections still need to be further explored by expanding the number of species. Nucleotide diversity analysis showed that section Ilex had the highest nucleotide diversity, which was consistent with other studies in chloroplast markers17,46.

Phylogeny of Quercus section Ilex

Due to complex evolutionary problems such as convergent evolution and hybridization introgressions, studies on the phylogenetic relationships of Quercus face great challenges47,48,49. The cp genomes are valuable tools for understanding phylogenetic relationships in angiosperms50. In our study, we analyzed cp genomes from 32 Quercus species to reconstruct their phylogenetic relationships, which were inconsistent with classification systems based on nuclear markers1,51. The different evolutionary patterns revealed by the oak chloroplast and nuclear genomes are likely due to historical introgression of ancestral lineages and recent or ongoing gene flow between closely related species46. Notably, sections Cyclobalanopsis and Cerris were found within the section Ilex. This finding suggests possible introgression among ancestral taxa or chloroplast capture. Similar phenomena have been observed in white oaks52 and South American Nothofagus53, attributed to cytoplasmic nuclear inconsistencies. A similar unresolved complex phylogenic relationship within subgenus Cerris was also found in other studies using cpDNA markers46. The chloroplast capture events within a genus are mostly due to hybridization53,54. However, the hybridization between the extant sections in oaks is extremely rare in the wild55. Although the phylogenetic relationships of the Quercus section Ilex are complex, the development of sequencing techniques will allow for the inclusion of more taxa and samples in future studies. Such progress will facilitate further exploration of the interspecific relationships and phylogenomics of the Quercus section Ilex. These results are likely to further our understanding of taxonomy, phylogenetic evolution, and conservation efforts related to Chinese Quercus species.

Adaptive evolution of Quercus section Ilex

The Quercus section Ilex is widely distributed in habitats across humid to semi-arid regions in the Eurasian tropics and subtropics, making it an important regional vegetation group56,57. While the genetic homogeneity of section Cyclobalanopsis implicates selection as a driving force in its plastid evolution, section Ilex exhibits the hallmarks of geographic isolation17. At the cp genome level, 11 protein-coding genes (PCGs) were positively selected during evolution, which may be related to environmental suitability. Four of these are photosynthesis-related genes, including three NADH dehydrogenase genes (ndhB, ndhD, and ndhF) and the RubisCO large subunit (rbcL). These genes are known to be adept to light environmental stress and are commonly positively selected across many species58,59,60. Further, four genes are involved in the transcription and translation processes: rpl22, rpl32, rps12, and rpoC2. The first three encode ribosomal proteins and control protein synthesis, while rpoC2 encodes RNA polymerase and is essential for its function. The ccsA is responsible for C-type cytochrome synthesis, is thought to be linked to the binding of the C-type cytochrome and heme, potentially driving adaptive evolution61. Finally, ycf1 and ycf2 are the two largest hypothetical chloroplast reading frames within the cp genome. While their encoded products are believed to be crucial for chloroplast function62, their specific roles and their evolutionary significance remain unclear. These genes serve as a valuable gene set for further investigation into the mechanisms of adaptive evolution in Chinese Quercus species63.

Conclusions

In this study, the cp genomes of two threatened section Ilex species were described and compared with those of other Quercus species in NCBI. The results showed that the Quercus cp genomes were similar in quadripartite structure, GC content, codon usage features and gene order. Despite significant conserved overall cp genome structure and gene content, significant sequence differences were found in alternating regions of these genomes. Seven highly divergent regions (rpoC2, rps19, ycf1, rps16/psbK, psbM/trnD-GUC, psbZ/trnG-UCC, and rpl32/trnL-UAG) might be used as phylogenetic molecular markers. The nine positively selected sites identified in this analysis included photosynthesis-related genes (ndhB, ndhD, ndhF, and rbcL), transcription and transcript processing genes (rpl22, rpl32, rps12, and rpoC2) and could facilitate understanding of the adaptive evolution of Quercus genus. Overall, the data obtained will contribute to further studies on the diversity, ecology, taxonomy, phylogenetic evolution and conservation of Quercus genus.

Methods

Plant material, DNA extraction, and sequencing

Tender, healthy leaf samples of Q. marlipoensis (98°93′E, 18°79′N; Altitude, 2250 m) and Q. kingiana (104°83′E, 23°15′N; Altitude, 1789 m) were harvested from Malipo county and Simao city in Yunnan province, China. Silica gel was used to desiccate the collected material. Voucher specimens were deposited in the herbarium of the Shanghai Chenshan Botanical Garden. High-quality DNA was extracted using a modified cetyltrimethyl ammonium bromide (CTAB) protocol64 and its purity, concentration, and integrity were determined. Double-terminal sequencing was performed using a high-throughput sequencing platform DNBSEQ (http://www.bgitechsolutions.com/), with a read length of 150 bp. The raw data were filtered using SOAPnuke v1.3.0 to obtain 20 GB of clean data65.

Chloroplast genome assembly and annotation

The cp genomes were assembled de novo into rings using GetOrganelle version 1.7.6.1 in a Linux system66. The assembled cp genomes were annotated using the GeSeq online site (https://chlorobox.mpimp-golm.mpg.de/geseq.html)67. Finally, cp genome graphs were drawn using the online program Organellar GenomeDRAW (OGDRAW) version 1.3.1 (https://chlorobox.mpimp-golm.mpg.de/OGDraw.html)68. We used Geneious R9.0.2 software and cloud platform tools (http://cloud.genepioneer.com:9929) to perform statistics on the basic characteristics of the cp genomes. The newly cp genome sequences were uploaded and saved to the NCBI database under GenBank accession numbers OR966903 and OR966904.

Repeated sequences and codon usage bias

Simple sequence repeats (SSRs) were detected via batch analysis using MISA Perl scripts (https://webblast.ipk-gatersleben.de/misa/)69. The minimum number of repeats of mononucleotides, dinucleotides, trinucleotides, tetranucleotides, pentanucleotides, and hexanucleotides were set to ten, five, four, three, three, and three, respectively. We then used Tandem Repeats Finder (TRF) v4.09 software (https://tandem.bu.edu/trf/trf.html) to detect tandem repeats70. The alignment parameters match, mismatch, and indels were set to two, seven, and seven. respectively, and the minimum alignment score to report repeats was set to 80. In addition, we also used the website REPuter (https://bibiserv.cebitec.uni-bielefeld.de/reputer) to predict scattered repeats within each cp genome71. The scattered repeats included forward/direct repeats (F), reverse repeats (R), complement repeats (C), and palindromic repeats (P). Three and 30 were set as the hamming distance and minimum repeat size, respectively.

CodonW version 1.4.2, a common tool for codon bias analysis, was used to perform codon bias analysis and statistics on the coding sequences (CDSs) in these two cp genomes72. We screened CDSs with ≥ 300 bp for subsequent analysis. The GC content at the first, second, and third codon sites (GC1, GC2, and GC3) and the overall mean value (GC_all) of the two cp genomes were calculated. The effective number of codon (ENC), codon bias index (CBI), codon adaptation index (CAI), frequency of optical codons (FOP), and codon numbers were calculated using CodonW version 1.4.2. RSCU values were calculated and presented graphically using the R package (ggplot2). Neutrality, ENC-plot and PR2-bias-plot analyses were performed with different data to analyze the influencing factors and degree of codon usage bias.

Comparative genomic analyses of Quercus section Ilex

The currently published 17 cp genomes of Quercus section Ilex species, were downloaded from the National Center of Biotechnology Information (NCBI) database for comparative genomic analyses along with the two species in this study. The four boundary regions of each cp genome were visualized using IRscope (https://irscope.shinyapps.io/irapp/). The expansion and contraction of the IR regions were compared to identify differences in the cp genomes of Quercus section Ilex73. We used the Mauve plugin in Geneious R9.0.2 software to analyze structural changes in the cp genomes of Quercus section Ilex74. To evaluate the similarities and differences among cp genomes, the cp genome of Q. bawanglingensis was used as the reference sequence, and the online analysis tool mVISTA (http://genome.lbl.gov/vista/mvista/submit.shtml) was used to visualize the cp genomes of other species of Quercus section Ilex. DnaSP v6.12.03 software was used to detect polymorphic sites and nucleotide diversity (Pi values). A sliding window analysis with a step size of 200 bp and a window length of 800 bp was performed to identify regions with high variability in the cp genomes.

Comparison of chloroplast genome among different sections of Quercus

We downloaded all cp genomes of the four sections of Quercus from the NCBI database, including 19 of section Quercus, 37 of section Cyclobalanopsis, three of section Cerris, and 24 of section Ilex (available until July 2024). Combined with the two newly sequenced species of section Ilex in our study, the expansion and contraction of the IR regions and nucleotide diversity analysis of each section were carried out in the four sections, respectively. Then we performed statistics and significance analysis to explore the consistency and differences among the sections in the software Origin.

Phylogenetic analysis of Quercus

The phylogenetic relationships of Quercus were inferred using the maximum likelihood (ML) method based on the complete cp genomes. Fagus engleriana (NC_036929) was selected as the outgroup for tree rooting purposes. The whole cp genomes of 33 species were aligned using the multi-sequence alignment program MAFFT v7.45075. The best nucleotide substitution model was determined, and a phylogenetic tree was built with 1000 ultrafast bootstrap replicates using IQ-tree v2.1.3 software76. The resulting phylogenetic tree was constructed and saved using the FigTree v.1.4.4 software (http://tree.bio.ed.ac.uk/software/figtree/).

Selection pressure analysis

The Codeml program in the PAML v4.9j software was used for positive selection analysis, aiming to explore the adaptive evolution of protein-coding genes (PCGs) in the cp genomes of Quercus section Ilex77. The non-repeating PCGs were aligned using MEGA-X v10.2.6 software, followed by the removal of their stop codons. Subsequently, the aligned sequences were converted into a recognizable PAML format using DAMBE v7.3.32 software78. A phylogenetic tree was then constructed based on PCGs using IQ-tree v2.1.3 software. The selection pressure of 77 common PCGs was identified using six site models (seqtype = 1, model = 0, NSsites = 0, 1, 2, 3, 7, and 8). The posterior probability of each site was calculated using Bayes empirical Bayes (BEB) to identify positively selected sites (P > 0.95). For genes exhibiting positive selection sites, a likelihood ratio test (LRT) was carried out through pairwise comparison of M0 (single ratio) vs. M3 (discrete), M1 (near neutral) vs. M2 (positive selection), and M7 (beta) vs. M8 (beta & ω). The chi-square test (chi2) was used to detect statistical significance, with genes exhibiting P < 0.05 considered as positively selected genes79.