Introduction

Lardizabalaceae, a small family with approximately 50 species in 9 genera, is a core component of Ranunculales and belongs to the basal eudicots1, 2 . Decaisnea insignis (Griffith) Hook. f. & Thomson, which is widely distributed from central to south-western China and the Himalayan foothills, is the only species in the genus Decaisnea; it is nicknamed “dead man’s fingers” as it possesses racemes of strikingly deep purplish-blue elongated fruits3. This plant is economically important, as it is readily cultivated as an ornamental plant and its fruits are deemed a delicacy. It has also been used in traditional Chinese medicine as an antirheumatic and antitussive drug for a long time4. D. insignis is a type of wild resource plant and has a wide range of uses; thus, D. insignis is worthy of development and utilization. To support the development and utilization of this species, markers that are variable at the population level need to be developed from genomic resources. However, despite its importance, few studies have described the DNA sequences of D. insignis. Therefore, the molecular techniques is required to analyse the genetic diversity and phylogenetic relationship of this plant.

Chloroplast genomes have a typical quadripartite structure consisting of a large single copy region (LSC), a small single copy region (SSC) and a pair of inverted repeats (IRs) in most plants. The chloroplast genome is a highly conserved circular DNA ranging from 115 to 165 kb with a stable genome, gene content and gene order5,6,7. The substitution rates in plant chloroplast genomes are much lower than those in nuclear genomes8, 9. The angiosperms’ chloroplast genome has a uniparental inheritance and stable structure, providing sufficient genetic markers for genome-wide evolutionary studies at different taxonomic levels10,11,12. Although the chloroplast genome shows evolutionary conservation in plants, an accelerated rate of evolution has been widely observed in particular genes or some lineages5, 13. For example, rbcL, matK, and ycf1 have been used as DNA barcodes for barcoding plants14, 15. As a result of these characteristics, chloroplast genomes are considered to be good models for testing lineage-specific molecular evolution. With the development of high-throughput sequencing technologies, new approaches for chloroplast genome sequencing have been gradually proposed due to their high-throughput, time-saving and low-cost16.

In the present study, we reconstructed the whole chloroplast genome of D. insignis by using next-generation sequencing and applying a combination of de novo and reference-guided assembly. The objectives of this study were to establish and characterize the organization of the complete chloroplast genome of D. insignis and conduct comparative genomic studies to gain in-depth insights into the overall evolutionary dynamics of chloroplast genomes. Our data will also provide genomic resources for this species to determine its phylogenetic relationship with related species as well as a genetic diversity evaluation and plant molecular identification.

Results and Discussion

Chloroplast genome assembly

A total of 2.48 × 107 reads with an average read length of 150 bp were obtained. The de novo assembled contigs were analysed locally by BlANSTN using the Akebia quinata genome as a reference; seven contigs were retained. The gaps between the de novo contigs were checked by amplification. The total reads were re-mapped to the chloroplast genome, and correction of the sequences was confirmed. The coverage of the chloroplast genome was 2166×, and the sequence of the chloroplast genome was deposited in GenBank (accession number: KY200671).

Organization and gene content

The chloroplast genome of D. insignis was 158,683 bp in length (Fig. 1). The genome presented a typical quadripartite structure with two inverted repeats (each 26,167 bp in length) separated by one small and one large single-copy region (19,162 and 87,187 bp in length, respectively). These values were similar to those of Akebia (Table 1)17. The GC content of the chloroplast DNA was 38.5%. The GC content of the LSC (36.9%) and SSC regions (33.3%) was lower than that in IR regions (43.1%).

Figure 1
figure 1

Chloroplast genome map of D. insignis. The genes drawn outside of the circle are transcribed clockwise, while those inside are counterclockwise. Small single copy (SSC), large single copy (LSC), and inverted repeats (IRa, IRb) are indicated.

Table 1 Characteristics of the chloroplast genomes in three species.

The genome consisted of 113 different coding genes, of which 79 were protein-coding genes, 30 were distinct tRNA genes, and 4 were rRNA genes (Fig. 1, Supplementary Table S1). Of these, five protein-coding, four rRNA, and seven tRNA genes were duplicated in the IR regions. The LSC region comprised 62 protein-coding and 22 tRNA genes, whereas the SSC region comprised 12 protein-coding genes and one tRNA gene. Twelve genes contained introns, clpP and ycf3 comprised two introns, and the rest of the genes had one intron. In rps12, a trans-splicing event was observed, with the 5′ end located in the LSC region and the duplicated 3′ end in the IR region, as previously reported18, 19.

Among the 79 protein-coding genes, 75 genes had the standard AUG as the initiator codon, but psbC and rps19 used GUG, while rpl2 and ndhD used ACG. RNA editing events of the AUG initiation site to GUG had been reported for psbC 20 and rps19 21. Previous studies on non-canonical translational mechanisms suggested that the translational efficiency of the GUG codon was relatively higher compared with the canonical AUG as the initiation codon22. In Brassicaceae, psbC and rps19 also used AUG as the initiation codon21. ACG and GUG were used as the start codons for rpl2 and rps19, respectively, as reported in Oryza minuta 23.

Codon usage

We calculated the codon usage frequency and relative synonymous codon usage frequency (RSCU) in the D. insignis chloroplast genome. Codon usage plays an important part in shaping chloroplast genome evolution. Mutational bias has been reported to have an essential role in shaping this evolutionary phenomenon24, 25. The total protein coding genes comprised 78,375 bp that encoded 26,325 codons. Of these codons, 2,697 (10.3%) encoded leucine, whereas only 311 (1.2%) encoded cysteine (Fig. 2 and Supplementary Table S2), which were the most and the least frequently used amino acids in the D. insignis chloroplast genome, respectively. The AT content was 53.94%, 61.25%, and 68.70% at the 1st, 2nd, and 3rd codon positions, respectively. The preference for a high AT content at the 3rd codon position was similar to the A and T concentrations reported in other plants26. A general excess of A- and U-ending codons was noted. Except for TGA, CTA, and ATA, all preferred synonymous codons (RSCU > 1) ended with an A or U (Supplementary Table S2). Usage of the start codon AUG and tryptophan UGG had no bias (RSCU = 1).

Figure 2
figure 2

Codon content of 20 amino acid and stop codon of 82 coding genes of D. insignis chloroplast genome. Color of the histogram is the proportion of codon usage for amino acid and stop codon.

Repeat sequence

SSRs have been described as a major tool that can be used to unravel genome polymorphisms across species and perform population genetics of species on the basis of repeat length polymorphisms in plant molecular studies27,28,29. Because the chloroplast genome sequences are highly conserved, SSR primers for chloroplast genomes are transferable across species and genera. Considering the role of chloroplast genome SSRs as important phylogenetic markers, we applied a length threshold of greater than 10 repeats for mono-, 5 repeats for di-, 4 repeats for tri-, and 3 repeats for tetra-, penta-, and hexanucleotide repeat patterns. Eighty-three perfect microsatellites were analysed in D. insignis (Fig. 3). Mononucleotide SSRs were the richest with a proportion of 72.29%, and the mononucleotide A and T repeat units occupied the highest portion of 96.67%. Our findings agreed with the observation that chloroplast SSRs were generally composed of polyadenine (poly A) and polythymine (poly T) and rarely contained tandem guanine (G) and cytosine (C) repeats30, 31. Furthermore, there were 9 di-, 7 tri-, 6 tetra-, and one hexanucleotide repeats in the D. insignis chloroplast genome. Most SSRs were present in the noncoding regions of this chloroplast genome, and only one coding gene, ycf1, contained 4 SSRs. Forty-six spacer regions and eight intron regions harboured SSRs; the trnK-rps16 and petA-psbJ spacers had the highest number of indels (four), followed by the atpF intron and trnT-trnL (three). Most of these SSRs (95.18%) were present in the single copy region (Fig. 3 and Supplementary Table S3). Interestingly, the number of identified SSRs in the D. insignis chloroplast genome was low compared with the previously characterized SSRs. Additionally, we did not find a larger abundance of di- and tri-nucleotide repeats. Slipped strand mispairing (SSM) and intramolecular recombination had been suggested as the likely mechanism that led to most SSRs32. Based on the identified SSRs, we designed 79 primer pairs (except the four repeat SSRs in IR region), which could be used for future in-depth studies of phylogeography and the population structure pattern of this species (Supplementary Table S3).

Figure 3
figure 3

The distribution, type and presence of simple sequence repeats(SSRs) in the chloroplast genome of D. insignis. (A) Number of different SSRs types. (B) Proportion of SSRs in the LSC, SSC, and IR regions. (C) Number of identified SSR motifs in different repeat class types.

In addition to the SSRs, we explored the role of long repeats, as identified by REPuter33, with a minimal repeat size of 30 bp and a Hamming distance of 90. We identified 18 repeats, including 7 forward and 11 palindromic repeats (Table 2). The length of the repeats ranged from 30 to 73 bp. Approximately 40% of these repeats fell exclusively within genes, mainly in ycf2 (Table 2). Similar to the SSRs, D. insignis contained a lower number of repeat elements compared with other plants34, 35. The presence of these repeats indicated that the locus was a crucial mutation hotspot in the genome because the repeat sequences led to sequence variations and genome rearrangements due to the slipped-strand mispairing and improper recombination36. Additionally, these repeats played an important role in developing genetic markers for phylogenetic studies.

Table 2 Distribution and localization of tandem repeats in D. insignis chloroplast genome. F, forward repeat; P, palindrome repeat.

Comparative analysis of the D. insignis chloroplast genome and two Akebia species

A comparative analysis based on mVISTA was performed between the chloroplast genomes of D. insignis and two Akebia species (A. quinata and A. indica) to investigate the levels of sequence divergence (Fig. 4). The organization of the chloroplast genome between the D. insignis and Akebia genomes revealed a high degree of synteny and gene order conservation, suggesting an evolutionary conservation of these genomes at the genome-scale level. As expected, the IR region was more conserved than the LSC and SSC regions among the three genomes. Meanwhile, as seen in other flowering plants, the coding region was more highly conserved than the non-coding regions. The most dissimilar coding regions of the three chloroplast genomes were matK, ndhF, and ycf1, which were located in the LSC, SSC, and SSC regions, respectively. The matK and ycf1 coding regions had been observed to be divergent in chloroplast genomes and could serve as markers for DNA barcoding and phylogenetic analysis12, 15. The most divergent regions were localized in the intergenic spacers and introns, including the trnH-psbA, rps16-trnQ, ycf3-trnS, petA-psbJ, ndhF-rpl32, rps32-trnL as well as rps16 introns.  TrnH-psbA loci were highly variable in most plant groups, and inversions or mononucleotide repeats occurred within these loci, which might result in incorrect alignments or sequencing difficulties37, 38. Rps16-trnQ, ycf3-trnS, petA-psbJ, rps16 introns, ndhF-rpl32 and rps32-trnL had been used in previous phylogenetic studies39, 40. The most variable one of the identified loci was ycf1 encoding a protein of approximately 1,800 amino acids15. These non-coding regions could be used to assess phylogenetic relationships within the Lardizabalaceae species.

Figure 4
figure 4

Visualization alignment of chloroplast genome sequences of D. insignis, A. quinata and A. indica. VISTA based similarity graphical information portraying sequence identity of D. insignis with reference A. indica chloroplast genomes. Grey arrows above the alignment indicate the orientation of genes. Purple bars represent exons, blue ones represent introns, and pink ones represent non-coding sequences (CNS). A cut-off of 50% identity was used for the plots. The Y-scale axis represents the percent identity within 50–100%.

IR expansion/contraction also represents a highly variable pattern, which can be used to study the phylogenetic classification of plants. Moreover, the IR boundary expansion/contraction is regarded to be an evolutionary event and has been shown to be the reason for size variation in chloroplast genomes. Detailed comparisons of the IR-SSC and IR-LSC boundaries among the three Lardizabalaceae chloroplast genomes were presented in Fig. 5. The LSC/IRb border was located within the coding region of rps19 and created a pseudogene of 39 or 87 bp at the IRb/LSC border. The IRa/SSC border extended into ycf1, resulting in a pseudogene in the three compared chloroplast genomes. The length of the ycf1 pseudogene was 1,039 bp in the two Akebia species and 1,043 bp in D. insignis. Furthermore, ndhF deviated from the IRb/SSC in A. trifoliata, A. quinata and D. insignis by 166, 123, and 237 bp, respectively. The trnH-GUG gene was located in the LSC, which ranged from 27 to 79 bp from the IRa/LSC border. Overall, the IR boundary regions of the D. insignis chloroplast genome were slightly different from those of the other genomes in Lardizabalaceae.

Figure 5
figure 5

Comparison of chloroplast genome borders of LSC, SSC, and IRs among three Lardizabalaceae species.

Synonymous and nonsynonymous substitution rate

The synonymous and nonsynonymous nucleotide substitution patterns are very important markers in gene evolution studies. Estimation of these mutations plays a pivotal role in understanding the dynamics of molecular evolution41. In most genes, nonsynonymous nucleotide substitutions occur less frequently than synonymous substitutions due to the action of purifying selection. Accordingly, the ratio ω = dN/dS has become a standard measure of selective pressure with ω = 1, >1, <1 signifying neutral evolution, positive selection, and negative or purifying selection, respectively42. Using D. insignis as the outgroup, the nonsynonymous substitution (dN), synonymous substitution (dS) and dN/dS of gene groups and some genes in A. trifoliata and A. quinata were computed and compared (Fig. 6). As expected, a rate of heterogeneity existed among genes and gene groups. After sorting the genes into functional categories, significant differences were revealed among the groups. Analysis of gene groups indicated that the photosynthetic apparatus genes (psa, psb, pet) and atp had the lowest dN values relative to the other gene groups. Moreover, the photosynthetic apparatus genes and atp had the lowest dN/dS. The ycf1 and matK genes (dN/dS > 0.5) had the highest ratios of dN and dS, indicating that these genes were selected for sequence diversity. For dS values, there were no notable differences among the genes, except in the matK gene. It was noteworthy that the matK, ycf1, ccsA and ribosomal protein genes (rpl, rpo, rps) evolved faster than other genes. The gene function and locus-specific variation, selection pressure and gene expression level had been shown to influence the rates of sequence evolution in chloroplast genomes43, 44.

Figure 6
figure 6

Nonsynonymous substitution (dN), synonymous substitution (dS), and dN/dS values for individual genes or gene groups.

Phylogenetic inference

Chloroplast genome sequences have been widely used to reconstruct plant phylogenies, and the rapid improvements in sequencing technologies have led to the routine sequencing of complete chloroplast genomes45, 46. D. insignis belongs to the Lardizabalaceae family in the early diverging eudicots. To identify its phylogenetic position, 82 common genes were extracted from the chloroplast genomes of 33 species from all families of early diverging eudicots and others47. Ceratophyllum demersum was set as the outgroup.

After concatenating the alignment, all positions containing gaps and missing data were eliminated, and the sequence alignment comprised 68,668 characters. In the maximum likelihood (ML) and Bayesian inference (BI) tree, most of the nodes had a 100% bootstrap value and 1.0 Bayesian posterior probability. Both the ML and BI trees had similar phylogenetic topologies, which strongly supported the position of D. insignis as a sister of the closely related Akebia species in the family Lardizabalaceae (Fig. 7). The early diverging eudicot lineages, including the five major lineages (Ranunculales, Sabiales, Proteales, Buxales, and Trochodendrales), formed monophyly clades with 100% bootstrap support in the ML analyses and 1.0 Bayesian posterior probability in BI. Lardizabalaceae was a member of Ranunculales and was present at the basal position of the order. The phylogenetic positions of this group were in agreement with recent studies48. Although taxon sampling was inadequate and we could not perform a deeper phylogenetic analysis of Lardizabalaceae, our data would provide an example for future genome-scale phylogenetic studies in Lardizabalaceae.

Figure 7
figure 7

Phylogenetic tree reconstruction of 33 taxa using maximum likelihood and bayesian inference based on concatenated sequences of 82 genes. ML topology was shown with ML bootstrap support value/Bayesian posterior probability given at each node.

Conclusions

Using Illumina high-throughput sequencing technology, we obtained the complete sequence of the D. insignis chloroplast genome. The genomic organization and gene order of D. insignis were in agreement with the previously reported chloroplast genomes in Lardizabalaceae. The ML and BI phylogenetic trees strongly supported the position of Lardizabalaceae as a member of the order Ranunculales. The data obtained in this study will be beneficial for further investigations on D. insignis. Moreover, the availability of the chloroplast genomes provides a powerful genetic resource for the molecular phylogeny and biological study of this wild resource plant.

Methods

Taxon sampling, DNA extraction and sequencing

Fresh leaves of D. insignis were collected from a tree at the Research Institute of Forestry, Chinese Academy of Forestry. Fresh leaves were immediately dried with silica gel before DNA extraction. Total genomic DNA was extracted and purified following the method of Li et al.49. DNA was randomly fragmented into 400–600 bp using an ultrasonicator. An Illumina paired-end cpDNA library was constructed using the NEBNext® Ultra™DNA Library Prep Kit following the manufacturer’s instructions. Paired-end sequencing (2 × 150 bp) was carried out on an Illumina HiSeq 4000 platform.

Genome assembly and genome annotation

The paired-end reads were qualitatively assessed and assembled with SPAdes 3.6.150. Chloroplast genome sequence contigs were selected from the initial assembly by performing a BLAST search using the Akebia quinata chloroplast genome sequence as a reference (GenBank accession KX611091)17. The selected contigs were assembled with Sequencher 5.4.5. Ambiguous nucleotides, or gaps in the chloroplast genome sequences were filled by PCR amplification and Sanger sequencing. The four junctions between the inverted repeats (IRs) and small single copy (SSC)/large single copy (LSC) regions were checked by amplification with specific primers followed by Sanger sequencing51. Chloroplast genome annotation was performed with Plann52  based on chloroplast genome sequence of Akebia quinata . A chloroplast genome map was drawn using Genome Vx software53.

Codon usage

Codon usage was determined for all protein-coding genes. To examine deviations in synonymous codon usage by avoiding the influence of the amino acid composition, the relative synonymous codon usage (RSCU) was determined using MEGA 6 software54.

Analysis of single sequence repeats and tandem repeats

Perl script MISA (MIcroSAtellite; http://pgrc.ipk-gatersleben.de/misa) was used to detect single sequence repeats (SSR) within the chloroplast genome, with the parameters set at >10 for mononucleotide, >5 for dinucleotide, >4 for trinucleotide, and >3 for tetranucleotide, pentanucleotide, and hexanucleotide SSRs. REPuter was used to visualize the repeat sequences in D. insignis by forward vs. reverse complement (palindromic) alignment33. The minimal repeat size was set at 30 bp, and the identity of repeats was ≥90%.

Comparative genome analysis

The complete cp genome of D. insignis was compared with two Akebia species, A. quinata and A. indica, using the mVISTA program in a Shuffle-LAGAN mode55. A. indica was set as a reference. The chloroplast genome borders of LSC, SSC, and IRs were compared according to their annotations.

Synonymous and nonsynonymous substitution rate analysis

The relative rates of sequence divergence were analysed using the PAML v4.4 package56. D. insignis was used as an outgroup. The program yn00 was employed to estimate dN, dS, and dN/dS under a F3 × 4 substitution matrix using the Nei–Gojobori method. Genes with the same functions were grouped following previous studies5, 57,58,59. Analyses were carried out on datasets corresponding to the same functions, i.e., for atp, pet, ndh, psa, psb, rpl, rpo, and rps, and datasets corresponding to singular genes, i.e., for cemA, matK, ccsA, and ycf1.

Phylogenetic analysis

The chloroplast genome is uniparentally inherited and does not undergo recombination; thus, its constituent genes should track the same evolutionary history45, 46, 60. Therefore, in this study, we concatenated the 82 chloroplast genes for phylogenetic analysis without concern for strongly conflicting phylogenetic signals. A molecular phylogenetic tree was constructed using 33 angiosperms. The 30 completed chloroplast genome sequences representing the lineages of angiosperms, especially early diverging eudicots, were downloaded from the NCBI Organelle Genome Resource database. GenBank information for all of the chloroplast genomes used for the present phylogenetic analyses can be found in Supplementary Table S4. The sequences were aligned using MAFFT v761, and the alignment was manually adjusted.

The best-fitting model of sequence evolution was identified with jModeltest62 based on the Akaike Information Criterion (AIC). Maximum likelihood (ML) analysis was performed using the RAxML v 8.0.5 software package63 with 1,000 non-parametric bootstrap replicates.

Bayesian inference (BI) was implemented with MrBayes 3.2.264. Two independent Markov chain Monte Carlo (MCMC) chains were run, each with three heated and one cold chain for 10 million generations. The trees were sampled every 1,000 generations, with the first 25% discarded as burn-in. The remaining trees were used to build a 50% majority-rule consensus tree. Analysis was run to completion, and the average standard deviation of the split frequencies was <0.01.