Complete chloroplast genome of the medicinal plant Amomum compactum: gene organization, comparative analysis, and phylogenetic relationships within Zingiberales

Fresh A. compactum leaves were acquired from cultivated bases in Guangdong Province, China. The total cp DNA was extracted from roughly 100 g of leaves through an improved method by Li et al. [10]. The quality of cp DNA was checked by Nanodrop-2000 spectrometer (Nanodrop Technologies, Wilmington, DE, USA), and agarose gel electrophoresis. Pure cp DNA was used for shotgun library construction with 454 GS FLX Titanium platform. The obtained SFF file was preprocessed by trimming short (L < 50 bp) and low-quality (Q < 20) reads. Trimmed reads were assembled using Newbler V2.6 (GS FLX De Novo Assembler Software). In order to verify the assembly, the four junctional regions were further confirmed by Sanger sequencing.

Preliminary gene annotation of this cp genome was performed by CpGAVAS, a program available online (http://​www.​herbalgenomics.​org/​0506/​cpgavas) [11]. The position of each gene was then manually corrected by Apollo [12] after alignment to the reference genomes by MEGA 5.0. In addition, according to start and stop codons, minor revisions were performed. The tRNAs were further confirmed by the online tool tRNAscan-SE with default settings (http://​lowelab.​ucsc.​edu/​tRNAscan-SE/​). [13]. Then, the circular map of this cp genome was accomplished by OrganellarGenomeDRAW program (http://​ogdraw.​mpimp-golm.​mpg.​de/​) [14]. Finally, the complete cp genome of A. compactum was submitted to NCBI GenBank database (Accession Number: MG000589).

Relative synonymous codon usage (RSCU) values, which were used to research the features of variations in synonymous and nonsynonymous codon usage by disregarding the composition impact of amino acid, were determined using MEGA 6.0 [15]. Additionally, GC content and codon usage were determined by MEGA 6.0. SSRs (simple sequence repeats) loci were detected by MISA software (http://​pgrc.​ipk-gatersleben.​de/​misa/​), with following thresholds: ten, six, five, five, five, and five repeat units for mono-nucleotide, di-nucleotide, tri-nucleotide, tetra-nucleotide, penta-nucleotide, and hexa-nucleotide SSRs, respectively. To analyze the repeat structure, REPuter [16] (http://​bibiserv.​techfak.​uni-bielefeld.​de/​reputer/​) was performed to detect forward (direct) and palindromic (inverted) repeats in the cp genome. The minimum repeat unit was set to 30 bp in length, the identity of repeats was set to > 90%, and the Hamming distance equals three. All identified results were verified and redundant repeats were manually removed.

Pairwise alignments of several cp genome sequences were conducted by MUMmer [17], and the dot plots were drawn using a Perl script. The complete cp genomes of A. compactum and three other Zingiberaceae species (Additional file 1), Curcuma flaviflora (KR967361), Curcuma roscoeana (KF601574), and Zingiber spectabile (JX088661), were used for comparative analysis by mVISTA program (http://​genome.​lbl.​gov/​vista/​index.​shtml) [18] in Shuffle-LAGAN mode. A. compactum was set as the reference.

To examine the phylogenetic position of A. compactum, 14 complete chloroplast genomes were downloaded from NCBI. The 67 shared protein-coding gene sequences were extracted using a Python script and aligned separately by ClustalW2. Phylogenetical trees were reconstructed based on 67 concatenated protein-coding gene sequences by MP and ML methods. The best-fitting model was filtrated by jModelTest 2.1.7 through the Akaike information criterion (AIC) [19]. The MP tree was reconstructed by PAUP ver. 4.0b10 [20] with a heuristic search, while ML analysis was calculated by RAxML-HPC 2.7.6.3 on XSEDE in the CIPRES Science Gateway (http://​www.​phylo.​org/​) with default parameters. Based on APGIII, Fritillaria cirrhosa was set as an outgroup. Both MP and ML analyses used 1000 bootstrap replicates.

The Minimum Standards of Reporting Checklist includes details of the experimental design, statistics, and resources used in this study.

The complete cp genome sequence of A. compactum is 163,553 bp in length with a obvious quadripartite structure (Fig. 1). A pair of inverted region (IR) with 29,824 bp in length partition the rest sequence into a LSC region (88,535 bp) and a SSC region (15,370 bp) (Table 1). The universal GC content of this cp sequence was 36.0%, which has been reported to act a significant role in evolution of genomic structures. Nevertheless, the overall GC content is unequally distributed across the cp genome, which is lowest in SSC region (29.8%) but highest in IR regions (41.1%), followed by LSC region (33.7%).

As shown in Fig. 1, the A. compactum cp genome totally encodes 135 functional genes, that 113 are unique, containing eighty protein-coding genes, twenty-nine tRNAs and four rRNAs (Table 2). Among the functional genes, all rRNAs, eight tRNAs and seven protein-coding genes are duplicated in IR regions. The LSC region includes 60 protein-coding genes and 21 tRNAs, whereas the SSC region includes 11 protein-coding genes and one tRNA gene. Among the protein-coding genes, 72 are single-copy, whereas eight are duplicated. Among the tRNA genes, 20 are single-copy genes and nine are duplicated. Among the 113 unique genes, 13 include one intron (eight protein-coding and five tRNAs) and three (ycf3, clpP, and rps12) include two introns (Table 2). Unusually, the rps12 gene is trans-spliced, of which the 5′ end is situated in LSC region whereas two replicative 3′ ends are located in IRa and IRb regions respectively. What’s more, the ndhA gene contains the longest intron region (1033 bp).

The protein-coding gene sequences are 79,701 bp in length, which comprise 26,567 codons. And the usage frequency of codon was counted and exhibited in Table 3. In protein-coding sequences (CDSs), the AT content are 55.3% at the first codon positions, 62.0% at the second codon positions and 72.1% at the third codon positions, respectively (Table 1). Most protein-coding genes in land plant cp genomes use the standard ATG as the initiation codon. However, in the A. compactum cp genome, two genes use alternatives to ATG as start codon, as following: ATC for ndhD and ATA for rpl2. Relative synonymous codon usage (RSCU) is a statistics of uneven usage of synonymous and nonsynonymous codons in the coding sequences. An RSCU value < 1.00 indicates that the use of a codon is less frequent than expected, whereas a codon used more frequently will attain an RSCU value > 1.00. A total of 96.7% (29/30) of preferred synonymous codons, i.e., RSCU values > 1, end with A/U, whereas 90.6% (29/32) of non-preferred synonymous codons, i.e., RSCU values < 1, end with G/C. This codon usage pattern is similar with other reported cp genomes [21, 22], which might be driven by the high proportion of A/T. The usage of the start codon (ATG) and UGG (coding TRP) show no bias (RSCU value = 1).

SSRs are a class of tandemly repeated sequences that consists of 1–6 nucleotide repeat units. SSRs are important in plant typing and widely developed as molecular genetic markers for species identification. Total 58 SSRs loci were found in the A. compactum cp genome (Table 4), and 47 SSRs were only composed of A/T bases. Furthermore, 10 SSRs were composed of di-nucleotide (AT/TA) repeats, and one SSR was composed of trinucleotide (ATA) repeats. Obviously, the SSRs in the A. compactum cp genome were rich in A/T, which has been reported in many plant families [23‐25]. Among these SSRs, 17 SSRs were situated in protein-coding genes and one was located in a tRNA gene. Furthermore, five were in coding regions and 12 in intronic regions. No tetra-, penta- or hexa-nucleotide repeats over 15 bp long was detected. REPuter allowed us to identify 24 repeats, including 13 forward and 11 palindromic repeats (Table 5). Almost all repeats were situated in the intronic and intergenic regions, although few of them were situated in protein-coding regions [26]. As reported in other genomes, the gene richest in repeats was ycf2, carrying two direct and two palindromic repeats.

The variations of angiosperm cp genomes in length are mainly because of the contraction and expansion of boundary regions between the IR regions with single copy (SC) regions. A minute comparison of junctional regions between the IR and SC boundaries among A. compactum, C. flaviflora, C. roscoeana, and Z. spectabile is presented in Fig. 2. In addition, a size comparison of cp genome among the four Zingiberaceae species is shown in Additional file 2: Table S1. In spite of the alike lengths of IR regions in these four species (from 25,618 to 29,824 bp), few IR contractions/expansions were still detected. rpl22, ycf1 and rps19 pseudogenes with various lengths were situated in IRb/LSC or IRb/SSC boundaries. The borderline of the IRb/LSC junction was situated in left side of the rps19 gene in examined cp genomes, except in Z. spectabile, which resulted from the contraction of the IRa region in the Z. spectabile cp genome. By contrast, the ycf1 pseudogene was situated in the left side of the IRa-SSC border and was 4–5 bp away from the IRa-SSC borderline, except in the A. compactum cp genome. The size of the ycf1 pseudogene was 918 bp in A. compactum, 1068 bp in C. flaviflora and C. roscoeana, and 924 bp in Z. spectabile. In addition, in the A. compactum cp genome, the ycf1 pseudogene was 2969 bp away from the IRa-SSC borderline, that indicated the expansion of the IR region. The trnH gene was situated in LSC region, except in Z. spectabile cp genome, where it was situated in SSC region and was 136 bp away from the IRb-LSC borderline.

Three sequences representing the Zingiberaceae (C. flaviflora, C. roscoeana and Z. spectabile) were selected for comparison with A. compactum. Pairwise cp genome alignments between A. compactum and other three cp genomes regained a high degree of synteny (Additional file 3: Figure S1, Additional file 4: Figure S1 and Additional file 5: Figure S3). To detect the divergent regions in the cp genome, this study compared the sequence identities among four Zingiberaceae cp genomes by mVISTA, using the annotation of A. compactum as a reference. The multiple sequences alignment showed the coding regions are highly conserved, however the non-coding regions are divergent (Fig. 3). As an example, the intergenic sequences between the trnT-GGU–psbD, rps16–trnQ-UUG, atpH–atpI, trnE-UUC–trnT-GGU, trnT-UGU–trnL-UAA, petA–psbL and psaC–ndhE regions were highly divergent, parts of which have been also reported as divergent sequences in other plant. Obviously, the LSC region and SSC region were more divergent than IR regions.

Cp genomes are widely employed in the study of evolution through phylogenetics. To examine the phylogenetic position of A. compactum and its relationship within Zingiberales, MP and ML phylogenetical analyses were performed based on 67 protein-coding gene sequences from 15 plant taxa, including A. compactum, as sequenced in the study. The total alignment was 51,452 bp in length. The results are presented in Figs. 4 and 5. The basic topologies were similar in the MP and ML analyses, but there were few differences. Bootstrap values were all extremely high, and nine of the 12 nodes with bootstrap values of ≥ 90% were found in MP tree, whereas eight of 12 nodes were found in ML tree with 100% bootstrap values. The Zingiberaceae species A. compactum, C. flaviflora, C. roscoeana and Z. spectabile were grouped in both MP and ML phylogenetic trees with 100% bootstrap values. In the MP trees, the four Zingiberaceae species composed a unique clade and were separated from the rest of Zingiberales with high bootstrap values in every node. By contrast, the ML tree was mainly separated into two clades, one of which included Strelitziaceae, Heliconiaceae, Musaceae and Lowiaceae species, whereas another included Zingiberaceae, Costaceae, Cannaceae and Marantaceae species. However, the Zingiberaceae and Costaceae species were grouped with a very low bootstrap value (15%) in the ML tree. These phylogenetic results strongly support the position of A. compactum and provide some helpful hints about relationships within the order Zingiberales.