Comparative analysis of the complete chloroplast genomic sequence and chemical components of Cinnamomum micranthum and Cinnamomum kanehirae

Plant material and DNA extraction:

During winter (October–January) in Taiwan, 23 fresh leaf samples of C. micranthum were collected for chemical analyses from Taipei, Ilan, and Natou, while 23 C. kanehirae samples were collected from 13 sites throughout the Taiwan island, including sampling areas covered by previous studies (Lin 1993; Liao et al. 2010). Wood materials (with height over 3 m) were collected from 12 individuals from Fushan Botanical Garden in Ilan and Liouguei Research Center in Kaohsiung city of Taiwan. The diameters of wood discs collected were over 60 mm. The C. micranthum wood samples were fresh or air-dried for 2 years; the C. kanehirae wood samples were air-dried for 2 or 4 years.

The DNA extraction of leaf samples followed the CTAB method (Doyle and Doyle 1990). The extraction of sapwood DNA followed the DNeasy Plant Mini Kit (Qiagen, Hilden, Germany) protocol according to Jiao et al. (2014). The 2-μg DNA sample was subjected to quality checks before next-generation sequencing (NGS) using a Qubit fluorometer with a Quant-iT dsDNA BR assay kit (Thermo Fisher Scientific, Waltham, MA, USA) and agarose gel electrophoresis assay.

Next-generation sequencing and chloroplast genome assembly:

Genomic DNAs were sequenced with the Illumina Miseq system platform (Illumina, San Diego, CA, USA) and a MiSeq Reagent Kit V3 (600 cycles) at National Yang-Ming University, Taiwan. The paired-end libraries were prepared with Illumina TruSeq DNA polymerase chain reaction (PCR)-free sample preparation (550 bp insert size). The adapter sequences and low-quality score of the raw reads were trimmed by the CLC quality trimming tool with default parameters, and all the trimmed reads were used for de novo assembly with the CLC genome assembler (Genomics Workbench 8.0.1, CLC, Rarhus, Denmark). The assembled contigs were then submitted to BLASTn (cut-off e-value=1e-5) with the reference cp genome (C. kanehirae, KR01425), (Wu et al. 2016). The contigs, which have high similarity to the C. kanehirae cp DNA, were manually assembled into a circular genomic sequence. Validations of the overlapping regions between contigs were done by PCR and all of the primers and PCR conditions are described in Supplementary Table S1. BWA v0.6.2 (Li and Durbin 2009) and IGV v2.3.25 (Robinson et al. 2011) were also consulted for mapping all raw reads onto the final assembly and visually inspected the genome coverage levels.

Genome annotation:

Protein coding, ribosomal (r)RNA, and transfer (t)RNA genes of the cp genome were predicted with DOGMA (Wyman et al. 2004) and CpGAVAS (Liu et al. 2012). The start and stop codons of all predicted genes were manually corrected. The C. kanehirae cp genome (KR014245) served as a reference to adjust the gene annotation. Eventually, a complete C. micranthum cp genome was obtained and deposited in the NCBI database (KT833081). The cp genome circle map was drawn with OGDRAW (Lohse et al. 2007).

Comparative analysis within the chloroplast genomic sequence:

A WebSat online analysis (Martins et al. 2009) was used to find simple sequence repeat (SSR) regions, with the repeat units being 12 for mono-nucleotides, six for di-nucleotides, five for tri-ncleotides, and three for tetra-, penta-, and hexa-nucleotides. A repetitive structure analysis was conducted with REPuter (Kurtz et al. 2001), (minimum size of repeat event was 20 bp). InDel events were calculated with the DNaSP 5.0 software (Librado and Rozas 2009). Long InDel regions were found with MEGA 6.0 by alignment (Tamura et al. 2013).

Phylogenetic tree construction:

Sixteen complete cp genome sequences related to C. micranthum served for constructing the phylogenetic tree, including all available complete cp genomes in Lauraceas. The complete cp sequences were aligned with MAFFT (Katoh and Standley 2013) with default setting. A maximumlikelihood phylogeny was constructed using MEGA 6.0 (Tamura et al. 2013) with the GTR+I+G model. Bootstrap-values were calculated with 1000 replicates. Eight Magnoliales, Magnoliaceae species: Magnolia tripetala (KJ408574), Magnolia liliiflora (JN867576), Magnolia denudate (NC018357), Magnolia yunnanensis (KF753638), Magnolia kwangsiensis (NC015892), Magnolia grandiflora (NC020318), Magnolia officinalis (NC020316), and Liriodendron tulipifera (DQ899947), and eight Laurales, Lauraceae species: C. kanehirae (KR014245), C. micranthum (KT833081), Machilus yunnanensis (KT348516), Machilus balansae (KT348517), Phoebe omeiensis (KX437772), Phoebe sheareri (KX43773), Persea americana (KX43771), and Calycanthus fertilis (AJ42841) formed the basis for the phylogenetic tree construction. The Drimys granadensis (DQ887676) (Canellales, Winteraceae) served as the outgroup.

PCR amplification and sequencing for validating the InDel molecular markers:

For InDel validation, primer sets were designed with the Primer 3.0 software (Untergasser et al. 2012). The PCR contained 10 ng of genomic DNA and was performed with a Taq polymerase kit (GenetBio, Daejeon, Korea) following the manual in a total volume of 50 μl. The thermocycler Veriti (Applied Biosystems, Foster city, CA, USA) was used for all PCRs with the following cycling parameters: 94°C for 5 min; 32 cycles of 94°C for 30 s, the annealing temperature for 30 s (see Supplementary Table S1), and 72°C for 1 min; and a final extension at 72°C for 10 min. The details of the PCR primer sets are shown in Table 2 and in Supplementary Table S1. Forty-six leaf and 12 wood samples were used for InDel primers’ validation. PCR products were analyzed by 2–3% agarose gel electrophoresis (Agarose SFR™, AMRESCO, Solon, OH, USA) and detected by ethidium bromide under UV light for visualization.

Solvent extraction and the ultra-performance liquid chromatography analysis:

Fresh leaves of C. micranthum and C. kanehirae trees were sampled in the same winter season. They were cut into pieces, and 1 g of fresh weight was sampled with three replications. Each sample was put into a 50-ml centrifugal tube, 10-ml MeOH was added, the mixture was ground to a fine powder in a homogenizer, and then rinsed with 10-ml MeOH. These tubes were cooled in an ice bath and ultrasonicated for 60 min. The supernatant was collected in a centrifuge at 2500 rpm. The residue was re-extracted with 15-ml MeOH, and a 2^nd supernatant was obtained with the same procedure. Clear supernatants were combined and passed through a 0.2-μm filter (Xiboshi, Hydrophilic PTFE, Tianjin, China). The UPLC analyses were carried out with Acquity UPLC Waters (Waters, Milford, MA, USA) equipped with a C-18 column (150x2.1 mm, 1.8-μm particle size; Acquity UPLC HSS T3, Waters). The mobile phase for elution was distilled-deionized water (DDW): acetonitrile (70:30) for 23 min, DDW graded down to 30% for 25 min, increased up to 70% for 5 min, and maintained for 5 min at a flow rate of 0.3 ml min⁻¹, followed by UV detection at 200 nm. A standard curve was obtained based on authentic terpinen-4-ol (Acros Organic, Geel, Belgium), linalool (Acros Organic), sesamin (Sigma-Aldrich, St. Louis, MO, USA), and safrole standards (Acros Organic). Concentrations of extracts were expressed as % b.o. dry wt. For dry-weight measurements, residues were dried at 60°C for 48 h after extraction.

De novo assembly and gene organization

Raw data obtained from NGS technology of a total 3634256229 bp (12491350 reads) were from C. micranthum sequencing. After trimming, 12196250 clean reads with 1806905010 bp were obtained and used for de novo assembly (Table 1). Three long contigs which covered the entire cp genome sequence of C. kanehirae (KR014245) were used for BLASTn and showed high similarity (99.61–99.9%) to the C. kanehirae large single copy (LSC), small single copy (SSC), and inverted repeat (IR) region. It can be inferred that these three long contigs are the LSC, IR, and SSC, respectively. The two IR sequences were reversed and complementary; the assembly software could not combine these three contigs as one contig. This result is similar to the assembly of the Tartary buckwheat (Fagopyrum tataricum) cp genome based on NGS (Cho et al. 2015). The three long contigs revealed about 10–15-bp overlap between each of their flanking regions and then were manually assembled as a single circular cp genome. The Sanger sequencing of overlapping regions (LSC-IRa, IRa-SSC, SSC-IRb, and IRb-LSC) was 100% identical with the NGS-based assembled cp genome, and these sequences formed the basis for drawing a circle map (Figure S1). The final assembly resulted in an average of 425-fold coverage of pair-end reads with a mapping quality of 60. The complete C. micranthum cp genome was 152675 bp in length (GenBank no. KT833081), which was 25 bp shorter than the C. kanehirae cp genome. Both cp genomes consisted of one LSC, one SSC, and two IRs. The respective lengths of the LSCs in C. kanehirae and C. micranthum were 93642 and 93662 bp. The SSC of C. kanehirae was 18845 bp and that of C. micranthum was 18875 bp. The total lengths of the two IRs in C. kanehirae and C. micranthum were 40212 and 40138 bp, respectively. Two genes, ndhF and trnH, are located at the IR/LSC and SSC/IR junction areas and have different lengths of expansion and contraction. IR contraction and expansion at the junction of the SSC or LSC lead to cp size variations in different plants and contribute to the formation of pseudogene (Lin et al. 2012). The ycf1 and ycf2 span the junctions of the IRa/SSC and LSC/IRa, respectively, showed incomplete duplication and caused the contraction of the IR regions. This result is consistent with other Lauraceae cp genomes (Song et al. 2015, 2016). The overall GC content and gene contents of the C. micranthum cp genome are listed in Table 1. The overall GC content also falls in the ranges of vascular plants (38~39%) (Palmer and Stein 1986). The value 39.1% (Table 1) is almost the same as that in other plants in the Lauraceae (Song et al. 2016). The gene contents between C. micranthum and C. kanehirae were the same (Table 1). Eight of 112 genes (trnV-UAC, psaA, rps2, atpF, trnK-UUU, ycf1, ycf2, and ndhB) had different sequence orientations from those of C. kanehirae. Among these annotated genes, 97 genes were a single copy and 15 genes were duplicated. Among the 78 protein coding genes in the C. micranthum cp genome, 63 genes were located in the LSC region, 11 genes in the SSC region, three genes in the IRa, and five genes in the IRb. Ten protein-coding genes and five tRNA genes contained introns in C. micranthum, and among these, 10 protein-coding genes and eight genes (atpF, petB, ndhA, ndhB, rpoC1, rps12, rps16, and rpl2) had one intron, and two genes (ycf3 and clpP) had two introns.

Distribution of SSRs and repetitive structures

The cp SSR regions may be variable and typically non-recombination, uniparentally inherited and effectively haploid, and have been used for population genetic studies in many plant species (Navascues and Emerson 2005; Desiderio et al. 2012). Within the two cp genomes of C. micranthum and C. kanehirae, the SSRs were analyzed. In total, 43 SSR loci accounted for 595 bp of the total sequence in C. micranthum (0.39%). In the C. kanehirae cp genome, 39 SSR loci were found which accounted for 520 bp of the total sequence (0.34%). In total, the 43 SSR loci of C. micranthum cp genome can be subdivided into 24 mono-nucleotides, five di-nucleotides, three tri-nucleotides, eight tetra-nucleotides, two penta-nucleotides, and one hexa-nucleotide. Among the Lauraceae species, the number of SSR showed little variation and almost all were located at inter-genic spacer (IGS) regions as has been seen in other plants (Do et al. 2013; Ma et al. 2013; Zhang et al. 2014; Song et al. 2015, 2016). In the C. micranthum cp genome, 43 SSR primer sets were designed (Table S1), and these SSR loci can be used for further population genetic research in the Cinnamomum genus.

In addition, 41 repetitive structures were found in C. micranthum containing three complement matches, eight reverse matches, 13 forward matches, and 17 palindromic matches. Compared to the repetitive structures in C. kanehirae, 30 of 41 repeat events (80.5%) were located close to the same location. Other repeat events (eight in C. micranthum and seven in C. kanehirae) revealed different locations between the two genomes. Most of the repeat events occurred in IGS in both species. This is in accordance with findings in other plants (Do et al. 2013; Ma et al. 2013; Su et al. 2014; Zhang et al. 2014; Cho et al. 2015). In comparison with two other Lauraceae species, Machilus balansae and M. yunnanensis, the number of repeats (with lengths of >20 bp) varied from 39 to 41 (Song et al. 2015). About 61% of C. micranthum sequence repeats belonged to reverse or palindromic repeats. The palindromic repeats may form secondary structures (hairpin structure) and contribute duplicated sequence, and then induce InDel sequences (Johnson et al. 2016). Thus, four of the six validated InDels in this study were located around the palindromic sequence region. The IRs were the largest palindromic structure in the cp genome, and larger IRs can stabilize the structure of the cp genome and reduce the inversion probability of the sequence (Ma et al. 2013).

Phylogenetic tree analysis of C. micranthum and other Lauraceae species

In the present study, all available complete cp genomes were selected in Lauraceae and relative Magnoliaceae species to determine the phylogenetic relationship of C. micranthum. The phylogenetic tree (Figure 1) revealed two major clades, Magnoliales and Laurales. Within the Laurales, the Calycanthaceae formed the first branch, followed by the Lauraceae. In the clade of the Lauraceae, genus Cinnamomum was as a sister group to the other Lauraceae species. Persea americana is closest to Machilus and Phoebe. Phylogenetic tree is in accordance with angiosperm phylogenic group classification (Chanderbali et al. 2001; Bremer et al. 2009; Renner, 2011; Song et al. 2016). This phylogenetic tree is also consistent with APG III and provides the basic framework for studying Lauraceae phylogenetics (Bremer et al. 2009) and also supports the relationship of C. micranthum cp genome being close to C. kanehirae in Cinnamomum.

Figure 1:

Maximum-likelihood phylogeny of complete cp genomes in the Magnoliales, Laurales and Canellales. Bootstrap supports are shown above the branches. The C. micranthum genome is denoted by an asterisk.

InDel predictions and identification of C. micranthum and C. kanehirae

In total, 72 InDel events were detected between the C. micranthum and C. kanehirae cp genome sequences. Six InDel length of >10 bp had been validated for distinguishing between C. micranthum and C. kanehirae (Table 2). Three InDels were located at the SSC, one in the LSC and two at IRs. Most of the InDels were found within the IGS area. InDels 2, 3, 4, and 6 were also found at repetitive structure regions. InDels 2 and 4 were palindromic matching structures. InDels 3 and 6 were insertion sequences of C. kanehirae with a forward matching structure. InDels 3 and 6 were palindromic sequences with each other at the ycf1 gene, because they were located at IRs. In this study, we successfully extracted DNA from the sapwood and utilized as InDel biomarkers to authenticate wood and leaf of C. micranthum and C. kanehirae. Six InDels show identical amplicon sizes and clear polymorphism in all wood and leaf samples between two species. Six InDel markers can be successfully applied to identify wood and leaf samples. The results of agarose gel electrophoresis showed six wood collections of C. micranthum and C. kanehirae (Figure 2). Cinnamomum kanehirae wood samples that were air-dried for 4 years were used in the validation of our data: it revealed that the amount of wood DNA extraction was enough for PCR amplification. This might be because cp DNA was a high copy number in the cells based on sapwood. Many studies reported that the less potential inhibitors of DNA isolation and PCR amplification are present in the sapwood (Rachmayanti et al. 2006; Tnah et al. 2012; Yu et al. 2016). There are many more living parenchyma cells in the sapwood than in the heartwood (Jiao et al. 2014; Yu et al. 2016). Thus, the six InDel primers are useful for wood differentiation between C. micranthum and C. kanehirae and can be helpful in forensic investigations in the court. Compared to morphological characteristics, molecular markers are more universal, which are not affected by environmental and developmental factors (Zhao et al. 2015). The features of C. micranthur and C. kanehirae are so similar that the adulteration of these two species happened in the seedling and wood trade. These six useful validated cp InDel markers can be used to identify C. micranthum and C. kanehirae and detect the adulteration problem.

Figure 2:

Confirmation of insertion/deletion (InDel) markers between C. micranthum and C. kanehirae wood by PCR amplification. a–f: PCR amplifications with InDel nos. 1–6. The 1–6 indicate C. micranthum wood collections; the 7–12 indicate C. kanehirae wood collections. DNA maker ladder: 100-bp DNA ladder (Cat# GTL101, GeneTeks BioSicence, Inc., Taipei, Taiwan.). The detail DNA band sizes are labeled next to the figure.

Differences in chemical constituents of C. micranthum and C. kanehirae leaves

Chemotaxonomical studies are also very helpful to investigate the relationships at intergenus, interspecies, and intraspecies levels (Angioni et al. 2004; de Sena Filho et al. 2012; Rasheed et al. 2012). The results of the chemical profiles further suggested that C. micranthum is different from C. kanehirae. The C. micranthum leaves do not contain terpinen-4-ol, linalool, safrole, and sesamin, but sesamin and linalool are abundant in the C. kanehirae leaves. In the investigated C. kanehirae samples, the sesamin content varied from 1771 to 15082 ppm with a mean of 12870 ppm; the linalool content varied from 1286 to 41026 ppm with a mean of 6696 ppm. The safrole and terpinen-4-ol contents were scant, and some contents were detected in specimens of C. kanehirae (Table 3). Previous studies mentioned that the major component of C. micranthum leaf essential oil is decylaldehyde and a trace component is linalool, while the major components in the C. kanehirae essential oil include linalool and terpinen-4-ol (Fujita, 1952, 1960, 1967). C. kanehirae methanol extraction also contained sesamin in the previous report (Hsieh et al. 2005). Our study provides the comparative chemical profiles of methanol extracts between these two species. Terpinen-4-ol was not found in the extract, perhaps because methanol was used for extraction. Compared with a previous study (Cheng et al. 2015), the plant samples investigated here were collected from throughout the islands of Taiwan. Thus, it is reasonable that variable linalool concentrations occurred in the UPLC analysis and it might be further classified into five different chemotypes (Cheng et al. 2015). The chemical profiles showed obvious discrimination between C. micranthum and C. kanehirae. Thus, it is confirmed that the natural chemical characters are different between these two species. Our study provides the first report comparing the chemical constituents of C. micranthum and C. kanehirae leaves based on methanol extraction. These chemotaxonomical data are useful and can be an important supplement to those of genetic identification.