Structure and features of the complete chloroplast genome of Melastoma dodecandrum

Abstract

Melastoma dodecandrum, the only creeping species in the Melastoma genus, serves as a medicinal herb in southeast China. It belongs to the huge family Melastomataceae, which contains over 5000 species worldwide. In this study, we used next-generation sequencing to determine the complete chloroplast genome sequences of M. dodecandrum, which is a circular molecule of 156,611 bp in length. After annotation, we identified 131 putative genes in total, comprised of 85 protein-coding genes, 38 transfer RNA genes and 8 ribosomal RNA genes. Genome structure, GC content, repeat sequences and codon usage were investigated to gain a comprehensive understanding of this genome. Furthermore, we conducted comparative genome analyses between the M. dodecandrum genome and that of four other Melastomataceae species. Additionally, a phylogenetic analysis was performed based on available chloroplast genomes of Melastomataceae species and several Myrtaceae species, revealing the taxonomic relationships between M. dodecandrum and related species. In conclusion, our study represents the first look into the complete chloroplast genome of M. dodecandrum, providing abundant information for further studies such as species identification, taxonomy and phylogenetic resolution of Melastomataceae species.

Electronic supplementary material

The online version of this article (10.1007/s12298-019-00651-x) contains supplementary material, which is available to authorized users.

Introduction

Melastoma dodecandrum Lour. is a creeping shrub widely distributed in southeast China, which prefers to grow in acidic soils (Sciences 1984). It is used in the clinic as an alternative medicine to cure several diseases in south China, due to its multiple pharmaceutical properties, including hemostatic, anti-inflammatory (Ishii et al. 1999; Yang et al. 2014) and antibacterial (Liu et al. 2014). Compared to other species that are erect shrubs, M. dodecandrum is the only creeping species within the Melastoma genus. In addition, it was reported that M. dodecandrum could hybridize with three other congeneric species, indicating that strong reproductive isolation exists between M. dodecandrum and these species (Zou et al. 2017). Previously, microsatellite markers were designed and evaluated to study the genetic diversity and population structure of M. dodecandrum and its congeneric species (Liu et al. 2013; Renner and Meyer 2001; Zou et al. 2017). However, phylogenetic analysis based on one or several plastid markers has proven to be unsatisfactory, especially for the analysis of lineages within the Melastomataceae, which contains over 5000 species worldwide (Reginato et al. 2016). In contrast, phylogeny based on the complete chloroplast genome represents a significant improvement over phylogenies based on a single or a few plastid loci (Reginato et al. 2016).

The chloroplast (cp) is the characteristic organelle of a plant cell, which is believed to have descended from an ancestral cyanobacterium via endosymbiosis (Dyall et al. 2004). The chloroplast genome contains genetic information about the enzymic machinery necessary for photosynthesis (Green 2011) as well as the gene expression system of the chloroplast (Allen 2015). Moreover, cp genomes are generally small in size and conserved in structure and gene content (Wicke et al. 2011). For this reason, cp genomes have been increasingly used for multiple molecular studies on organelle biology (Jarvis and Lopez-Juez 2013; Smith and Keeling 2015), synthetic biology (Scharff and Bock 2014), development of molecular markers (Wu et al. 2010), phylogenetic analysis (Nikiforova et al. 2013) and so on. Due to increasing sequencing depth and decreasing costs, next-generation sequencing (NGS), has proven to be a powerful tool for the acquisition of new and complete cp genomes (Daniell et al. 2016; Moore et al. 2006). To date, there are more than 2000 cp genomes available in the National Center for Biotechnology Information (NCBI).

The objective of this study was to gain a comprehensive understanding of the cp genome of M. dodecandrum by describing its genome structure and features, and by comparing it to that of several congeneric species.

Results and discussion

Genome assembly

Using the Illumina HiSeq 2000 platform, a total of 25,225,380 paired-end raw reads were generated from a sequencing library of genomic DNA of M. dodecandrum, accounting for 7.57 GB data in total. After data filtering by trimming adaptors and removing low-quality reads, 24,967,763 clean reads were retained for the following analysis. With the cp-DNA sequence of a closely related species, M. candidum, serving as a reference, cp-like reads were extracted from the clean reads and then assembled into the complete cp genome sequence of M. dodecandrum (designated as Md_cp genome). The Md_cp genome was annotated and deposited into GenBank with the Accession Number MH748092.

Analysis of genome features

The Md_cp genome is a circular molecule of 156,611 bp with the classical quadripartite organization commonly observed for plant chloroplasts, which contains one large single copy (LSC), one small single copy (SSC), as well as two inverted repeat (IR) regions. A circular map of the Md_cp genome is presented in Fig. 1. The GC content of the Md_cp genome is 37.1 mol%, which is consistent with other members of Melastomataceae (Ng et al. 2018; Reginato et al. 2016). A further analysis of the four regions of the Md-cp genome (see Table 1) revealed that the GC content of the IR regions is greater than that of the two single copy regions. In addition, we observed a bias towards a high AT representation at the third codon position, similar to what has been found in other land plant cp genomes (Shen et al. 2017), which is believed to be beneficial for distinguishing cp-DNA from nuclear and mitochondrial DNA (Clegg et al. 1994).

Fig. 1.

Map of the Melastoma dodecandrum chloroplast genome. Genes in different functional groups are color coded as shown in the legend (bottom left). Genes shown outside the outer circle are transcribed counterclockwise, while those inside are transcribed clockwise. The shaded area inside the inner circle indicates the GC content, with dark shading indicating a higher GC mol percent

Table 1.

Base composition in the Melastoma dodecandrum chloroplast genome

Region	T/U (%)	C (%)	A (%)	G (%)	Total (bp)
LSC	33.2	17.9	31.8	17.2	86,015
IRA	28.5	20.5	29.0	22.1	26,750
SSC	34.0	16.2	34.8	15.0	17,096
IRB	28.9	22.1	28.5	20.5	26,750
Total	31.7	18.8	31.1	18.3	156,611
CDS	31.5	17.6	30.8	20.0	80,310
1st position	24.1	18.9	30.8	26.2	26,770
2nd position	32.8	19.8	30.0	17.5	26,770
3rd position	37.7	14.3	31.6	16.4	26,770

After annotation, we identified a total of 131 predicted genes (including duplicates and ycf genes) in the Md_cp genome, comprised of 114 unique genes and 17 duplicates in the IR regions. The unique genes included 79 protein-coding genes (CDS), 31 transfer RNAs (tRNA) and 4 ribosomal RNAs (rRNA), while 6 CDS, 7 tRNA and 4 rRNA gene duplicates were found in the IR regions (see Table 2). In general, the Md_cp genome shares similarity to that of its sister species M. candidum in terms of gene content and gene order (Ng et al. 2018). Among the predicted genes, nine CDS (atpF, ndhA, ndhB, petB, petD, rpl16, rpl2, rpoC1 and rps16) and five tRNAs contain one intron, while three CDS (clpP, rps12 and ycf3) contain two introns (see Table 3). Based on the annotation of the reference M. candidum cp genome, homologous sequences of ycf15 were found in the Md_cp genome. Nonetheless, those sequences share poor homology to other reported sequences (Shi et al. 2013). Thus, we eliminated those putative ycf15 annotations in the Md_cp genome.

Table 2.

Gene contents in the chloroplast genome of Melastoma dodecandrum

Classification of genes	Gene names	Number of genes
Photosystem I	psaA, psaB, psaC, psaI, psaJ	5
Photosystem II	psbA, psbB, psbC, psbD, psbE, psbF, psbH, psbI, psbJ, psbK, psbL, psbM, psbN, psbT, psbZ	15
Cytochrome b/f complex	petA, petB, petD, petG, petL, petN	6
ATP synthase	atpA, atpB, atpE, atpF, atpH, atpI	6
NADH dehydrogenase	ndhA, ndhB (× 2), ndhC, ndhD, ndhE, ndhF, ndhG, ndhH, ndhI, ndhJ, ndhK	12
RubisCO large subunit	rbcL	1
RNA polymerase	rpoA, rpoB, rpoC1, rpoC2	4
Ribosomal proteins (in SSC region)	rps2, rps3, rps4, rps7 (× 2), rps8, rps11, rps12 (× 2), rps14, rps15, rps16, rps18, rps19	14
Ribosomal proteins (in LSC region)	rpl2 (× 2), rpl14, rpl16, rpl20, rpl22, rpl23 (× 2), rpl32, rpl33, rpl36	11
Ribosomal RNAs	rrn 4.5 (× 2), rrn 5 (× 2), rrn 16 (× 2), rrn 23 (× 2)	8
Protein of unknown function	ycf1 (× 2), ycf2 (× 2), ycf3, ycf4	6
Transfer RNAs	tRNAs (7 contain an intron, 7 in the inverted repeats region)	38
Other genes	accD, ccsA, cemA, clpP, matK	5
Total		131

Table 3.

Genes with introns in the chloroplast genome of Melastoma dodecandrum

Gene	Location	Exon I (bp)	Intron I (bp)	Exon II (bp)	Intron II (bp)	Exon III (bp)
atpF	LSC	144	826	411
clpP	LSC	71	951	292	632	228
ndhA	SSC	549	1015	543
ndhB	IR	780	680	756
petB	LSC	6	774	642
petD	LSC	8	794	475
rpl16	LSC	9	1039	399
rpl2	IR	391	662	434
rpoC1	LSC	430	726	1619
rps12	IR	114	–	232	540	26
rps16	LSC	39	817	216
trnA-UGC	IR	38	801	35
trnI-GAU	IR	42	842	35
trnK-UUU	LSC	37	2537	35
trnL-UAA	LSC	37	538	50
trnV-UAC	LSC	39	611	37
ycf3	LSC	126	740	226	823	155

Analysis of codon usage and RNA editing sites

Based on the sequences of CDS, codon usage frequency of the Md_cp genome was evaluated and summarized (see Supporting Information Table S1). In general, all 61 codons, as well as three stop codons, were found in the Md_cp genome, encoding 20 amino acids. In sum, the CDS of the Md_cp genome is encoded by 26,770 codons. Among them, 2885 and 311 codons encode leucine and cysteine, respectively, representing the most and the least frequently used amino acids. Relative synonymous codon usage (RSCU), whose value reflects the ratio between the usage frequency and expected frequency of a particular codon, was introduced to evaluate the codon usage of the Md_cp genome. As illustrated in Fig. 2, the RSCU value of the Md_cp genome exhibits a pattern similar to that seen in four other Melastomataceae species. We found a strong AT bias in codon usage, with an RSCU value > 1, including UUA (1.90) in leucine, the stop-codon UAA (1.80) and UAU (1.64) in tyrosine, which is a common phenomenon in land plant plastomes (Morton 1998).

Fig. 2.

Codon usage of 20 amino acid and stop codons based on all protein-coding genes in the chloroplast genome of Melastoma dodecandrum and four related species. There are five bars for every amino acid and the stop codons, representing the RSCU value of chloroplast genome of M. dodecandrum, M. candidum, Allomaieta villosa, Barthea barthei and Tigridiopalma magnifica from left to right. Different codons that encode the same amino acid are colored differently

Additionally, a total of 48 putative RNA editing sites were identified and summarized (see Supporting Information Table S2), based on 35 CDS sequences. The most frequent occurrence (15) was the amino acid conversion from leucine (L) to serine (S). Meanwhile, the conversion from methionine (M) to threonine (T) was observed twice, in ndhD and rpl2. It has been reported that the ACG codon, with which ndhD starts, may be restored to a canonical start codon AUG by RNA editing (Hoch et al. 1991). Meanwhile, we found that rps19 starts with GUG, consistent with what has been previously observed in date palm and 14 other higher plants (Yang et al. 2010). Furthermore, the non-canonical start codon GUG was also found in psbC of the tobacco plastid, suggesting that a strong interaction between the extended Shine-Dalgarno-like sequence (GAGGAGGU, nine nucleotides upstream of the GUG) and the 3′ end of 16S rRNA facilitates translation initiation from the GUG (Kuroda et al. 2007).

Analysis of sequence repeats

Sequence repeats, including simple sequence repeats (SSRs) and long repeats, of the Md_cp genome were also investigated. Simple sequence repeats, also known as DNA microsatellites, refer to short (1–6 bp) tandemly repeated DNA sequences, which are widely distributed throughout nuclear genome. SSRs have been utilized for multiple genetic analyses, including plant population genetics, as well as evolutionary and ecological studies (Huang et al. 2014; Zhao et al. 2015). We found 308 SSRs found in the Md_cp genome in total (see Table 4), among which mono-nucleotide SSRs account for 53.6%. It is worth noting that these SSRs are generally composed of tandem adenine (A) or thymine (T), which contributes to the AT richness of the cp genome and is consistent with other species within the Melastomataceae family.

Table 4.

Types and amounts of SSRs in the chloroplast genomes of Melastoma dodecandrum and M. candidum

SSR type	Repeats unit	Amount		Ratio
		M. dodecandrum	M. candidum	M. dodecandrum (%)	M. candidum (%)
Mono	A/T	163	165	98.8	98.8
	C/G	2	2	1.2	1.2
Di	AC/GT	1	1	1.6	1.6
	AG/CT	20	20	32.8	32.8
	AT/AT	40	40	65.6	65.6
Tri	AAC/GTT	8	8	11.3	10.8
	AAG/CTT	19	18	26.8	24.3
	AAT/ATT	29	33	40.8	44.6
	ACC/GGT	2	2	2.8	2.7
	ACT/AGT	3	3	4.2	4.1
	AGC/CTG	4	4	5.6	5.4
	AGG/CCT	1	1	1.4	1.4
	ATC/ATG	5	5	7.0	6.8
Tetra	AAAT/ATTT	7	5	77.8	71.4
	AATT/AATT	1	1	11.1	14.3
	AGAT/ATCT	1	1	11.1	14.3
Penta	AAAAT/ATTTT	1	0	50.0	0.0
	AATAT/ATATT	1	0	50.0	0.0

In addition to the SSRs, we also examined long repeats of the Md_cp genome, including forward, reverse, complement and palindromic repeats, with a minimum detected length set at 30 bp. The cp genomes of four other Melastomataceae species were added to the comparison in this analysis, and the results are shown in Fig. 3. In general, there are 49 long repeats in both M. dodecandrum and M. candium, while less than half of this number is found in Allomaieta billos and Tigridiopalma magnifica. Most of these long repeats are forward, palindromic, and 30–49 bp in length. These sequence repeats provide valuable information for further phylogenetic studies on Melastomataceae species.

Fig. 3.

Distribution of long repeat sequences in the chloroplast genome of Melastoma dodecandrum and four related species. F forward repeats, P palindrome repeats, R reverse repeats and C complement repeats. Repeats with different lengths are colored differently as indicated (color figure online)

Expansion and contraction of the inverted repeats regions

The IRs are the most conserved regions in cp genomes. Therefore, expansion and contraction of their borders, which are common evolutionary events, are often investigated to assess the structural differences among cp genomes (Kim and Lee 2004; Wang et al. 2008). Genes that span or are in close proximity to the neighboring regions of the cp genomes of M. dodecandrum and four other Melastomataceae species were compared and are illustrated in Fig. 4. The boundaries of LSC-IRB are located in the rps19 genes, while those of IRB-SSC spanned the ndhF genes. As a result, two gene fragments, namely ψndhF and ψrps19, were also found to be located in the IRA region. Moreover, the 5′ truncated rps19 gene was located downstream of rpl2 in the IRA region and was followed by an intact trnH in the LSC, indicating that the IRA-LSC junction in Melastomataceae is in agreement with those of other eudicots (Wang et al. 2008). At the same time, we observed that ycf1, which spans across the SSC-IRA junction, produces a corresponding pseudogene, 5′ truncated ψycf1, in the IRB region. Judging from Fig. 4, a very similar pattern of expansion and contraction of the IR regions was observed among the Melastomataceae species, indicating that structural conservation is common within the cp genomes of Melastomataceae.

Fig. 4.

Distribution of genes located in/near the borders of two IR regions in the chloroplast genome of Melastoma dodecandrum and four related species. The number around the gene features indicates the distance from the borders site to the other pole. Ψ stands for pseudogene

Genome comparison analysis

Comparative analysis of different genomes allows the examination of how DNA sequences diverge between and/or among related species. We performed comparative analysis of the cp genome of M. dodecandrum with four reported Melastomataceae species, including M. candidum (NC_034716), Allomaieta villosa (NC_031875), Barthea barthei (NC_035661) and Tigridiopalma magnifica (NC_036021), using the mVISTA program (see Fig. 5). Among them, the cp genome of M. dodecandrum represents the second largest, which is 71 bp shorter than that of M. candidum with a length of 156,682 bp. As shown in Fig. 5, there are slight sequence differences between the cp genomes of the two Melastoma species. Furthermore, we observed more sequence divergence in the LSC and SSC regions of the other three cp genomes, which mainly occur in the intergenic spacers. Moreover, divergence also occurs in several protein-coding genes, such as clpP, matK, rpoC2, ycf1 and ycf3. The sequence differences identified in the comparison provide potential spots for biomarker discovery and species authentication. In addition, both IR regions of the compared cp genomes, which contain four rRNAs and ycf2, appear to be very conserved in sequence.

Fig. 5.

Genome comparison of Melastoma dodecandrum and four Melastomataceae species using the mVISTA program. Gray arrows indicate the position and orientation of genes; purple bars represent exons; pink bars represent non-coding sequences. Y axis scales represent percentage identity ranging from 50 to 100% (color figure online)

Phylogenetic analysis

As mentioned above, cp genomes are useful for phylogenetic analyses in plants. In order to investigate the taxonomic relationship of M. dodecandrum and other related species, we performed a phylogenetic study based on the cp genome sequences of 20 Melastomataceae species, along with three Myrtaceae species as an outgroup, using both maximum-likelihood (ML) and neighbor-joining (NJ) trees. As shown in Fig. 6, both trees exhibit the same topological pattern. In detail, the two Melastoma species grouped together, and further clustered with the other genus of Melastomataceae. The Melastomataceae branch could then be clearly distinguished from the three outgroup species of Myrtaceae. Bootstrap support of most nodes in this phylogenetic tree is high. This phylogeny is consistent with that of a previous report (Reginato et al. 2016), which was reconstructed based on 78 protein-coding genes, and thus extends to a broader taxonomic range with two Melastoma species and one Tigridiopalma species. Furthermore, the consistency of these two phylogeny results demonstrates that whole cp genome sequences (including one IR region) are useful for phylogenic analyses. It may be more convenient to use whole cp genome sequences rather than just coding region sequences, which require additional data extraction. Melastomaceae has been historically treated as a major core family of the Myrtales in taxonomic order. However, the limited number of available whole cp genome sequences has hampered a comprehensive assessment of the evolutionary position of Myrtales plants.

Fig. 6.

Phylogenetic trees of Melastoma dodecandrum and 22 related species based on their chloroplast genome sequences. Maximum-likelihood tree (a); neighbor-joining tree (b). Bootstrap support value are given adjacent to the nodes, representing the percentages of resampling. Filled circle indicates the target species of this study; open circle indicates outgroup species. The bar (bottom left) indicates 0.01 changes per nucleotide position

Materials and methods

Plant material, DNA extraction and sequencing

Genomic DNA (gDNA) was extracted from fresh tender leaves of a M. dodecandrum plant grown in the Medicinal Botanical Garden of Guangzhou University of Chinese Medicine, using a DNeasy Plant Mini Kit (Qiagen, Germany). The purity and integrity of the extracted gDNA were measured by ultraviolet spectrophotometry and gel electrophoresis, respectively. DNA samples of good integrity and with both OD260/280 and OD260/230 ratios greater than 1.8 were used in subsequent experiments. After being sheared into 250 bp fragments, a sequencing library was constructed with this gDNA and submitted for NGS on an Illumina HiSeq 2000 platform to generate paired-end reads.

Genome assembly and annotation

Approximately 7.57 GB of resulting sequencing data was filtered and trimmed to clean reads with an average length of 130 bp. Then, the complete sequence of the M. candidum chloroplast genome was used as a reference, to extract cp-like reads from the clean reads of M. dodecandrum. Next, those cp-like reads were assembled using the Abyss 2.0 program (Jackman et al. 2016) to form a complete chloroplast genome sequence of M. dodecandrum. Self-alignment was performed using BLASTn to locate the precise position of the quadripartite structure. To verify the assembly, four junction regions between the IR regions and the LSC/SSC region were confirmed by PCR amplification. Annotation of the cp genome was performed using the GeSeq online tool (Tillich et al. 2017) (https://chlorobox.mpimp-golm.mpg.de/geseq.html) with default parameters, which annotates not only protein-coding genes, but also tRNAs and rRNAs. Further examination and revision of the annotation information were manually performed under the assistance of the CLC Sequence Viewer (version 8). A circular map of the M. dodecandrum cp genome was drawn using the OGDRAW program (Lohse et al. 2013) (https://chlorobox.mpimp-golm.mpg.de/OGDraw.html).

Genome features analysis and genome comparison

After annotation, the Md_cp genome was analyzed for structural characteristics, including GC content and codon usage, using MEGA (version 7) (Kumar et al. 2016). Analysis of SSRs and long sequence repeats analysis was carried out using MISA (Hennequin et al. 2001) and REPuter (Kurtz et al. 2001) programs, respectively. Another online tool, the Predictive RNA Editor for Plants (PREP) suite (Mower 2009), was used to identify potential RNA editing sites, with a cutoff value set at 0.8. The mVISTA program (Frazer et al. 2004) was used for whole-genome alignment of the cp genomes of M. dodecandrum and four other Melastomataceae species.

Phylogenetic analysis

Phylogenetic analysis was performed using the whole cp genome sequences of M. dodecandrum, along with other 19 complete cp genome sequences within the same family and three of Myrtaceae as an outgroup, obtained from the NCBI Organelle Genome Resources database (see Supporting Information Table S3). Those cp genomes were trimmed manually to remove the IRA region, and then aligned with the MAFFT (version 7) (Katoh et al. 2002) program using the FFT-NS-i × 1000 strategy. The alignment information was then used to reconstruct both maximum-likelihood and neighbor-joining trees using MEGA, employing the GTR+G model and the Kimura 2-parameter model, respectively. Support was estimated through 1000 bootstrap replicates to assess the reliability of the phylogenetic tree.

Conclusions

The chloroplast genome serves as a useful tool for plant phylogenetic studies due to high conservation of gene content and order, and the virtual lack of recombination (Ravi et al. 2008). Here, we provide the complete chloroplast genome sequence of M. dodecandrum, which is a circular molecule of 156,611 bp with a classic quadripartite structure. Compared to other species within the family, the cp genome of M. dodecandrum exhibits high structural conservation, in terms of gene content, gene order and other structural features, which is consistent with the results of other reported Melastomataceae species (Reginato et al. 2016), and even matches that of the Myrtales (Gu et al. 2016). Importantly, the full sequence and annotation of the M. dodecandrum chloroplast genome will accelerate phylogenetic, population and evolutionary studies of this species, and further contribute to a better understanding of plastid biology among Melastomataceae.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Abbreviations

cp

Chloroplast

gDNA

Genomic DNA

IR

Inverted repeat

LSC

Large single copy

ML

Maximum likelihood

NCBI

National Center for Biotechnology Information

NGS

Next generation sequencing

NJ

Neighbor-joining

SSC

Small single copy

SSR

Simple sequence repeats