Assembly and comparative analysis of the first complete mitochondrial genome of Acer truncatum Bunge: a woody oil-tree species producing nervonic acid

Background Acer truncatum (purpleblow maple) is a woody tree species that produces seeds with high levels of valuable fatty acids (especially nervonic acid). The species is admired as a landscape plant with high developmental prospects and scientific research value. The A. truncatum chloroplast genome has recently been reported; however, the mitochondrial genome (mitogenome) is still unexplored. Results We characterized the A. truncatum mitogenome, which was assembled using reads from PacBio and Illumina sequencing platforms, performed a comparative analysis against different species of Acer. The circular mitogenome of A. truncatum has a length of 791,052 bp, with a base composition of 27.11% A, 27.21% T, 22.79% G, and 22.89% C. The A. truncatum mitogenome contains 62 genes, including 35 protein-coding genes, 23 tRNA genes and 4 rRNA genes. We also examined codon usage, sequence repeats, RNA editing and selective pressure in the A. truncatum mitogenome. To determine the evolutionary and taxonomic status of A. truncatum, we conducted a phylogenetic analysis based on the mitogenomes of A. truncatum and 25 other taxa. In addition, the gene migration from chloroplast and nuclear genomes to the mitogenome were analyzed. Finally, we developed a novel NAD1 intron indel marker for distinguishing several Acer species. Conclusions In this study, we assembled and annotated the mitogenome of A. truncatum, a woody oil-tree species producing nervonic acid. The results of our analyses provide comprehensive information on the A. truncatum mitogenome, which would facilitate evolutionary research and molecular barcoding in Acer. Supplementary Information The online version contains supplementary material available at 10.1186/s12870-021-03416-5.

and damaged tissues. Previous studies have indicated that nervonic acid is potentially useful for treatment of schizophrenia, psychosis, and attention deficit disorder [5,6]. It has been detected in several plant species [2,7,8], but issues related to their nervonic acid content and growth adaptability have limited the utility of these species. The characteristics of rapid growth, wide geographic distribution, and high adaptability, thus A. truncatum is a novel potential plant source of nervonic acid for treating human cerebral and neurological problems.
The main function of mitochondria, the "energy factories" of cells, is the conversion of biomass energy into chemical energy in living cells [9,10]. In most seed plants, nuclear hereditary information is inherited biparentally, whereas DNA of both mitochondria and chloroplasts is maternally derived [9,10]. In addition, recent researches have revealed that intergenomic gene transfer between nuclear and organellar genomes, which was a common phenomenon during plant evolution [11][12][13]. Along with rapid developments in sequencing and genome assembly technologies, an increasing amount of information on mitogenomes has been uncovered. At present, 6026 complete land plant organelle genomes, including 5735 chloroplast and 291 plant mitogenomes have been assembled and deposited in GenBank Organelle Genome Resources (https:// www. ncbi. nlm. nih. gov/ genome/ browse/), as the mitochondrial genome is more complex and harder to assemble than that of other organelles [9,14].
Plant mitogenomes are species specific [15,16] and vary considerably in length, gene order, and gene content [9,10,14,17]. Genome size is extremely variable, ranging from 66 kb (Viscum scurruloideum) [18] to 11.3 Mb (Silene conica) [19], and most genomes are 200-800 kb in size [20]. This wide variation in mitogenome size can be attributed to the repetitive sequences and the foreign DNA from other organisms during evolution [21,22]. Repetitive sequences, including simple sequence repeats (SSRs), tandem repeats and dispersed repeats, are abundant in the mitogenomes of seed plants. SSRs are frequently used as molecular markers for identifying species in plant mitogenomes [14,23]. In addition, insertions/ deletions (indels) and single nucleotide polymorphisms (SNPs) within mitogenomes also have been applied to rapidly distinguish species and for phylogenetic analyses [24,25].
The mitochondrial gene content of land plants varies considerably, ranging from 32 to 67 genes. Some genes, including those related to NADH dehydrogenase, ATP synthase, ubiquinol cytochrome, and cytochrome c biogenesis [14], are highly conserved, whereas others, such as sdh3, sdh4, rps11, and cox2 have been lost [26,27].
Mitogenomes in the genus Acer, except for the mitogenome sequence of A. yangbiense released in 2019, have not been analyzed in detail [28]. In this study, we first assembled the complete mitogenome of A. truncatum and analyzed its gene content, repetitive sequences, RNA editing sites, selective pressure, and phylogenetic relationships. We also surveyed gene transfer among nuclear, chloroplast, and mitochondrial genomes of A. truncatum. Moreover, we developed a marker based on an indel in the NAD1 intron to distinguish seven Acer species (A. buergerianum, A. truncatum, A. henryi, A. negundo, A. ginnala, A. yangbiense and A. tonkinense). The data presented herein expand genetic information available for the genus Acer and provide an opportunity to conduct further important genomic breeding studies on A. truncatum.

Features of the A. truncatum mitogenome
The A. truncatum genome sequence generated was submitted to the GenBank database (accession number MZ318049) in this study. The complete mitogenome of A. truncatum is 791,052 bp in length and has the typical circular structure of land plant genomes (Fig. 1). The nucleotide composition of the complete mitogenome is 27.11% A, 27.21% T, 22.79% G, and 22.89% C, with a GC content of 45.68% (Table 1). Protein-coding genes (PCGs) and cis introns account for 4.31 and 2.94% of the whole mitogenome, while tRNA and rRNA genes comprise only 0.22 and 0.67%, respectively. A total of 62 unique genes, including 35 protein-coding, 23 tRNA, and 4 rRNA genes, were identified in the A. truncatum mitogenome (Table 2). Interestingly, two copies of cox1 genes were found. Additionally, five tRNA and one rRNA gene(s) located in repeat sequences were found to be present in two or four copies (trnN-GTT , trnM-CAT , trnP-TGG , trnH-GTG , trnW-CCA , and rrn5) (Fig. 1).

Codon usage analysis of PCGs
The total length of PCGs in A. truncatum was 34,059 bp. Most PCGs had the typical ATG start codon, whereas atp6, nad1, and nad4L had ACG as the start codonpresumably a consequence of C-to-U RNA editing of the second site (Table 2). Three types of stop codons were identified, namely, TAA, TGA, and TAG, the C to U RNA editing phenomenon was not found in the stop codons. As shown in Fig. 2, the codon usage analysis revealed the most frequent amino acids to be leucine (Leu) (11.2-11.3%), serine (Ser) (10.6-11.0%), and arginase (Arg) (8.1-8.4%), whereas cysteine (Cys) and tryptophan (Trp) were rarely found.
We also analyzed the relative synonymous codon usage (RSCU) of 35 PCGs in the A. truncatum mitogenome. As shown in Fig. 3, the 35 PCGs comprised 33,948 bp encoding 11,316 codons excluding termination codons. We found that nearly all of the RSCU values of NNT and NNA codons were higher than 1.0 with the exception of Ile (AUA, 0.82), Leu (CUA, 0.93), and Ser (UCA, 0.97). Codon usage was generally strongly biased toward A or T(U) at the third codon position in the A. truncatum mitogenome, which is very common in mitogenomes of land plant species.

Analysis of synonymous and nonsynonymous substitution rates
In genetics, the nonsynonymous-to-synonymous substitution ratio (Ka/Ks) is used to understand the evolutionary dynamics of genes. In this study, the Ka/Ks ratio was determined for 26 protein-coding genes common to A. truncatum, A. yangbiense, A. thaliana and C. sinensis mitogenomes (Fig. 4). The PCGs shared between A. truncatum and A. yangbiense were close homologs, as the Ka/ Ks ratio of 21 PCGs was 0. In addition, nearly all Ka/Ks ratios were less than 1.0, which suggested that most of the PCGs were subject to stabilizing selection during evolution. Conversely, the Ka/Ks ratios of nine genes (atp6, cob, cox1, nad2, ccmFn, nad4, nad6, nad7 and rpl5) were greater than 1.0, which indicateed these genes had been under positive selection during evolution. Finally, three genes (atp4, ccmB and rps4) had Ka/Ks ratios close to 1, thus suggested that they had experienced neutral evolution since the divergence of their common ancestor.

Prediction of RNA editing sites in PCGs
In plants, RNA editing is necessary for gene expression, with cytidine (C)-to-uridine (U) RNA editing enriched in mitochondrial and chloroplast genomes. In this study, we predicted the RNA editing sites of 26 PCGs common to mitogenomes of four angiosperm species. The number of RNA editing sites predicted for A. truncatum, A. yangbiense, A. thaliana, and C. sinensis-421, 427, 342 and 288, respectively-suggests that these sites are extremely conserved in PCGs in Acer. A total of 421 RNA editing sites were predicted in A. truncatum, all exhibiting C-to-U RNA editing. Among the 421 sites, 32.07 and 67.93% were predicted at the first and the second positions of codons, respectively, whereas none were found at the third position (Fig. 5).
RNA editing can change PCG initiation and termination codons. As shown in Table 2, atp6, nad1 and nad4L genes use ACG as their initiation codons, we infer that they may have been altered by RNA editing. The number of RNA editing sites in different genes was found to vary greatly, with the largest predicted numbers detected in cytochrome c biogenesis (ccmB, ccmC, ccmFn, and ccmFc), Complex I (NADH dehydrogenase) and nad4 genes. In contrast, no RNA editing sites were found in atp9 and nad3 genes in A. truncatum and A. yangbiense.

Phylogenetic analysis
To determine the phylogenetic position of A. truncatum, we downloaded 25 plant mitogenomes from GenBank (https:// www. ncbi. nlm. nih. gov/ genome/ browse/) ( Table   Table 1 ) and constructed a phylogenetic tree based on a set of 25 conserved single-copy orthologous genes present in all 26 analyzed mitogenomes. As shown in Fig. 7, 21 of 23 nodes in the generated tree had bootstrap support values over 70%, including 12 nodes with 100% support. The phylogenetic tree strongly supports (100% bootstrap support) the close phylogenetic relationship between A. truncatum and A. yangbiense. In addition, it also revealed that both species were closely related to C. sinensis, which is similar to conclusions inferred using the nuclear genome [2]. Overall, the results of our analysis of mitogenomes provide a valuable foundation for future analyses of the phylogenetic affinities of Acer species.

Plastid-derived and nuclear-shared sequence transfer events
DNA fragment transfers among nuclear and organellar genomes are common events during plant evolution. Six directions of gene transfer are possible among the three types of genomes. To further understand the characteristics of sequence transfer events in A. truncatum, the A. truncatum nuclear and chloroplast genomes [2,29] were searched by using its mitogenome sequences as queries.
We obtained 393 hits covering 230.0 kb of sequences of nuclear genome transferred into the mitogenome. According to the nuclear-mitochondrial alignment, hits occurred on every A. truncatum chromosome (Fig. 8A), however, the total lengths of the hits and the percent coverage on the chromosomes were different. Chromosome 1 had the maximum total length of hits (25.30 kb), which was much larger than on other chromosomes, whereas the highest percent coverage (0.05%) occurred on chromosomes 5, 6, and 13. In addition, fragment lengths were mainly between 200 bp and 400 bp (Fig. 8B). A total of 62,241 bp of sequences (7.87% of the A. truncatum mitogenome) were found to be shared between nuclear and mitochondrial genomes. The shared sequences contained seven complete genes (trnN-GTT , rpl5, trnS-GCT , trnF-GAA , trnQ-TTG , atp1, and trnH-GTG ) as well as partial gene sequences of matR, ccmFN, cox2, rps3, rps4, atp8, sdh4, nad4 and atp6. The A. truncatum mitogenome sequence (791,052 bp) was approximately five times longer than the chloroplast genome (156,492 bp). Forty-one fragments with a total length of 18,637 bp, corresponding to 2.36% of the mitogenome, were observed to have migrated from the chloroplast genome to the mitogenome in A. truncatum (Table 4). Six intact chloroplast genes (psbJ, trnP-UGG , trnW-CCA , trnN-GUU , trnD-GUC , and trnH-GUG ) were located on these fragments. The remaining fragments were partial sequences of transferred genes or intergenic spacer regions in the chloroplast genome. Interestingly, we found

Development of an NAD1 intron indel marker
Among Acer species, only the mitogenome of A. yangbiense has currently been reported. To further characterize the NAD1 intron, we compared its sequence between A. truncatum and A. yangbiense, and detected a 33-bp indel. The following seven Acer species were selected for characterization of the NAD1 intron sequence: A. truncatum, A. buergerianum, A. ginnala, A. yangbiense, A. palmatum, A. pubipalmatum, and A. tonkinense. To develop indel markers, primers were designed to anneal to conserved regions   (Table S5). The predicted amplification products were successfully obtained using these NAD1-intron-F/R primers in all seven tested samples (Fig. 9A). In all six species, the length of the amplified NAD1 intron sequence was identical (808 bp) and highly conserved. The corresponding sequence in A. yangbiense was indeed longer (841 bp) because of the 33-bp putative insertion ( Fig. 9A and B). Several species close to A. truncatum in the phylogenetic tree (A. yangbiense, Populus tremula, Salix suchowensis and C. sinensis) were selected to verify  whether the 33-bp sequence was an insertion or a deletion. According to the sequence alignment, the sequence was indeed an insertion (Fig. S3). In previous studies, indel markers have frequently been used to distinguish closely related species; however, Acer species have not been identified on the basis of their mitogenomes using this approach. Our firstever characterization of the NAD1 intron in Acer may therefore be applicable for classification and identification of Acer species.

Characterization of the A. truncatum mitochondrial genome
Mitochondria, which produce the energy required to carry out life processes, are the powerhouses of plants.
Because of factors such as size variation and repeated sequences, plant mitogenomes are more complex than those of animals [9,14,30,31]. The emergence of rapid, cost-effective genome sequencing technologies has accelerated understanding of mitogenomes. Our study has produced the first detailed characterization of a complete mitogenome in Acer. The size of the A. truncatum mitogenome is similar to that of A. yangbiense [28], both of which are moderate in size relative to most genomes [32]. GC content is an important factor for assessing species. The GC content of the A. truncatum mitogenome is 45.68%, which is comparable to that of other sequenced plant mitogenomes (A. thaliana, 44.8% [33]; Phaseolus vulgaris, 45.11% [14]; Beta vulgaris, 43.9% [34], but higher than the A. truncatum chloroplast genome (37.90%) assembled by our research group [29]. Similar to most other mitogenomes, most sequences in the A. truncatum mitogenome are non-coding. Protein-coding genes account for only 4.31%, which is probably the result of a gradual increase in sequence duplication during evolution. Most PCGs were the typical ATG start codon, and the distribution of amino acid compositions was similar to other angiosperms [28,33]. while the atp6, nad1 and nad4L genes use ACG as initiation codons, this phenomenon also has been found in other studies, which were considered to be altered by RNA editing modification [9,14]. The cox1 is one of the most reported mitochondrial genes involved in horizontal gene transfer among sngiosperms. In our study, two copies of cox1 genes were found. The previous studies reported that the cox1 copies existed in different species and different populations of a species [35]. Ka/Ks ratios > 1 have also been reported for some other mitochondrial genes [9,14,26]. In our study, the high Ka/Ks ratios of genes observed were very important for further studies in the gene selection and evolution of Acer species, including atp6, cob, cox1, nad2, ccmFn etc.

Identification of repeat sequences and RNA editing sites
Repeats are important sources of information for developing markers for population and evolutionary analyses [23,36,37]. Including tandem, short and large repeats, they are widely present in mitogenomes [14,38,39]. Repeats in mitochondrial DNA are generally vital for intermolecular recombination, which can generate structural variations and extreme mitogenome sizes [20,40]. In this study, five pair of large repeats (> 1 kb) by rearrangements could produce two subgenomic circles in A. truncatum mitogenome, comprising of 457,840 bp and 333,212 bp, respectively. This phenomenon also reported in Soybean [41]. we also found major differences between the repeat sequences of A. truncatum and A. yangbiense mitogenomes. In particular, the proportion of long repeat sequences in the A. truncatum mitogenome (18.24%) was higher than that of A. yangbiense (17.20%), and the longest repeats were 28,452 bp and 27,124 bp, respectively. These repeats may have contributed to the increase in the mitogenome size of A. yangbiense. This finding also suggests that intermolecular recombination has frequently occurred in the mitogenome during Acer evolution [14,31]. RNA editing, a post-transcriptional process that occurs in chloroplast and mitochondrial genomes of higher plants, contributes to improved protein folding [9,14,26]. Previous researches had uncovered approximately 491 RNA editing sites within 34 genes in rice [42] and 486 RNA editing sites within 31 genes in P. vulgaris [14]. In the present study, we predicted RNA editing sites in 26 PCGs common to A. truncatum, A. yangbiense, A. thaliana and C. sinensis mitogenomes. We found that the number of RNA editing sites in PCGs was extremely conserved in Acer but differed in the other two species. Although the number of RNA editing sites varies greatly among genes, cytochrome c biogenesis and NADH dehydrogenase genes harbor the largest number, which is similar to P. vulgaris [14]. In addition, all identified RNA editing sites are located at first and second codon positions. Previous researchers have speculated that the lack of RNA editing sites at the third codon position is probably due to the limitations of the PREP-Mt predictive methodology used rather than an actual absence [14,43]. Further analysis using experimental methods is thus needed.

DNA fragment transfer events
Information pertaining to DNA transfer events between different genomes (mitochondrial, nuclear and chloroplast) has been uncovered by sequencing analysis [21,44,45]. Previous studies have determined that the most prominent transfer direction in angiosperms is from organellar genomes into the nuclear genome, followed in importance by transfer from nuclear and plastid genomes into the mitogenome [13,21,[46][47][48]. The total length of transferred DNA varies among plant species in higher plants, lengths range from 50 kb (A. thaliana) to 1.1 Mb (O. sativa subsp. japonica) [49]. According to our study, 230.0 kb of nuclear DNA has been transferred into the mitogenome of A. truncatum. Although the nuclearmitochondrial transfer process has occurred on every A. truncatum chromosome, the total lengths of transferred material and the percent coverage differs among chromosomes. In total, 62, 241 bp of sequences (7.87% of the A. truncatum mitogenome) is shared between nuclear and mitochondrial genomes. Most genes with transferred sequences shared between nuclear and mitochondrial genomes are tRNA genes, such as trnN-GTT, trnH-GTG , and trnH-GTG . Chang et al. [41] have reported similar results in soybean. In regards to chloroplast genome to mitogenome migration events, a total of 18,637 bp of transferred fragments were observed, accounting for 2.36% of the A. truncatum mitogenome. In comparison, the proportion in S. suchowensis and Suaeda glauca is 2.8 and 5.18%, respectively [31]. We identified 41 fragments that had been transferred from the chloroplast genome to the mitogenome, these fragments included six integrated genes, namely, five tRNA genes and psbJ. Transfer of tRNA genes from chloroplast to mitochondrial DNA is common in angiosperms [21,26,31]. Interestingly, we also observed that DNA migration often occurred in the inverted repeat region of the A. truncatum chloroplast genome.

Development of a mitochondrial NAD1 intron marker for Acer species
Because indel regions are relatively easy to detect, they are often used to develop markers for identifying species [50]. The genus Acer comprises more than 200 species grown in China [2,51]; however, the highly similar shapes of some species present a challenge for identification, and a molecular approach would be beneficial. NAD1 intron indel markers have been useful for identification of some plant species [52][53][54]. In Acer, only the mitogenome of A. yangbiense has been previously reported [28]. In the present study, we first identified a 33-bp sequence difference by aligning the NAD1 intron regions of A. truncatum and A. yangbiense. Amplification of the NAD1 intron with specific primers revealed that a 33-bp indel was present in A. yangbiense, whereas the amplified NAD1 intron sequence was of the same length and highly conserved in the other six species. We verified that this 33-bp indel was an insertion in Acer by analyzing several species close to A. truncatum in our phylogenetic tree (A. yangbiense, P. tremula, S. suchowensis, and C. sinensis). The development of mitogenome-based molecular markers has not been previously reported for Acer. Although only a few Acer species were used in this study, our findings should nonetheless contribute to species classification in Acer.

Conclusions
In this study, we assembled and annotated the mitogenome of A. truncatum and performed extensive analyses based on DNA and amino acid sequences of annotated genes. The A. truncatum mitogenome is circular, with a length of 791,052 bp. We annotated 62 genes, including 35 protein-coding, 23 tRNA and 4 rRNA genes. In addition, the codon usage, sequence repeats, RNA editing and selective pressure were also analyzed in the A. truncatum mitogenome. The evolutionary status of A. truncatum was verified by phylogenetic analysis based on the mitogenomes of this species and 25 other taxa. Gene conservation between chloroplast and mitochondrial genomes and between nuclear and mitochondrial genomes were also detected in A. truncatum by analyzing gene migration. Finally, a newly developed NAD1 intron indel marker was used to distinguish Acer species. Our study has yielded extensive information about the A. truncatum mitogenome. The data presented herein supplement the genetic knowledge available for the genus Acer, provide novel insights into A. truncatum evolution, and form an important theoretical basis for increasing A. truncatum seed yield.

Plant materials and DNA sequencing
A. truncatum plants were grown at our Aceraceae seed base of Jangsu Academy of Agricultural Sciences (Lishui District, Nanjing, China; 31°65 N, 119°02E) under natural conditions. Fresh leaves were frozen in liquid nitrogen and stored at 80 °C. DNA extraction and sequencing were performed using methods described in our previous de novo genome sequencing study [2].

Mitogenome assembly and annotation
For the A. truncatum mitogenome, PacBio RS II reads (59.42 GB) sequenced in our previous study [2] were de novo assembled using Canu v1.4 [55]. The obtained contigs were mapped to core mitochondrial genes by mini-map2 [56], then extended. The assembled contigs were polished (Pilon v1.18) with Illumina reads (75.0 GB) to correct read errors [57]. Finally, five large contigs were assembled into mitogenome, to verify the quality and accuracy of our assemblies, we further verified the junctions by Sanger sequencing (Table S6 and Fig. S4). The GE-Seq tool on the MPI-MP CHLOROBOX website (https:// chlor obox. mpimp-golm. mpg. de) was used for the mitogenome annotation, with the A. yangbiense mitogenome (CM017774.1) serving as a reference. Mitochondrial protein-coding genes were predicted using the MITOFY webserver [40]. All tRNA genes were confirmed by using tRNAscan-SE with default settings [9,14,58]. ORFfinder (https:// www. ncbi. nlm. nih. gov/ orffi nder/) was used to analyze open reading frames longer than 300 bp. RSCU values and the amino acid composition of PCGs were calculated in MEGA X [59]. A circular mitochondrial map was drawn using Organellar Genome DRAW [60].

Analysis of repeat structures and SSRs
Forward, reverse, palindromic and complementary repeats were identified with REPuter [61]. SSRs were analyzed with the MISA program [62]. The motif size of one-to six-nucleotide SSRs was set as 8, 4, 4, 3 and 3, respectively.

Selective pressure analysis
We calculated the nonsynonymous (Ka) and synonymous (Ks) substitution rates of each PCG between A. truncatum and A. yangbiense, A. thaliana, and C. sinensis. Orthologous gene pairs were separately aligned in MEGA 6.0. Ka, Ks, and Ka/Ks values were calculated using DnaSP [63].

Genome alignments
The A. truncatum mitogenome was searched against the chloroplast genome of A. truncatum (MH638284) using BLASTN 2.9.0+ according to the following screening criteria: matching rate ≥ 70%, E-value ≤1e − 6 , and length ≥ 40 [31]. To identify regions of potential nuclear origin in the mitogenome of A. truncatum, we also performed a BLASTN search (maximum E-value = 1e − 50 ) of the complete mitogenome against all contigs from the A. truncatum nuclear genome sequenced in our previous study. BLASTN results of sequences longer than 250 bp and a pairwise similarity > 80% were inspected for sequence features.

Prediction of RNA editing sites
RNA editing sites in the PCGs of A. truncatum and other three mitogenomes (A. yangbiense, A. thaliana and C. sinensis) were predicted using the the online PREP-Mt suite of servers (http:// prep. unl. edu/). To obtain a more accurate prediction, the cutoff value was set as 0.2 [43]..

Verification of the Nad1 insertion in Acer
Primers were designed with Primer 5. PCR amplifications were carried out in 15-μl volumes containing 20 ng genomic DNA, 0.4 μl dNTPs (2.5 mM each), 2.5 μl of 10× Ex Taq buffer (Mg 2+ ), 0.4 μl Ex Taq DNA polymerase (Takara,Tokyo, Japan), and 1.0 μl of each primer (10 mM). The amplification conditions were 94 °C for 5 min, followed by 30 cycles of 94 °C for 30 s, 56 °C for 30 s, and 72 °C for 30 s, with a final extension of 72 °C for 10 min. The PCR products were purified and linked to the pMD19-T easy plasmid (Takara) for sequencing to confirm the accuracy of PCR product sizes. Three samples per species were sequenced by the General Biology Company (Nanjing, Jiangsu, China).