Shifei Sng, Hongto Cheng, Desheng Mei, Li Fu, Hui Wng, Ji Liu,Wenxing Wng, Qmr U. Zmn, Kede Liu, Qiong Hu,

aOil Crops Research Institute of Chinese Academy of Agricultural Sciences/Key Laboratory for Biology and Genetic Improvement of Oil Crops,Ministry of Agriculture and Rural Affairs, Wuhan 430062, Hubei, China

bNational Key Laboratory of Crop Genetic Improvement, Huazhong Agricultural University, Wuhan 430070, Hubei, China

ABSTRACT Sinapis arvensis, belonging to the genus Sinapis of the family Brassicaceae, has good agronomic characters that make it a valuable genetic resource for crop improvement and is a cytoplasmic source of heterologous cytoplasmic male sterility (CMS). In addition, S.arvensis has played an important role in the evolution of the six major cultivated Brassica species involved in the triangle of U.Using next-generation sequencing,we assembled and revealed the gene composition of S. arvensis cytoplasmic genome. The chloroplast genome comprises 153,590 bp,with 112 individual genes,including 4 rRNA,29 tRNA,and 79 proteincoding genes.The mitochondrial genome comprises 240,024 bp,with 54 genes,including 18 tRNA,three rRNA and 33 protein-coding genes.Genome structure and evolutionary analysis indicated that the sequences of the S. arvensis organellar genomes were more similar to those of Brassica nigra and B. carinata than to those of other Brassicaceae species. Four mitochondrial open reading frames displaying chimeric structural features and encoding hypothetical proteins with transmembrane domains may account for the infertility of Nsa CMS previously derived from somatic cell hybridization between B. napus and S. arvensis.These results will not only contribute to utilize the germplasm resource of S. arvensis, and comprehend the evolution of organelle genomes within the Brassicaceae family, but also help to identify genes conditioning the alloplasmic male sterility of Nsa CMS in B.napus.

1. Introduction

Sinapis arvensis (2n = 18) belongs to the genus Sinapis of the family Brassicaceae. It is wide distributed in Africa, Asia, and Europe, as well as North America, South America, and Australia. Because S. arvensis develops a robust root system and grows rapidly,it competes with crops for water,light,and fertilizer and reduces the yields of cultivated Brassica species.Thus, in Canada, S. arvensis is often regarded as a noxious weed [1]. However, S. arvensis carries resistances to disease,pod shattering, insects, drought and other agronomic traits,rendering it a valuable germplasm resource for the breeding of B. napus varieties [2-5]. In addition, a novel cytoplasmic male sterility (Nsa CMS) line with complete anther abortion has been developed by somatic hybridization between S. arvensis and B. napus [6-8]. The sterility gene of Nsa CMS has two possible sources. One is a direct derivation from S. arvensis cytoplasm, without recombination in the mitochondrial gene of somatic cell hybrids generated from S.arvensis and B.napus.Another possibility is that new sterility genes are formed by recombination or rearrangement in the mitochondrial genomes of S.arvensis and B.napus.Ogura CMS lines of B.napus generated by somatic cell hybridization with R.sativus carried both cytoplasm and restorer genes from R. sativus [9,10].Previous results[11]showed that the restorer gene of Nsa CMS was derived from S.arvensis.Given that the sterility gene and the restorer gene are always co-evoluted in the process of plant propagation,the sterility gene of Nsa CMS is more likely to be directly derived from the mitochondrial genome of S.arvensis. However, genomic information for S. arvensis is lacking, hindering the genetic exploitation of its germplasm.Investigation of the cytoplasmic genome may shed light on the genetic variation in S. arvensis.

Studying the classification status of S.arvensis in Brassicaceae species may facilitate the introduction of beneficial traits into B. napus [12]. Wu et al. [13] suggested that S. arvensis and Chinese B. juncea have a close kinship, and proposed that S.arvensis of Chinese origin (i.e., Xinjiang wild rape) is the ancestor of B. nigra based on integrated findings on morphology,cytology,and biochemical classification.The triangle of U[14]clarifies the relationship of B.napus,B.nigra,B.oleracea,B.rapa,B.juncea,and B.carinata.Among them,B.napus,B.juncea,and B. carinata are composite species derived from natural hybridization of the three primary species B.nigra,B. oleracea,and B. rapa. The genomes of many species have been completely sequenced since the recent improvements in sequencing technology [15,16]. Organelle genomes of many Viridiplantae species have also been sequenced, including major crops such as corn, rapeseed, pearl millet, and wheat[17-23]. These cytoplasmic genome resources have also been exploited for the development of allo-cytoplasmic malesterile lines.

In the present study, the chloroplast and mitochondrial genomes of a S. arvensis accession, the original S. arvensis parent of the Nsa CMS line developed in B. napus, were sequenced and assembled. The structure and composition of the S. arvensis cytoplasmic genomes are described.

2. Materials and methods

2.1. Plant materials

The study employed S. arvensis Yeyou 18,a wild species from Xinjiang, China. Seeds were provided by the Institute of Economic Crops of Xinjiang Academy of Agricultural Sciences.Seeds were grown in 1% (m/v) agar and cultured at 28 °C in dark.

2.2. Organelle DNA extraction

Organellar DNA was extracted from seven-day-old seedlings following the protocol described by Chen et al. [24], with minor modifications. A sucrose gradient with mass fractions of 23%:33% was used for ultracentrifugation at 72,000 ×g for 1.5 h. The middle layer containing purified organelles was collected for DNA extraction. The purified organelles were gently resuspended in 3 mL lysis buffer containing 100 mmol L-?1Tris-HCl (pH 8.0), 50 mmol L?1EDTA (pH 8.0), 1.0 mol L?1NaCl, 2% (m/v) SDS, and 200 mg mL?1proteinase K (Sigma-Aldrich, St Louis, Missouri, USA), then incubated at 37 °C for 3 h, followed by addition of 1/3 volume of NaCl solution (5 mol L?1). The suspension was incubated for an additional 20 min on ice before centrifugation at 10,000 ×g for 15 min. The supernatant was mixed with 1/5 volume of CTAB solution (100 mmol L?1Tris-HCl, pH 8.0 50 mmol L?1EDTA, pH 8.0; 1.0 mol L?1NaCl; 10% CTAB) and incubated at 65 °C for 10-20 min. One volume of chloroform:isoamyl alcohol (24:1) was added and mixed evenly, followed by centrifugation at 10,000 ×g for 15 min. The supernatant was collected and DNA was recovered by ethanol precipitation. After 2× washing with 70% ethanol,the DNA was dissolved in 100 μL DNAse-free water. DNA quality was evaluated by 1% agarose gel electro-phoresis.Primers were designed for the plastid-specific psbA gene,mitochondrion-specific rrn26S gene, and nuclear β-actin gene sequences. The specific primers for the plastid genome were psbA-F, 5′-AAGCGAAAGCCTATGGGGTC-3′; psbA-R, 5′-CAACAGCAATCCAAGGACGC-3′. The specific primers for the mitochondrial genome were rrn26S-F, 5′-TTTTCAAGTGTCAG TAGCGC-3′; rrn26S-R, 5′-TTGACTATGACAAGAGTCGC-3′. The β-actin gene-specific primers were β-actinF, 5′-ATGGCCGATGG TGAGGAC ATTC-3′, and β-actinR, 5′-CCGATGGTGATGACTTG TCCA-3′. DNA samples were stored at ?20 °C until use.

2.3.Organelle genome sequencing and sequence assembly

Sequencing libraries were constructed using good quality DNA, and sequencing was performed by Roche 454 FLX+pyrosequencing [21]. Sequencing services were provided by Personal Biotechnology Co., Ltd., Shanghai, China. Low quality reads were filtered and sorting of the raw reads yielded 94,790 good quality reads.The mean read length was 502 bases.Sequences with read length ＞500 bases accounted for 56.9% of total sequences. These high quality reads were assembled using Newbler v2.8 (Illumina, Life Technologies,Roche, Basel, Switzerland). After chloroplast and mitochondrial genome contigs were assembled, they were joined by Sanger sequencing of PCR products, and the primers (Table S1) designed from contig end sequences were employed for closing gaps.

2.4.Detection of rps3 gene mutations

Total DNA was extracted from seedlings of S. arvensis and B. napus using the CTAB method [25]. Primers were designed using Primer Premier 5.0 and Primer-BLAST (http://www.ncbi.nlm.nih.gov/tools/primer-blast/index.cgi?LINK_LOC=BlastHome), and designed in the region of a deletion in exon 2 of the ribosomal protein S3 (rps3) gene: rps3F-33 bp (5′-AAACTAGACCCGGAAAGGAGAAG-3′), rps3R-33 bp (5′-AACGCTATGTCATTTACAACCCC-3′). PCR amplification was performed using the 2× Es Taq MasterMix polymerase purchased from CWBiotech Co., Ltd. (Beijing, China)in accordance with the recommended PCR reaction system. PCR products were separated by 2% agarose gel electrophoresis at 130 V and viewed for fragment patterns specific to S.arvensis or B.napus.

2.5. Annotation and analysis of genomic sequences

Sequence alignment was performed using BLAST [26] with evalues of ＜0.001. ORFs longer than 303 bases were selected using ORF Finder (https://www.ncbi.nlm.nih.gov/orffinder/).Plastid and mitochondrial genomes were annotated using DOGMA and MITOFY [27]. All t-RNA genes were predicted using tRNAscan-SE [28]. To improve the accuracy of annotation, the identified genes were further verified by alignment with genes of other species. The annotated sequences were submitted to GenBank. The GenBank accession numbers of the mitochondrial and plastid genome sequences are KM851044 and KU050690, respectively. The circular maps of organelle genomes were drawn using OGDRAW v1.2 [29].Transmembrane structures of protein-coding genes were predicted using TMHMM 2.0 [30]. Progressive Mauve [31]software was used to identify complete genome single nucleotide polymorphisms (SNPs) in the homologous regions of organelle genome in different species. Two phylogenetic trees were constructed based on SNPs shared by organelle genomes. The phylogenetic analysis was performed using MEGA 7.0 [32]. A maximum-likelihood tree was constructed using the Kumar method [32]. Progressive Mauve [31] was used for multiple alignments of the mitochondrial genomes in Brassica. The GenBank accession numbers of organelle genome sequences of plant species used for phylogenetic analysis are listed in Table S2.

Fig.1-Plastid genome map of Sinapis arvensis.Thick lines indicate inverted repeats(IRa and IRb)that separate the genome into large and small single-copy regions(LSC and SSC).Features on the clockwise-and counter-clockwise-transcribed strands are drawn on respectively the inside and outside of the circle.

3. Results

3.1.The plastid genome of S. arvensis

The full set of Roche GS FLX+ reads was used for de novo assembly of the S. arvensis organelle genome. The coverage of chloroplast genome sequencing was 28×. The plastid genome was 153,590 bp in length (Fig. 1), including a pair of inverted repeats (IR) regions of 26,239 bp long (IRa and IRb).The IRs were separated on one side of the circular genome by a 17,720-bp small single-copy (SSC) region and on the other side by an 83,393-bp large single-copy (LSC) region.Coding gene sequences accounted for 50.2% of the plastid genome, with a total of 128 genes, including four rRNA, 29 tRNA, and 79 protein-coding genes. The IR regions contained 16 duplicate genes. The rps19 gene spanned the IRa and LSC regions.

3.2. The mitochondrial genome of S.arvensis

Three mitochondrial sequence contigs were obtained by de novo assembly of the S. arvensis organelle genome. The mitochondrial genome sequencing coverage was 53×. The mitochondrial genome of S. arvensis is 240,024 bp in length with a 45.2% G + C content (Fig. 2). The 54 identified genes included 33 known protein-coding genes,18 tRNA genes,and 3 rRNA genes. The sequence of full-length protein-coding genes was 30,059 bp in length,accounting for 19.2% of the full genome.The mitochondrial genome of S.arvensis was similar to those of Raphanus sativus, B. nigra, Eruca sativa, and B.carinata in harboring a deletion of the cox2-2 gene, and also similar to B.nigra in harboring a 33-bp deletion in exon 2 of the rps3 gene (Fig. 3). Sequence variation in this exon was confirmed by PCR amplification using primers designed to amplify the expected fragments of 280 bp and 313 bp from different haplotypes of this gene. A 33-bp deletion in S.arvensis rps3 was absent in B.napus(Fig.4).

Forty-five open reading frames (ORFs) ＞300 bp in length were identified in the intergenic regions of the S. arvensis mitochondrial genome (Table S3).One of them was predicted to lie within the mitochondrial membrane (orf115c), and 17 ORFs were identified with transmembrane structure(Table 1),whereas other ORFs were located outside the membrane.Four ORFs showed chimeric structures of a known mitochondrial gene and encoded hypothetical proteins with transmembrane domains.

Fig.2- The mitochondrial genome of Sinapis arvensis.Genes with forward and reverse orientation are shown respectively outside and inside the circle.Features on the clockwise-and counter-clockwise-transcribed strands are drawn on respectively the inside and outside of the circle.

Fig.3-Sequence alignment of rps3 genes from S.arvensis and B.napus.The labels arv-rps3,nig-rps3,rap-rps3,jun-rps3,naprps3,ole-rps3,and car-rps3 represent the rps3 genes of S. arvensis,B. nigra,B.rapa,B. juncea,B. napus,B. oleracea,and B. carinata.

Fig. 4-PCR amplification of rps3 from B.napus and S. arvensis.Lane 1, marker DL2000;lane 2,B. napus;lane 3, S.arvensis.

3.3. Phylogenetic analysis of organelle genomes

Two phylogenetic trees were constructed based on SNPs shared by organelle genomes of representative monocots and dicots. The phylogenetic tree of the mitochondrial genomes included 13 Brassicaceae species, while that of the chloroplast genome included 12 Brassicaceae species.

In the two constructed phylogenetic trees, monocots and dicots were separated in different groups, and the Brassicaceae species were clustered in a separate group (Fig.5). S. arvensis was located in the circle of the species involved in the triangle of U [14], indicating that S. arvensis has a close genetic relationship with Brassica species. Furthermore, the phylogeny of the mitochondrial genomes from 12 Brassicaceae species strongly supported a close relationship between S.arvensis and B.nigra and B.carinata(Fig.5-A).The chloroplast genome of S.arvensis was more similar to those of the Brassica species than to those of other Brassicaceae species such asCapsella rubella, A. thaliana, Limnanthes floccose, and Carica papaya(Fig.5-B).

Table 1-Transmembrane structure prediction of ORFs in the Sinapis arvensis mitochondrial genome.

3.4. Comparative analysis of organelle genomes

Because evolutionary analysis indicated that S. arvensis was closely related to B.nigra,B.carinata,and R.sativus,we selected the mitochondrial genomes of these species for comparative genome synteny analysis.

The syntenic relationship between S. arvensis and its relatives based on mitochondrial genome sequence was highly collinear (Fig. 6). Very little recombination of the mitochondrial genomes among S. arvensis, B. nigra, and B.carinata was found, with only one block inverted. However,compared with the R. sativus mitochondrial genome, the S.arvensis mitochondrial genome showed a wider range of recombination. These results showed that the genetic relationship between S. arvensis and B. nigra and B. carinata was much closer than that between S. arvensis and R.sativus.

Comparison of the S. arvensis chloroplast genome with those of related species such as Arabidopsis thaliana showed that these species had a more distant genetic relationship with S. arvensis than did B. nigra and B. carinata. However, all the chloroplast genomes showed high collinearity, without recombination or rearrangement (Fig. 7). The comparative maps of the S. arvensis chloroplast genome with those of related species were the same as in Fig. 7. This result indicated that the structure of the chloroplast genome was more conserved in evolution than that of the mitochondrial genome.

4. Discussion

4.1.The organelle genomes of S. arvensis

As in other sequenced Brassica species, the plastid genome,including that of S. arvensis, showed a greater degree of conservation than the mitochondrial genome. The 154-kb plastid genome of S. arvensis was larger than those(120-160 kb) [33] of most sequenced higher plants. Compared to other chloroplast genomes of sequenced Brassicaceae species, S. arvensis showed a high level of conservation of protein-coding sequences, indicating that the chloroplast genome underwent evolution at a very slow rate. The 240-kb mitochondrial genomes was smaller than those(200-2400 kb)of most sequenced higher plants.The mitochondrial gene rps3 of S. arvensis and B. nigra contained a 33-bp exonic deletion and lacked the cox2-2 gene harbored by most Brassicaceae species [34], indicating that mitochondrial gene variation occurred during evolution. The position of each species was consistent between the phylogenetic trees constructed using chloroplast and mitochondrial genomes, indicating that the evolution of the chloroplast genome was consistent with that of the mitochondrial genome.

Studies have shown that most cytoplasmic male sterile genes are chimeric open reading frames (ORFs) that contain chimeric structure with functional genes, such as cox1, atp8,and atp6, and most CMS genes encode transmembrane proteins [35]; for example, orf288 of Hau CMS [36] and orf138 of Ogu CMS in B. napus [29] and orf463 of DCGMS CMS in R.sativus [37]. In the present study, 45 ORFs were identified in the mitochondrial genome of S. arvensis, of which 17 ORFs have shown transmembrane structure features, including 4 ORFs that display a chimeric structure with known mitochondrial genes. These four ORFs displaying chimeric structures and encoding hypothetical proteins with transmembrane domains are likely associated with the infertility of Nsa CMS which derived from somatic cell hybridization between B.napus and S. arvensis. Further functional studies may identify the regulationary mechanism that cause of sterility in Nsa CMS.

4.2.Phylogenetic analysis of genomes

Phylogenetic analysis successfully clustered the six species in U’s triangle together based on organelle genomic sequences.Each composite species clustered together with one of its base species based on the organelle genome, supporting the correctness of U’s triangle from a cytoplasmic perspective.The composite species showed a certain preference in cytoplasm selection in the course of evolution. For example,B.napus is a natural hybrid between B.rapa and B.oleracea,and its cytoplasm is closer to that of B.rapa,whereas B.juncea is a hybrid between B. nigra and B. rapa, and its cytoplasm is also closer to B. rapa. This finding suggests that in the natural hybridization process, hybridization was more likely to succeed when B. rapa was used as the female parent. The hybrid produced from a B. rapa female parent thus showed strong adaptability to the environment and could be used in agriculture. Yamagishi et al. [34] proposed that the B. carinata mitochondrion originated in B. nigra, based on the comparative analysis of mitochondrial genome sequences of B.nigra,B.oleracea, and B. carinata. We grouped S. arvensis together with B. nigra and B. carinata based on mitochondrial genomic sequences and S. arvensis together with B. nigra based on plastid genomic sequences other than the B. carinata plastid genome. The rps3 gene sequence of B. carinata is clearly distinct from that of B.nigra and S.arvensis,but is the same as that of B. oleracea, B. napus, and B. rapa. This finding suggests that the mitochondrial genome of S.arvensis is more similar to that of B. nigra than to that of B. carinata, and that the evolutionary path should be first from S. arvensis to B. nigra and then to B.carinata.

Fig.5- Phylogenetic analysis based on organelle genome.(A)Phylogenetic tree constructed using mitochondrial genome.(B)Phylogenetic tree constructed using chloroplast genomes.Unrooted phylogenetic tree was created in MEGA 7.0 software by the maximum likelihood method with 1000 bootstrap iterations.

Fig.6- Comparative analysis of mitochondrial genomes of Sinapis arvensis and its relatives. The alignment was made using Mauve.Blocks of the same color denote homologous regions;blocks above or below the midline of a genome represent respectively direct or inverted blocks.

In a previous karyotypic investigation [38], S. arvensis was shown to contain the B genome of B. nigra. Our phylogenetic analysis showed that S. arvensis underwent earlier differentiation than B. nigra. The mitochondrial genome of S. arvensis(240,024 bp in length) was around 8 kb longer than that of B.nigra (232,145 bp) and B. carinata (232,241 bp), indicating that the mitochondrial genome of S. arvensis was partially lost during the evolution towards B.nigra.Thus,integration of the results of karyotypic and organelle genome analysis of S.arvensis supports a hypothesis that it is the ancestral species of B.nigra and B.carinata.

Fig.7-Comparison of syntenic region of Sinapis arvensis chloroplast genome(X-axis)with the Arabidopsis thaliana chloroplast genome(Y-axis).Alignment of the two genomes showed that they were consistent and no rearrangement was found,apart from SNPs.

5. Conclusions

We have described the organelle genome sequences of S.arvensis, a member of the Brassicaceae family. The S. arvensis chloroplast genome is 153,590 bp in length and contains 112 individual genes, including 4 rRNA, 29 tRNA, and 79 proteincoding genes. The mitochondrial genome is 240,024 bp in length and contains 54 genes,including 33 protein-coding,18 tRNA,and 3 rRNA genes.The organelle genome sequence of S.arvensis is more similar to that of B. nigra and B. carinata than to those of the other Brassicaceae species.The mitochondrial genome contains multiple ORFs; of which four display chimeric structural features and encode proteins with transmembrane domains, possibly accounting for the infertility of Nsa CMS constructed by somatic cell hybridization between B.napus and S. arvensis. These findings may be used as a foundation for identifying favorable variations in the organelle genomes of S. arvensis. Our results also provide further evidence that S.arvensis is the ancestral species of B.nigra and B. carinata. This study improves us to develop and utilize the germplasm resources of the S. arvensis and comprehend the evolution of organelle genomes in the Brassicaceae.

Declaration of competing interest

All authors declare no competing financial interests.

Acknowledgments

We thank professor Guan Rongzhan in Nanjing Agriculture University for guidance in organelle extraction. This work was supported by the National Natural Science Foundation of China (30871553), the Agricultural Science and Technology Innovation Program of Chinese Academy of Agricultural Sciences (Group No. 118), and the Hubei Agricultural Science and Technology Innovation Center(201620000001048).

Author contributions

HQ conceived and designed the research, SSF performed the analysis, interpreted the results and drafted the manuscript,MDS grew plant material, WH, LJ, WWX assisted in data analysis, FL assisted in DNA preparation, and CHT, QUZ, LKD,and HQ revised the manuscript. All authors read and approved the final version of the manuscript.

Additional information

The datasets generated during the current study are available in the GenBank with accession numbers KM851044 (mitochondrial) and KU050690 (plastid) genome sequences,respectively.

Appendix A. Supplementary data

Supplementary data for this article can be found online at https://doi.org/10.1016/j.cj.2019.12.001.