Characterization and Phylogenetic Analyses of the Complete Mitochondrial Genome of Sugarcane ( Saccharum spp. Hybrids) Line A1

: Modern sugarcane cultivars are highly polyploid with complex nuclear genomic genetic background, while their mitochondrion (mt) genomes are much simpler, smaller and more man-ageable and could provide useful phylogenetic information. In this study, the mt genome of a modern commercial cultivar A1 was sequenced via Illumina Hiseq XTen and PacBio Sequel platform. The assembled and annotated mitochondrial genomes of A1 were composed of two circular DNA molecules, one large and one small, which were named Chromosome 1 and Chromosome 2. The two distinct circular chromosomes of mitogenome construct is consisted with other sugarcane cultivars i.e., Saccharum ofﬁcinarum Khon Kaen 3 and Saccharum spp. hybrids ROC22 and FN15. The Chromosome 1 of A1 mitogenome is 300,822 bp in length with the GC content of 43.94%, and 7.14% of Chromosome 1 sequences (21,468 nucleotides) are protein coding genes (PCGs) while 92.86% (279,354 nucleotides) are intergenic region. The length of Chromosome 2 is 144,744 bp with the GC content of 43.57%, and 8.20% of Chromosome 2 sequences (11,865 nucleotides) are PCGs while 91.80% (132,879 nucleotides) are intergenic region. A total of 43 genes are located on Chromosome 1, which contains 22 PCGs (six nad genes, four rps genes, four atp genes, three ccm genes, three cox genes, one mat gene and one mtt gene) and 21 non-coding genes including 15 tRNAs and 6 rRNAs. Chromosome 2 includes 18 genes in total, which contains 13 PCGs (four nad genes, three rps genes, two atp genes, one ccm gene, one cob gene, one cox gene and one rpl gene) and ﬁve non-coding genes (tRNA genes). Analysis of codon usage of 35 PCGs showed that codon ending in A/U was preferred. Investigation of gene composition indicated that the types and copy numbers of CDS genes, tRNAs and rRNAs of A1 and FN15 were identical. The cox1 gene has two copies and the trnP gene has one copy in A1, FN15 and ROC22 three lines, while there is only one copy of cox1 and two copies of trnP in S. ofﬁcinarum Khon Kaen 3. In addition, S. ofﬁcinarum Khon Kaen 3 have no nad1 gene and rps7 gene. 100 sequence repeats, 38 SSRs and 444 RNA editing sites in A1 mt genome were detected. Moreover, the maximum likelihood phylogenetic analysis found that A1 were more closely related to S. spp. hybrid (ROC22 and FN15) and S. ofﬁcinarum (Khon Kaen 3). Herein, the complete mt genome of A1 will provide essential DNA molecular information for further phylogenetic and evolutionary analysis for Saccharum and Poaceae.


Introduction
Mitochondria (mt) are vital organelles in cell, which are involved in a large number of metabolic processes associated with energy production and the synthesis of some compounds [1,2]. Mitochondria are known as the energy factories which are major sites of ATP synthesis by oxidative phosphorylation [3]. Mitochondria are semiautonomous organelle due to they have their own genome and protein-synthesizing mechanism, and thus, mitochondria are involved in multitudinous other living processes, such as the biosynthesis of amino acids, vitamin cofactors and fatty acids [3].
Though plants can obtain energy from sunlight by the chloroplasts, plant mitochondria remain the energy centers of their metabolism through plant cell respiration [4]. Plant mitochondria are typically 1-3 µm in length and~0.5 µm in diameter, meaning that they are the same size and shape as some common bacteria because they are developed from endosymbiotic bacteria [4]. Mitochondrial has long been thought to tend to integrate DNA from different sources, especially bacterial genomic DNAs, through intracellular and horizontal transfer [5]. It is difficult to observe by light microscopy due to tiny genome size and this postponed their recognition [4]. In addition, it is difficult to purify plant mitochondria and its purification is often interfered with by chloroplasts and other plastids, which also leads to a lag in the study of mitochondrial genome compared with that of animals.
Plant mitochondrial genomes range from 42 Kb (Mesostigma viride) to 11.3 Mb (Silene conica) in size [6,7]. The mt genomes of plants have distinct differences in length, GC content, sequence and gene content and could contain potential phylogenetic messages [8]. Moreover, the mt genome exhibits many features, such as a faster frequency of evolution than nuclear DNA, the characteristics of conservation gene functions and high AT content [9]. The mt genome is typically inherited maternally, and the maternal inheritance limits outcrossing recombination [10]. Thus, the mt genome sequences usually were employed to phylogenetic and evolutionary analysis and retention of maternal inheritance in crossbreeding [11,12].
Sugarcane, which ranks amongst the top ten crops worldwide, is the most important crop for sugar and biofuel [13,14]. It accounts for more than 80% of sugar in the world, approximately 90% of sugar in China, and about 60% of ethanol global production [15,16]. Modern sugarcane cultivars are complex hybrids at least from the three species: Saccharum spontaneum, Saccharum robustum and Saccharum officinarum [17]. S. officinarum (called "noble cane") has much sugar content and had the largest cultivated area in the world before the 1920s [18]. However, they were susceptible to disease and insects. To introduce high resistance characters and reserve high sugar content trait, noble canes were constantly crossbreeding with other Saccharum [19,20] such as S. spontaneum. Hence, S. officinarum contributes to high sugar content, and S. spontaneum provides disease resistance, stress resistance and ratooning ability [21]. In other words, modern sugarcane varieties have an extremely complex interspecific polyploid genome and the complete genome is still remaining to be explored, whether nuclear genome or organelle genome (chloroplast and mitochondrial genome) [22]. At present, the whole genome of AP85-441, which belongs to S. spontaneum (allele-defined genome of tetraploid), was assembled by Zhang et al. in 2018 [20]. Modern sugarcane cultivar R570 based on bacterial artificial chromosome (BAC) clones have been assembled [23]. The complete chloroplast genome of two sugarcane ancestors S. officinarum and S. spontaneum were assembled and analyzed [11]. With the development of next-generation sequencing technologies (NGS), a growing number of mt genomes have been assembled. At present, more than 300 complete mt genomes have been submitted to GenBank Organelle Genome Resources. The complete mitochondrial genome of modern commercial sugarcane cultivars such as S. spp. hybrid (ROC22 and FN15) and S. officinarum Khon Kaen 3 were obtained and analyzed [15,24,25].
In the present study, to better comprehend the evolutionary information of modern sugarcane, the complete mitochondrial genome of S. spp. hybrid line A1 were sequenced and assembled using a combination of Illumina Hiseq XTen and PacBio Sequel platform. The mitogenome characteristics, codon usage bias, repetitive element during genome, the prediction of RNA editing and phylogenetic relationship with in Poaceae were investigated. The mitochondrial genome herein will facilitate the sugarcane breeding, molecular marker selection and further phylogenetic and evolutionary analysis.

Plant Sampling, Mitochondria DNA Sequencing and Genome Assembly
Sugarcane line A1 were provided by the Fujian Agriculture and Forestry University (geographic coordinates: 26 • 9 8" N, 119 • 24 24" E), Fujian, Fuzhou, China. The specimen of A1 was stored in the Key Laboratory of Sugarcane Biology and Genetic Breeding, Fujian Agriculture and Forestry University with store number A1-FJ2016004. A1 were nursed in the greenhouse at 32 • C. After one week of dark culture, fresh yellowing seedlings were harvested. The A1 fresh yellowing seedlings were frozen in liquid nitrogen immediately after sampling and stored at −80 • C until mtDNA extraction. MtDNA was extracted and purified by an improved protocol as described by Chen et al. [26].
After mtDNA isolation, one microgramme of purified DNA was fragmented to construct short-insert libraries (insert size 430 bp) according to the manufacturer's instructions (Illumina, San Diego, CA, USA), then sequenced on the Illumina Hiseq XTen and PacBio Sequel platform by Shanghai BIOZERON Co., Ltd. (Shanghai, China). The raw sequenced reads from Illumina Hiseq were filtered firstly [27]. Then the mitochondria genome was reconstructed and assembled by SPAdes v3.10.1 using a combination of both Illumina and Pacbio data [28].

RNA Editing Analyses
Based on the proverbial principle of calculating predictive RNA editing sites, where editing in plant organelles increases protein preservation across species, PREP suite software (http://prep.unl.edu/, accessed on 4 November 2021) with a cut off value of 0.2 was performed [39].

Characteristics of the Mitogenome of Sugarcane Line A1
Mitochondrial genomes of sugarcane line A1 were sequenced via Illumina sequencing (7099 Mb raw data, Q20 = 97.28%) and PacBio sequencing (188,008 subreads, N50 = 1553 bp). The assembled and annotated mitochondrial genomes of sugarcane line A1 were composed of two circular DNA molecules, one large and one small, which were named Chromosome 1 and Chromosome 2 (Chr1 and Chr2). The Chr1 is 300,822 bp in length with the GC content of 43.94%, and 7.14% of Chr1 genome sequences (21,468 nucleotides) are PCGs (proteincoding genes) while 92.86% (279,354 nucleotides) are intergenic region. The length of Chr2 is 144,744 bp with the GC content of 43.57%, and 8.20% of Chromosome 2 sequences (11,865 nucleotides) are PCGs while 91.80% (132,879 nucleotides) are intergenic region (Table S1).
A total of 43 genes are located on Chromosome 1, which contains 22 PCGs (six nad genes, four rps genes, four atp genes, three ccm genes, three cox genes, one mat gene and one mtt gene) and 21 non-coding genes including 15 tRNAs and 6 rRNAs. Chromosome 2 includes 18 genes in total, which contains 13 PCGs (four nad genes, three rps genes, two atp genes, one ccm gene, one cob gene, one cox gene and one rpl gene) and five non-coding genes (tRNA genes) ( Figure 1).

Characteristics of the Mitogenome of Sugarcane Line A1
Mitochondrial genomes of sugarcane line A1 were sequenced via Illumina sequencing (7099 Mb raw data, Q20 = 97.28%) and PacBio sequencing (188,008 subreads, N50 = 1553 bp). The assembled and annotated mitochondrial genomes of sugarcane line A1 were composed of two circular DNA molecules, one large and one small, which were named Chromosome 1 and Chromosome 2 (Chr1 and Chr2). The Chr1 is 300,822 bp in length with the GC content of 43.94%, and 7.14% of Chr1 genome sequences ( (Table S1).
A total of 43 genes are located on Chromosome 1, which contains 22 PCGs (six nad genes, four rps genes, four atp genes, three ccm genes, three cox genes, one mat gene and one mtt gene) and 21 non-coding genes including 15 tRNAs and 6 rRNAs. Chromosome 2 includes 18 genes in total, which contains 13 PCGs (four nad genes, three rps genes, two atp genes, one ccm gene, one cob gene, one cox gene and one rpl gene) and five non-coding genes (tRNA genes) ( Figure 1).  We further investigated the differences of gene composition between sugarcane line A1 and other Saccharum complexes (S. spp. hybrid ROC22, S. spp. hybrid FN15 and S. officinarum Khon Kaen 3) (Table S2). The types and copy numbers of CDS genes, tRNAs and rRNAs of A1 and FN15 were identical. The cox1 gene has two copies and the trnP gene has one copy in A1, FN15 and ROC22, while there is only one copy of cox1 and two copies of trnP in S. officinarum Khon Kaen 3. Moreover, S. officinarum Khon Kaen 3 have no nad1 gene and rps7 gene.
In Chromosome 1, almost all the PCGs use ATG as the initiation codon, while gene nad1 starts with ACG and matR begins with ATA. In Chromosome 2, however, all the PCGs use the initiation codon ATG except for nad2 (TTG) and nad5 (CCA). Regarding the stop codon, whether in Chromosome 1 or Chromosome 2, most of the PCGs stop with TAA, TAG and TGA, except for gene nad2 and nad5 in Chromosome 1 terminate with CGG and GTA, respectively (Table S3).

Codon Usage Bias
We analyzed the codon usage frequency of the A1 mitochondrial genome. A total of 11,320 codons were calculated, which included 20 amino acids and three stop codons, indicating a stronger coding capacity (Figure 2 and Table S4). The most abundant amino acid is leucine (Leu) and the next is isoleucine (Ile), with the number of 1246 and 1052 codons, respectively. The least amount is cysteine (Cys) with only 313 codons. Both methionine (AUG) and tryptophan (UGG) had only one codon type with number of 288 and 198, respectively, and showed no bias (RSCU = 1.00). The codon CAA for glutamine with the highest RSCU (relative synonymous codon usage) values (1.48). While the codon CGC for arginine with the minimal RSCU values (0.46). Spectacularly, all codons whose RSCU values were higher than one ended with U or A, except for AUC-isoleucine, UUG-leucine, UCC-serine, and UAG-termination codon.
genes outside the ring are located on the direct strand, while the genes inside the ring are on the reverse strand. The gray color of inner circle represented the GC content of mt geno We further investigated the differences of gene composition between sugarca A1 and other Saccharum complexes (S. spp. hybrid ROC22, S. spp. hybrid FN15 officinarum Khon Kaen 3) (Table S2). The types and copy numbers of CDS genes, and rRNAs of A1 and FN15 were identical. The cox1 gene has two copies and t gene has one copy in A1, FN15 and ROC22, while there is only one copy of cox1 a copies of trnP in S. officinarum Khon Kaen 3. Moreover, S. officinarum Khon Kaen no nad1 gene and rps7 gene.
In Chromosome 1, almost all the PCGs use ATG as the initiation codon, whi nad1 starts with ACG and matR begins with ATA. In Chromosome 2, however, PCGs use the initiation codon ATG except for nad2 (TTG) and nad5 (CCA). Regard stop codon, whether in Chromosome 1 or Chromosome 2, most of the PCGs sto TAA, TAG and TGA, except for gene nad2 and nad5 in Chromosome 1 termina CGG and GTA, respectively (Table S3).

Codon Usage Bias
We analyzed the codon usage frequency of the A1 mitochondrial genome. A 11,320 codons were calculated, which included 20 amino acids and three stop cod dicating a stronger coding capacity (Figure 2 and Table S4). The most abundant acid is leucine (Leu) and the next is isoleucine (Ile), with the number of 1246 an codons, respectively. The least amount is cysteine (Cys) with only 313 codons. Bo thionine (AUG) and tryptophan (UGG) had only one codon type with number of 2 198, respectively, and showed no bias (RSCU = 1.00). The codon CAA for glutami the highest RSCU (relative synonymous codon usage) values (1.48). While the codo for arginine with the minimal RSCU values (0.46). Spectacularly, all codons whose values were higher than one ended with U or A, except for AUC-isoleucine, UUG-l UCC-serine, and UAG-termination codon.

Repetitive Element and Simple Sequence Repeat (SSR) Analysis
A total of 100 repetitive sequences of three types were detected in the A1 mi drial genome. F (forward repeat), P (palindromic repeat) and R (reverse repeat) acc

Repetitive Element and Simple Sequence Repeat (SSR) Analysis
A total of 100 repetitive sequences of three types were detected in the A1 mitochondrial genome. F (forward repeat), P (palindromic repeat) and R (reverse repeat) accounted for 64%, 19% and 17%, respectively (Table S5). There were no complement repeat sequences, while reverse repeat sequences only existed on chromosome 2. In addition, the longest forward repeat was 12,241 bp and the longest palindromic repeat was 4058 bp. It also showed that the 30-90 bp repeats are most common in the two chromosomes.
A total of 38 SSRs were found in the mt genome of A1, including five types of SSR, and most were located in intergenic regions ( Table 1). The mono-nucleotide repeats were the most common (27,71.05%), followed by penta-nucleotide (four, 10.54%) and compound-Diversity 2022, 14, 333 6 of 12 nucleotide (three, 7.89%). Two additional types of SSRs were less abundant: di-nucleotide (two, 5.26%) and tri-nucleotide (two, 5.26%). In addition, A/T-containing motifs are most frequent among the SSRs, accounting for 76.3%. This phenomenon is consistent with the previous study in plant mitochondrial genomes [39].

The Prediction of RNA Editing
RNA editing widely exists in eukaryotes, including higher plants. In mitochondrion, the conversion of specific cytosine into uridine changes the genomic information. In this analysis, the software PREP was employed to predict the RNA edit site, 444 RNA editing sites within 35 PCGs were predicted in the two chromosomes of A1, using the PREP-MT program (Table 2 and Figure 3). Among these PCGs, nad5 have no possible editing site, while ccmC (36) has the maximum editing sites on Chromosome 1. In addition, 35.36% (157) were located at the first position of the triplet codes, 61.49% (273) occurred with the second base of the triplet codes. There were two particular editing cases in which the first and second positions of the triplet codes were all edited, resulting in two amino acids changing from the proline (CCT, CCC) to phenylalanine (TTT, TTC).

The Prediction of RNA Editing
RNA editing widely exists in eukaryotes, including higher plants. In mitochondrion, the conversion of specific cytosine into uridine changes the genomic information. In this analysis, the software PREP was employed to predict the RNA edit site, 444 RNA editing sites within 35 PCGs were predicted in the two chromosomes of A1, using the PREP-MT program ( Table 2 and Figure 3). Among these PCGs, nad5 have no possible editing site, while ccmC (36) has the maximum editing sites on Chromosome 1. In addition, 35.36% (157) were located at the first position of the triplet codes, 61.49% (273) occurred with the second base of the triplet codes. There were two particular editing cases in which the first and second positions of the triplet codes were all edited, resulting in two amino acids changing from the proline (CCT, CCC) to phenylalanine (TTT, TTC).    After the RNA editing, 42.34% of amino acids did not change hydrophobicity or hydrophilicity. The proportion of amino acids that changed from hydrophilic to hydrophobic was 45.05%, while the proportion of amino acids that changed from hydrophobic to hydrophilic was 2.16%. In addition, the amino acid of predicted editing codons showed a leucine (170 sites) bias after RNA editing.

Phylogenetic Analysis
To understand the genetic status of A1, the phylogenetic tree was performed based on the whole mtDNAs of 11 species (nine Poaceae species and two outgroups) using PhyML v3.0 (Figure 4). The phylogenetic tree showed that S. spp. hybrid A1 is very close to S. spp. hybrid (FN15 and ROC22) and S. officinarum (Khon Kaen 3). The complete mitochondrial genome herein will provide important and fundamental DNA molecular information of evolutionary analysis for Saccharum and Poaceae.
After the RNA editing, 42.34% of amino acids did not change hydrophobicity or hydrophilicity. The proportion of amino acids that changed from hydrophilic to hydrophobic was 45.05%, while the proportion of amino acids that changed from hydrophobic to hydrophilic was 2.16%. In addition, the amino acid of predicted editing codons showed a leucine (170 sites) bias after RNA editing.

Phylogenetic Analysis
To understand the genetic status of A1, the phylogenetic tree was performed based on the whole mtDNAs of 11 species (nine Poaceae species and two outgroups) using PhyML v3.0 ( Figure 4). The phylogenetic tree showed that S. spp. hybrid A1 is very close to S. spp. hybrid (FN15 and ROC22) and S. officinarum (Khon Kaen 3). The complete mitochondrial genome herein will provide important and fundamental DNA molecular information of evolutionary analysis for Saccharum and Poaceae.

Discussion
Mitochondria produce the energy for life processes and are named as the powerhouses. Plant mitochondria have more complicated genomes than animals with a wide

Discussion
Mitochondria produce the energy for life processes and are named as the powerhouses. Plant mitochondria have more complicated genomes than animals with a wide range of size variations, sequence arrangement, and a dynamic structure with various conformations [41,42]. In this study, after the algorithm predicts, the mt genome of a modern cultivated sugarcane line A1 was assembled into two distinct circular chromosomes. According to the published data, most mt genomes are single circular, while the mt genome of sugarcane has two circular chromosomes. For example, S. spp. hybrid ROC22, S. spp. hybrid FN15 and Saccharum officinarum Khon Kaen 3 all have two distinct circular chromosomes [15,24,25]. In addition, in some species like Cucumis sativus [43], Glycine max [44] and Allium cepa [45], mitochondria also consist of two or more circular chromosomes. However, there are no detailed studies to elucidate why the mitochondrial genome of sugarcane has two chromosomes.
Codons play a vital role in the course of transformation of genetic information. There are 11,320 codons in A1 mt genome. The most and least abundant amino acids are leucine and cysteine. This phenomenon is also found in Mangifera mitochondrial genomes [46]. In addition, codon usage bias is universal in plant species. In this study, codon usage bias was calculated by the RSCU value. The results suggested an intense A or U bias in the third position of the codon in the PCGs and were different with Mangifera mitochondrial genomes and Solanales (A or T bias in the third position), indicating that different plants have formed different codon predilection during the long periods of evolution [46,47].
The repeat sequences diffusely exist in the mt genome. The majority of differences in the mt genome size can be elucidated by distinction in the size of the repeat sequences in plants [48]. Previous studies also have shown that repeats in mitochondria are connected with rearrangement and recombination of the mt genome [49]. In this study, there are numerous repeats in sugarcane mitochondrial genome, which may imply that intermolecular recombination occurs hourly in sugarcane mitochondrial genome. Moreover, a total of 38 SSRs were found in the mt genome of A1, and most were located in IGS, which is similar to Suaeda glauca mt genome [50]. Single base repeats were the most, accounting for 76.3%, which was different with the sunflower. While the result of A/T-containing motifs are most frequent among the SSRs was consistent with sunflower [51]. RNA-editing is a post-transcriptional step that exists in the chloroplast and mitochondrial genome of plants, conducing to the proteins fold better [52]. The investigation of RNA editing sites contributes to understanding the expression of CP and MT genes in plants. In rice mitochondria, 491 editing sites have been identified [53]. In this study, 444 RNA-editing sites within 35 genes were identified, which can provide available information for forecasting gene functions with novel codons.
We also made a series of comparisons with related cultivars (FN15, ROC22, and Khon Kaen 3), including chromosomes structure, gene composition, codon usage and repetitive elements. The results showed that the differences between them were extremely small. Firstly, A1, FN15, ROC22, and Khon Kaen 3 four lines all have two distinct circular chromosomes (a larger one and a smaller one). Secondly, the gene composition differences of them were tiny. The CDS genes compositions of A1, FN15 and ROC22 were coincident, while Khon Kaen 3 has no nad1 gene and rps7 gene and only one copy of cox1 gene. Four lines all have two 5S rRNA, two 18S rRNA and two 26S rRNA. In addition, the tRNA compositions of them are identical except the trnP, A1 and FN15 only have one copy of trmP, while ROC22 and Khon Kaen 3 have two copies (Table S2). Thirdly, the CDS genes of A1, FN15 and ROC22 were identical, so the codon usage analysis of them have no difference, while Khon Kaen 3 have differences caused by the three genes lost (Table S4). Fourthly, repetitive element analysis showed little difference between the A1, FN15, and ROC22. Forward repeat, Palindromic repeat and Reverse repeat account for 64, 19 (Table S5).
Plant mitochondria genomes are usually dynamic, resulting in heterogeneity, largescale genomic reorganization, and gene mosaicism in the mitochondrial genomes of various species [54,55]. Size and structural changes of plant mt genomes are individual. Here, the whole genome sequences are used to construct phylogenetic tree to explore the lineage between Saccharum and Gramineae species. The result showed the phylogenetic relationship of S. spp. hybrid A1 is very close to FN15, ROC22 and Khon Kaen 3.

Conclusions
Here, the mt genome of a modern commercial cultivar A1 was sequenced, assembled, and annotated. Interestingly, A1 mitogenome contains two distinct circular chromosomes, one large and one small, which were named Chromosome 1 and Chromosome 2. The two distinct circular chromosomes of mitogenome construct is consisted with other sugarcane cultivars i.e., Saccharum officinarum Khon Kaen 3 and Saccharum spp. hybrids ROC22 and FN15. Investigation of gene composition indicated that the CDS gene contents of A1 mt genome were identical with ROC22 and FN15, while Khon Kaen 3 lacks three CDS genes (cox1, nad1 and rps7). Analysis of codon usage of 35 PCGs showed that codon ending in A/U was preferred. RNA editing sites and SSRs analysis showed that 100 sequence repeats, 38 SSRs and 444 RNA editing sites in A1 mt genome were detected. Maximum likelihood phylogenetic analysis found that A1 were more closely related to S. spp. hybrid ROC22, S. spp. hybrid FN15 and S. officinarum Khon Kaen 3. The complete mitochondrial genome herein will provide essential genetic resources for further phylogenetic and evolutionary analysis for Saccharum and Poaceae.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/d14050333/s1; Table S1: Characteristics of mitochondrial genome of sugarcane line A1; Table S2: The genes composition of A1 and related species. Table S3: Initiation codon and termination codon in mitochondrial genome of sugarcane line A1; Table S4: Codon usage analysis of the A1 mitochondrial genome and related species; Table S5: The distribution of repeats in the A1 mt genome and related species.