*3.4. Prediction of RNA Editing Sites in PCGs*

Many previous studies have documented that RNA editing is one of the necessary steps for gene expression in the mitochondrial and chloroplast genomes of higher plants [64–67]. RNA editing is a post-transcriptional modification that converts specific cytidines (C) to uridines (U) and uridines to uridines in the transcripts of nearly all mitochondrial PCGs. Based on the web-based PREP-mt program, we predicted a total of 486 RNA editing sites in 31 PCGs and 100% C-to-U RNA editing. Among the 486 RNA editing sites, 34.57% (168 sites) were predicted at the first base position of the codon and 65.43% (318 sites) were found in the second position, while none were found in the third position. The lack of predicted RNA editing sites in the silent position is probably due to the limitation of the PREP-Mt predictive methodology rather than there being no RNA editing in this position. Since most of the RNA editing sites in third codon positions did not change the amino acid encoded by the codon, the tie-breaking rules used by PREP-Mt could not select the edited state [68]. Therefore, RNA editing in the silent editing position needs to be further identified by experimental methods.

The occurrence of RNA editing can cause alteration of initiation and termination codons in PCGs, and the frequency of their generation is much higher than that of their removal. As shown in Table 3, *mttB*, *nad1*, *nad4L*, and *rps10* genes use ACG as their initiation codons, which may be altered to the normal AUG by RNA editing modification. Additionally, the *ccmFC* gene uses CGA as its termination codon, which may be altered to UGA by RNA editing modification. As shown in Figure 6, the number of RNA editing sites in different genes varies greatly, and the Complex I (NADH dehydrogenase) and Cytochrome *c* biogenesis genes (*ccmB*, *ccmC*, *ccmFC*, and *ccmFN*) encode the most predicted RNA editing sites. Based on a comparison of the predicted RNA editing sites in five leguminous plants, the *nad4* gene encodes the most RNA editing sites, while *atp1* encodes the fewest (Figure 6).

**Figure 6.** Predicted RNA editing sites of the *P. vulgaris* mitogenome compared with four other leguminous plants. RNA-editing sites are predicted on PREP-Mt sites (http://prep.unl.edu/) with the cut-off value of 0.2.

Previous studies have shown that the frequency and type of RNA editing in each organelle is highly lineage-specific [26,65,69,70]. As shown in Figure 6, the number of predicted RNA editing sites in different papilionoid legume mitogenomes is very conserved, ranging from 486 sites in *P. vulgaris* to 503 sites in *Lotus japonicus*, suggesting that they share extremely conserved PCGs. In angiospermous mitogenomes, nearly all of the RNA editing sites are C to U, and the number of RNA editing sites is concentrated between 400 to 500. For example, 463 and 444 RNA-editing sites were found in the *C. lanatus* and *C. pepo* mitogenomes, of which 394 are shared [8]; 441 and 427 RNA-editing sites were found in *A. thaliana* and *B. napus* mitogenomes, of which 358 are shared [10]. In the gymnosperm *Cycas taitungensis*, 1084 RNA editing sites were found in its mitogenome [71]. The clearly descending number of RNA editing sites is in accordance with gene losses from gymnosperms to angiosperms. In contrast to angiosperms and gymnosperms, both types of C-to-U and U-to-C conversions are found in the mitochondrial transcripts of ferns and hornworts [69,72].
