Moss Genome Study Reveals Surprising RNA Editing and Codon Usage Patterns

A medicinal moss from the mountains of Yunnan, China, has delivered its complete mitochondrial genome to science for the first time, and the results are turning some long-held assumptions about plant organelle evolution on their head. Rhodobryum laxelimbatum, a member of the genus Rhodobryum in the family Bryaceae, is one of the well-known Chinese medicinal mosses, sharing traditional uses with its congeneric relatives Rhodobryum roseum and Rhodobryum giganteum, which are collectively referred to as Huixincao. Although direct pharmacological data on R. laxelimbatum remain scarce, the literature indicates that it possesses comparable traditional medicinal efficacy to its relatives, which contain amino acids, organic acids, phenolic compounds, and flavonoids with antioxidant, antibacterial, anti-inflammatory, and cardiovascular activities. Until now, however, genomic research on the genus has lagged far behind its pharmacological profile.

The new study, published in the journal Ecology and Evolution, reports the first complete mitochondrial genome for any species in the genus Rhodobryum and compares RNA editing and codon usage bias between its mitochondrial and chloroplast genomes. The research team collected whole moss individuals from Haba Snow Mountain in Shangri-La, Yunnan, China, at coordinates 100.117 degrees E and 27.313 degrees N, depositing the voucher specimen in the Herbarium of Hebei Normal University. Approximately 5 milligrams of green vegetative tissue from silica gel-dried samples was used for DNA extraction, and sequencing libraries were prepared from 300 to 500 base pair DNA fragments. The sequencing was performed on an Illumina HiSeq X Ten platform using 150-base pair paired-end reads, generating over 5.4 gigabases of raw data. After quality assessment with FastQC and adapter trimming with fastp, the mitochondrial genome was assembled using GetOrganelle with SPAdes as the core assembler and the embplant_mt seed reference database, achieving an average coverage depth of 307.8 times.

The assembled mitochondrial genome of R. laxelimbatum measures 103,071 base pairs in length with a GC content of 39.8 percent, closely resembling the mitochondrial genomes of the model moss Physcomitrella patens, which spans 105,340 base pairs with a GC content of 40.6 percent, and Anomodon rugelii, which spans 104,239 base pairs with a GC content of 41.2 percent. The genome contains 40 protein-coding genes, among which nad7 is identified as a pseudogene, along with 24 transfer RNA genes and three ribosomal RNA genes. The gene sequence accounts for 70,419 base pairs, representing 68.32 percent of the total genome, with exons comprising 39,107 base pairs, introns 31,312 base pairs, and intergenic spacers 32,668 base pairs. The GC content of the genes, at 40.99 percent, is slightly higher than the overall genome value, while the intergenic regions exhibit a lower GC content of 37.24 percent. Three pairs of overlapping genes were identified, with atp4 and atp8 overlapping by 8 base pairs, cox1 and atp4 by 4 base pairs, and ribosomal protein S12 and S7 by 4 base pairs.

The mitochondrial genome harbors 3 Group I introns and 22 Group II introns, a composition largely consistent with that reported for other mosses. Two open reading frames of at least 300 base pairs were detected within the introns of cox1 and nad5, measuring 1,869 and 1,602 base pairs respectively. Notably, no open reading frames were found within the 38 intergenic spacer regions that exceeded 300 base pairs, a result significantly different from observations in other mosses. Local BLASTN searches against the chloroplast and nuclear genomes of the model organism revealed no chloroplast-derived exogenous sequences in the mitochondrial intergenic regions, but 26 nuclear-derived sequences were identified, encompassing 8,625 base pairs and representing 8.37 percent of the complete mitochondrial genome. This proportion is significantly lower than those reported for angiosperms and other moss species, likely because nuclear genomic data from Bryales species were not used as references.

In total, 76 simple sequence repeats were identified in the mitochondrial genome, with mononucleotide repeats being the most abundant at 34 instances, or 44.74 percent, followed by dinucleotide repeats at 23 instances, or 30.26 percent, tetranucleotide repeats at 11 instances, complex repeats at 5 instances, trinucleotide repeats at 2 instances, and pentanucleotide repeats at a single instance. The mononucleotide repeats were predominantly composed of A and T bases, while dinucleotide repeats were primarily composed of AT motifs, a pattern consistent with SSR distributions in numerous plant species. Based on established criteria for reliable SSR primer design, seven feasible SSR loci were identified in R. laxelimbatum, providing a practical resource for SSR mining in mosses under limited conditions.

One of the most striking findings concerns RNA editing, the post-transcriptional process that alters RNA sequences after they are transcribed from DNA. The mitochondrial genome contained a total of 189 predicted RNA editing sites, with the majority attributed to C-to-U transitions at 128 sites, while the chloroplast genome contained a significantly higher number at 608 sites, split between 318 C-to-U and 290 U-to-C editing events. This result contradicts the conventional understanding that mitochondrial editing sites are generally more numerous than chloroplast ones in terrestrial plants. For the model bryophyte Physcomitrella patens, only 11 mitochondrial editing sites and a single chloroplast editing site have been annotated in previous studies. The cytochrome c biogenesis-related genes ccmFc and ccmFN exhibited relatively high numbers of editing sites in the mitochondrion, followed by nad5 and sdh3, while in the chloroplast genome, the hypothetical genes ycf2 and ycf1 showed the highest frequencies, followed by the RNA polymerase beta subunit 2 and maturase K genes. In both organellar genomes, 28 types of amino acid transitions were identified, with the conversion from hydrophilic to hydrophobic amino acids being predominant, aligning with previous findings in other plant species.

The codon usage bias analysis, conducted on 28 mitochondrial and 52 chloroplast protein-coding sequences, revealed that both organellar genomes exhibited the highest GC content at the first codon position, followed by the second and third positions, with an overall preference for A and T. Neutrality plot analysis showed no significant correlation between GC12 and GC3 for mitochondrial protein-coding genes, indicating strong constraints from natural selection on mitochondrial codon composition. In contrast, a significant correlation was observed in the chloroplast genome, suggesting that mutation pressure significantly contributes to shaping chloroplast codon usage bias. The ENC-plot analysis demonstrated that the distribution range of ENC values was wider in the mitochondrial genome, with nearly all values above 35, while the chloroplast genome exhibited a narrower range but contained more genes with ENC values below 35, indicating stronger codon usage bias in a subset of chloroplast genes. The PR2-plot revealed distinct base-composition imbalances at synonymous third codon positions, with mitochondria showing T exceeding A and comparable G and C frequencies, while chloroplasts showed both T exceeding A and G exceeding C.

Five optimal codons were identified in the mitochondrial genome, namely UUG, CCC, AAA, AGC, and GGC, while nine optimal codons were found in the chloroplast genome, all ending in A or U, consistent with the general pattern in land plant chloroplasts. Correspondence analysis revealed that mitochondrial codon variation is primarily driven by the magnitude of codon bias and synonymous third-position base composition, as the second principal axis was significantly correlated with ENC and GC3s. In chloroplasts, the first axis was significantly correlated with CAI, Fop, GC3s, and GC content, and the second axis exhibited a significant negative correlation with the Gravy value, indicating that chloroplast codon variation is affected by translational efficiency, base composition, and protein hydrophobicity. Collectively, the primary determinants of codon usage differ between the mitochondria and chloroplasts, suggesting that these organelles possess distinct replication-repair machinery, tRNA pools, and translational systems shaped by different evolutionary constraints.

Perhaps the most evolutionarily significant finding is the confirmation of nad7 pseudogenization in R. laxelimbatum, a non-basal moss species belonging to the Bryales family. Previous studies had documented nad7 pseudogeneization in the mitochondrial genome of Tetraphis pellucida and proposed that such events evolved independently in liverworts, mosses, and hornworts. Because nad7 pseudogeneization or gene loss was absent in non-basal species, its presence in R. laxelimbatum was unexpected, but it is consistent with a report documenting the same phenomenon in another non-basal moss species, Mielichhoferia elongata. Given that the affected species are phylogenetically distantly related and that nad7 functional loss also occurs in liverworts and hornworts, researchers have hypothesized that the functional loss of this gene evolved at least twice during moss evolution. The identification of nad7 pseudogenization in R. laxelimbatum further supports this hypothesis. Only two introns, nad7i140g2 and nad7i209g2, were absent from the R. laxelimbatum mitochondrial genome, closely associated with the functional impairment of the nad7 gene, while the intronic elements present in basal moss lineages and liverworts were gradually lost during the evolutionary diversification of mosses.

This study presents the entire mitochondrial genome of R. laxelimbatum, elucidating its genome size, GC content, gene composition, intron profile, repetitive sequences, and SSR loci, while providing a dual-organellar comparative perspective on RNA editing and codon usage bias. The predicted RNA editing sites in the chloroplast genome far exceeded those in the mitochondrion, contradicting conventional understanding and exhibiting significant differences from the editing characteristics of most moss species. The overall codon usage bias among the organelles was relatively weak and was regulated by multiple evolutionary factors, with mitochondria subjected to stronger constraints from natural selection and chloroplasts more influenced by mutation pressure alongside translational selection, base composition, and protein physicochemical properties. These findings provide crucial theoretical support for further elucidating the post-transcriptional regulatory differences between the chloroplasts and mitochondria of this species and for understanding the evolutionary mechanisms of organellar genomes, while also serving as a reference for optimizing expression systems and producing metabolites with pharmacological implications.

Subject of Research: Mitochondrial genome characteristics and comparative organelle genome analysis of the medicinal moss Rhodobryum laxelimbatum

Article Title: Mitochondrial Genome Characteristics of Rhodobryum laxelimbatum and Comparative Analysis of RNA Editing and Codon Usage Bias With the Chloroplast Genome

Article References: Niu, J., Shi, S., Li, S., Zhao, J., & Li, L. (2026). Mitochondrial Genome Characteristics of Rhodobryum laxelimbatum and Comparative Analysis of RNA Editing and Codon Usage Bias With the Chloroplast Genome. Ecology and Evolution, 16(10), Article e74421. https://doi.org/10.1002/ece3.74421

Image Credits: AI Generated

DOI: 10.1002/ece3.74421

Keywords: Rhodobryum laxelimbatum, mitochondrial genome, chloroplast genome, RNA editing, codon usage bias, moss genomics, organelle evolution, Mitochondrial, Genome, Characteristics, Rhodobryum, laxelimbatum