The effects of rapid mitochondrial gene loss on organellar proteomes

1 2 Mitochondrial genomes retain only a tiny number of genes from their bacterial progenitors, including key 3 components of protein translation machinery. The set of mitochondrially encoded tRNAs and ribosomal 4 subunits is highly variable across angiosperms, with many examples of mitochondrial gene loss, 5 replacement, and/or transfer to the nucleus. This dynamic history suggests large-scale remodeling of 6 mitochondrial translation machinery in some lineages, but such conclusions are largely inferred from 7 genomic sequence and protein targeting predictions. Here, we use proteomic (LC-MS/MS) analysis of 8 purified mitochondria and chloroplasts from angiosperm species with major differences in mitochondrial 9 gene content (Arabidopsis thaliana and Silene conica). Our analysis largely confirms the current 10 understanding of subcellular localization for nuclear-encoded proteins involved in tRNA metabolism and 11 ribosome function in A. thaliana, although some aminoacyl-tRNA synthetases (aaRSs) may have more 12 specialized subcellular roles than previously thought. In contrast, S. conica has undergone extensive 13 mitochondrial gene loss and numerous associated changes in the composition of its mitochondrial 14 proteome, including retargeting of aaRSs, replacement of ribosomal subunits, and loss of the glutamine 15 amidotransferase (GatCAB) complex. Overall, this analysis illustrates how the complex network of 16 molecular interactions necessary for mitochondrial translation are perturbed by gene loss, transfer, and 17 replacement. 18 .CC-BY-NC 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 25, 2025. ; https://doi.org/10.1101/2025.11.24.690252doi: bioRxiv preprint

19 20 Endosymbiotically derived organelles such as mitochondria and plastids evolved from free-living bacteria 21 and still retain their own genomes after more than a billion years (Gould et al. 2008; Roger et al. 2017). 22 Thus, protein synthesis occurs within each of these organelles, using machinery that is distinct from the 23 translation system responsible for synthesizing nuclear-encoded proteins in the cytosol. One common 24 theme in the evolution of obligate endosymbionts is extensive gene loss, functional replacement, and/or 25 transfer to the nucleus (McCutcheon and Moran 2012; Sloan et al. 2018). As such, the gene content in 26 ancient endosymbionts/organelles is whittled down to mostly just key components of the metabolic and 27 biosynthetic functions they provide to the host cell, such as cellular respiration in mitochondria and 28 photosynthesis in plastids. However, even the most ancient endosymbionts/organelles retain some of the 29 genes involved in translation, including ribosomal rRNAs (rRNAs), ribosomal protein subunits, and/or 30 transfer RNAs (tRNAs) (Timmis et al. 2004; Salinas-Giegé et al. 2015; McCutcheon et al. 2024). 31 Mitochondrial gene content has largely stabilized in some eukaryotic lineages. For example, most 32 bilaterian animals retain the same set of 37 genes that were ancestrally present prior to the Cambrian 33 Explosion (Boore 1999). In contrast, mitochondrial gene content is highly dynamic in angiosperms, often 34 differing even among closely related species (Adams, Qiu, et al. 2002). Angiosperm mitochondrial 35 genomes (mitogenomes) can contain anywhere from 19 to 41 protein-coding genes and from 1 to 19 types 36 of tRNA genes (excluding gene duplicates) (Richardson et al. 2013; Skippington et al. 2015; Yu et al. 2025). 37 Multiple evolutionary process can facilitate the loss of genes from the mitogenome. Recent losses 38 of mitochondrial protein-coding genes in angiosperms are typically associated with transfer of those genes 39 to the nucleus. Prior to the loss of the native mitochondrial gene copy, the transferred nuclear copy must 40 be expressed, and the resulting protein must be imported back into the mitochondria (Adams et al. 1999; 41 Sloan et al. 2018). In addition to mitochondrial-to-nuclear gene transfer, there are also cases where 42 mitochondrial protein-coding genes are functionally replaced by homologs of plastid or nuclear origin 43 (Adams, Daley, et al. 2002). Losses of mitochondrial tRNA genes appear to follow this latter route. 44 Specifically, they are replaced by import of existing nuclear-encoded tRNAs from the cytosol (Salinas-45 Giegé et al. 2015; Warren et al. 2021). To our knowledge, there are no documented cases in which a 46 mitochondrial tRNA gene was functionally transferred to the nucleus and targeted back for import into the 47 mitochondria (although one example of plastid tRNA gene transfer was recently discovered in the 48 lycophyte Selaginella; Berrissou et al. 2024). 49 A fundamental challenge in the field of cytonuclear coevolution is to understand the process by 50 which functional gene replacements occur and whether they perturb the intimate and coevolved 51 interactions among gene products encoded in two different genomes. For example, tRNAs must be 52 .CC-BY-NC 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 25, 2025. ; https://doi.org/10.1101/2025.11.24.690252doi: bioRxiv preprint recognized by their cognate aminoacyl-tRNA synthetases (aaRSs) to be charged with the correct amino 53 acid, and ribosomal proteins and rRNAs physically interact with dozens of subunits within a massive 54 enzyme complex. It is remarkable that functional replacement of a (bacterial-like) mitochondrial gene with 55 its (archaeal-like) nuclear counterpart is possible given that their divergence spans the very deepest split in 56 the tree of life and that even small sequence changes have the potential to disrupt these coevolved 57 interactions (Meiklejohn et al. 2013; Sloan et al. 2023). 58 The mitogenomes found in the angiosperm genus Silene are highly variable among species and 59 characterized by many unusual features. Perhaps the most extreme is found in S. conica. It has a highly 60 accelerated mutation rate, the largest size of any known angiosperm mitogenome (>11 Mb), and a 61 fragmented multichromosomal structure (Sloan, Alverson, Chuckalovcak, et al. 2012; Broz et al. 2021). 62 Despite its large size, the S. conica mitogenome has an unusually small gene content with just 25 protein-63 coding, 3 rRNA, and 2 tRNA genes (excluding gene duplicates). The near-complete loss of tRNA genes from 64 this mitogenome has been accompanied by extensive import of cytosolic-like tRNA counterparts (Warren 65 et al. 2021). Our previous analysis based on in silico targeting predictions and fluorescence microscopy 66 indicated that these tRNA replacement events in S. conica have had differing effects on coevolved 67 relationships with aaRSs (Warren et al. 2023). All plant aaRSs are nuclear-encoded, but they differ in where 68 they are localized within the cell. One of the most common patterns is that the plant expresses two aaRSs 69 for a given amino acid – one that functions in the cytosol and another that is dual-targeted and imported 70 into both the mitochondria and plastids (Duchêne et al. 2005). For half of the replaced tRNA genes in S. 71 conica, we found evidence that a corresponding cytosolic aaRS also gained targeting to the mitochondria. 72 Therefore, the ancestral pairing of a cytosolic tRNA and aaRS was apparently maintained in these cases 73 and simply relocated to an additional cellular compartment. In contrast, we did not find evidence of aaRS 74 retargeting for the other half of replaced tRNA genes, suggesting that (bacterial-like) organellar aaRSs are 75 responsible for charging the cytosolic-like tRNAs now being imported into the mitochondria. However, 76 these inferences have not been investigated by direct analysis of mitochondrial protein content, as the 77 mitochondrial proteome of S. conica remains entirely unexplored. 78 Here, we perform proteomic analysis of S. conica and the model angiosperm Arabidopsis thaliana 79 using liquid chromatography-tandem mass spectrometry (LC-MS/MS). The resulting datasets allow us to 80 directly investigate how the composition of the mitochondrial proteome and its translational machinery 81 have changed in association with the extensive gene loss from the S. conica mitogenome. 82 83

and Discussion 84 85 Purification of mitochondrial and chloroplast proteomes 86 .CC-BY-NC 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 25, 2025. ; https://doi.org/10.1101/2025.11.24.690252doi: bioRxiv preprint We analyzed LC-MS/MS data from extracted protein content from two biological replicates of purified 87 mitochondria, purified chloroplasts, and total leaf tissue from both A. thaliana and S. conica. A similar 88 dataset was also generated for a third species (Agrostemma githago). However, preliminary analysis of the 89 A. githago dataset indicated that it had limited detection of the translation machinery that was the focus of 90 this study. Therefore, we have deposited data from all three species (see Data Availability), but only A. 91 thaliana and S. conica were analyzed in the present study. 92 We used peptide-spectrum matches (PSMs) as a semi-quantitative metric of protein abundance. 93 Overall, we detected PSMs from 5599 and 5491different proteins in A. thaliana and S. conica, respectively 94 (Table 1). To verify the efficacy of our organelle purifications, we measured the enrichment of each 95 mitochondrial-encoded and plastid-encoded protein in the purified organelles relative to total leaf tissue 96 based on PSM counts. Although we did detect PSMs for mitochondrial-encoded proteins in chloroplast 97 samples and vice versa, we observed a strong quantitative difference between the sample types that 98 confirmed we had effectively enriched for mitochondrial and chloroplast subcellular fractions (Figure 1). 99 We also found this same pattern of strong enrichment using ion intensity (i.e., peak area) as a measure of 100 protein abundance instead of PSM counts (Figure S1). 101 Figure 1. Enrichment of proteins encoded by the organellar genomes in purified chloroplast fractions (x-axis) and mitochondrial fractions (y-axis). Enrichment ratios are quantified based on summed counts of PSMs from two replicates relative to counts in leaf tissue samples. Each point represents a protein, and point sizes are scaled proportionally to the total number of PSMs observed for the corresponding protein across all samples. Plastid-encoded proteins and mitochondrial-encoded proteins are represented by green circles and gold triangles, respectively. Subcellular specialization of Arabidopsis aaRSs 102 Numerous studies have been conducted to catalog the subcellular localization of aaRSs in A. thaliana, 103 primarily by fusing putative organelle-targeting transit peptides from aaRSs to fluorescent proteins or other 104 reporters, which could be visualized within cells or tested for import into isolated organelles in vitro 105 (Mireau et al. 1996; Uwer et al. 1998; Souciet et al. 1999; Peeters et al. 2000; Duchêne et al. 2001; Duchêne 106 Arabidopsis Silene −1.0 −0.5 0.0 0.5 1.0 1.5 −1.0 −0.5 0.0 0.5 1.0 1.5 −1 0 1 log10 (Chloroplast / Total Leaf PSM Ratio) log10 (Mitochondrial / Total Leaf PSM Ratio) Genome Chloroplast Mitochondrial .CC-BY-NC 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 25, 2025. ; https://doi.org/10.1101/2025.11.24.690252doi: bioRxiv preprint et al. 2005). Collectively, these studies established a consensus for the subcellular localization of all 107 aaRSs in A. thaliana (Table 2), which we will refer to as the Duchêne classification because the most 108 extensive sampling was performed by Duchêne et al. (2005), and the cumulative body of work was later 109 summarized by Duchêne et al. (2009). For most amino acids, A. thaliana expresses two different aaRSs – 110 one that is targeted to the cytosol (cyto-only) and another that is dual-targeted to both the chloroplasts 111 and mitochondria (chloro-mito). However, some aaRSs exhibit atypical patterns of subcellular 112 localization, such as dual targeting to both the cytosol and mitochondria (cyto-mito) or to all three cellular 113 compartments (Duchêne et al. 2009). One LeuRS enzyme (AT4G04350) was also found to be specific to the 114 chloroplasts (chloro-only). 115 Overall, our patterns of aaRS enrichment in mitochondrial and chloroplast fractions strongly 116 aligned with the Duchêne classification (Figure 2; Table 2). All aaRSs that had been previously described as 117 cyto-only showed little or no detection in our purified mitochondrial and chloroplasts samples from A. 118 thaliana. In contrast, all of these aaRSs were detected in total leaf tissue (except for four that presumably 119 had low overall expression due to the presence of close paralogs with higher expression; Table 2), 120 supporting the inference that their functional role is limited to the cytosol. Likewise, the four aaRSs 121 previously identified cyto-mito aaRSs (GlyRS, LeuRS, ValRS, ThrRS) were found at higher abundance in 122 mitochondria than all the cyto-only aaRSs and were not detected in chloroplasts. As expected, previously 123 identified chloro-mito aaRSs were generally detected at high abundance in both organelle fractions, and 124 the chloro-only LeuRS was found in high abundance in our chloroplast fractions but not detected in 125 mitochondria (Figure 2; Table 2). In addition, our data support previous inferences that one ArgRS enzyme 126 (AT4G26300) functions in all three cellular compartments. In prior GFP localization studies, this ArgRS 127 protein was only observed to be targeted to the chloroplasts (Duchêne et al. 2005), but it was predicted to 128 function in the cytosol and mitochondria as well because mutants lacking the only other known A. thaliana 129 ArgRS gene (AT1G66530) are still viable (Berg et al. 2005; Duchêne et al. 2009). Indeed, we detected 130 AT4G26300 in both mitochondrial and chloroplast fractions (Figure 2). In contrast, AT1G66530 was only 131 detected in total leaf samples and at very low abundance (only a single PSM in each replicate sample). 132 Therefore, it is likely that both ArgRS enzymes are expressed in the cytosol, but whether AT1G66530 makes 133 any contribution to cytosolic translation is not clear. 134 Despite the overall congruence between our proteomic analysis and the Duchêne classification, 135 there were some discrepancies that point to potential refinements in our current understanding of aaRS 136 targeting in A. thaliana (Table 2). First, a previous analysis has indicated that there is one chloro-mito AlaRS 137 (AT5G22800) and another AlaRS (AT1G50200) targeted to all three cellular compartments (Duchêne et al. 138 2005). However, we did not detect AT1G50200 in purified chloroplasts, and we did not detect AT5G22800 139 in purified mitochondria (Figure 2). Therefore, it is possible that AlaRSs follow a division of labor similar to 140 .CC-BY-NC 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 25, 2025. ; https://doi.org/10.1101/2025.11.24.690252doi: bioRxiv preprint the pattern observed for LeuRSs, with distinct cyto-mito and chloro-only enzymes. This scenario would be 141 consistent with the finding that plastid tRNA-Ala is a poor substrate for the cytosolic AlaRS in spinach 142 (Steinmetz and Weil 1986) and that the fusion of the AT1G50200 transit peptide to b-glucuronidase (GUS) 143 failed to localize reporter activity to chloroplasts (Mireau et al. 1996). However, it is difficult to reconcile 144 with the lines of evidence supporting dual-organellar localization of both AlaRSs reported by Duchêne et 145 al. (2005), so it is possible that AT1G50200 and AT5G22800 are indeed present in chloroplasts and 146 mitochondria, respectively, but at levels we were unable to detect with our analysis. 147 Second, the ThrRS enzymes might offer yet another example of division of labor between cyto-mito 148 and chloro-only enzymes in A. thaliana, even though previous analysis has suggested two different ThrRSs 149 functioning in the mitochondria (one chloro-mito and one cyto-mito). Fusion of the transit peptide from the 150 putative chloro-mito ThrRS (AT2G04842) to reporter genes was previously shown to drive localization and 151 import into both chloroplasts and mitochondria (Duchêne et al. 2005), but we only detected this ThrRS in 152 chloroplasts. As expected, we did not detect the previously classified cyto-mito ThrRS (AT5G26830) in 153 chloroplasts. A third ThrRS gene (AT1G17960) has been identified in A. thaliana and presumed to have 154 cyto-only function due to the apparent lack of any transit peptide (Duchêne et al. 2005), but we did not 155 detect this protein in any of our samples. 156 Third, our results may provide insights into the unusual cases of GlyRS and ValRS function in A. 157 thaliana. Although there is a cyto-mito and a chloro-mito aaRS reported in each case, the cyto-mito aaRS 158 has been suggested to not function in aminoacylation within the mitochondria for both Gly (Duchêne et al. 159 2001) and Val according to a personal communication reported by Duchêne et al (2009). We detected the 160 chloro-mito GlyRS (AT3G48110) at substantial abundance in mitochondria (Figure 2), supporting the earlier 161 interpretation that it could be the sole or primary enzyme responsible for aminoacylation of mitochondrial 162 tRNA-Gly even though the cyto-mito GlyRS (AT1G29880) is also present in the mitochondrial fraction. In 163 contrast, we detected the putative chloro-mito ValRS (AT5G16715) at only very low abundance in A. 164 thaliana mitochondria (only a single PSM in one of the two replicate samples; Figure 2). Coupled with the 165 fact that there has been conflicting or inconsistent evidence for targeting of AT5G16715 to mitochondria 166 (Duchêne et al. 2005; Duchêne et al. 2009), this observation raises the possibility that ValRS represents a 167 fourth example (along with AlaRS, LeuRS, and ThrRS) where the division of labor is primarily between a 168 cyto-mito and a chloro-only enzyme. 169 .CC-BY-NC 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 25, 2025. ; https://doi.org/10.1101/2025.11.24.690252doi: bioRxiv preprint Figure 2. Summary of mitochondrial and chloroplast enrichment of Arabidopsis thaliana aaRSs relative to total leaf samples based on ratios of PSM counts combined across two biological replicates. Color coding of points reflects whether the aaRS was previously classified as being targeted to the cytosol, chloroplasts, and/or mitochondria (Duchêne et al. 2005; Duchêne et al. 2009). Changes in aaRS targeting and tRNA interactions associated with massive mitochondrial tRNA gene loss in 170 Silene conica 171 With only two tRNA genes (tRNA-Ile and tRNA-fMet), the S. conica mitogenome represents one of the most 172 extreme cases of mitochondrial tRNA gene loss in plants (Sloan, Alverson, Chuckalovcak, et al. 2012), with 173 potential widespread effects on aaRS-tRNA interactions. Only six of the 20 aaRS types are not expected to 174 be affected by recent mitochondrial tRNA gene loss in the Silene lineage – either because S. conica has 175 retained the corresponding tRNA gene in its mitochondrial genome (IleRS) or because the corresponding 176 tRNA gene had already been lost prior to the most recent common ancestor of angiosperms (AlaRS, ArgRS, 177 LeuRS, ThrRS, and ValRS). Note that even though the S. conica mitogenome retains a gene for the initiator 178 tRNA-fMet, it has lost the elongator tRNA-Met gene, so interactions involving MetRS may still have been 179 perturbed. For four of these six aaRS types (AlaRS, ArgRS, IleRS, and LeuRS), mitochondria from A. thaliana 180 and S. conica exhibit the same bias with respect to cytosolic-like vs. organellar-like aaRSs (Figure 3). 181 However, the bias towards cytosolic-like ThrRS and ValRS in A. thaliana mitochondria (see preceding 182 section) is not shared by S. conica. In these cases, it is likely that S. conica has retained the ancestral state 183 (an organellar-like aaRS in the mitochondria) and that the Arabidopsis lineage has evolved to import a 184 cytosolic-like enzyme. 185 ThrRS AlaRS AlaRS ArgRS ArgRS AsnRS AsnRS AspRS AspRS AspRS CysRS CysRS GlnRS GluRS GluRS GlyRS GlyRS HisRS HisRS IleRS IleRS LeuRS LeuRS LysRS LysRS MetRS MetRS PheRS PheRS PheRS ProRS ProRS SerRS SerRS ThrRS TrpRS TrpRS TyrRS TyrRS ValRS ValRS 0 0.3 1 3 0 0.3 1 3 Chloroplast / Total Leaf PSM Ratio Mitochondrial / Total Leaf PSM Ratio Cytosolic Chloroplast Mitochondrial Duchêne et al. Classification .CC-BY-NC 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 25, 2025. ; https://doi.org/10.1101/2025.11.24.690252doi: bioRxiv preprint Figure 3. Summary of whether mitochondrial samples were biased towards containing organellar-like vs. cytosolic-like aaRSs based on proteomic analysis. The mitochondrial enrichment bias metric was calculated as described in the Methods. Note that there is no organellar-like GlnRS because mitochondria and plastids use GluRS and the indirect aminoacylation pathway for tRNA- Gln (Pujol et al. 2008). Therefore, enrichment bias for GlnRS was simply reported as “organellar” if the cytosolic-like GlnRS was not detected in mitochondria and cytosolic if the cytosolic-like GlnRS was detected in mitochondria. Overall, the results strongly align with previous predictions (Warren et al. 2023) for which cytosolic aaRSs were or were not retargeted to mitochondria with the exception of SerRS, which showed clear evidence for retargeting despite the lack of any predicted transit peptide. The other 14 aaRS types are potentially affected by the large-scale loss/replacement of 186 mitochondrial tRNA genes in S. conica. Previous analysis based on in silico prediction and fusions of 187 putative transit peptides to GFP (Warren et al. 2023) found that a cytosolic-like aaRS likely gained targeting 188 to the mitochondria in seven of these 14 cases (GlnRS, GlyRS, LysRS, MetRS, ProRS, TrpRS, and TyrRS) 189 where a native mitochondrial tRNA gene was replaced by import of its cytosolic counterpart. Proteomic 190 data supported these predictions in all seven cases (Figure 3), indicating that the ancestral pairing of 191 cytosolic-like tRNAs and aaRSs has been preserved and simply retargeted to also function in the 192 mitochondria. Notably, cytosolic-like and organellar-like MetRS enzymes were both detected in S. conica 193 mitochondria (Figure 3), likely reflecting the contrasting pattern of loss/replacement of the elongator tRNA-194 Met gene but retention of the initiator tRNA-fMet gene in the mitogenome (Sloan, Alverson, Chuckalovcak, 195 et al. 2012). 196 TyrRS TrpRS ProRS MetRS LysRS GlyRS GlnRS SerRS PheRS HisRS GluRS CysRS AspRS AsnRS IleRS ArgRS ValRS ThrRS LeuRS AlaRS Arabidopsis Silene 0.00 0.25 0.50 0.75 1.00 aaRS Enrichment Bias in Mitochondria No Change to Silene mt-tRNAsLoss of Silene mt-tRNAs No Predicted Retargeting of Cytosolic aaRS (Warren et al. 2023) Predicted Retargeting of Cytosolic aaRS (Warren et al. 2023) Organellar Cytosolic Enrichment Bias .CC-BY-NC 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 25, 2025. ; https://doi.org/10.1101/2025.11.24.690252doi: bioRxiv preprint The remaining seven aaRS types (AsnRS, AspRS, CysRS, GluRS, HisRS, PheRS, and SerRS) were not 197 previously predicted to show retargeting of the cytosolic-like aaRS in S. conica (Warren et al. 2023). Our 198 analysis here generally supported these predictions. However, SerRS was a notable exception, showing 199 clear evidence of retargeting of the cytosolic-like SerRS to the mitochondria (Figure 3). This finding resolves 200 a mystery because previous transit peptide analysis had suggested that the ancestral organellar-like SerRS 201 had lost targeting to mitochondria and was localized exclusively to chloroplasts in the Silene lineage, 202 making it unclear which enzyme was providing SerRS function in mitochondria (Warren et al. 2023). Our 203 proteomic results indicate that SerRS represents an eighth case of retargeting a cytosolic-like aaRS to the 204 mitochondria. Our analysis also detected a mix of both organellar-like and cytosolic-like GluRS proteins in 205 S. conica mitochondria (Figure 3). It is possible that the cytosolic-like protein is a truncated version of the 206 enzyme that was previously predicted based on full-length mRNA sequencing and in silico targeting 207 analysis (Warren et al. 2023). If so, it is not clear whether this protein would have a functional role in 208 aminoacylation in the mitochondria given that it lacks a substantial N-terminal portion (129 amino acids) 209 of the enzyme body. 210 211 Duplication and subfunctionalization of the organellar PheRS in Silene conica 212 PheRS is one of the aaRS types that shows no evidence of cytosolic enzyme retargeting despite the loss of 213 the mitochondrial tRNA-Phe and apparent functional replacement with its cytosolic counterpart (Figure 3) 214 (Warren et al. 2023). The persistence of the organellar-like PheRS in S. conica mitochondria follows a 215 broader evolutionary pattern, as this aaRS appears to be one of the very last to be functionally replaced in 216 the mitochondria even in cases of complete loss of all tRNA genes from the mitogenome (Pett and Lavrov 217 2015; DeTar et al. 2024). The recalcitrance of the organellar-like PheRS is thought to result from its 218 cytosolic counterpart being divided into two subunits, making it improbable for both subunits to 219 independently gain import into the mitochondria and serve as a viable functional replacement. In the 220 Silene lineage, the organellar-like PheRS gene has been duplicated, and fusion of putative transit peptides 221 to GFP suggested that the paralogs have subfunctionalized with one specializing on the mitochondria and 222 the other specializing on the chloroplasts (Warren et al. 2023). Our proteomic analysis supported this 223 subfunctionalization model, as the S. conica chloroplast fraction contained only the putative chloroplast 224 specialist, whereas the mitochondrial fraction was dominated by the putative mitochondrial specialist 225 (Figure 4A). 226 The organellar (mitochondrial) PheRS in humans has been found to have the capacity to charge 227 tRNA-Phe substrates from diverse organisms (Klipcan et al. 2012). If plant organellar PheRSs share this 228 capacity, it may have predisposed them to charge imported cytosolic tRNA-Phe in the mitochondria of S. 229 conica. In addition, the division of labor observed between the two organellar-like PheRSs in S. conica 230 .CC-BY-NC 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 25, 2025. ; https://doi.org/10.1101/2025.11.24.690252doi: bioRxiv preprint could be a response to the challenges associated with a single organellar-like aaRS having to recognize 231 both a cyanobacterial-like tRNA-Phe in chloroplasts and a cytosolic-like tRNA-Phe imported into the 232 mitochondria. 233 To identify amino acid substitutions that might potentially have improved the ability of the S. conica 234 mitochondrial PheRS to charge cytosolic tRNA-Phe, we compared the S. conica organellar PheRS 235 sequences to each other and to orthologs from a broad sampling of species that represented angiosperms 236 (including eudicots, monocots, magnoliids, and Amborella), gymnosperms (Ginkgo), bryophytes 237 (Physcomitrium), and green algae (Micromonas and Ostreococcus). The enzyme bodies (i.e., after removal 238 of putative transit peptide sequences) of the S. conica mitochondrial and chloroplast PheRSs differed by 239 39 amino acid substitutions (out of 374 positions). Four of the substitutions in the mitochondrial PheRS 240 (I184V, S343T, A363G, and R411I) were at sites that were otherwise universally conserved across our 241 sampling of plant and algal taxa (Figure S2A). No substitutions were found at any such conserved sites in 242 the S. conica chloroplast PheRS sequence. 243 Figure 4. Duplication of organellar PheRS has led to subfunctionalized copies with specialized chloroplast and mitochondrial localization. (A) PSM counts in Silene conica samples (with A and B replicates shown separately for each sample type) indicate that the putative chloroplast PheRS was the only one present in the chloroplast fraction, whereas the putative mitochondrial type dominated the mitochondrial fraction, aligning with previous predictions (Warren et al. 2023). (B) AlphaFold3 model of the structural interaction between S. conica mitochondrial PheRS and cytosolic tRNA-Phe. Four substitutions at PheRS residues that are otherwise broadly conserved across green plants are labeled and highlighted with “sphere” representation. Two of those substitutions (I184V) and (R411I) are emphasized with a darker color because the exact same substitutions evolved in parallel in another angiosperm (Sapria himalayana) with an organellar-like PheRS that is expected to function exclusively in the mitochondria and charge imported cytosolic tRNA-Phe. To further explore the potential relevance of these four substitutions, we compared the PheRS 244 sequence to a representative of the Rafflesiaceae (Sapria himalayana) because this family of parasitic 245 (non-photosynthetic) plants has also lost the native tRNA-Phe gene from its mitogenome, and it has lost its 246 plastid genome entirely (Molina et al. 2014; Smith and Asmail 2014; DeTar et al. 2024). Therefore, its 247 Total Leaf Isolated Chloroplasts Isolated Mitochondria A B A B A B0 5 10 15 20 25 Sample # PSMs Cytosolic Chloroplast Mitochondrial Silene PheRS TypeA B Cytosolic tRNA-Phe Anticodon Loop Acceptor Stem Mitochondrial PheRS Catalytic Domain Anticodon Binding Domain I184V R411I A363G S343T .CC-BY-NC 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 25, 2025. ; https://doi.org/10.1101/2025.11.24.690252doi: bioRxiv preprint organellar PheRS is expected to function exclusively in the mitochondria (due to lack of translation in the 248 plastid) and likely charges imported cytosolic tRNA-Phe. Strikingly, the S. himalayana PheRS has 249 independently evolved the exact same amino acid substitutions at two of the four sites identified in the S. 250 conica mitochondrial PheRS (I184V and R411I; Figure S2A). The R411I substitution is at a residue predicted 251 to interact with the G37 position in cytosolic tRNA-Phe (Figure 4B), which is situated immediately adjacent 252 to the anticodon and represents the only sequence difference in the anticodon loop between cytosolic and 253 mitochondrial tRNA-Phe (Figure S2B). The I184V substitution is at a position abutting the 3¢ side of the 254 tRNA-Phe acceptor stem (Figure 4B), which contains multiple nucleotide substitutions that distinguish 255 cytosolic and mitochondrial tRNA-Phe (Figure S2B). We also compared to the organellar PheRS from 256 another parasitic plant (Balanophora fungosa) that has lost the mitochondrial tRNA-Phe gene. The B. 257 fungosa organellar PheRS is expected to function in both the mitochondria and plastids because the 258 Balanophoraceae lineage still retains a plastid genome (Su et al. 2019; Ceriotti et al. 2021). However, this 259 plastid genome has also lost its copy of the tRNA-Phe gene, so the B. fungosa organellar PheRS is expected 260 to exclusively charge cytosolic tRNA-Phe. We found that B. fungosa PheRS also carried the R411I 261 substitution. It also had a substitution at residue 184, although it was an Ile-to-Thr change rather than the 262 Ile-to-Val substitution observed in S. conica and S. himalayana. 263 Given the recurrence of substitutions in S. conica, S. himalayana, and B. fungosa at these two sites 264 that are otherwise highly conserved in green plants, we speculate that they are involved in the 265 specialization of these enzymes to charging an imported cytosolic tRNA-Phe substrate. In the S. conica 266 lineage, it appears that these substitutions occurred prior to the divergence of multiple Silene species that 267 were previously sampled (Warren et al. 2023). However, another member of this family (Agrostemma 268 githago) shares the organellar PheRS paralogs with Silene (Warren et al. 2023), but it does not share these 269 two amino acid substitutions. Because A. githago independently lost the mitochondrial tRNA-Phe gene and 270 must use its imported cytosolic counterpart (Warren et al. 2021), the lack of these two substitutions in the 271 A. githago mitochondrial PheRS implies that they are not essential for charging cytosolic tRNA-Phe. 272 Likewise, it is possible that other amino acid substitutions distinguishing the S. conica mitochondrial and 273 plastid PheRSs might contribute to their specialization on different tRNA substrates even if they are not at 274 positions that are widely conserved in other taxa. Furthermore, it is possible this gene duplication has 275 occurred for reasons that are entirely unrelated to changes in tRNA substrates. Performing aminoacylation 276 assays with recombinantly expressed proteins (Gamper and Hou 2020) would be a promising route to 277 assess the effect of specific substitutions on charging efficiency with different tRNA substrates and to 278 distinguish among these alternative hypotheses. 279 280 Loss of the GatCAB complex in Silene conica mitochondria 281 .CC-BY-NC 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 25, 2025. ; https://doi.org/10.1101/2025.11.24.690252doi: bioRxiv preprint Like most bacteria, plant mitochondria and plastids generally lack GlnRS activity and instead use an 282 indirect pathway in which GluRS indiscriminately charges tRNA-Gln with Glu followed by enzymatic 283 conversion of Glu to Gln by the glutamyl-tRNA amidotransferase enzyme complex (GatCAB) (Pujol et al. 284 2008). In S. conica, the loss of the mitochondrial tRNA-Gln gene has been associated with import of 285 cytosolic-like tRNA-Gln and GlnRS into the mitochondria (Figure 2). Therefore, we would predict that 286 GatCAB activity is no longer necessary in S. conica mitochondria because they now use the direct Gln 287 aminoacylation pathway typical of cytosolic translation (Rogers and Söll 1995). Our proteomic analysis 288 supported this prediction. All three GatCAB subunits were detected in both the mitochondria and 289 chloroplast fractions from A. thaliana but only in the chloroplast fraction from S. conica (Table S1). 290 Therefore, the GatCAB complex and the indirect Gln aminoacylation pathway appears to no longer 291 function in S. conica mitochondria. 292 293 Retention of other components of tRNA metabolism machinery in Silene conica mitochondria 294 The two remaining tRNA genes in the S. conica mitogenome (tRNA-fMet and tRNA-Ile) are noteworthy 295 because they both have distinctive bacterial-like features. Bacterial translation is initiated with an N-296 formylmethionine (fMet), which is synthesized by the methionyl-tRNA formyltransferase (MTF). After tRNA-297 fMet is initially charged with Met, the MTF enzyme formylates the amino group of the Met residue to 298 produce fMet (Ibba and Söll 2004). Meanwhile, a class of bacterial tRNA-Ile genes have a CAT anticodon, 299 which would typically correspond to an ATG (Met) codon. However, the C base in this anticodon is 300 modified to lysidine by tRNA-Ile lysidine synthetase (TilS), which results in decoding of Ile codons (Suzuki 301 and Miyauchi 2010). Homologs of bacterial MTF and TilS are both found in plant nuclear genomes and 302 expected to function in mitochondrial and plastid translation systems (Warren and Sloan 2020). Given the 303 retention of tRNA-fMet and tRNA-Ile genes in the S. conica mitogenome, we would predict that MTF and 304 TilS are also functional in S. conica mitochondria. Accordingly, we detected both proteins in the 305 mitochondrial fraction (Table S1). TilS was only supported by a single PSM in one of the S. conica 306 mitochondrial samples, which is likely due to low overall expression, as it was not detected in chloroplast 307 or total-leaf samples from S. conica or in any A. thaliana samples. Nonetheless, the detection of both of 308 these enzymes suggests that the bacterial-like translation features associated with tRNA-fMet and tRNA-309 Ile have been retained in S. conica mitochondria, which contrasts with the functional loss of GatCAB and 310 many aaRSs. 311 Mitochondrial tRNA metabolism also relies on additional enzymes that are typically shared with 312 other subcellular compartments (von Braun et al. 2007; Canino et al. 2009; Gobert et al. 2010; Gutmann et 313 al. 2012; Salinas-Giegé et al. 2015), including those responsible for cleaving primary transcripts to remove 314 5¢ ends (protein-only RNase P [PRORP]) or 3¢ ends (tRNase Z) and for adding a 3¢ CCA tail (CCAse). In 315 .CC-BY-NC 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 25, 2025. ; https://doi.org/10.1101/2025.11.24.690252doi: bioRxiv preprint general, we had limited sensitivity to detect these enzymes with our dataset (Table S1). Although we did 316 confirm the expected presence of CCAse in S. conica mitochondria, the general lack of signal for PRORP or 317 tRNase Z enzymes across all samples precludes any interpretation of whether there has been retargeting 318 of these enzymes associated with loss of mitochondrial tRNA genes. 319 320 Mitochondrial ribosomal subunit gene loss, transfer, and replacement in Silene conica 321 The S. conica mitogenome has lost most of the genes that encode ribosomal protein subunits, retaining 322 only three (rpl5, rps3, and rps13) of the 15 that were present in most recent common ancestor of 323 angiosperms (Adams, Qiu, et al. 2002; Kubo and Arimura 2010; Sloan, Alverson, Chuckalovcak, et al. 324 2012). Although rps3 was originally annotated as a pseudogene in the S. conica mitogenome due to the 325 apparent loss of the first exon and the presence of large indels (Sloan, Alverson, Chuckalovcak, et al. 326 2012), we detected Rps3 peptides via LC-MS/MS, indicating that it remains an expressed, functional gene. 327 More generally, our proteomic dataset provides insights into how the extensive gene loss from the 328 mitogenome occurred without disrupting function of the mitochondrial ribosome (mitoribosome). Our 329

point to diverse mechanisms by which these genes were replaced (Figure 5; Table 3). 330 Figure 5. Structure of A. thaliana mitoribosome (PDB accession 6XYW; Waltz et al. 2020). Ribosomal proteins are shown with surface renderings, and rRNAs are shown in blue with cartoon renderings. Subunits that were encoded by genes in the ancestral angiosperm mitogenome are highlighted and colored according to their evolutionary history in S. conica. Note that the Rps1 subunit is not pictured because it has been lost in A. thaliana (Skaltsogiannis et al. 2024) even though the gene is retained in most angiosperms, including the copy that has been transferred to the nucleus in S. conica (Table 3). Nine of the twelve genes that have been lost (rpl2, rps1, rps2, rps4, rps10, rps11, rps12, rps14, and 331 rps19) have followed the typical route of intracellular gene transfer from the mitogenome to the nucleus 332 (Adams and Palmer 2003). In two of these cases (rpl2 and rps4), the transfer appears to have occurred in 333 two pieces, resulting in different nuclear genes encoding separate, non-overlapping portions of each 334 subunit (Table 3). 335 Two more genes absent from the S. conica mitogenome (rpl10 and rpl16) also appear to have been 336 functionally replaced through import of a nuclear-encoded protein. However, in these cases, the nuclear 337 .CC-BY-NC 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 25, 2025. ; https://doi.org/10.1101/2025.11.24.690252doi: bioRxiv preprint gene is plastid-like rather than an intracellular transfer of the mitochondrial gene itself. The nuclear gene 338 encoding the plastid-targeted Rpl10 subunit in S. conica has been duplicated, with one copy now targeted 339 exclusively to the mitochondria based on our LC-MS/MS data. Although many angiosperms retain the rpl10 340 gene in the mitogenome, similar duplications and neofunctionalization of plastid-targeted homologs were 341 previously identified in monocots and in the Brassicaceae (Mower and Bonen 2009; Kubo and Arimura 342 2010). In the case of rpl16, the S. conica nuclear genome contains a plastid-like gene copy encoding a 343 protein that we exclusively detected in the mitochondrial fraction, and the S. conica plastid genome still 344 contains a typical rpl16 gene. As such, it appears that a copy of the plastid rpl16 gene was transferred to 345 the nucleus but gained targeting to the mitochondria, facilitating the loss of its mitochondrial homolog. 346 This scenario differs from the case of rpl10 because there is no indication that the transferred rpl16 gene 347 was ever targeted back to the plastid. Therefore, this evolutionary transition may have required overcoming 348 multiple barriers almost simultaneously. Specifically, the transferred nuclear gene would have had to gain 349 expression and targeting to the mitochondria but also adapt to function in the novel context of the 350 mitoribosome. 351 The one other ribosomal gene that has been lost from the S. conica mitogenome (rps7) has no 352 detectable homolog in the nuclear genome other than the distantly related family member that encodes a 353 subunit of the cytosolic ribosome (RPS5; Sconica_v3_48856-RA). The plastid rps7 homolog is still retained 354 in the plastid genome itself. Across angiosperm diversity, the rps7 gene has been subject to an unusually 355 large number of losses from the mitogenome (Adams, Qiu, et al. 2002). Although there are cases where 356 loss of the mitochondrial copy of rps7 appears to have been accompanied by transfer to the nucleus (Liu et 357 al. 2009), outright loss of the gene may also be common (Adams, Qiu, et al. 2002). The S. conica cytosolic-358 like homolog (Sconica_v3_48856-RA) was detectable in all three sample types. However, proteomic 359 detection in the mitochondrial fraction should not be taken as strong evidence for function within the 360 mitochondria given the extremely high abundance of cytosolic ribosomes and the propensity for these 361 ribosomes to adhere to outer mitochondrial membranes (Chang et al. 2024; Dimnet et al. 2024). Therefore, 362 it is not clear if or how the Rps7 subunit has been replaced in the mitoribosome. 363 Even in angiosperms that have retained a larger number of ribosomal genes in their mitogenomes 364 than S. conica, the majority of mitoribosome subunits are encoded by nuclear genes. The composition of 365 the A. thaliana mitoribosome has been thoroughly characterized (Rugen et al. 2019; Waltz et al. 2019; 366 Waltz et al. 2020; Skaltsogiannis et al. 2024), so we used proteomic data from the S. conica mitochondrial 367 fraction to infer whether these subunits are conserved in the S. conica mitoribosome. Of the 70 nuclear-368 encoded subunits of the A. thaliana mitoribosome (Table S2), only bL32m (AT1G26740) lacks a detectable 369 ortholog among the annotated S. conica protein set, and this appears to be an annotation issue because 370 homologous sequence is detectable with a TBLASTN search against the S. conica nuclear genome. Only 371 .CC-BY-NC 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 25, 2025. ; https://doi.org/10.1101/2025.11.24.690252doi: bioRxiv preprint three of the annotated proteins (bS21m, mS38, and bTHXm) were not detected in any of our samples, and 372 every one of the proteins that was detected had higher PSM counts in the mitochondrial fraction than in 373 chloroplast or total leaf samples. Indeed, more than 90% of these proteins were exclusively found in the 374 mitochondrial fraction (Table S2). Therefore, orthologous gene content and mitochondrial proteomes 375 suggest a high degree of stability in the ancestral nuclear-encoded components of the S. conica 376 mitoribosome, although confirming that they are assembled as part of the ribosome complex would 377 require purification and structural analysis of the mitoribosomes themselves. 378 379 380

381 382 The ongoing loss of angiosperm mitochondrial genes has been well characterized for more than two 383 decades, with especially pronounced effects on the genes encoding translational machinery such as 384 tRNAs and ribosomal proteins (Adams, Qiu, et al. 2002; Richardson et al. 2013). However, direct proteomic 385 analysis of mitochondrial translation machinery has been lacking in species such as S. conica that show 386 extreme reductions in mitochondrial gene content. Our study details the extensive changes that 387 accompany mitochondrial gene loss, highlighting alternative evolutionary pathways for functionally 388 replacing genes and responding to perturbations in the network of tRNA-interacting enzymes. 389 For example, our work confirms that numerous cytosolic-like aaRSs have been retargeted to the 390 mitochondria in S. conica (Figure 3), thereby preserving ancestral charging relationships with cytosolic 391 tRNAs that are newly imported into the mitochondria (Warren et al. 2023). We found that this set of 392 retargeted cytosolic-like aaRSs includes SerRS even though in silico predictions suggested that it has not 393 gained an N-terminal extension that could serve as a mitochondrially targeting transit peptide (Warren et 394 al. 2023). This finding illustrates why direct proteomic analysis is an important complement to in silico 395 predictions and assays based on fusing reporters to N-terminal peptides (Mireau et al. 1996; Uwer et al. 396 1998; Souciet et al. 1999; Peeters et al. 2000; Duchêne et al. 2001; Duchêne et al. 2005; Warren et al. 397 2021). On the other hand, our work also shows that many cytosolic-like aaRSs were not retargeted to the 398 mitochondria despite import of their cognate cytosolic tRNAs. Therefore, organellar-like aaRSs are 399 presumably charging novel substrates (cytosolic-like tRNAs) in these cases. 400 Likewise, functional replacement of mitochondrial ribosomal protein genes appears to have 401 followed multiple pathways (Figure 5), including relocation of the gene to the nucleus, replacement by a 402 plastid-like counterpart, or (in the case of rps7) outright loss or replacement with a subunit lacking 403 detectable homology. Overall, the recent evolutionary changes in S. conica highlight that the composition 404 .CC-BY-NC 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 25, 2025. ; https://doi.org/10.1101/2025.11.24.690252doi: bioRxiv preprint of plant mitochondrial translation machinery is still highly dynamic despite billions of years since the 405 establishment of mitochondria in the eukaryotic lineage. 406 407 408

409 410 Plant growth and organelle isolations 411 Arabidopsis thaliana Col-0 seeds were stratified at 4 °C in water for 3 days prior to sowing in 3-inch pots 412 with 5 seeds per pot. Agrostemma githago KEW0053084 (Warren et al. 2023) and S. conica ABR (Fields et 413 al. 2023) were sown in 4-inch pots with 6 seeds per pot. Pots contained Pro-Mix BX potting media and were 414 covered with clear plastic domes for ~1 week until seedlings emerged. All plants were grown on shelves 415 with fluorescent lighting (~100 µE m-2 sec-1) at ~22-23 °C under short-day conditions (10-hr light / 14-hr 416 dark). Tissue for total leaf protein samples for all species was harvested at ~12 weeks after sowing. 417 Mitochondrial and chloroplast isolations were performed with tissue harvested either ~9 weeks after 418 germination (A. thaliana and A. githago) or ~12 to 13 weeks after germination (S. conica). 419 Two biological replicates were performed for each species. Mitochondrial and chloroplast 420 fractions were isolated with Percoll gradients. Mitochondrial isolations were performed as described 421 previously (Warren et al. 2021) based on a modified protocol from Meyer et al. (2009), using ~70 g of 422 rosette leaf tissue per replicate. Chloroplast isolations were performed using a modified version of 423 published protocols (van Wijk et al. 2007; Kley et al. 2010). Briefly, ~2 g of rosette leaf tissue was ground in 424 a prechilled mortar with 20 ml of ice-cold chloroplast grinding and wash buffer (cpGW: 50mM HEPES pH 425 7.5, 5 mM EDTA, 0.3 M sorbitol, 10 mM NaHCO3, 0.5 mM DTT), filtered through miracloth and centrifuged 426 at 1,300 rcf for 5 min at 4 °C. The resulting chloroplast pellet was resuspended in ~2 ml cpGW, applied to a 427 Percoll step gradient (40-80%) and centrifuged at 2,500 rcf in a swinging bucket rotor for 10 min at 4 °C. The 428 chloroplast band at the 40-80 interface was removed and washed three times with cpGW buffer. 429 Chloroplasts were resuspended in 1 ml of cpGW buffer containing 1X Halt protease inhibitor (Thermo 430 Scientific 78430), aliquoted and centrifuged at 1,000 rcf for 5 min at 4 °C. Detailed versions of the isolation 431 protocols are available via GitHub (https://github.com/dbsloan/silene_proteomics). Whole leaf tissue and 432 isolated mitochondrial/chloroplast pellets were flash frozen in liquid N2 and stored at -80 °C before 433 shipment to the Proteome Exploration Laboratory at the California Institute of Technology for protein 434 extraction, digestion, and LC-MS/MS analysis. 435 436 Protein extraction and digestion 437 .CC-BY-NC 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 25, 2025. ; https://doi.org/10.1101/2025.11.24.690252doi: bioRxiv preprint For leaf and purified chloroplast samples, 100 µl of lysis buffer from a PreOmics iST Kit was added and 438 followed by processing with a PreOmics BeatBox Tissue Homogenizer (Preomics, Germany) for 10 min on 439 the high setting. Samples were then centrifuged at 21,000 rcf for 2 min to remove the insoluble fraction. 440 Protein concentration was evaluated using Pierce BCA Protein Assay Kit. Aliquots containing 100 µg of 441 protein from each sample were digested with ProtiFi S-trap according to the manufacturer’s protocol. 442 Briefly, the protein was reduced and alkylated with tris(2-carboxyethyl)phosphine and chloroacetamide 443 and digested overnight with trypsin. Following elution, the resulting peptides were dried and resuspended, 444 using 2% acetonitrile, 0.2% formic acid in water. 445 For purified mitochondria samples, which were smaller and contained less protein content, pellets 446 were resuspended in 120 µl of 8 M urea with 50 mM HEPES. Samples were reduced with tris(2-447 carboxyethyl)phosphine (10 min, 60 °C) and chloroacetamide (15 min, room temperature). Samples were 448 then treated with 2 µl 0.1 mg/ml LysC endopeptidase (Wako Pure Chemical) for 2 hr at 37 °C. Following 449 sample dilution with 360 µl of 50 mM HEPES, 5 µl of 100 mM CaCl2 and 3 µl of the 0.1 mg/ml Trypsin 450 (Pierce) were added for overnight digestion at 37 °C. The digested peptides were desalted using Pierce C18 451 Spin Columns according to the manufacturer’s protocol. The eluates were dried and resuspended using 452 2% acetonitrile, 0.2% formic acid in water. 453 454 LS-MS/MS 455 For each sample, 500 ng was loaded onto a Thermo Scientific EASY-nLC 1200 connected to an Q Exactive 456 HF Quadrupole-Orbitrap Hybrid Mass Spectrometer. Peptides were separated on an Aurora UHPLC 457 Column (25 cm × 75 µm, 1.6 µm C18, AUR2-25075C18A, Ion Opticks) with a flow rate of 0.35 µl/min for a 458 total duration of 160 min, including washing and re-equilibration. The gradient was composed of 2% 459 Solvent B from the start, 2-6% B for 3.5 min, 6-25% B for 97 min, 25-40% B for 19.5 min, 40-98% B for 2 min, 460 98% B for 3 min, and 98-2% B for 2 min. The gradient was followed by 3 "see-saw" cycles (2% B for 3 min, 2-461 98% B for 2 min, 98% B for 3 min, and 98-2% B for 2 min) for cleaning, and the column was re-equilibrated 462 at 2% B for 3 min. Solvent A consisted of 97.8% H2O, 2% acetonitrile, and 0.2% formic acid, and solvent B 463 consisted of 19.8% H2O, 80% acetonitrile, and 0.2% formic acid. Peptides were ionized via electrospray 464 ionization (NSI) at 2.0 kV. MS1 scans were acquired with a range of 350–1600 m/z at 60K resolution. The 465 maximum injection time was 15 ms with an AGC target of 3 × 106. MS2 scans were acquired at 30K 466 resolution with a scan range of 200-2000 m/z. The maximum injection time was 45 ms with a minimum 467 AGC target of 4.5 × 103. The isolation window was 1.2 m/z, collision energy was 28 NCE, and loop count 468 was set to 12. Mass spectrometer method modification and data collection were performed using Thermo 469 Scientific Xcalibur software. 470 .CC-BY-NC 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 25, 2025. ; https://doi.org/10.1101/2025.11.24.690252doi: bioRxiv preprint Data analysis was performed using Proteome Discoverer 2.5, reference databases from the 471 respective species (see below), and Sequest HT with Percolator validation. Percolator FDRs were set at 472 0.01 (strict) and 0.05 (relaxed). Peptide FDRs were set at 0.01 (strict) and 0.05 (relaxed), with medium 473 confidence and a minimum peptide length of 6. Carbamidomethyl (C) was set as a static modification; 474 oxidation (M) was set as a dynamic modification; acetyl (protein N-term), Met-loss (Protein N-term M), and 475 Met-loss + acetyl (Protein N-term M) were set as dynamic N-Terminal modifications. PSM counts were not 476 reported separately by sample in our initial Proteome Discoverer run, so data were later reanalyzed with 477 Proteome Discover 3.1 to obtain these counts with equivalent settings as above except that Met-loss 478 modifications were not used, and minimum peptide confidence was set to high. 479 Sequence for the A. thaliana reference protein database were obtained from the 2023-10 release of 480 PeptideAtlas (van Wijk et al. 2021) and included the Araport11 (Cheng et al. 2017) nuclear-encoded protein 481 sequences (longest isoform only) combined with the mitochondrial-encoded and plastid-encoded 482 proteins curated by van Wijk et al. (2024). Loci in the large insertion of mitochondrial DNA in nuclear 483 chromosome 2 (Stupar et al. 2001; Fields et al. 2022) that are annotated as functional protein-coding 484 genes in the Araport11 database were removed from the reference to avoid ambiguity in mapping to the 485 true mitochondrial-encoded proteins. For S. conica, annotated protein sequences were taken from the 486 published nuclear (Fields et al. 2023), mitochondrial (Sloan, Alverson, Chuckalovcak, et al. 2012), and 487 plastid genomes (Sloan, Alverson, Wu, et al. 2012). As with A. thaliana, annotated genes found in recent 488 insertions of mitochondrial or plastid DNA into the nucleus were removed from the reference. We also 489 added three aaRS protein sequences that were previously identified with full-length RNA-seq transcripts 490 (Warren et al. 2023) but were not annotated in the published S. conica genome (Iso-Seq AspRS, HisRS, and 491 ValRS). Mitochondrial-encoded and plastid-encoded protein sequences for A. githago were taken from 492 published genomes (Sloan et al. 2014; Warren et al. 2021). Because no reference-quality annotation was 493 available for the A. githago nuclear genome (although one has since been published; Mian and Leitch 494 2024), we used previously published full-length RNA-seq transcripts (Warren et al. 2023) for nuclear-495 encoded protein references. The longest open-reading frame for each transcript was identified with 496 TransDecoder v5.7.1, and the corresponding protein-coding sequences with >99% identity were collapsed 497 with CD-HIT v4.8.1 (Fu et al. 2012). This transcriptome reference for A. githago was then filtered to avoid 498 duplicating annotated proteins from the mitochondrial and plastid reference genomes. 499 500 Enrichment analysis of mitochondrial-encoded and plastid-encoded proteins 501 For each species, enrichment of mitochondrial-encoded and plastid-encoded proteins in the organellar 502 fractions was calculated as the PSM count ratio in the corresponding organellar sample relative to total 503 leaf tissue. PSM counts for the two biological replicates from each sample type were summed for these 504 .CC-BY-NC 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 25, 2025. ; https://doi.org/10.1101/2025.11.24.690252doi: bioRxiv preprint calculations. Proteins with fewer than 5 PSMs in total across all samples from a species were excluded. In 505 cases where no PSMs were detected in the leaf tissue (but were detected in an organellar fraction), a 506 minimum count of 1 was applied for the leaf sample to avoid dividing by 0. Enrichment ratios were also 507 calculated based on ion intensity by using the reported abundance values from Proteome Discoverer. For 508 these calculations, proteins were only considered to be detected in a sample if they were assigned “high” 509 confidence based on identification of a corresponding PSM in that sample. Proteins that were only 510 identified in a sample based on comparisons of peak positions to other samples with detected PSMs were 511 treated as not found. Low or missing abundance values were set to a minimum floor value (0.1 percentile 512 in the corresponding sample), and all values were normalized to the median abundance in the sample. 513 514 Analysis of aaRSs and other tRNA metabolism enzymes 515 To investigate the subcellular localization of previously identified aaRSs (Duchêne et al. 2005; Duchêne et 516 al. 2009; Warren et al. 2023), PSM enrichment ratios were calculated as described above for the 517 organellar-encoded proteins. To assess how relative aaRS abundance within the mitochondria differed 518 between species, we calculated a metric of enrichment bias between organellar-like and cytosolic-like 519 counterparts. Specifically, the mitochondrial enrichment ratio (see above) for the cytosolic-like aaRS was 520 divided by the sum of the mitochondrial enrichment ratios for the organellar-like and cytosolic-like aaRSs. 521 Therefore, a value of 1 for this metric would indicate that only the cytosolic-like aaRS was found in the 522 mitochondria, whereas a values of 0 would indicate that only the organellar-like aaRS was found. When 523 both types were detected in the mitochondria, this metric produces an intermediate value that reflects the 524 relative bias of organellar-like vs. cytosolic-like types. By using the enrichment ratios in these calculations, 525 we were able to normalize to the total leaf samples and avoid directly comparing raw PSM counts between 526 different proteins. We summed PSM counts in cases where there were multiple proteins or subunits for the 527 same aaRS class. For some proteins, we found very low PSM counts in leaf tissue, resulting in large and 528 highly variable enrichment ratios. Therefore, we capped enrichment ratios at a value of 3 to avoid 529 obscuring signal of co-existing aaRS types based on these inflated values. 530 PheRS and tRNA-Phe sequences were obtained via SHOOT (Emms and Kelly 2022), BLASTP 531 searches of the NCBI RefSeq database (Pruitt et al. 2004), and previously curated datasets (Warren et al. 532 2021; Warren et al. 2023; DeTar et al. 2024). Alignments were performed with MAFFT v7.526 (Katoh and 533 Standley 2013) under default parameters and visualized with Geneious (Kearse et al. 2012). The structure 534 of the S. conica PheRS/tRNA-Phe complex was predicted using the AlphaFold3 Server (Abramson et al. 535 2024) under default parameters. The mitochondrial PheRS and cytosolic tRNA-Phe sequences were 536 obtained from previous studies (Warren et al. 2021; Warren et al. 2023). The first 49 amino acids were 537 removed from the PheRS sequence as a putative transit peptide. 538 .CC-BY-NC 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 25, 2025. ; https://doi.org/10.1101/2025.11.24.690252doi: bioRxiv preprint The A. thaliana genes encoding GatCAB, MTF, TilS, PRORP, tRNase Z, and CCAse proteins were 539 previously identified (von Braun et al. 2007; Pujol et al. 2008; Canino et al. 2009; Gutmann et al. 2012; 540 Warren and Sloan 2020). We used reciprocal BLASTP searches against the annotated proteins sequences 541 from the S. conica genome to identify orthologs and inferred presence of the enzymes in organellar 542 fractions based on PSM counts as described above. 543 544 Analysis of mitoribosome subunits 545 The set of A. thaliana mitoribosome subunits was taken from the published structure (Waltz et al. 2020). 546 We also included the uL1m (AT2G42710) and bL12m (AT3G06040) subunits, which were not captured in 547 this structure presumably because they are located in highly mobile parts of the ribosome (Waltz et al. 548 2020). We used the mitochondrial-encoded Rps1 subunit from Carica papaya because this protein is 549 known to have been lost entirely from A. thaliana (Skaltsogiannis et al. 2024). Reciprocal BLASTP searches 550 were used to identify S. conica orthologs, and we inferred presence of subunits in the mitochondrial 551 fraction based on PSM data as described above. Some ribosomal subunits were encoded by two or more 552 closely related paralogs, and some reciprocal BLASTP searches failed due to this paralogy. In other cases, 553 no PSMs were detected in any sample for the named subunit in the A. thaliana mitoribosome structure. In 554 these cases, closely related paralogs were manually checked and substituted in where necessary. The 555 mitoribosome structure was visualized with PyMol v3.1.3, using Protein Data Bank accession 6XYW (Waltz 556 et al. 2020). 557 558 559 Data Availability 560 LC-MS/MS data have been deposited to the ProteomeXchange Consortium via the PRIDE (Pérez-Riverol et 561 al. 2024) partner repository with the dataset identifier PXD071069. Code and processed data are available 562 via GitHub (https://github.com/dbsloan/silene_proteomics). 563 564 565

566 We thank Baiyi Quan and the Proteome Exploration Laboratory at the California Institute of Technology for 567 performing LC-MS/MS runs and data processing. This work was supported by a grant from the National 568 Science Foundation (MCB-2048407) and an HHMI Hanna H. Gray Fellowship. 569 .CC-BY-NC 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 25, 2025. ; https://doi.org/10.1101/2025.11.24.690252doi: bioRxiv preprint Table 1. Detected proteins in LC-MS/MS dataset Detected Proteinsa Sample Type Arabidopsis Silene Mitochondria 2832 2494 Chloroplasts 2118 1974 Leaf Tissue 4371 4294 Total (any sample) 5599 5491 a Proteins were only included in counts if a corresponding PSM was identified in either or both replicates for that sample type. Identifications based solely on a peak position that matches a PSM from another sample type were not included. .CC-BY-NC 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 25, 2025. ; https://doi.org/10.1101/2025.11.24.690252doi: bioRxiv preprint Table 2. Comparison between aaRS detection in LC-MS/MS dataset and previous classification of subcellular localization (Duchêne et al. 2005; Duchêne et al. 2009) Accession Type Duchêne Classification Potential Refinements Notes AT5G22800 AlaRS Chloro-Mito Chloro-only No mito localization detected AT1G50200 AlaRS Cyto-Chloro-Mito Cyto-Mito No chloroplast localization detected AT4G26300 ArgRS Cyto-Chloro-Mito(?) Cyto-Chloro-Mito Mito localization confirmed AT1G66530 ArgRS Cyto-only(?) No change AT4G17300 AsnRS Chloro-Mito No change AT5G56680 AsnRS Cyto-only No change AT1G70980a AsnRS Cyto-only No change AT4G33760 AspRS Chloro-Mito No change AT4G31180 AspRS Cyto-only No change AT4G26870 AspRS Cyto-only No change AT2G31170 CysRS Chloro-Mito No change AT5G38830 CysRS Cyto-only No change AT3G56300a CysRS Cyto-only No change AT1G25350 GlnRS Cyto-only No change AT5G64050 GluRS Chloro-Mito No change AT5G26710 GluRS Cyto-only No change AT3G48110 GlyRS Chloro-Mito No change AT1G29880 GlyRS Cyto-Mito* No change AT3G46100 HisRS Chloro-Mito No change AT3G02760 HisRS Cyto-only No change AT5G49030 IleRS Chloro-Mito No change AT4G10320 IleRS Cyto-only No change AT4G04350 LeuRS Chloro-only No change AT1G09620 LeuRS Cyto-Mito No change AT3G13490 LysRS Chloro-Mito No change AT3G11710 LysRS Cyto-only No change AT3G55400 MetRS Chloro-Mito No change AT4G13780 MetRS Cyto-only No change AT3G58140 PheRS Chloro-Mito No change AT4G39280 PheRS Cyto-only No change AT1G72550 PheRS Cyto-only No change AT5G52520 ProRS Chloro-Mito No change AT3G62120 ProRS Cyto-only No change AT1G11870 SerRS Chloro-Mito No change AT5G27470 SerRS Cyto-only No change AT2G04842 ThrRS Chloro-Mito Chloro-only No mito localization detected AT5G26830 ThrRS Cyto-Mito No change AT1G17960a ThrRS Cyto-only No change AT2G25840 TrpRS Chloro-Mito No change AT3G04600 TrpRS Cyto-only No change AT3G02660 TyrRS Chloro-Mito No change AT1G28350 TyrRS Cyto-only No change AT2G33840a TyrRS Cyto-only No change AT5G16715 ValRS Chloro-Mito Chloro-only(?) Only weak mito localization detected AT1G14610 ValRS Cyto-Mito* Cyto-Mito(?) May be functional in mito given low detection of AT5G16715 in mito a Four aaRSs were not detected in our proteomic dataset (all predicted to be cyto-only and close paralogs of aaRSs that were detected). .CC-BY-NC 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 25, 2025. ; https://doi.org/10.1101/2025.11.24.690252doi: bioRxiv preprint Table 3. Genes encoding subunits of the mitoribosome that were ancestrally present in the angiosperm mitogenone Silene PSM Counts Protein Silene gene IDa Arabidopsis gene IDa Mt1 Mt2 Cp1 Cp2 Leaf1 Leaf2 Rpl2 N-term (uL2m) Chr03-anno2.g23183.t1 mito rpl2 6 7 0 0 0 0 Rpl2 C-term (uL2m) Sconica_v3_41290-RA AT2G44065 9 11 0 0 0 0 Rpl5 (uL5m) mito rpl5 mito rpl5 6 2 0 0 0 0 Rpl10 (uL10m) Sconica_v3_43288-RAb AT3G12370b 6 3 0 0 0 0 Rpl16 (uL16m) Sconica_v3_30744-RAb mito rpl16 8 3 0 0 0 0 Rps1 (uS1) Sconica_v3_17748-RA Lost in A. thaliana 9 6 0 0 0 0 Rps2 (uS2m) Sconica_v3_17930-RA AT3G03600 13 7 0 0 0 0 Rps3 (uS3m) mito rps3 mito rps3 24 17 0 0 0 0 Rps4 N-term (uS4m) FUN_045851-T1/FUN_045790-T1 mito rps4 5 7 0 0 0 0 Rps4 C-term (uS4m) Sconica_v3_40630-RA mito rps4 10 3 0 0 0 0 Rps7 (uS7m) Lost in S. conica mito rps7 NA NA NA NA NA NA Rps10 (uS10m) Sconica_v3_23770-RA AT3G22300 7 5 0 0 0 0 Rps11 (uS11m) Sconica_v3_11125-RA AT1G31817 8 2 0 0 0 0 Rps12 (uS12m) Sconica_v3_11465-RA mito rps12 5 2 0 0 0 0 Rps13 (uS13m) mito rps13 AT1G77750b 3 5 0 0 0 0 Rps14 (uS14m) FUN_029269-T1 AT2G34520 2 5 0 0 0 0 Rps19 (uS19m) Sconica_v3_16857-RA AT5G47320 1 2 0 0 0 0 aNames in bold italics correspond to genes in the mitogenome. All others are in the nuclear genome. bThe genes encoding Rps13 in Arabidopsis, Rpl16 in Silene, and Rpl10 in both species are duplications of their plastid counterparts and not derived from the ancestral mitochondrial gene (Adams, Daley, et al. 2002; Mower and Bonen 2009; Kubo and Arimura 2010). .CC-BY-NC 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 25, 2025. ; https://doi.org/10.1101/2025.11.24.690252doi: bioRxiv preprint Figure S1. Enrichment of proteins encoded by the organellar genomes in purified chloroplast fractions (x-axis) and mitochondrial fractions (y-axis). Enrichment ratios are quantified based on summed abundance values (peak areas) from two replicates relative to abundances in leaf tissue samples. Each point represents a protein, and point sizes are scaled proportionally to the total number of PSMs observed for the corresponding protein across all samples. Plastid-encoded proteins and mitochondrial-encoded proteins and are represented by green circles and gold triangles, respectively. Arabidopsis Silene −2 0 2 −2 0 2 −2 0 2 log10 (Chloroplast / Total Leaf Abundance Ratio) log10 (Mitochondrial / Total Leaf Abundance Ratio) Genome Chloroplast Mitochondrial .CC-BY-NC 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 25, 2025. ; https://doi.org/10.1101/2025.11.24.690252doi: bioRxiv preprint Figure S2. PheRS and tRNA-Phe alignments. (A) Amino acid alignments showing four positions in the PheRS sequence that have a derived change in the Silene conica mitochondrial PheRS but are otherwise conserved across a diverse sampling of green plants/algae. In two of these cases (I184V and R411I), parallel substitutions are observed in the parasitic plant taxa Sapria himalayana and Balanophora fungosa. (B) Alignment of cytosolic, mitochondrial, and plastid tRNA-Phe. I184V S343T A363G R411I GCGGGGAUAGCUCAGUUGG-GAGAGCGUCAGACUGAAGAUCUGAAGGUCGCGUGUUCGAUCCACGCUCACCGCACCA GUUCAGGUAGCUCAGCUGGUUAGAGCAAAGGACUGAAAAUCCUUGUGUCAGUGGUUCGAAUCCACUUCUAAGCGCCA GUCGGGAUAGCUCAGCUGG-UAGAGCAGAGGACUGAAAAUCCUCGUGUCACCAGUUCAAAUCUGGUUCCUGGCACCA Cytosolic tRNA-Phe (Silene conica) Mitochondrial tRNA-Phe (Beta vulgaris) Plastid tRNA-Phe (Silene conica) 5¢ Acceptor Stem Anticodon Loop 3¢ Acceptor Stem A B .CC-BY-NC 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 25, 2025. ; https://doi.org/10.1101/2025.11.24.690252doi: bioRxiv preprint Table S1. PSM counts by sample for additional enzymes involved in tRNA metabolism in mitochondria and other compartments Arabidopsis thaliana Silene conica Enzyme AGI ID Mt1 Mt2 Cp1 Cp2 Leaf1 Leaf2 Gene ID Mt1 Mt2 Cp1 Cp2 Leaf1 Leaf2 GatA AT3G25660 6 7 7 2 6 3 Sconica_v3_10523-RA 0 0 6 10 3 2 GatB AT1G48520 8 7 12 3 7 5 Sconica_v3_48968-RA 0 0 12 15 5 1 GatC AT4G32915 4 3 2 1 0 0 Sconica_v3_42040-RA 0 0 0 5 0 0 CCAse AT1G22660 0 0 0 0 0 0 FUN_019134-T1 3 2 0 1 1 0 MTF AT1G66520 1 0 0 0 0 0 Sconica_v3_07812-RA 7 4 7 10 6 2 TilS AT3G24560 0 0 0 0 0 0 Chr10-anno1.g29341.t1 1 0 0 0 0 0 tRNase Z1a AT1G74700 0 0 0 0 0 0 FUN_038515-T1 2 0 0 0 0 0 tRNase Z2 AT2G04530 0 0 0 0 0 0 Chr03-anno1.g57321.t1 0 0 1 3 0 0 tRNase Z3 AT1G52160 0 0 0 0 0 0 Multiple 0 0 0 0 0 0 tRNase Z4 AT3G16260 0 0 0 0 0 0 Multiple 0 0 0 0 0 0 PRORP1 AT2G32230 0 0 1 0 0 0 FUN_011644-T1 0 0 2 1 0 0 PRORP2 AT2G16650 0 0 0 0 0 0 No gene found NA NA NA NA NA NA PRORP3 AT4G21900 0 0 0 0 0 0 No gene found NA NA NA NA NA NA aA second S. conica gene had a top BLAST hit to A. thaliana tRNase Z1, but no PSMs were detected for it. .CC-BY-NC 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 25, 2025. ; https://doi.org/10.1101/2025.11.24.690252doi: bioRxiv preprint Table S2. Genes encoding subunits of the A. thaliana mitoribosome (Waltz et al. 2020) and their counterparts in S. conica Silene PSM Counts Protein Silene gene IDa Arabidopsis gene IDa Mt1 Mt2 Cp1 Cp2 Leaf1 Leaf2 uL1m Sconica_v3_19193-RA AT2G42710 15 11 1 0 0 0 uL2m N-term Chr03-anno2.g23183.t1 mito rpl2 6 7 0 0 0 0 uL2m C-term Sconica_v3_41290-RA AT2G44065 9 11 0 0 0 0 uL3m Sconica_v3_37834-RA AT3G17465 8 8 0 0 0 0 uL4m Sconica_v3_19156-RA AT2G20060 8 5 0 0 0 1 uL5m mito rpl5 mito rpl5 6 2 0 0 0 0 uL6m Chr08-anno2.g6040.t1 AT2G18400 0 3 0 0 0 0 bL9m FUN_038434-T1 AT5G53070 5 5 0 0 0 0 uL10m Sconica_v3_43288-RA AT3G12370 6 3 0 0 0 0 uL11m Chr03-anno1.g57427.t1 AT4G35490 7 3 0 0 0 0 bL12m Sconica_v3_40269-RA AT3G06040 11 12 0 0 0 0 uL13m Chr04-anno1.g50273.t1 AT3G01790 8 5 0 0 0 0 uL14m FUN_046256-T1 AT5G46160 5 4 0 0 0 0 uL15m Sconica_v3_17785-RA AT5G64670 9 7 0 0 0 0 uL16m Sconica_v3_30744-RA mito rpl16 8 3 0 0 0 0 bL17m Sconica_v3_10014-RA AT5G09770 3 3 0 0 0 0 uL18m Sconica_v3_57073-RA AT5G27820 7 5 0 0 0 0 bL19m FUN_019669-T1 AT1G24240 6 7 0 0 0 0 bL20m Sconica_v3_23041-RA AT1G16740 6 5 0 0 0 0 bL21m Sconica_v3_47180-RA AT4G30930 8 6 0 0 0 0 uL22m Sconica_v3_06702-RA AT1G52370 9 7 0 0 0 0 uL23m Chr08-anno1.g34018.t1 AT4G39880 8 6 0 0 0 0 uL24m Sconica_v3_17854-RA AT5G23535 3 5 0 0 0 0 bL25m Sconica_v3_43645-RA AT5G66860 7 8 0 0 0 0 bL27m Sconica_v3_22742-RA AT2G16930 3 2 0 0 0 0 bL28m Chr10-anno1.g32626.t1 AT4G31460 11 8 0 0 0 0 uL29m Sconica_v3_24808-RA AT1G07830 4 1 0 0 0 0 uL30m Sconica_v3_42097-RA AT5G55140 3 1 0 0 0 0 bL31m Chr03-anno2.g23587.t1 AT5G55125 2 1 0 0 0 0 bL32m Not annotated AT1G26740 N/A N/A N/A N/A N/A N/A bL33m Sconica_v3_46773-RA AT5G18790 2 0 0 0 1 0 bL35m Chr02-anno1.g23806.t1 AT5G45590 3 2 0 0 0 0 bL36m Chr09-anno2.g40880.t1 AT5G20180 1 0 0 0 0 0 mL40 Chr04-anno1.g46970.t1 AT4G05400 3 3 0 1 0 0 mL41 FUN_023568-T1 AT5G40080 1 3 0 0 0 0 mL43 Chr10-anno1.g28670.t1 AT3G59650 4 2 0 0 0 0 mL46 Sconica_v3_41121-RA AT1G14620 13 9 0 0 1 0 mL53 Sconica_v3_53235-RA AT5G39600 8 3 0 0 0 0 mL59/mL64 Chr06-anno1.g19483.t1 AT4G22000 6 1 0 0 0 0 mL60 Sconica_v3_30462-RA AT1G27435 1 3 0 0 0 0 mL80 Chr01-anno1.g62765.t1 AT1G73940 6 1 0 0 0 0 mL87 Sconica_v3_10159-RA AT3G51010 3 2 0 0 0 0 mL101 (rPPR4) Sconica_v3_25967-RA AT1G60770 9 7 0 0 0 0 mL102 (rPPR5) Sconica_v3_09229-RA AT2G37230 27 23 0 0 0 0 mL104 (rPPR9) Sconica_v3_02474-RA AT5G60960 19 17 0 0 0 0 uS1 Sconica_v3_17748-RA Lost in Arabidopsis 9 6 0 0 0 0 uS2m Sconica_v3_17930-RA AT3G03600 13 7 0 0 0 0 uS3m mito rps3 mito rps3 24 17 0 0 0 0 .CC-BY-NC 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 25, 2025. ; https://doi.org/10.1101/2025.11.24.690252doi: bioRxiv preprint uS4m N-term FUN_045851-T1/FUN_045790-T1 mito rps4 5 7 0 0 0 0 uS4m C-term Sconica_v3_40630-RA mito rps4 10 3 0 0 0 0 uS5m Sconica_v3_25633-RA AT1G64880 18 20 0 0 0 0 bS6m Sconica_v3_05402-RA AT3G18760 3 2 0 0 0 0 uS7m Lost in Silene mito rps7 N/A N/A N/A N/A N/A N/A uS8m Sconica_v3_37372-RA AT4G29430 3 4 0 0 0 0 uS9m Sconica_v3_19137-RA AT3G49080 10 5 0 0 0 0 uS10m Sconica_v3_23770-RA AT3G22300 7 5 0 0 0 0 uS11m Sconica_v3_11125-RA AT1G31817 8 2 0 0 0 0 uS12m Sconica_v3_11465-RA mito rps12 5 2 0 0 0 0 uS13m mito rps13 AT1G77750 3 5 0 0 0 0 uS14m FUN_029269-T1 AT2G34520 2 5 0 0 0 0 uS15m Sconica_v3_25605-RA AT1G15810 19 14 0 0 0 0 bS16m Sconica_v3_11133-RA AT5G56940 10 5 0 0 0 0 uS17m Sconica_v3_53755-RA AT1G49400 11 8 0 0 0 0 bS18m Chr02-anno1.g21175.t1 AT1G07210 10 9 0 0 0 0 uS19m FUN_042562-T1 AT5G47320 9 6 0 0 0 0 bS21m FUN_011913-T1 AT3G26360 0 0 0 0 0 0 mS23 Sconica_v3_11989-RA AT1G26750 16 12 0 0 0 0 mS26 Sconica_v3_22695-RA AT5G49210 4 6 0 0 0 0 mS29 Chr04-long_reads1.PB.1859.1 AT1G16870 19 14 0 0 0 0 mS33 Sconica_v3_48073-RA AT5G44710 4 3 0 0 0 0 mS34 Chr02-anno1.g21262.t1 AT5G52370 7 8 0 0 0 0 mS35 Sconica_v3_02574-RA AT3G18240 12 7 0 0 0 0 mS37 FUN_006674-T1 AT1G47278 5 4 0 0 0 0 mS38 Sconica_v3_12210-RA AT5G63150 0 0 0 0 0 0 mS41 Sconica_v3_30872-RA AT5G26800 2 1 0 0 0 0 mS45 Chr09-anno1.g45083.t1 AT5G62270 7 8 0 0 0 0 mS47 Chr08-anno1.g33382.t1 AT4G31810 15 7 0 0 0 0 mS80 (rPPR6) Chr10-anno1.g29827.t1 AT3G02650 21 13 0 0 0 0 mS83 (rPPR10) Sconica_v3_49004-RA AT4G15640 6 6 0 0 0 0 bTHXm Chr01-anno1.g64414.t1 AT2G21290 0 0 0 0 0 0 aNames in bold italics correspond to genes in the mitogenome. All others are in the nuclear genome. .CC-BY-NC 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 25, 2025. ; https://doi.org/10.1101/2025.11.24.690252doi: bioRxiv preprint

Abramson J, Adler J, Dunger J, Evans R, Green T, Pritzel A, Ronneberger O, Willmore L, Ballard AJ, Bambrick J, et al. 2024. Accurate structure prediction of biomolecular interactions with AlphaFold 3. Nature 630:493–500. Adams KL, Daley DO, Whelan J, Palmer JD. 2002. Genes for two mitochondrial ribosomal proteins in flowering plants are derived from their chloroplast or cytosolic counterparts. Plant Cell 14:931– 943. Adams KL, Palmer JD. 2003. Evolution of mitochondrial gene content: gene loss and transfer to the nucleus. Mol. Phylogenet. Evol. 29:380–395. Adams KL, Qiu YL, Stoutemyer M, Palmer JD. 2002. Punctuated evolution of mitochondrial gene content: high and variable rates of mitochondrial gene loss and transfer to the nucleus during angiosperm evolution. Proc. Natl. Acad. Sci. U. S. A. 99:9905–9912. Adams KL, Song K, Roessler PG, Nugent JM, Doyle JL, Doyle JJ, Palmer JD. 1999. Intracellular gene transfer in action: dual transcription and multiple silencings of nuclear and mitochondrial cox2 genes in legumes. Proceedings of the National Academy of Sciences 96:13863–13868. Berg M, Rogers R, Muralla R, Meinke D. 2005. Requirement of aminoacyl-tRNA synthetases for gametogenesis and embryo development in Arabidopsis: Aminoacyl-tRNA synthetase knockouts. Plant J. 44:866–878. Berrissou C, Cognat V, Koechler S, Bergdoll M, Duchêne A-M, Drouard L. 2024. Extensive import of nucleus-encoded tRNAs into chloroplasts of the photosynthetic lycophyte, Selaginella kraussiana. Proc. Natl. Acad. Sci. U. S. A. 121:e2412221121. Boore JL. 1999. Animal mitochondrial genomes. Nucleic Acids Res. 27:1767–1780. von Braun SS, Sabetti A, Hanic-Joyce PJ, Gu J, Schleiff E, Joyce PBM. 2007. Dual targeting of the tRNA nucleotidyltransferase in plants: not just the signal. J. Exp. Bot. 58:4083–4093. Broz AK, Waneka G, Wu Z, Sloan DB. 2021. Detecting de novo mitochondrial mutations in angiosperms with highly divergent evolutionary rates. Genetics 218:iyab039. Canino G, Bocian E, Barbezier N, Echeverría M, Forner J, Binder S, Marchfelder A. 2009. Arabidopsis encodes four tRNase Z enzymes. Plant Physiol. 150:1494–1502. Ceriotti LF, Roulet ME, Sanchez-Puerta MV. 2021. Plastomes in the holoparasitic family Balanophoraceae: Extremely high AT content, severe gene content reduction, and two independent genetic code changes. Mol. Phylogenet. Evol. 162:107208. Chang Y-T, Barad BA, Rahmani H, Zid BM, Grotjahn DA. 2024. Cytoplasmic ribosomes on mitochondria alter the local membrane environment for protein import. bioRxivorg [Internet]. Available from: http://biorxiv.org/lookup/doi/10.1101/2024.07.17.604013 Cheng C-Y, Krishnakumar V, Chan AP, Thibaud-Nissen F, Schobel S, Town CD. 2017. Araport11: a complete reannotation of the Arabidopsis thaliana reference genome. Plant J. 89:789–804. .CC-BY-NC 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 25, 2025. ; https://doi.org/10.1101/2025.11.24.690252doi: bioRxiv preprint DeTar RA, Chustecki JM, Martinez-Hottovy A, Ceriotti LF, Broz AK, Lou X, Sanchez-Puerta MV, Elowsky C, Christensen AC, Sloan DB. 2024. Photosynthetic demands on translational machinery drive retention of redundant tRNA metabolism in plant organelles. Proc. Natl. Acad. Sci. U. S. A. 121:e2421485121. Dimnet L, Salinas-Giegé T, Pullara S, Moyet L, Genevey C, Kuntz M, Duchêne A-M, Rolland N. 2024. Isolation of cytosolic ribosomes associated with plant mitochondria and chloroplasts. Methods Mol. Biol. 2776:289–302. Duchêne A-M, Giritch A, Hoffmann B, Cognat V, Lancelin D, Peeters NM, Zaepfel M, Maréchal-Drouard L, Small ID. 2005. Dual targeting is the rule for organellar aminoacyl-tRNA synthetases in Arabidopsis thaliana. Proceedings of the National Academy of Sciences 102:16484–16489. Duchêne AM, Peeters N, Dietrich A, Cosset A, Small ID, Wintz H. 2001. Overlapping destinations for two dual targeted glycyl-tRNA synthetases in Arabidopsis thaliana and Phaseolus vulgaris. J. Biol. Chem. 276:15275–15283. Duchêne A-M, Pujol C, Maréchal-Drouard L. 2009. Import of tRNAs and aminoacyl-tRNA synthetases into mitochondria. Curr. Genet. 55:1–18. Emms DM, Kelly S. 2022. SHOOT: phylogenetic gene search and ortholog inference. Genome Biol. 23:85. Fields P, Weber MM, Waneka G, Broz A, Sloan DB. 2023. Chromosome-level genome assembly for the angiosperm Silene conica. Genome Biol. Evol. 15:evad192. Fields PD, Waneka G, Naish M, Schatz MC, Henderson IR, Sloan DB. 2022. Complete Sequence of a 641- kb Insertion of Mitochondrial DNA in the Arabidopsis thaliana Nuclear Genome. Genome Biol. Evol. 14:evac059. Fu L, Niu B, Zhu Z, Wu S, Li W. 2012. CD-HIT: accelerated for clustering the next-generation sequencing data. Bioinformatics 28:3150–3152. Gamper H, Hou Y-M. 2020. A label-free assay for aminoacylation of tRNA. Genes 11:1173. Gobert A, Gutmann B, Taschner A, Gössringer M, Holzmann J, Hartmann RK, Rossmanith W, Giegé P. 2010. A single Arabidopsis organellar protein has RNase P activity. Nat. Struct. Mol. Biol. 17:740–744. Gould SB, Waller RF, McFadden GI. 2008. Plastid evolution. Annu. Rev. Plant Biol. 59:491–517. Gutmann B, Gobert A, Giegé P. 2012. PRORP proteins support RNase P activity in both organelles and the nucleus in Arabidopsis. Genes Dev. 26:1022–1027. Ibba M, Söll D. 2004. Aminoacyl-tRNAs: setting the limits of the genetic code. Genes Dev. 18:731–738. Katoh K, Standley DM. 2013. MAFFT multiple sequence alignment software version 7: improvements in performance and usability. Mol. Biol. Evol. 30:772–780. Kearse M, Moir R, Wilson A, Stones-Havas S, Cheung M, Sturrock S, Buxton S, Cooper A, Markowitz S, Duran C. 2012. Geneious Basic: an integrated and extendable desktop software platform for the organization and analysis of sequence data. Bioinformatics 28:1647–1649. .CC-BY-NC 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 25, 2025. ; https://doi.org/10.1101/2025.11.24.690252doi: bioRxiv preprint Kley J, Heil M, Muck A, Svatos A, Boland W. 2010. Isolating intact chloroplasts from small Arabidopsis samples for proteomic studies. Anal. Biochem. 398:198–202. Klipcan L, Moor N, Finarov I, Kessler N, Sukhanova M, Safro MG. 2012. Crystal structure of human mitochondrial PheRS complexed with tRNAPhe in the active “open” state. J. Mol. Biol. 415:527– 537. Kubo N, Arimura S. 2010. Discovery of the rpl10 gene in diverse plant mitochondrial genomes and its probable replacement by the nuclear gene for chloroplast RPL10 in two lineages of angiosperms. DNA Res. 17:1–9. Liu SL, Zhuang Y, Zhang P, Adams KL. 2009. Comparative analysis of structural diversity and sequence evolution in plant mitochondrial genes transferred to the nucleus. Mol. Biol. Evol. 26:875–891. McCutcheon JP, Garber AI, Spencer N, Warren JM. 2024. How do bacterial endosymbionts work with so few genes? PLoS Biol. 22:e3002577. McCutcheon JP, Moran NA. 2012. Extreme genome reduction in symbiotic bacteria. Nat. Rev. Microbiol. 10:13–26. Meiklejohn CD, Holmbeck MA, Siddiq MA, Abt DN, Rand DM, Montooth KL. 2013. An Incompatibility between a mitochondrial tRNA and its nuclear-encoded tRNA synthetase compromises development and fitness in Drosophila. PLoS Genet. 9:e1003238. Meyer EH, Tomaz T, Carroll AJ, Estavillo G, Delannoy E, Tanz SK, Small ID, Pogson BJ, Millar AH. 2009. Remodeled respiration in ndufs4 with low phosphorylation efficiency suppresses Arabidopsis germination and growth and alters control of metabolism at night. Plant Physiol. 151:603–619. Mian S, Leitch IJ. 2024. The genome sequence of the corn cockle, Agrostemma githago L., 1753 (Caryophyllaceae). Wellcome Open Res. 9:590. Mireau H, Lancelin D, Small ID. 1996. The same Arabidopsis gene encodes both cytosolic and mitochondrial alanyl-tRNA synthetases. Plant Cell 8:1027–1039. Molina J, Hazzouri KM, Nickrent D, Geisler M, Meyer RS, Pentony MM, Flowers JM, Pelser P, Barcelona J, Inovejas SA. 2014. Possible loss of the chloroplast genome in the parasitic flowering plant Rafflesia lagascae (Rafflesiaceae). Mol. Biol. Evol. 31:793–803. Mower JP, Bonen L. 2009. Ribosomal protein L10 is encoded in the mitochondrial genome of many land plants and green algae. BMC Evol. Biol. 9:265. Peeters NM, Chapron A, Giritch A, Grandjean O, Lancelin D, Lhomme T, Vivrel A, Small I. 2000. Duplication and quadruplication of Arabidopsis thaliana cysteinyl- and asparaginyl-tRNA synthetase genes of organellar origin. J. Mol. Evol. 50:413–423. Pérez-Riverol Y, Bandla C, Kundu DJ, Kamatchinathan S, Bai J, Hewapathirana S, John NS, Prakash A, Walzer M, Wang S, et al. 2024. The PRIDE database at 20 years: 2025 update. Nucleic Acids Res. 53:D543–D553. Pett W, Lavrov DV. 2015. Cytonuclear interactions in the evolution of animal mitochondrial tRNA metabolism. Genome Biol. Evol. 7:2089–2101. .CC-BY-NC 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 25, 2025. ; https://doi.org/10.1101/2025.11.24.690252doi: bioRxiv preprint Pruitt K, Tatusova T, Maglott D. 2004. NCBI Reference Sequence (RefSeq): a curated non-redundant sequence database of genomes, transcripts and proteins. Nucleic Acids Research 33:D501–D504. Pujol C, Bailly M, Kern D, Marechal-Drouard L, Becker H, Duchene AM. 2008. Dual-targeted tRNA- dependent amidotransferase ensures both mitochondrial and chloroplastic Gln-tRNAGln synthesis in plants. Proceedings of the National Academy of Sciences 105:6481–6485. Richardson AO, Rice DW, Young GJ, Alverson AJ, Palmer JD. 2013. The “fossilized” mitochondrial genome of Liriodendron tulipifera: ancestral gene content and order, ancestral editing sites, and extraordinarily low mutation rate. BMC Biol. 11:29. Roger AJ, Muñoz-Gómez SA, Kamikawa R. 2017. The origin and diversification of mitochondria. Curr. Biol. 27:R1177–R1192. Rogers KC, Söll D. 1995. Divergence of glutamate and glutamine aminoacylation pathways: providing the evolutionary rationale for mischarging. J. Mol. Evol. 40:476–481. Rugen N, Straube H, Franken LE, Braun H-P, Eubel H. 2019. Complexome profiling reveals association of PPR proteins with ribosomes in the mitochondria of plants. Mol. Cell. Proteomics 18:1345–1362. Salinas-Giegé T, Giegé R, Giegé P. 2015. tRNA biology in mitochondria. Int. J. Mol. Sci. 16:4518–4559. Skaltsogiannis V, Nguyen T-T, Corre N, Pflieger D, Blevins T, Hashem Y, Giegé P, Waltz F. 2024. Structural insights into maturation and translation of a plant mitoribosome. bioRxiv [Internet]. Available from: http://dx.doi.org/10.1101/2024.10.28.620559 Skippington E, Barkman TJ, Rice DW, Palmer JD. 2015. Miniaturized mitogenome of the parasitic plant Viscum scurruloideum is extremely divergent and dynamic and has lost all nad genes. Proceedings of the National Academy of Sciences [Internet] In Press. Available from: http://dx.doi.org/10.1073/pnas.1504491112 [doi] Sloan DB, Alverson AJ, Chuckalovcak JP, Wu M, McCauley DE, Palmer JD, Taylor DR. 2012. Rapid evolution of enormous, multichromosomal genomes in flowering plant mitochondria with exceptionally high mutation rates. PLoS Biol. 10:e1001241. Sloan DB, Alverson AJ, Wu M, Palmer JD, Taylor DR. 2012. Recent acceleration of plastid sequence and structural evolution coincides with extreme mitochondrial divergence in the angiosperm genus Silene. Genome Biol. Evol. 4:294–306. Sloan DB, Triant DA, Forrester NJ, Bergner LM, Wu M, Taylor DR. 2014. A recurring syndrome of accelerated plastid genome evolution in the angiosperm tribe Sileneae (Caryophyllaceae). Mol. Phylogenet. Evol. 72:82–89. Sloan DB, Warren JM, Williams AM, Kuster SA, Forsythe ES. 2023. Incompatibility and interchangeability in molecular evolution. Genome Biol. Evol. 15:evac184. Sloan DB, Warren JM, Williams AM, Wu Z, Abdel-Ghany SE, Chicco AJ, Havird JC. 2018. Cytonuclear integration and co-evolution. Nat. Rev. Genet. 19:635–648. Smith D, Asmail S. 2014. Next-generation sequencing data suggest that certain nonphotosynthetic green plants have lost their plastid genomes. New Phytol. 204:7–11. .CC-BY-NC 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 25, 2025. ; https://doi.org/10.1101/2025.11.24.690252doi: bioRxiv preprint Souciet G, Menand B, Ovesna J, Cosset A, Dietrich A, Wintz H. 1999. Characterization of two bifunctional Arabdopsis thaliana genes coding for mitochondrial and cytosolic forms of valyl-tRNA synthetase and threonyl-tRNA synthetase by alternative use of two in-frame AUGs: A. thalianagene characterization using two in-frame AUGs. Eur. J. Biochem. 266:848–854. Steinmetz A, Weil J. 1986. Isolation and characterization of chloroplast and cytoplasmic transfer RNAs.

in Enzymology 118:212–231. Stupar RM, Lilly JW, Town CD, Cheng Z, Kaul S, Buell CR, Jiang J. 2001. Complex mtDNA constitutes an approximate 620-kb insertion on Arabidopsis thaliana chromosome 2: implication of potential sequencing errors caused by large-unit repeats. Proc. Natl. Acad. Sci. U. S. A. 98:5099–5103. Su H-J, Barkman TJ, Hao W, Jones SS, Naumann J, Skippington E, Wafula EK, Hu J-M, Palmer JD, DePamphilis CW. 2019. Novel genetic code and record-setting AT-richness in the highly reduced plastid genome of the holoparasitic plant Balanophora. Proceedings of the National Academy of Sciences 116:934–943. Suzuki T, Miyauchi K. 2010. Discovery and characterization of tRNAIle lysidine synthetase (TilS). FEBS Lett. 584:272–277. Timmis JN, Ayliffe MA, Huang CY, Martin W. 2004. Endosymbiotic gene transfer: Organelle genomes forge eukaryotic chromosomes. Nat. Rev. Genet. 5:123–135. Uwer U, Willmitzer L, Altmann T. 1998. Inactivation of a glycyl-tRNA synthetase leads to an arrest in plant embryo development. Plant Cell 10:1277–1294. Waltz F, Nguyen TT, Arrivé M, Bochler A, Chicher J, Hammann P, Kuhn L, Quadrado M, Mireau H, Hashem Y. 2019. Small is big in Arabidopsis mitochondrial ribosome. Nature Plants 5:106–117. Waltz F, Soufari H, Bochler A, Giegé P, Hashem Y. 2020. Cryo-EM structure of the RNA-rich plant mitochondrial ribosome. Nat. Plants 6:377–383. Warren JM, Broz AK, Martinez-Hottovy A, Elowsky C, Christensen AC, Sloan DB. 2023. Rewiring of aminoacyl-tRNA synthetase localization and interactions in plants with extensive mitochondrial tRNA gene loss. Mol. Biol. Evol. 40:msad163. Warren JM, Salinas-Giegé T, Triant DA, Taylor DR, Drouard L, Sloan DB. 2021. Rapid shifts in mitochondrial tRNA import in a plant lineage with extensive mitochondrial tRNA gene loss. Mol. Biol. Evol. 38:5735–5751. Warren JM, Sloan DB. 2020. Interchangeable parts: The evolutionarily dynamic tRNA population in plant mitochondria. Mitochondrion 52:144–156. van Wijk KJ, Bentolila S, Leppert T, Sun Q, Sun Z, Mendoza L, Li M, Deutsch EW. 2024. Detection and editing of the updated Arabidopsis plastid- and mitochondrial-encoded proteomes through PeptideAtlas. Plant Physiol. 194:1411–1430. van Wijk KJ, Leppert T, Sun Q, Boguraev SS, Sun Z, Mendoza L, Deutsch EW. 2021. The Arabidopsis PeptideAtlas: Harnessing worldwide proteomics data to create a comprehensive community proteomics resource. Plant Cell 33:3421–3453. .CC-BY-NC 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 25, 2025. ; https://doi.org/10.1101/2025.11.24.690252doi: bioRxiv preprint van Wijk KJ, Peltier JB, Giacomelli L. 2007. Isolation of chloroplast proteins from Arabidopsis thaliana for proteome analysis. Methods Mol. Biol. 355:43–48. Yu R, Zhi X, Ceriotti LF, Skippington E, Rice DW, Su H-J, Barkman TJ, Sun C, Liu Y, Fang D, et al. 2025. A record-setting mitogenome in the holoparasitic plant Balanophora yakushimensis accompanied by exceptional loss of organellar DNA repair and recombination genes. BMC Biol [Internet] 23. Available from: http://dx.doi.org/10.1186/s12915-025-02449-8 .CC-BY-NC 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 25, 2025. ; https://doi.org/10.1101/2025.11.24.690252doi: bioRxiv preprint