Switching a conflicted bacterial DTD-tRNA code is essential for the emergence of mitochondria

Optimization of DTD-tRNA code and mito-tRNA(Gly) discriminator base is important for emergence of mitochondria.

Abstract

Mitochondria emerged through an endosymbiotic event involving a proteobacterium and an archaeal host. However, the process of optimization of cellular processes required for the successful evolution and survival of mitochondria, which integrates components from two evolutionarily distinct ancestors as well as novel eukaryotic elements, is not well understood. We identify two key switches in the translational machinery—one in the discriminator recognition code of a chiral proofreader DTD [d-aminoacyl–transfer RNA (tRNA) deacylase] and the other in mitochondrial tRNA^Gly—that enable the compatibility between disparate elements essential for survival. Notably, the mito-tRNA^Gly discriminator element is the only one to switch from pyrimidine to purine during the bacteria-to-mitochondria transition. We capture this code transition in the Jakobida, an early diverging eukaryotic clade bearing the most bacterial-like mito-genome, wherein both discriminator elements are present. This study underscores the need to explore the fundamental integration strategies critical for mitochondrial and eukaryotic evolution.

INTRODUCTION

Primary endosymbiosis is the most accepted model for the origin of mitochondria in eukaryotes (1–3). The transformation of the protobacterial endosymbiont (4) into an integrated organelle for adenosine triphosphate (ATP) generation, mitochondria, in an Asgard archaeal host (5), required drastic rewiring of its proteome. Of the ~1000 identified proteins of the mito-proteome, 40% are of bacterial origin and the remaining proteins are either archaeal or are eukaryotic-specific innovations (6). Therefore, one of the imminent requirements pertaining to endosymbiosis is that both the bacterial- and archaeal-derived macromolecules should be devoid of toxic incompatibilities among them. Here we provide one of the first instances of how such an incompatibility in reconstituting the translational machinery is resolved. A eukaryotic protein of bacterial origin, d-aminoacyl-tRNA deacylase (DTD), and mitochondrial tRNA^Gly rewired their tRNA code with a concomitant discriminator base switch, to avoid untoward hydrolysis of an essential ingredient for protein translation, Gly-tRNA^Gly, in both cytoplasm and mitochondria.

The chiral proofreading enzyme DTD discriminates between l- and d-aminoacyl-tRNAs through L-chiral rejection (7, 8). Owing to the DTD’s L-chiral rejection strategy, its catalytic pocket also fits achiral Gly charged tRNAs and can hydrolyze it (9). The activity of DTD on achiral glycine was shown to be advantageous in avoiding Ala-to-Gly mistranslation in bacteria (10). A discriminator code was also elucidated for bacterial DTD according to which pyrimidine at the N73rd position in Gly-tRNAs acts as a negative determinant for DTD and thereby prevents Gly-tRNA^Gly misediting, whereas Gly-tRNAs bearing purine at the N73rd position as in Gly-tRNA^Ala is hydrolyzed robustly by DTD. DTD from bacteria has been shown to have 100-fold more activity on Gly-tRNA^Gly bearing A73 than the ones having U73 (11).

Unlike bacterial tRNA^Gly, eukaryotes harbor adenine at the N73rd position of cytoplasmic and mitochondrial tRNA^Gly. How DTD avoids misediting of Gly-tRNA^Gly in eukaryotes is largely unexplored. Here, we show that eukaryotic DTDs have switched their tRNA code, compared to bacterial DTDs, to be compatible with the A73 containing eukaryotic tRNA^Glys. Owing to the reciprocal tRNA code, cross-domain expression of bacterial DTD in eukaryotes and eukaryotic DTD in bacteria is toxic to the host because of Gly-tRNA^Gly depletion. Notably, despite mitochondria having a bacterial origin, mito-tRNA^Gly does not follow the bacterial discriminator code. The N73rd base of mito-tRNA^Gly has switched from uracil to adenine to be compatible with the eukaryotic DTD. On the basis of bioinformatics analysis, we could trace this N73 transition of mito-tRNA^Gly in one group of protists, the Jakobida, which have the most bacterial-like mito-genome among eukaryotes. This study underscores the importance of optimization among biomolecules derived from diverse origins, viz., bacteria, archaea, and eukaryotic innovations to acquire the compatibility necessary for emergence of mitochondria and therefore eukaryotes.

RESULTS

Eukaryotic DTD’s tRNA-code switch is necessary for compatibility with tRNA^Gly

DTD is present universally across bacterial and eukaryotic domains but not in archaea, where a structurally and evolutionarily distinct enzyme DTD2 performs the DTD function (12). The conservation of DTD across all the eukaryotic lineages suggests that it was present in the Last Eukaryotic Common Ancestor (LECA) and came through the bacterial endosymbiont that became mitochondria (Fig. 1A). From a comprehensive bioinformatic analysis of tRNA^Gly in eukaryotes, we found that they harbor adenine at the N73rd position in both cytoplasm and mitochondria (Fig. 1A). It was puzzling how DTD in eukaryotes avoids misediting of Gly-tRNA^Gly, which harbors adenine at the 73rd position. We earlier noted a switch in eukaryotic DTD’s discriminator code preference using Mus musculus DTD and Escherichia coli tRNA^Gly U73 (wild type, WT) and (Ec)tRNA^Gly U73A mutant (11); however, the universality of such a shift across eukaryotes and its physiological necessity in avoiding Gly-tRNA^Gly misediting was unexplored.

Fig. 1. Bacterial DTD causes toxicity in eukaryotes.

(A) Distribution of DTD and discriminator (N73rd) base status of tRNA^Gly in three domains of life. Kingdom Archaea lacks DTD and has an evolutionarily distinct enzyme called DTD2 for the d-aminoacyl-tRNA deacylation function. (B) Graphs showing deacylation of Gly-(Sc)tRNA^Gly WT and A73U mutant by ScDTD (0.1, 1, 10, and 100 nM). ScDTD is active on Gly-(Sc)tRNA^Gly WT at 10 nM concentration, whereas on Gly-(Sc)tRNA^Gly A73U mutant, ScDTD is active at 1 nM enzyme concentration. (C) Graphs showing deacylation of Gly-(Sc)tRNA^Gly WT and A73U mutant by EcDTD (0.1, 1, 10, and 100 nM). EcDTD is active on Gly-(Sc)tRNA^Gly WT at 0.1 nM, whereas on Gly-(Sc)tRNA^Gly A73U mutant, EcDTD is active at 10 nM enzyme concentration. (D) Spot dilution assay showing toxicity in S. cerevisiae BY4741 Δdtd strain upon expression of bacterial DTD but not eukaryotic DTD. (E) Top: Northern blot showing in vivo level of aminoacylation of tRNA^Gly in S. cerevisiae BY4741 Δdtd strains. Strain expressing EcDTD has only 20% Gly-(Sc)tRNA^Gly compared to empty vector, and EcDTD A102F and ScDTD have normal levels of Gly-(Sc)tRNA^Gly. Bottom: In vivo aminoacylation levels of tRNA^Ala is unperturbed in all the S. cerevisiae BY4741 Δdtd strains expressing EcDTD, EcDTD A102F, and ScDTD.

We checked the N73 preference of multiple eukaryotic DTDs, viz., Saccharomyces cerevisiae (ScDTD), Leishmania donovani (LdDTD), and Danio rerio (DrDTD), by doing deacylation assay on glycylated tRNA^Gly from S. cerevisiae [Gly-(Sc)tRNA^Gly] bearing A73 (WT) as well as on the A73U mutant version. We found that eukaryotic DTDs have 10- to 100-fold more activity on Gly-(Sc)tRNA^Gly U73 than compared to Gly-(Sc)tRNA^Gly A73 (Fig. 1B and fig. S1A). In contrast, the bacterial DTDs, viz., E. coli DTD and B. subtilis DTD, have 10- to 100-fold more activity on Gly-(Sc)tRNA^Gly A73 than Gly-(Sc)tRNA^Gly U73 (Fig. 1C and fig. S1B). These in vitro deacylations also show that eukaryotic DTDs have 10- to 100-fold reduced activity on Gly-(Sc)tRNA^Gly A73 than bacterial DTDs (Fig. 1, B and C). To probe the physiological relevance of the reduction in preference for Gly-tRNA^Gly A73 by eukaryotic DTDs compared to bacterial DTDs, we checked the effect of their expression in cytoplasm of S. cerevisiae BY4741 Δdtd strain. Bacterial DTD from E. coli (EcDTD) was expressed under copper-inducible promoter in the BY4741 ∆dtd strain. The inactive mutant of EcDTD A102F was used as a negative control based on our earlier study (7) and the yeast DTD (ScDTD) was used as a representative of eukaryotic DTD. To assess the effect of these proteins on the eukaryotic system, BY4741 ∆dtd strains expressing EcDTD-WT, EcDTD-A102F, and ScDTD (fig. S1C) were checked for the cell viability by spot dilution and growth curve assays. The cells expressing EcDTD WT showed severe growth defect, whereas the yeast strains expressing ScDTD and EcDTD-A102F had normal growth comparable to the empty vector control (Fig. 1D and fig. S1D). Furthermore, to ascertain the molecular basis of the cellular toxicity upon expression of EcDTD in yeast, we checked for the misediting of cytoplasmic Gly-tRNA^Gly by probing its in vivo aminoacylation levels. The in vivo levels of aminoacylated and unacylated tRNA^Gly were determined by isolating total RNAs from respective strains and then Northern blot of tRNA^Gly using sequence-specific radiolabeled DNA oligo probe. In vivo levels of cytoplasmic Gly-tRNA^Gly in the strain expressing EcDTD were depleted to as low as 20% (Fig. 1E). However, the levels of aminocylated tRNA^Gly were above 90% in ScDTD and EcDTD-A102F expressing BY4741 ∆dtd strains comparable to the empty vector control strain. In addition, the levels of L-Ala-tRNA^Ala in the same samples were unperturbed in all the strains, which confirms that the toxicity seen upon expression of EcDTD WT in BY4741 ∆dtd strain is a Gly-(Sc)tRNA^Gly misediting specific phenomenon (Fig. 1E). Together, the in vitro and in vivo experiments highlight the implications of the discriminator code switch that has occurred in DTDs across the eukaryotic branch compared to the bacterial DTDs in avoiding the deleterious misediting of Gly-tRNA^Gly in the cytoplasm.

Eukaryotic DTD’s tRNA-code switch rendered it incompatible in bacteria

The flip in the eukaryotic DTD’s discriminator code led to increase in its activity on Gly-tRNA^Gly bearing U73, which is reciprocal to the bacterial DTD’s discriminator code, implying a cross-domain incompatibility of DTDs. We checked the deacylation of WT Gly-tRNA^Gly from E. coli–bearing U73 and on its U73A mutant by bacterial DTDs (EcDTD and BsDTD) (Fig. 2A and fig. S2A) and eukaryotic DTDs (ScDTD and DrDTD) (Fig. 2, B and C, and fig. S2, B and C). Eukaryotic DTDs deacylated Gly-(Ec)tRNA^Gly with 10-fold more activity than the bacterial DTDs (Fig. 2, A to C). Therefore, eukaryotic dtd, owing to the discriminator base switch, would not genetically complement dtd in bacteria because of misediting of Gly-tRNA^Gly as bacterial tRNA^Gly harbor uracil at the N73rd position. To investigate this, we opted for E. coli K12 Δdtd strain as our in situ experimental system as it contains tRNA^Gly with uracil at the N73rd position (Fig. 1A). We expressed eukaryotic DTDs (ScDTD and DrDTD) and bacterial DTD (EcDTD) in K12 Δdtd strain (fig. S2D) and performed spot dilution–based toxicity assays. Toxicity assays have shown that ScDTD and DrDTD are toxic in bacteria whereas bacterial DTD, i.e., EcDTD, is not toxic when expressed in K12 Δdtd strain (Fig. 2D). Thus, eukaryotic DTDs owing to their tRNA-code switch are no longer compatible in the ancestral bacteria as they are toxic because of Gly-tRNA^Gly misediting. The cross-domain toxicity caused by bacterial and eukaryotic DTDs upon reciprocal expression is, to the best of our knowledge, the first example of interdomain incompatibility that exists in components of translation apparatus through actively depleting an essential substrate for protein biosynthesis.

Fig. 2. Cross-domain toxicity of eukaryotic-DTDs in bacteria.

(A) Graph showing deacylation of Gly-(Ec)tRNA^Gly WT by B. subtilis DTD (1, 10, and 100 nM). BsDTD deacylated Gly-(Ec)tRNA^Gly at 10 nM enzyme concentration. (B and C) Graph showing deacylation of Gly-(Ec)tRNA^Gly by eukaryotic DTDs—ScDTD and DrDTD (0.1, 1, and 10 nM). ScDTD and DrDTD are deacylated Gly-(Ec)tRNA^Gly at 1 nM enzyme concentration. (D) Spot dilution assay in E. coli K12 Δdtd strain showing that expression of eukaryotic DTDs is toxic to bacteria, but expression of bacterial DTD (EcDTD) does not cause toxicity.

Mitochondrial tRNA^Gly U73A switch confers compatibility with eukaryotic DTD

The misediting of Gly-tRNA^Gly in bacteria by eukaryotic DTD poses a pertinent question as to whether eukaryotic DTD will be compatible in the mitochondrial compartment, which has an independent translation apparatus of bacterial origin (13). First, we wanted to check whether DTD localizes to mitochondrial compartment. We have generated green fluorescent protein (GFP)–tagged ScDTD constructs under its native promoter and performed confocal microscopy to track its localization using MitoTracker Red CMXRos as a marker for active mitochondria. Confocal microscopy images clearly show dual localization (Pearson’s R value: 0.59) of ScDTD in both mitochondria and cytosol (Fig. 3A and fig. S3A). This observation is corroborated by the mass spectrometry–based high-confidence yeast mito-proteomics data (14) and in CYCLoPs database (15) wherein DTD is reported to be present in mitochondria. We then analyzed the discriminator base status of mitochondrial tRNA^Gly. To our surprise, mitochondrial tRNA^Glys harbor adenine at the N73rd position instead of uracil, which it is expected to harbor owing to the bacterial origin of mitochondria (Fig. 3B). Notably, tRNA^Gly is the only tRNA to have switched its discriminator preference during the transformation from an ancestral bacteria to mitochondria (table S1). To probe the reason for such a switch, we performed in vitro deacylations on S. cerevisiae mitochondrial Gly-tRNA^Gly substrate using bacterial DTDs (EcDTD and BsDTD) (Fig. 3C) and eukaryotic DTDs (ScDTD, LdDTD, and DrDTD) (Fig. 3D and fig. S3B). We found that bacterial DTDs deacylated mitochondrial Gly-(Sc)tRNA^Gly 10- to100-fold more actively than eukaryotic DTDs (Fig. 3, C and D). Thus, although mitochondria is of bacterial origin, its tRNA^Gly has switched its discriminator base from uracil to adenine to avoid misediting by eukaryotic DTD—which has 10- to100-fold more activity on U73-harboring Gly-tRNA^Gly.

Fig. 3. Bacterial DTD causes toxicity in mitochondria.

(A) Dual localization of C-terminal GFP-tagged ScDTD of both cytoplasm and mitochondria visualized using confocal microscopy. (B) Discriminator base status of all mitochondrial tRNAs in eukaryotes. (C) Deacylation of mitochondrial Gly-(Sc)tRNA^Gly WT by bacterial DTDs (EcDTD and BsDTD). (D) Deacylation of mitochondrial Gly-(Sc)tRNA^Gly WT by eukaryotic DTDs (ScDTD and DrDTD). (E) Spot dilution assay of S. cerevisiae BY4741 Δdtd strain with empty vector, MTS-EcDTD, MTS-EcDTD A102F, and MTS-ScDTD showing toxicity in the MTS-EcDTD strain but not in MTS-EcDTD A102F and MTS-ScDTD strains in plates containing dextrose as carbon source. (F) Further enhancement of mitochondrial toxicity upon performing the spot dilution assay of S. cerevisiae BY4741 Δdtd strain with empty vector, MTS-EcDTD WT, MTS-EcDTD A102F, and MTS-ScDTD in plates containing only glycerol (nonfermentable carbon source) only in the MTS-EcDTD WT strain, but the MTS-EcDTD A102F and MTS-ScDTD WT strains’ growth was unaffected. (G) Histogram of fluorescence-activated cell sorting analysis (N = 3) with MitoTracker Red CMXRos showing reduction in amount of active mitochondria in MTS-EcDTD WT strain, but strains with MTS-EcDTD A102F and MTS-ScDTD have normal amount of mitochondria like the empty vector control.

We propose that the U73A switch in mitochondrial tRNA^Glys resulted in its compatibility with eukaryotic DTDs, but concomitantly became incompatible with the DTDs from its bacterial ancestors. To test this in situ, we expressed and targeted both bacterial and eukaryotic DTDs to the mitochondria in S. cerevisiae BY4741 Δdtd strain using mitochondrial targeting signal (MTS) under copper-inducible promoter. The mitochondrial localization and expression of the MTS-tagged DTD-GFP constructs were confirmed using confocal microscopy (fig. S3C). However, to exclude the effect of GFP tag on DTD activity, we performed the viability assays with strains transformed with DTD genes without GFP tag. The spot dilution and growth curve assays showed that indeed the strains where MTS-EcDTD targeted to mitochondria showed toxicity but the strains with MTS-EcDTD A102F, and MTS-ScDTD grew normally like the empty vector control strain (Fig. 3E and fig. S3D). We also checked the growth of these mitochondrial targeting yeast strains in a medium containing nonfermentable carbon source, glycerol, to enhance the effect of mitochondrial toxicity. Indeed, upon culturing in nonfermentable carbon source, the growth defect was elevated in BY4741 Δdtd strain where EcDTD is targeted to mitochondria but the strains expressing EcDTD A102F and ScDTD grew normally like the empty vector control (Fig. 3F and fig. S3D). Furthermore, we quantified the amount of active mitochondria in the above strains by staining them with MitoTracker Red CMXRos dye and then subjecting them to fluorescence-activated cell sorting analysis (16). The cells of BY4741 Δdtd MTS-EcDTD strain had notably lesser amount of active mitochondria compared to MTS-EcDTD A102F and MTS-ScDTD strains, which had normal amount of mitochondria similar to the empty vector control (Fig. 3G and fig. S4). The effect of discriminator base switch in the mitochondrial tRNA^Gly is so profound a change that they are no more compatible with bacterial DTDs, despite having bacterial ancestry. Thus, discriminator base switch in the mitochondrial tRNA^Gly is an important optimization event that led to its compatibility with the eukaryotic DTD.

Jakobids have the relics of the mitochondrial-tRNA^Gly discriminator base switch

In our thorough analysis of mitochondrial tRNA^Gly discriminator base, we found uracil occupancy at the N73rd position—a hallmark of bacterial type tRNA^Gly—in plants and jakobids. The basis for the presence of U73 containing tRNA^Glys in plants is not clear and warrants further investigations as it harbors archaea-derived DTD2 in addition to DTD (17), whereas the presence of U73 at the N73rd position of tRNA^Gly in jakobid mitochondria is of special interest as they are considered as the closest lineage to the ancestral proto-mitochondria owing to their bacterial-like genome (18, 19). We found that among those jakobids whose mito-genome sequences are available, some members have a chimeric set of mito-tRNA^Gly with both U73 and A73 discriminator base while others have all A73 containing mito-tRNA^Glys (Fig. 4A and table S2). Unlike most mito-genomes of higher eukaryotes, which code for only one tRNA^Gly isoacceptor—TCC—the mito-genomes of jakobid members encode for two isoacceptors of tRNA^Gly—TCC and GCC. Notably, the mitochondrial tRNA^Glys with TCC anti-codon have A73 throughout while GCC containing tRNA^Gly has both A and U in the discriminator position (Fig. 4A). In a phylogenetic analysis of tRNA^Glys, jakobids indeed branched from an intermediate node position between bacterial tRNA^Glys and mitochondrial tRNA^Glys (Fig. 4B). However, why the tRNA^Gly with GCC anti-codon was eventually lost outside the jakobid clade during the course of evolution remains to be elucidated (Fig. 4C). Thus, we could capture a contemporary group of protists—the Jakobida—which have a mitochondrial tRNA^Gly isoacceptor in a state of transition between a bacterial type with U73 and a eukaryotic type A73 containing eukaryotic-type mito-tRNA^Gly. Furthermore, we performed a deacylation assay on mitochondrial Gly-tRNA^Gly from the jakobid Andalucia godoyi [mito-Gly-(Ag)tRNA^Gly] with bacterial DTD (EcDTD) (Fig. 4D) and eukaryotic DTDs (ScDTD and DrDTD) (Fig. 4E and fig. S3B). Deacylation assays showed that bacterial DTDs have 10- to 100-fold more activity on mito-Gly-(Ag)tRNA^Gly than eukaryotic DTDs. We further sought to investigate DTD from jakobid protists. We predicted DTD from the only available nuclear genome sequence of a species from the Jakobida—A. godoyi DTD1 (AgDTD) (20). Next, we did phylogenetic analysis of the AgDTD by inferring a Maximum Likelihood phylogenetic tree, which shows that AgDTD is of eukaryotic type (fig. S5). Moreover, we found the A. godoyi mitochondrial GlyRS to be a bacterial ɑ₂-type, which can glycylate tRNA^Gly bearing adenine or uracil at the N73rd position indiscriminately (fig. S6, A and B). Hence, the presence of a bacterial ɑ₂-type mitochondrial GlyRS in the Jakobida rules out any role of mito-GlyRS in mitochondrial tRNA^Gly U73-to-A73 switch in early mitochondrial lineages. Thus, we identified a deep eukaryotic branch of protists, the Jakobida, containing the relics of the switch in the discriminator base of mitochondrial tRNA^Gly to become compatible with eukaryotic DTDs.

Fig. 4. Jakobids have the relics of the mitochondrial-tRNA^Gly discriminator base transition.

(A) Organisms belonging to jakobid protists with corresponding mitochondrial tRNA^Gly discriminator base. Multiple jakobid species have mitochondrial tRNA^Glys enriched with both A73 and U73. (B) ML phylogenetic tree of tRNA^Gly from bacteria and mitochondria showing jakobid mitochondrial tRNA^Gly branching from a node position at the transition between bacterial and other mitochondrial-tRNA^Glys. (C) Schematic depiction of the transition of mitochondrial tRNA^Gly N73 status from bacterial type “U73” to eukaryotic type “A73.” Mitochondrial tRNA^Gly bearing anticodon GCC is lost beyond the Jakobida, which has chimeric enrichment of both U73 and A73. (D and E) Deacylation of A. godoyi (Ag) mitochondrial Gly-tRNA^Gly by EcDTD and ScDTD. (F) A model depicting optimizations of mitochondrial tRNA^Gly and eukaryotic DTD important for the emergence of mitochondria. The engulfment of a bacterium by an archaeal host led to the presence of two redundant chiral proofreaders—archaeal DTD2 and bacterial DTD in the same system. During the course of evolution, bacterial DTD was recruited in both cytoplasmic and mitochondrial compartments and the archaeal DTD2 was lost. The recruitment of bacterial DTD in eukaryotes necessitated its tRNA-code switch triggered by its incompatibility with archaeal-derived cytoplasmic tRNA^Gly. The eukaryotic DTD with switched tRNA code in turn mediated a U73A transversion in the mitochondrial tRNA^Gly. The DTD-tRNA code switch and mito-tRNAGly U73A transversion illustrates the molecular optimizations that were important for the emergence of mitochondria.

DISCUSSION

Our studies show that the eukaryotic DTD tRNA-code switch—reciprocal to their ancestral counterparts in bacteria—is universal and prevents misediting of Gly-tRNA^Gly bearing A73 present in the translation apparatus of eukaryotic cytoplasm and mitochondria. The conservation of DTD across extant branches of eukaryotes and its absence in archaea suggest its recruitment to the eukaryotes via the endosymbiont alpha-proteobacteria during the symbiotic origin of mitochondria. The selection pressure for the tRNA-code switch in eukaryotic DTD came from the toxic nature of the incompatibility of the bacterial DTDs with the archaeal-derived tRNA^Gly harboring adenine at the N73rd position in the cytoplasm of eukaryotes (Fig. 3B). The fixation of adenine at the N73rd position of cytoplasmic tRNA^Gly got precedence over the tRNA code of bacterial-derived DTD because of the archaeal-derived GlyRS, which requires A73 for aminoacylation in the cytoplasm of eukaryotes (21).

The discriminator base switch in mitochondrial tRNA^Gly was an important optimization that was enforced by the eukaryotic DTD to prevent the misediting of mitochondrial Gly-tRNA^Gly. Notably, we traced the tRNA^Gly discriminator base switch in the deep branch of mitochondria from the Jakobida. On the basis of phylogenetic analysis, we found that the jakobid DTD is of eukaryotic type, so we propose that the eukaryotic DTD’s tRNA-code switch has already occurred at the stage of the LECA, before the earliest branches of eukaryotic lineages diverged. The study thus demonstrates that it is an essential aspect of the optimization of translational machinery for compatibility among components recruited from different branches of life about 2 billion years ago.

The two optimizations shown in this study in tRNA^Gly and DTD are part of numerous others yet to be explored—important for successful integration of mitochondria into the physiology of eukaryotes (Fig. 4F). This becomes even more important considering the emerging data that nearly 1000 nuclear encoded proteins of diverse origin are targeted to mitochondria. Our study highlights compatibility among archaeal and bacterial components as an important factor that influenced the integration strategies for successful emergence of mitochondria. In this context, it is worth noting emerging evidence of migration of proteins across compartments in higher systems like nuclear localization of mitochondrial tricarboxylic acid (TCA) cycle enzymes in mammalian zygotic genome activation (22). Furthermore, the transition of the Proteobacteria endosymbiont into the mitochondria entailed myriad gene loss and nuclear gene transfer (6, 23); the former is proposed to be driven by genetic redundancy between endosymbiont and nuclear genomes (24–27). Here, on the basis of our results, we propose molecular compatibility in key cellular processes as another driver apart from gene redundancy that influenced the genetic rewiring of the evolving mitochondria. Therefore, molecular optimization could yet be another mode of appearance of the mosaic mitochondrial proteome.

MATERIALS AND METHODS

Bioinformatics analysis

To understand the distribution of tRNA^Gly discriminator base (N73) in opisthokonta, all the tRNA^Gly sequences (in both cytoplasm and mitochondria) available in the GtRNAdb database (http://gtrnadb.ucsc.edu/) (28) and mitotRNAdb database (http://mttrna.bioinf.uni-leipzig.de/mtDataOutput/), respectively, were downloaded and analyzed for the N73 base.

The phylogenetic trees were constructed using IQ-tree server (http://iqtree.cibiv.univie.ac.at/) (29). The multiple sequence alignment of all available tRNA^Gly sequences of bacteria and eukaryota was performed using the T-coffee (http://tcoffee.crg.cat/) MSA alignment tool (30).

In vitro experiments

Cloning, expression, and protein purification

DTD genes from the genome of E. coli (Ec) and B. subtilis (Bs) and cDNAs of S. cerevisiae (Sc), L. donovani (Ld), and D. rerio (Dr) were cloned in pET28b (except ScDTD, cloned in pET21a) vectors with C-terminal 6×-Histidine tags. The primer list is given in table S3. The proteins were expressed and purified as described previously (7, 9, 31). These constructs (EcDTD, BsDTD, ScDTD, LdDTD, and DrDTD) were transformed and overexpressed into E. coli BL21 (DE3) cells. Purification of these 6×-His-tagged proteins was performed by Ni-NTA (nitrilotriacetic acid) affinity chromatography, followed by size exclusion chromatography (SEC). Purification method and buffers for Ni-NTA and SEC were used as previously described (7). The lysis buffer contained 50 mM tris-HCl (pH 8.0), 150 mM NaCl, and 10 mM imidazole. After loading the Ni-NTA column with the cell lysate, the column was washed successively with wash buffers containing 50 mM tris-HCl (pH 8.0), 300 mM NaCl, and 30 mM imidazole, and 50 mM tris-HCl (pH 8.0), 150 mM NaCl, and 50 mM imidazole. Protein was eluted with elution buffer containing 50 mM tris-HCl (pH 8.0), 150 mM NaCl, and 250 mM imidazole. Fractions containing the protein were pooled, concentrated, and subjected to further purification with SEC (Superdex-75) in buffer containing 100 mM tris-HCl (pH 7.5) and 200 mM NaCl. Last, the fractions containing the purified protein were pooled, concentrated, and mixed thoroughly with an equal volume of 100% glycerol before aliquoting and storing at −30°C for future use.

Biochemical assays

E. coli, A. godoyi, and S. cerevisiae tRNA^Glys were prepared using an in vitro transcription method through the MEGAshortscript T7 Transcription Kit (Thermo Fisher Scientific, USA). These tRNAs were then end-labeled at the 3′ end using CCA adding enzyme (32). These 3′ end-labeled tRNA^Glys were charged with glycine in a reaction mix containing 100 mM Hepes (pH 7.2), 30 mM KCl, 10 mM MgCl₂, 2 mM ATP, 50 mM glycine, 1 μM tRNA^Gly, and 2.2 μM T. thermophilus glycyl-tRNA synthetase (UniProt ID: P56206) incubated at 37°C for 15 min. Deacylations of these glycylated-tRNAs were performed by varying concentrations of both EcDTD, BsDTD, ScDTD, LdDTD, and DrDTD enzymes (11). Briefly, deacylation experiments were performed by incubating different DTD enzymes of varying concentrations (as mentioned in the figures) with 0.2 μM of different substrates (α-32P–labeled aminoacyl-tRNAs) in a buffer containing 20 mM tris (pH 7.2), 5 mM MgCl₂, 5 mM dithiothreitol, and bovine serum albumin (0.2 mg/ml) at 37°C. One microliter of the reaction mixture was withdrawn at different time points and subjected to 2.5 μl of S1 nuclease (2 U/μl) (Thermo Fisher Scientific, USA) digestion and incubated for 30 min at 22°C. One microliter of S1 nuclease digested samples was spotted on to cellulose thin-layer chromatography (TLC) plates (Merck KGaA, Germany). These TLC plates were developed using a mobile phase consisting of 100 mM ammonium chloride and 5% glacial acetic acid; the mobile phase front was allowed to move up to about three-fourths of the vertical length of the TLC plate before air-drying the plate. The developed TLC plates were exposed overnight to imaging plates (Fujifilm, Japan). Typhoon FLA 9000 biomolecular imager (GE Healthcare) was used for phosphor imaging of the exposed image plates. Percentages of aminoacylation were assessed using Image Gauge V4.0 software.

The percentage of aa-AMP present at the 0-min time point was considered as 100%. The percentage of aa-AMP remaining after deacylation was plotted against the respective deacylation time points (10, 11). The deacylation assays were done in independent triplicates and plotted using GraphPad prism software where the error bars show the deviation from the average values.

In vivo experiments in yeast and bacteria

Our aim was to understand the effect of bacterial DTD in a eukaryotic system and vice versa. We chose the well-studied S. cerevisiae BY4741 and E. coli K12 as the eukaryotic and bacterial model system for our in vivo studies.

Construction of E. coli K12 Δdtd::Kan deletion strain

E. coli K12 Δdtd::Kan was generated using P1 lysate from Keio collection (JW 3858–2) (9).

Bacterial toxicity assays

Cell toxicity assays were done with E. coli K12Δdtd::Kan, which were transformed with empty vector, EcDTD, ScDTD, and DrDTD cloned in pTrc99 vector with C-terminal 6×-His Tag, respectively. Cell viability assays were carried out in minimal media–agar plate–based assay; a primary culture was grown in LB medium and then a secondary culture was grown in minimal medium (9) at 37°C until the optical density of 600 nm (OD₆₀₀) reached 0.6. Subsequently, eight 10-fold serial dilutions were made in phosphate-buffered saline, and 3 μl of each was spotted on minimal media–agar plates supplemented with 0, 0.5, and 1 mM isopropyl-β-d-thiogalactopyranoside (IPTG). The plates were incubated at 37°C for 12 to 16 hours.

Construction of Δdtd S. cerevisiae BY4741 strain

The WT BY4741 strain was obtained as a gift from R. Palanimurugan’s laboratory, Centre for Cellular and Molecular Biology, Hyderabad, India. The dtd gene was replaced by kanamycin cassette by the polymerase chain reaction (PCR) method using Kan MX plasmid as a template. The PCR product was generated bearing flanking chromosomal sequences of dtd gene with the Kan cassette. The PCR product was purified and transformed through direct carrier DNA/polyethylene glycol (PEG) method–based transformations (33), and Δdtd deletion mutant colonies were screened for kanamycin selection and were assessed by genotyping using the flanking and kanamycin primers and confirmed the deletion (34).

Cloning and transformation of dtd genes into yeast

EcDTD, EcDTD (A102F) catalytic mutant, and ScDTD genes with C-terminal flag tag were cloned under a CUP (metallothionine gene) promotor containing vector (pLE124, obtained as a gift from R. Palanimurugan’s laboratory) using restriction digestion and ligation cloning method. To target the DTD proteins into yeast mitochondria, an MTS was inserted on the N-terminal region of the respective DTD genes. The MTS of aconitase enzyme was selected as it shows the maximum efficiency of mitochondrial targeting (94:6) (35). Apart from the CUP promotor, to investigate the localization of native ScDTD using microscopy, another construct of ScDTD gene was constructed with the native ScDTD promotor and a C-terminal GFP tag.

These constructs were transformed in E. coli DH5-alpha cells and clones were confirmed by sequencing of the respective isolated plasmids. The confirmed plasmids were transformed into BY4741 Δdtd strain using the lithium acetate/single-stranded carrier DNA/PEG method (33). The colonies were selected on an SD/−Leu selective medium plate and were screened for expression of DTD levels under different copper concentrations by Western blotting.

Western blotting of yeast and bacterial strains of prokaryotic and eukaryotic DTD

The BY4741 Δdtd strains containing EcDTD (WT), EcDTD-Mut (A102F), empty vector, and ScDTD constructs were grown in 20 ml of −Leu synthetic-defined broth medium with varying Cu²⁺ concentrations overnight at 30°C. The pelleted cells were subjected to the NaOH-TCA precipitation method (36) for protein precipitation. These cultures were then pelleted and dissolved in SDS loading dye. The yeast samples were run on 12% SDS–polyacrylamide gel electrophoresis (SDS-PAGE). The gel was transferred onto the polyvinylidene difluoride (PVDF) membrane through the wet gel transfer method (12). The transferred protein bands were blocked in 5% nonfat milk solution in tris buffered saline with tween (TBST) buffer for 1 hour. The blocked membrane was cut for Flag-tagged DTD proteins and loading control (PGK1) and were probed using respective primary antibodies in 1:10,000 dilution in 2% nonfat milk solution in TBST buffer—Monoclonal ANTI-FLAG M2 mouse antibody (Sigma-Aldrich, catalog no. F3165) and PGK1 Monoclonal mouse antibody (Thermo Fisher Scientific, catalog no. 459250), respectively, for overnight incubation. The respective membranes were then washed thrice for 15 min each with TBST buffer and incubated with secondary antibody in 1:10,000 dilution in 2% nonfat milk solution in TBST buffer [Mouse immunoglobulin G (IgG) horseradish peroxidase (HRP)–Linked Whole Ab from GE Healthcare, catalog no. NA931V] for 1 hour at room temperature. After three 15-min washes, the blots were developed under chemiluminescence and analyzed by immunoblotting. In case of E. coli K12Δdtd::Kan, strains expressing EcDTD, ScDTD, and DrDTD in pTrc99 vector with C-terminal 6× Histidine-tag and empty vector of pTrc99, cultures were grown overnight in 3 ml of LB broth with and without 0.5 mM IPTG induction (37). The bacterial samples were run on 12% SDS-PAGE. The gel was transferred onto the PVDF membrane through the wet gel transfer method (12). The transferred protein bands were blocked in 5% nonfat milk solution in TBST buffer for 1 hour and probed by using anti-His antibody (Cell Signaling Technology, catalog no. 12698S) in 1:1000 dilution in 2% nonfat milk solution in TBST buffer as primary antibody and anti-rabbit IgG, HRP-linked antibody (Cell Signaling Technology, catalog no. 7074s) as secondary antibody. The loading control for bacterial Western blot—FtsZ was probed using anti-FtsZ sera (gift from M. Reddy’s lab, CSIR-CCMB, India) as primary antibody and anti-rabbit IgG, HRP-linked antibody (Cell Signaling Technology, catalog no. 7074s) as secondary antibody in 1:10,000 dilution in 2% nonfat milk solution in TBST buffer and then developed through the same procedure as mentioned above for yeast samples.

Yeast viability growth assays

Growth of BY4741 Δdtd strains containing EcDTD (WT), EcDTD-Mut (A102F), empty vector, and ScDTD constructs was tested on the basis of the protocol explained in (38). Briefly, the strains were initially inoculated in a 5-ml SD/−Leu selective medium broth and allowed to grow overnight in a 30°C incubator shaker. These cultures were then subcultured to a final OD₆₀₀ of 0.2 in 5 ml of SD/−Leu selective medium broth and then were 10-fold serially diluted (10⁻¹, 10⁻², 10⁻³, and 10⁻⁴). Of each serially diluted sample, 10 μl was spotted on SD/−Leu selective medium plates and were incubated at 30°C for 48 to 72 hours.

Growth curves were also performed to check the effect of EcDTD expression in yeast. The primary culture of EcDTD (WT), EcDTD-Mut (A102F), empty vector, and ScDTD constructs containing yeast strains (Δdtd) was inoculated in a 5-ml SD/−Leu selective medium broth and allowed to grow overnight in a 30°C incubator shaker. From those overnight cultures, secondary inoculation was done in a 20-ml SD/−Leu selective medium broth to maintain an equal initial OD₆₀₀ of 0.2 for all cultures. These cultures were allowed to grow in a 30°C incubator shaker, and the cell density at OD₆₀₀ was recorded after every 4 hours. These were done in triplicate and the OD₆₀₀ values were plotted using GraphPad Prism software version 8.0. Error bars represent deviation from average values.

Microscopy

To investigate the cellular localization of native DTD protein in a eukaryotic system S. cerevisiae, a construct of ScDTD under the native promotor tagged with C-terminal GFP was transformed in BY4741 Δdtd deletion strain as explained earlier. This strain was grown overnight in 5 ml of −Leu synthetic defined broth at 30°C. MitoTracker Red CMXRos (20 nM; Thermo Fisher Scientific, catalog no. M7512) was added to these overnight grown cells and incubated further for 30 min. The MitoTracker Red CMXRos incubated cells were then pelleted and washed in autoclaved water. Five microliters of those cells was immobilized on poly-l-lysine–coated glass slides and visualized under an FV3000 (OLYUMPUS) microscope. The details of the imaging process are as follows:

• 1)GFP excitation/emission laser: 488/510 nm
• 2)Mitotracker excitation/emission laser: 561/572 nm
• 3)Scan size: 1024 × 1024
• 4)Magnification: 100×-oil

Individual and merged images of differential interference contrast (DIC), GFP, and MitoTracker Red CMXRos were generated using ImageJ2 software (39), and the colocalization was tested using its JACOP plugin (40) that generated the cytofluorogram along with Pearson’s coefficient values.

Northern blotting

Northern blotting was performed as earlier reported (11). Briefly, the BY4741 Δdtd strains expressing EcDTD (WT), EcDTD-Mut (A102F), empty vector, and ScDTD were grown in 5 ml of -Leu synthetic defined broth medium as primary culture and later inoculated in 20 ml of secondary culture with 10 μM Cu²⁺ for 24 hours at 30°C. The culture was pelleted and total RNA was isolated under acidic conditions in ice (41).

This extracted RNA (0.15 to 0.25 A260 unit) was run on 6.5% acid–urea PAGE at 4°C for 20 to 24 hours. The RNA from the gel was electroblotted on Hybond-N⁺ nitrocellulose (GE Healthcare) membrane at 15 V, 3 A for 38 min. This membrane with nucleic acids transferred on was ultraviolet cross-linked for 4 min. This was followed by hybridization with [³²P]-labeled tRNA-specific DNA oligo probe. The probe was labeled using [Ƴ-³²P]ATP using polynucleotide kinase enzyme (NEB). The signal from the probe was recorded overnight on an image plate and quantified using a Typhoon FLA 9000 biomolecular imager (GE Healthcare).

MitoTracker red CMXRos-based estimation of active mitochondria in yeast using flow cytometry

We used MitoTracker Red CMXRos (Thermo Fisher Scientific, catalog no. M7512) to measure the amount of active mitochondria (16) in BY4741 Δdtd strain expressing EcDTD (WT), EcDTD-(A102F), empty vector, and ScDTD (WT) constructs. Initially, primary cultures (3 ml each in SD media) were grown overnight. From these primary cultures, an equivalent of 0.2 OD₆₀₀ cells were inoculated into a 5-ml synthetic glycerol medium with 100 μM copper for induction and grown at 30°C at 200 rpm until the OD₆₀₀ reaches 1.0. As the cells are in the growth phase, 10 nM concentration of MitoTracker Red CMXRos was added to cells, and cells were allowed to grow further for 30 min. Afterward, the cells were pelleted, washed, and dissolved in 1 ml of water. These liquid cultures were then subjected to flow cytometry with excitation and emission of 579, and 599 nm, respectively, using a BD LSRFortessa Flow Cytometer instrument. For each culture, 10,000 events were recorded to estimate the MitoTracker Red CMXRos fluorescence.

Supplementary Materials

This PDF file includes:

Figs. S1 to S6

Tables S1 to S3

References

View/request a protocol for this paper from Bio-protocol.