Functional Analyses of Mitoribosome 54S Subunit Devoid of Mitochondria-Specific Protein Sequences

Graphical Abstract

A) Structure of the 54S subunit of the mitoribosome. Mitochondria-specific proteins are colored in yellow and those with bacterial homologs are in blue. The 21S rRNA is in gray. B) The mitochondria-specific extensions present in the conserved proteins are now colored in red. The extensions of uL3(N), uL4(C), uL5(C), uL13(C), bL17(C), bL19 (C), bL21(N), uL22(N), uL22(C) bL27(C), and uL30(N) showed in the panel represent some of the mitochondria-specific extensions studied in this work.

Take away:

• →Study of the functional relevance of mit-specific extensions of mLSU proteins

• →Seven mit-specific extensions are not necessary for yeast respiratory growth.

• →Removal of extensions from 14 mtLSU resulted in unfunctional variants.

In Saccharomyces cerevisiae, mitoribosomes are composed of a 54S large subunit (mtLSU) and a 37S small subunit (mtSSU). The two subunits altogether contain 73 mitoribosome proteins (MRPs) and two ribosomal RNAs (rRNAs). Although mitoribosomes preserve some similarities with their bacterial counterparts, they have significantly diverged by acquiring new proteins, protein extensions, and new RNA segments, adapting the mitoribosome to the synthesis of highly hydrophobic membrane proteins. In this study, we investigated the functional relevance of mitochondria-specific protein extensions at the C-terminus (C) or N-terminus (N) present in 19 proteins of the mtLSU. The studied mitochondria-specific extensions consist of long tails and loops extending from globular domains that mainly interact with mitochondria-specific proteins and 21S rRNA moieties extensions. The expression of variants devoid of extensions in uL4(C), uL5(N), uL13(N), uL13(C), uL16(C), bL17(N), bL17(C), bL21(24), uL22(N), uL23(N), uL23 (C) uL24(C), bL27(C), bL28(N), bL28(C), uL29(N), uL29(C), uL30(C), bL31(C), and bL32(C) did not rescue the mitochondrial protein synthesis capacities and respiratory growth of the respective null mutants. On the contrary, the truncated form of the mitoribosome exit tunnel protein uL24(N) yields a partially functional mitoribosome. Also, the removal of mitochondria-specific sequences from uL1(N), uL3(N), uL16(N), bL9(N), bL19(C), uL29(C), bL31(N), did not affect the mitoribosome function and respiratory growth. The collection of mutants described here provides new means to study and evaluate defective assembly modules in the mitoribosome biogenesis process.

Mitoribosomes are ribonucleoprotein structures present in mitochondria of most eukaryotes organisms and are responsible for the synthesis of polypeptides necessary for the assembly of oxidative phosphorylation (OXPHOS) complexes. Similar to the respiratory complexes, the assembly of mitochondrial ribosomes depends on the coordination of nuclear and mitochondrial DNA (mtDNA) gene products.

The Saccharomyces cerevisiae mitoribosome is associated with the inner mitochondrial membrane (Ott et al., 2006; Prestele et al., 2009); it is composed of a 54S large subunit (mtLSU) and a 37S small subunit (mtSSU). The two subunits altogether contain 73 mitoribosome proteins (MRPs) and two ribosomal RNAs (rRNAs) (Desai et al., 2017). In S. cerevisiae, the 37S protein Var1p and the two rRNAs are mitochondrially encoded, while 72 proteins are nuclearly encoded (Terpstra et al., 1979; Tzagoloff and Myers 1986). Several auxiliary proteins are also required for mitoribosome assembly, including those necessary for rRNA processing, modification and folding, MRP-processing enzymes, GTPases, RNA helicases, and chaperones (Barrientos et al., 2003; Datta et al., 2005; Paul et al., 2012; De Silva et al., 2013; De Silva et al., 2017; Hillman and Henry, 2019). Although ongoing efforts in several labs over the last few years have improved our understanding of mitoribosome biogenesis (Zeng et al., 2018), its assembly process is still poorly understood compared to the assembly of the bacterial ribosome (Shajani et al., 2011). Mitoribosome biogenesis involves the transcription, processing, and modification of rRNA; the translation, proper folding, and modification of MRPs; proper folding of rRNA; and the binding and release of assembly factors (Shajani et al., 2011). In yeast, the process is assisted by chaperones and regulatory proteins, including three mitochondrial GTPases: Mtg1p, Mtg2p, and Mtg3p (Barrientos et al., 2003; Datta et al., 2005; Paul et al., 2012) and two DEAD-box helicases, Mrh4p and Mss116p (De Silva et al., 2013; De Silva et al., 2017). Also, the protein Mam33p binds and chaperones large subunit proteins to ensure proper mitoribosome assembly (Hillman and Henry, 2019). Many of these assembly factors have human homologues, such as DDX28 (Mrh4p), GTPBP5 and GTPBP10 (Mtg2p), GTPBP7 (Mtg1p), and p32 (Mam23p), involved in the assembly of the large subunit (Tu and Barrientos, 2015; Maiti et al., 2018, Maiti et al., 2020, Tobiasson et al., 2021), and C4orf14 (Mtg3p) involved in the assembly of the small subunit (He et al., 2012), which have been described in the last decade.

Mitoribosomes preserve some similarities with their bacterial counterparts; for example, the catalytic cores of the peptidyl transferase and decoding centers are conserved, resulting in the sensitivity to antibiotics such as chloramphenicol and erythromycin (Tzagoloff, 1977); however, one billion years after the endosymbiotic origin of the mitochondria, they have diverged greatly by acquiring new proteins, protein extensions and new RNA segments (Amunts et al., 2014; Desai et al., 2017; Mears et al., 2006). High-resolution cryo-EM structures of the mitoribosomal subunits have detailed the structural divergence from their bacterial ancestors (Brown et al., 2014; Desai et al., 2017). Most of the mitochondrion-specific proteins are located on the mtLSU surface, conferring to the mitoribosome a unique shape distinct from all other types of ribosomes (Amunts et al., 2014; Desai et al., 2017). The ratio content of proteins to rRNA is 1:1 in the yeast mitoribosome, 2:1 in the human mitoribosome, while in the bacterial counterpart, it is 1:2 (Brown et al., 2014). Overall, the yeast mitoribosome is 30% larger than the bacterial one and has an altered architecture of the polypeptide exit tunnel (Amunts et al., 2014). Along with the mitoribosome structure and on its exit tunnel, new mitochondrion-specific proteins, and new protein extensions of bacteria-like MRPs were selected along evolution, adapting the mitoribosome to the synthesis of highly hydrophobic membrane proteins, and consequently, the formation of the catalytic core of the OXPHOS complexes (Greber et al., 2014). In the assembly process, mitochondrion-specific membrane-binding is also required for proper anchoring of newly transcribed 21S rRNA to the inner membrane, where assembly of the mtLSU proceeds (Zeng et al., 2018). The inner membrane protein Mba1p and Mrx15p are mitoribosome receptors with overlapping roles (Möller-Hergt et al., 2018); Mba1p associates with 21S rRNA and aligns mitoribosomes with Oxa1p and the membrane insertion machinery (Pfeffer et al., 2015). In the human mitoribosome mL45, the homolog of Mba1p is an integral subunit of the mitoribosome, and together with OXA1L, it catalyses the delivery of newly synthesized mitochondrial polypeptides forming constriction sites that limit helix formation of the nascent chain (Itoh et al., 2021).

The mtLSU contains 46 proteins and the 21S rRNA; the large subunit catalyzes peptide bond formation during protein synthesis in the peptidyl transferase center, tRNA binding at the central protuberance, translational factors recruitment at L7/L12 stalk, the L1 stalk involved in RNA movement, and, finally, the tunnel exit for the growing nascent polypeptide chain. Thirty-seven proteins are essential for functional mtLSU assembly, three of which are critical for mtLSU 21S rRNA stability (Zeng et al., 2018). Like for its bacterial counterpart, mtLSU biogenesis involves the hierarchical assembly of proteins and pre-assembled modules (Zeng et al., 2018). Twenty-six proteins share homology with bacterial ribosomes, 22 of which have exclusive mitoribosome extensions at either the C-terminus (C) or the N-terminus (N) (Amunts et al., 2014).

These mitochondria-specific extensions consist of long tails and loops extending from globular domains associated with other protein extensions, mitochondria-specific proteins, and the 21S rRNA (Amunts et al., 2014). For instance, in the mtLSU central protuberance, the typical 5S rRNA core is missing, and the structure is remodeled with mitochondria-specific proteins (mL38, mL40, and mL46), 21S rRNA expansion clusters, and mitochondria-specific extensions from uL5, uL16, bL27, bL31, and bL33. In the tRNA exit-site, the extension in bL33 creates a different mechanism for tRNA stabilization. Finally, the bacterial typical tunnel exit is blocked by uL23 mitochondria-specific extension, which together with uL22 and uL24 extensions defines the boundaries of the mitoribosome tunnel exit (Amunts et al., 2014).

In this study, we investigated the functional relevance of mtLSU mitochondria-specific protein extensions of 19 proteins conserved in bacteria. Seventy MRP allele variants were constructed, 28 of them with a single mitochondria-specific extension removed from the coding sequence. From these 28, 20 variants failed to complement the corresponding null mutant strain, and 8 -uL1(N), uL3(N), uL16(N), bL9(N), bL19(C), uL29(C), bL31(N)- are fully functional.

1. Material and Methods

1.1. Plasmid Constructions

Nineteen genes and their allelic variants were cloned after PCR amplification. Primers were designed based on ClustalW alignments of the studied mitoribosome mtLSU components with bacterial counterparts. After PCR amplification, DNA fragments were cloned into YIp349 (McStay et al., 2013) or into YCp22NATP9 (Guedes-Monteiro et al., 2019) in which the gene was cloned in-frame to Neurospora crassa nuclear ATP9 mitochondrial targeting sequence (MTS) (Table S2).

1.2. Yeast Strains

The recombinant plasmids were used to transform the respective collection of mitoribosome null mutants (Zeng et al., 2018). The final strains are listed in Table S1. The compositions of the media used for the cultivation of yeast have been described elsewhere (Moda et al., 2016).

1.3. Mitochondrial Protein Synthesis

Mitochondrial gene products were labeled in whole cells with a mixture of [³⁵S] methionine and [³⁵S] cysteine (7 mCi/mmol) in the presence of cycloheximide. Unless otherwise indicated, yeast cells cultures were grown at 30 °C in rich galactose media. Near the end of the logarithmic phase, the cultures were centrifuged and washed twice with water, followed by additional 2-hours incubation in minimal galactose media. Cells were diluted in 0.5 mL at A₆₀₀ = 3, incubated for 10 minutes in the presence of 70 ng/μl of cycloheximide, and finally labeled with 5 μCi of labeling reagent containing 73% L-[³⁵S]-Methionine, 22% L-[³⁵S]-Cysteine, and 5% miscellaneous amino acids -American Radiolabeled Chemicals, Inc.– (Saint Louis, MO). Paromomycin sensitivity was assessed by a previous 1-hour incubation of the diluted cells in the presence of 15 mM of paromomycin salt.

Cells were lysed in 1.85M NaOH, 1M β-mercaptoethanol, 10 mM phenylmethylsulfonyl fluoride, and cold methionine-cysteine mixture were added to stop the labeling. Proteins were precipitated with 25% trichloroacetic acid and suspended in 45 μL of Laemmli buffer (Laemmli (1970). 15 μL of the radiolabeled protein extract was loaded on 12.5% polyacrylamide gel, 6 M Urea – Synth (Diadema, Brazil), representing therefore 1/3 of proteins of the started volume of cells. Radiolabeled proteins were transferred to a nitrocellulose membrane – Bio-Rad (Hercules, CA) and visualized by exposure to X-ray Amershan Hyperfilm^™ ECL – GE Healthcare (Buckinghamshire – England).

1.4. Network and Interaction analysis

The network of interactions of mtLSU ribosomal proteins was performed with Gephi (gephi.org) (Bastian et al., 2009). In this Gephi-network, edge thickness represents the solvent accessible surface buried in the interface, as calculated previously (Amunts et al., 2014). Modularity was calculated using the parameters: random, edge weights, and 0.8 of resolution. The modularity with resolution was 0.422, and the program found 12 communities. Hydrogen bonds, cation-Pi, and Pi-stacking interactions present in the studied mitochondria-specific extensions subunits were manually annotated based on the already identified structural and modeled interactions described in pdb: 5MRC (Desai et al. 2017) and available at https://www.rcsb.org/3d-view/5MRC/1.

1.5. Mitochondria isolation and fractionation

Yeast mitochondria were prepared by the method of Faye et al. (1974) using Zymolyase 20T –Zymo-Research (Irvine, CA) to obtain spheroplasts. Four mg of mitochondrial preparation were solubilized in 400 μL of an extraction buffer (20 mM Hepes pH 7.4, 25 mM KCl, 0.5 mM PMSF, 0.8% Triton X100, 5 mM EDTA, or 0.5 mM MgCl₂) and were centrifuged at 27,000 × g for 15 minutes and applied onto a linear sucrose gradient 0.3–1.0 M, with the same constitution of the correspondent extraction buffer. The linear sucrose gradients were centrifuged for 3 h at 40,000 rpm in the 55Ti Beckmann rotor – Beckmann Coultier (Brea, CA) (De Silva et al., 2013).

1.6. Miscellaneous procedures

Standard methods were used for plasmid manipulations and transformation of yeast (Schiestl and Gietz, 1989). Protein concentrations were determined by the method of Lowry et al (1951). Antibodies against mS37, Atp6, Cor1 and Cox1 were kindly provided by Dr. A. Tzagoloff (Columbia University – New York) anti-Prx1 by Dr. Luis Netto (IB-USP, Sao Paulo/Brazil) (Gomes et al., 2017); anti-uL-29 by Dr. Martin Ott (Stockholm University - Sweden) (Gruschke et. al., 2010) anti-uL3, anti-uL16, anti-uL24, and anti-bL31 were described elsewhere (Zeng et al., 2018); anti-Porin was obtained from Invitrogen (Waltham, MA). All reagents, unless otherwise indicated, were obtained from Sigma-Aldrich (Saint Louis, MO).

2. Results

Twenty-three MRPs from mtLSU 54S contain mitochondria-specific extensions that are not conserved in bacteria or cytosolic counterparts. Here we studied 19 of these MRPs through new allelic variants expressing proteins with different combinations of extensions. MRPs were aligned to their bacterial counterparts with ClustalW for primer’s designing (Figure S1). After PCR amplification, MRPs alleles fragments were cloned into YIp349 when the original N-terminus remained intact and thus retained the endogenous mitochondrial targeting sequence (endoMTS); or into YCp22NATP9 if the original N-terminus was truncated, and therefore the allele was cloned in-frame to the NcATP9 MTS (Table S2). Recombinant plasmids were sequenced and used to transform MRPs null mutant collection (Table S1) harbouring suppressor plasmids to diminish mtDNA loss (Zeng et al., 2018). The frequency of petite formation of each transformant was evaluated by genetic crosses with tester ρ⁰ strains after overnight growth in rich media. Non-covalent interactions of mitochondria-specific protein extensions with other mitoribosome proteins and with the 21S rRNA were analyzed based on yeast mitoribosome structure (PDB: 5MRC) (Desai et al., 2017) (Figure 1A, Table S3). The following sections are divided based on the gene product distribution and the Gephi-network of interactions of the large subunit mitoribosomal proteins described elsewhere (Amunts et al., 2014) (Figure 1B).

Figure 1 -.

A) Mitoribosome Large Subunit Structure (PDB: 5MRC – Desai et al., 2017), proteins relevant to this present study are coloured as indicated. B) Interaction network of mtLSU ribosomal proteins. Communities were detected using Gephi (see Materials and Methods). Communities containing single proteins are represented in white. Edge thickness represents the solvent-accessible surface buried in the interface (Amunts et al., 2014).

2.1. Growth properties and translation capacities of yeast deletion strains expressing the wild-type and truncated alleles of the corresponding gene

We assessed the growth on non-fermentable ethanol-glycerol media of the yeast deletion strains expressing the wild-type and truncated alleles of the corresponding gene. Translation of newly synthesized mitochondrial products of the different strains was subsequently assessed using cells grown in rich galactose media. In all tested mutants, the complete absence of growth on ethanol-glycerol media correlated with an evident impairment on the respective translation capacity. Altogether, these results indicate that the respiratory deficiency is due to a translation defect of the tested strains.

2.1.2. Exit tunnel proteins and interactors bL17, uL22, uL23, uL24, uL29, and bL32

Mitoribosome has a considerably remodelled polypeptide exit tunnel, mainly by the presence of mitochondria-specific extensions in proteins bL17, uL22, uL23, and uL24, which create a passageway that the nascent chain must cross to reach the inner mitochondrial membrane.

Extensions present in these proteins are responsible for blocking the bacterial exit tunnel and forming the walls of the new tunnel present in the yeast mitoribosome (Amunts et al., 2014). These combined changes adapted the mitoribosome exit tunnel for the transport of highly hydrophobic cargo. In the Gephi-network interactions analysis, these proteins are in different communities (Figure 1B): uL22 (blue), uL23, uL24, and uL29 (grey).

Different combinations of mitochondria-specific extensions from bL17, uL22, uL23, uL24, and uL29 were cloned to verify their essentiality for mitochondrial translation and mitoribosome biogenesis.

The null mutant for uL22 (Δmrpl22) was transformed with the constructs uL22–18, uL22–19, uL22–61, uL22–72, and uL22–84 (Table S1). The mature uL22 found in the mitoribosome starts at residue 85 (Ser), and the mitochondria-specific 77 residue-long N-terminus was removed in uL22–18 and uL22–19. In uL22–72, the endoMTS was replaced by NcATP9 MTS. This mitochondria-specific large N-terminus is necessary to block the bacterial exit tunnel path and form the typical mitoribosome tunnel (Amunts et al., 2015). Structural analyses have indicated at least twelve hydrogen bonds of uL22 mitochondria-specific N-terminus with other tunnel components (in parenthesis, the identified number of non-covalent interactions of each chain) uL23 (1), uL24 (3), the membrane-facing protein mL44 (7), and the 21S rRNA (1) (Table S3, Figure S2). The 39 residues of uL22 mitochondria-specific C-terminus were tested in the uL22–18 and uL22–84 constructs. This portion of uL22 is part of the membrane-facing protuberance (Amunts et al., 2015), and interacts through at least seventeen hydrogen bonds with 21S rRNA (4), mL43 (4), mL44 (2), mL58 (2), and uL13 (5) (Table S2). These connections indicate that uL22 extensions are part of the evolutionary remodelled structure of the mitoribosome and essential for its function. Indeed, constructs uL22–18, uL22–19, uL22–72, and uL22–84 (Table S2) expressing truncated versions of uL22 were unable to complement the uL22 null mutant (Figure 2B), presenting a complete impairment of mitochondrial translation and an elevated rate of petite formation (>95%) from the cells grown in rich galactose, a non-selective media for the retention of the petite suppressor plasmids (Figure 2C). As a positive control, the wild-type construct uL22–61 complemented the null mutant.

Figure 2 -.

A) uL22, uL23, and uL24 protein structures obtained from PDB -5MRC – in blue the conserved region and in red the indicated mitochondria-specific extension. B) Growth properties of uL22, uL24, and uL24 constructs. After overnight growth, cells were spotted on rich glucose (YPD) and rich non-fermentable ethanol-glycerol (YPEG) media. Plates were photographed after two days of growth C) Newly synthesized mitochondrial translation products were analysed from the indicated strains as described in the Material and Methods section. The radiolabeled bands corresponding to the mitochondrial gene products are marked in the margins as indicated: subunits 1 (Cox1), subunit 2 (Cox2), subunit 3 (Cox3) of cytochrome c oxidase; subunit 6 (Atp6), subunit 8 (Atp8) and subunit 9 (Atp9) of ATP synthase; cytochrome b subunit (Cyt b) of ubiquinol cytochrome c reductase; the ribosome Var1 protein. Note that uL23 samples shown in this display were run together with bL27 samples, shown in figure 5, and thereof the wild type control presented with uL23 and bL27 (Fig5) is the same. Membranes were also probed against anti-porin antibody for loading control.

The null mutant for uL23 (Δmrp20) was transformed with constructs uL23–20, uL23–21, uL23–22, uL23–23, uL23–52, and uL23–85 (Table S2). The uL23 mature protein found in the mitoribosome starts at residue 55 (Asn). Therefore, the mitochondria-specific N-terminus containing 27 residues was evaluated in uL23–21, while in uL23–23, the fusion with NcATP9 was conducted at position D46. The removal of the N-terminus in uL23–21 results in the loss of five non-covalent interactions with mL41 (2), mL67 (1), and the 21S rRNA (2) (Table S3). C-terminus functionality was tested in uL23–22 and uL23–85 with the removal of the last 89 C-terminus residues. uL23–20 expressed a truncated allele with both mitochondria-specific sequences removed. Growth and translation assays indicate that both extensions are essential for uL23 function (Figure 2B). The 89 long mitochondria-specific uL23 C-terminus has at least fourteen non-covalent interactions with uL29 (3), uL22 (7), and with 21S rRNA (4) (Table S3, Figure S2), blocking the path of the bacterial tunnel (Amunts et al., 2015). The C-terminal mitochondrial-specific domain of uL23 was already described as essential for the assembly of the mitoribosomes (Kaur and Stuart, 2011).

The construct expressing the wild-type version of uL23 (uL23–52) complemented the uL23 null mutant, as expected. Curiously, the allele uL23–23 has 9 extra residues after the NcATP9 MTS and showed partial complementation of the null mutant respiratory deficiency (Figure 2C). This partial rescue indicates that the extra residues at the N-terminus influence uL23 function or maturation. Nevertheless, the uL23–23 construct was selected for further phenotypic characterizations along with other respiratory competent mutants.

The last analyzed constituent of the mitoribosome tunnel exit was uL24. The null uL24 mutant (Δmrpl40) was tested with constructs uL24–79, uL24–80, uL24–81, and uL24–82. In these alleles, the C-terminus was removed in uL24–80, the N-terminus in uL24–81, and both in uL24–82. The mitochondria-specific N-terminus contains 60 residues and participates in the blockage of the bacterial tunnel exit path. For instance, the aromatic residue F33, together with uL22 F240, narrows the tunnel diameter (Amunts et al., 2015); it forms two hydrogen bonds with uL29 and uL23, one with bL34 and eleven with 21S rRNA. The mitochondria-specific uL24 C-terminal extension covers a long distance on the 54S subunit surface (Figure S2), forming at least 33 hydrogen bonds with bL28 (7), uL29 (7), mL41 (10), mL59 (5), and the 21S rRNA (4) one cation-Pi interaction with uL29 and one Pi-stacking with 21SrRNA (Table S3). Indeed, the physical interaction of uL24 C-terminus with mL41 was previously reported (Gruschke et al., 2010). On the one hand, removal of the mitochondria-specific uL24 N-terminus in uL24–81 allowed the transformed mutant slow growth on non-fermentable carbon sources (Figure2B) as well as the accumulation of newly synthesized mitochondrial products (Figure 2C). On the other hand, removing the mitochondria-specific C-terminus (uL24–80) did not restore the growth of the null mutant, generating an elevated rate of petite formation (>95%). The wild-type construct (uL24–79) complemented the null mutant tested phenotypes.

The protein bL17 is not part of the mitoribosome tunnel exit boundaries, but its mitochondria-specific C-terminus is located close to the bacterial peptide exit tunnel (Amunts et al., 2014). The null mutant for bL17 (Δmrpl8) was transformed with constructs bL17–12, bL17–13, bL17–57, and bL17–83 (Table S1). Only bL17–57 (wild-type gene) was able to rescue the respiratory growth of bL17 null mutant (Figure 3B) and produce newly synthesised products in the labelling experiment (Figure 3C). The mutants devoid of N-extension (bL17–13) or C-extension (bL17–83) generate high numbers of petite cells after overnight growth in galactose-rich media (>95%). Therefore, the six residues at the N-terminal and the 120 residues at C-terminal mitochondria-specific extensions (Figure 3A) are necessary for bL17 function. bL17 N-terminal connects with at least four hydrogen bonds with 21S rRNA (Table S3), while its mitochondria-specific C-terminal contains unstructured segments and form a complex compact domain with a four-helix bundle fold that extends interactions with bL32 in the vicinity of the bacterial exit tunnel. The amino acid residue K150 together with bL32 C120 coordinates a Zn²⁺ ion in this region (Figure S2); hydrogen bonds with bL19 (3), bL32 (3), uL3 (1), and 21S rRNA (8) are also observed (Table S3).

Figure 3 -.

A) bL17, uL29, and bL32 structures obtained as detailed in Figure 2. B) Growth properties bL17, uL29, and bL32 spotted on rich glucose (YPD) and rich non-fermentable ethanol-glycerol (YPEG) media processed and analysed as in Figure 2 C) Mitochondrial newly synthesized products from the indicated constructs were analysed as in Figure2.

Another protein containing mitochondria-specific extensions in the edges of the tunnel end is uL29 (Figure S2), being positioned close to the co-translational membrane insertion machinery (Desai et al., 2017). This protein presents a long mitochondria-specific N-terminus extension, which compensates in part the absence of 21S rRNA helix 63 (Amunts et al., 2014). The first 40 residues of this extension are unstructured, performing ten hydrogen bonds and two Pi-stacking interactions with 21S rRNA. A second part of the uL29 N-terminus extension is intricately connected by hydrogen bonds with 21S rRNA (3), uL23 (5), uL24 (4), bL32 (1), mL41 (3), and mL67 (7) (Table S3). On the other hand, uL29 C-terminus sticks out of the large subunit without any noticeable interaction with the mitoribosome structure (Figure S2), but cross-linking assays have previously demonstrated that uL29 interacts with Mba1p (Gruschke et al., 2010), a membrane-associated mitoribosome receptor (Ott et al., 2006). Constructs uL29–31 and uL29–32 that are devoid of the mitochondria-specific C-terminus were able to rescue the respiratory growth of the null mutant (Figure 3B) and produced reasonable amounts of newly synthesised labelled products (Figure 3C); inversely, uL29–34 misses the mitochondria-specific N-terminus, did not rescue the respiratory growth of the null mutant, and produced an elevate rate of petites after overnight culture in rich galactose media (>95%) (Figure 3B). uL29–54 expresses the wild-type gene (Figure 3A).

bL32 is a substrate of mAAA-proteases and is also present in the grey Gephi-network community (Figure 1B) along with uL23, uL24, and uL29. The bL32 null mutant (Δmrpl32) was transformed with bL32–41, bL32–42, and bL32–64 constructs. The mature protein starts at residue 72 and ends at 183. The deletion of 71 N-terminal amino acid residues of the MTS completely abolished the import of bL32 into isolated mitochondria (Nolden et al., 2005). This protein contains a mitochondria-specific C-terminus starting at Q141 in an α-helix structure, presenting non-covalent interactions with mL67(5), uL23(1), and uL29(1). The last 20 residues are unstructured, presenting at least seven hydrogen bonds, one cation-Pi interaction with 21S rRNA, and one hydrogen bond with uL22. Furthermore, in this region, residue K178 establishes non-covalent interactions with bL32 residues present in the conserved region of the protein, which is necessary for the coordination of a Zn²⁺ ion in the finger domain. The null mutant transformed with the truncated constructs bL32–41 and bL32–42 did not support growth on non-fermentable carbon sources (Figure 3B), despite having a reasonable rate of mtDNA maintenance after overnight culture in rich galactose media (~80%). In both constructs, the endogenous MTS at the N-terminus was replaced by NcATP9 MTS at position 70. The removal of the mitochondria-specific C-terminus in bL32–41 did not produce a significant difference compared to bL32–42, which retained this segment; the wild-type construct present in bL32–64 complemented the null mutant respiratory deficiency (Figure 3B). The absence of complementation by bL32–42 indicates that the truncation at position 70 and the replacement by NcATP9 MTS is not functional, and the protein needs its endogenous sequence for proper mitochondrial transport and maturation. Interestingly, both bL32–41 and bL32–42 mutants are able to accumulate newly synthesized mitochondrial products (Figure 3C) and present some albeit poor growth on ethanol-glycerol registered after seven days of incubation (Figure S3). The bL32 processing sequence is required for proper mature bL32 folding in post-translocational processing mediated by the m-AAA protease (Bonn et al., 2011). The bL32 folding process is also sensitive to oxidative stress, impairing respiratory growth (Bonn et al., 2011). The relevance of bL32 C-terminus in mitochondrial translation has also been studied before (Woogeng and Kitakawa, 2018).

2.1.3. Proteins uL4, uL13, bL28, and uL30

Adjacent to the exit tunnel are membrane-facing proteins such as mL43, mL44, mL50, mL57, and mL58. These proteins interact with mitochondrial inner-membrane proteins Mba1p and Mrx15p, which orient the insertion of newly synthesized polypeptides to the Oxa1p insertion machinery (Möller-Hergt et al., 2018; Pfeffer et al., 2015). uL4 locates in the exit tunnel entrance, and it is part of Gephi-network community coloured in light-green (Figure 1B), along with uL15, bL33, and bL35.

In the conserved region, uL4 functions as a molecular clamp together with uL24 to stabilize a 21S RNA segment (Geffroy et al., 2019). uL4 has a 93 amino acid-long mitochondrion-specific C-terminus that is involved in only five non-covalent interactions with other components of the mitoribosome, i.e., two with mL50, one with mL49, and another two with 21S rRNA (Table S3). Two constructs were obtained with different truncations at the C-terminus; the first construction uL4–77 deleted the mitochondria-specific extension, did not complement the null mutant, and showed an accumulation of petites after overnight culture in rich galactose media (>95%) (Figure 4B). The uL4–78 construct had removed only the last 10 residues at the C-terminus. The shorter C-terminus removal in uL4–78 resulted in partial rescue of the respiratory growth compared to the wild-type version obtained with the allele uL4–76, and an intermediate rate of mtDNA maintenance after overnight culture in rich galactose media (~50%) (Figure 4B) as well as comparable amounts of radiolabelled mitochondrial products (Figure 4C).

Figure 4 -.

A) uL4, uL13, bL28, and uL30 structures obtained as detailed in Figure 2. B) Growth properties of the indicated strains spotted on rich glucose (YPD) and rich non-fermentable ethanol-glycerol (YPEG) media processed and analysed as in Figure 2 C) Mitochondrial newly synthesized products from the indicated constructs were analysed as in Figure2.

In the blue Gephi-network interaction community (Figure 1B), uL13 interacts with mitochondria specific mL43 and tunnel-exit component uL22. It participates in the early events of the mitoribosome assembly pathway (Zeng et al., 2018). Although short, mitochondria-specific uL13 extensions present several intricate interactions with different large subunit proteins. uL13 N-terminus extension has only thirteen residues, and at least six hydrogen bonds with mL43(2), and the tunnel exit constituent uL22(4); while the 27-long C-terminus extension has at least 8 non-covalent interactions with uL3(3), mL43(1), mL44(3) and 21S rRNA(1) (Table S3). The constructs containing truncated versions of uL13 were unable to complement the null mutant and accumulated high rates of mtDNA loss after overnight culture in galactose-rich media (>95%). Altogether, both extensions are essential for uL13 function in mitoribosome biogenesis; the wild-type allele was tested in the uL13–62 (Figure 4B).

Also connecting with a tunnel exit component, bL28 is clustered along with bL9 in the pink Gephi-network community (Figure 1B); and presents C-terminus and N-terminus mitochondria-specific extensions, both interact with 21S rRNA and the tunnel exit component uL24 mitochondria-specific C-terminal extension (Table S2). The N-terminus is unstructured, presenting at least 22 non-covalent interactions with uL24 (4) and the 21S rRNA (18), ten of them are hydrogen bonds based on the backbones in mitochondria-specific residues. The mitochondria-specific C-terminus sticks out of the mitoribosome structure (Figure S2), nevertheless still presents at least eleven hydrogen bonds with uL24 (2) and 21S rRNA (9). Five different constructs were tested for the bL28 null mutant. In bL28–29 and bL28–30, the constructs are devoid of the mitochondria-specific N-terminal, while in bL28–27 the C-terminus was eliminated. None of these tested constructs was able to complement the bL28 null mutant (Figure 4B), and all produced a high rate of petites (>95%) after overnight culture in rich galactose media. The construct bL28–28 abolished only the endogenous MTS, which corresponds to the first 21 residues; this construct was able to grow on non-fermentable carbon source media as well as to accumulate newly synthesized mitochondrial products (Figure 4C). The wild-type allele was expressed in the bL28–63 construct.

Finally, uL30 was clustered in the orange Gephi-network interaction community with close interaction with the blue component bL21 (Figure 1B). uL30 has a mitochondria-specific C-terminus that starts connecting with mL60 and an α-helix presenting at least seven hydrogen bonds with 21S rRNA. The last 10 residues consist of an unstructured segment showing a close connection with bL21 (6 hydrogen bonds) and mL61(1), all of them involving backbones segments of the two proteins (Table S3, Figure S2). This extension showed to be essential for mitoribosome biogenesis, as construct uL30–35 did not complement the null mutant and produced a high rate of petites for the labelling experiment (>95%) (Figure 4B); this truncated version of the protein misses the last 30 residues. Constructs uL30–36, which removed only the two first residues, and uL30–65 (wild-type version) were able to complement the growth deficiency of the null mutant and its mitochondrial translation capacity (Figure 4C).

2.1.4. Central Protuberance components uL5, uL16, bL27, and bL31

The Central Protuberance facilitates communication between 54S functional sites; it is necessary for tRNA binding and coordinated movement along with the 37S small subunit. The conserved proteins uL5, uL16, bL27, and bL31 have mitochondria-specific extensions and were expressed in recombinant plasmid constructs and analysed here. In the Gephi-network interactions analyses, these proteins are clustered in the dark green community (Figure 1B).

The uL5 N-terminus mitochondria-specific extension undergoes several non-covalent interactions with at least 15 bases of 21S rRNA, and with six different proteins: uL16 (3), bL27 (2), bL31 (1), mL38 (2), mL40 (2), and mL46 (4) (Table S3, Figure S2). The network community centrality (Figure 1B), and the intricate connections of the uL5 N-terminus with other components of the central protuberance, indicate its importance for the mitoribosome structure. The null mutant for uL5 (Δmrpl7) was transformed with the truncated version devoid of the specific N-terminus uL5–7, without the endoMTS corresponding to the first 19 residues replaced by NcATP9 MTS. This construct was unable to complement the null mutant and showed mtDNA loss (>95% petites) after overnight growth on rich media. The wild-type allele was cloned in the uL5–56 construct, and the correct mitochondrial targeting was tested in uL5–86. As expected, both constructs rescued the respiratory growth (Figure 5B) and the translation capacity of the transformed cells (Figure 5C).

Figure 5 -.

A) uL5, uL16, bL27, and bL31 structures obtained as detailed in Figure 2. B) Growth properties of the indicated strains spotted on rich glucose (YPD) and rich non-fermentable ethanol-glycerol (YPEG) media processed and analysed as in Figure 2. C) Mitochondrial newly synthesized products from the indicated constructs were analysed as in Figure2. Note that bL31 samples shown in this display were run together with uL30 samples, shown in figure 4, and thereof the wild type control presented with bL31 and uL30 (Fig4) is the same.

The uL16 null mutant (Δmrpl16) was transformed with uL16–8, uL16–9, and uL16–60; in the constructs uL16–8 and uL16–9, the endogenous MTS was replaced by the NcATP9 MTS. The protein segment between residues K38 to P43, present in the mature wild-type protein, is missing in both constructs. The mitochondria-specific 44 residues long at the C-terminus were removed in the construct uL16–8, which was unable to complement the null mutant (Figure 5B) and showed a high rate of petite production after overnight growth on rich media (>95%). The remaining 6 residues in the N-terminus contain fifteen hydrogen bonds with 21S rRNA, but this small segment was not essential for uL16 function, as observed in uL16–9 growth (Figure 5B). The C-terminus has at least 12 non-covalent interactions with 21S rRNA (4), with the protein bL27 (5), and the aforementioned uL5 (3) mitochondria-specific N-terminus (Table S3, Figure S2). The uL16 wild-type construct uL16–60, and the N-terminus truncated uL16–9 transformants were able to grow on non-fermentable carbon source media (Figure 5B), as expected, and accumulated radiolabelled newly synthesized mitochondrial products (Figure 5C).

The mitochondria-specific C-terminus of uL16 is intimately connected to bL27, which also has a mitochondrion-specific C-terminus with 282 residues. The bL27 mitochondria-specific extension forms an extended domain, with no homology to known structures (Amunts et al., 2015), which sticks out of the mitoribosome and undergoes at least 54 con-covalent interactions with uL5 (2), uL16 (1), and mL46 (3), but preponderantly interacts with mL38 (21) and 21S rRNA (27 interactions) (Table S3, Figure S2). The bL27 null mutant (Δmrp7) was transformed with bL27–25, bL27–26, and the wild-type version bL27–51. As expected, this long mitochondria-specific C-terminus is essential for respiratory growth, mtDNA maintenance, and mitochondrial translation, as observed in the bL27–25 construct (Figure 5B). The wild-type (bL27–51) and bL27–26 alleles complemented the null mutant and showed normal protein translation (Figure 5C). The bL27–26 has the endogenous MTS replaced by NcATP9.

The last component of the central protuberance analyzed in this study was bL31. The bL31 null mutant (Δmrpl36) was transformed with bL31–37, bL31–38, bL31–39, bL31–40, and bL31–66. The mitochondria-specific C-terminus protrudes into the mitoribosome small subunit with close interaction with uS19 (Figure S2). The N-terminus is unstructured and buried in the 54S large subunit, interacting with 21S rRNA (10), uL5 (4), and bL33 (1). The endogenous MTS was removed in bL31–37 and bL31–38 and replaced by NcATP9 MTS. The bL31 null mutant transformed with constructs bL31–38 and bL31–40 that maintained the C-terminus were able to grow on ethanol-glycerol media, with high stability of mtDNA (>90% of ρ+ cells after overnight culture) and produced reasonable amounts of mitochondrial products (Figure 5C). In bL31–40, the mitochondria-specific N-terminus was also removed, indicating that it is not essential to bL31 function.

2.1.5. Surface proteins uL1, uL3, bL9, bL19, bL21

uL1 is universally conserved and interacts with the 21S rRNA to form the L1 stalk. A flexible L1 stalk is required for tRNA movement (Itoh et al., 2020). uL1 was not resolved in the S. cerevisiae 54S structure (Desai et al., 2017), but it was shown in the Neurospora crassa mitoribosome (Itoh et al., 2020). The uL1 null mutant (Δmrpl1) was transformed with a construct that eliminates its 48-residue mitochondria-specific N-terminus (uL1–1). The deletion strain transformed with this truncated allele grew on non-fermentable media as well as the null mutant transformed with the wild-type allele (uL1–53) (Figure 6B) and presented the same pattern of accumulation of newly synthesized products (Figure 6C). Altogether, these results show that the mitochondria-specific sequence of uL1 is not essential for its function (Figure 6B and 6C). Because uL1 is missing in the PDB 5MRC structure, it is not present in a community of Gephi-network interaction (Figure 1B), and it was not possible to manually annotate its non-covalent interactions.

Figure 6 -.

A) uL1, uL3, bL9, bL19, and bL21 structures obtained as detailed in Figure 2. B) Growth properties of the indicated strains spotted on rich glucose (YPD) and rich non-fermentable ethanol-glycerol (YPEG) media processed and analysed as in Figure 2. C) Mitochondrial newly synthesized products from the indicated constructs were analysed as in Figure 2. Note that uL1 samples shown in this display were run together with bL27 samples, shown in figure 5, and thereof the wild type control presented with uL1 and bL27 (Fig5) is the same

bL9 is not essential for mitoribosome assembly and mitochondrial protein synthesis (Zeng et al., 2018). It is located close to uL1 at the L1 stalk near the E site; its recruitment requires specific L1 stalk conformation that also depends on the peptidyl transferase center folding and absence of Atp25p, an ortholog of the bacterial ribosome silencing factor (Itoh et al., 2020). Nevertheless, we evaluated whether bL9 truncated alleles would interfere in the respiratory capacity of the null mutant, and all tested alleles presented the same growth capacity on non-fermentable carbon source media (Figure 4B), as well as comparable to the wild-type capacity of synthesis of mitochondrial products (Figure 4C). Respiratory growth properties of bL9 mutants remained unchanged when incubated at 37°C (Figure S4). The mitochondria-specific N-terminus starts with a short α-helix followed by an unstructured segment presenting six non-covalent interactions with 21S rRNA (Table S3).

uL3 and bL19 are part of the yellow Gephi-network interaction community (Figure 1B). uL3 is near the ribosomal peptidyl transferase center (Desai et al., 2017). The uL3 null mutant (Δmrpl9) was transformed with the constructs uL3–3, uL3–58, uL3–83, and uL3–84. uL3 has an MTS sequence of 20 residues and mitochondria-specific N-terminus with 60 residues that expand along the mitoribosome surface connecting at least 23 times with uL13 (8), uL14 (1), bL19 (3), mL44 (1), and 21S rRNA helix 0-ES2 (10) (Table S3, Figure S2) (Amunts et al., 2014). This extension showed to be non-essential for respiratory growth as observed in the uL3–3 construct (Figure 6B), which eliminates the entire N-terminus specific extension, mtDNA retention was also elevated (>90% after overnight culture in rich galactose media). The respiratory growth pattern and the accumulation of mitochondrial products are identical to uL3–58, which expresses the wild-type uL3; and to uL3–83, truncated in the middle of the mitochondria-specific sequence, and to uL3–84 with the endoMTS replaced by NcATP9 MTS (Figure 6B and 6C).

bL19 extends along the mitoribosome surface with the mitochondria-specific C-terminus expanding to the intersubunit-bridge connection with R166 and 15S rRNA A351 hydrogen bond (Figure S2) that is not present in the human mitoribosome (Desai et al., 2017). bL19 mitochondria-specific N-terminus Y16 hydrogen bonds with the membrane-facing component mL58 and the 21S rRNA helix 0–ES2 (Table S2 Figure S2) (Amunts et al., 2014). The C-terminus truncated allele bL19–16 showed to be unnecessary for respiratory growth (Figure 6B) and for the mitochondria translation system (Figure 6C). The allele with N-terminus MTS removed (bL19–17) was also able to complement the bL19 null mutant (Figure 6B). All mutant constructs did not show any considerable instability in mtDNA retention (petite formation < 5%). We were unable to obtain a bL19 allele totally devoid of its N-terminal.

bL21 stands next to membrane-facing protein mL43. bL21 is clustered with the blue Gephi-network interaction community (Figure 1B), and it shows an early connection with rRNA 21S in the mitoribosome assembly process (Zeng et al, 2018). bL21 null mutant (Δmrpl49) was transformed with bL21–24, bL21–68, and bL21–75 constructs. The 19-residues long mitochondria-specific N-terminus was removed in bL21–24, which resulted in a transformed strain unable to grow on non-fermentable carbon source media, with high instability of mtDNA retention (>95% petites after overnight culture on rich galactose media) (Figure 6B), as well as an incapacity to accumulate newly synthesized mitochondrial products (Figure 6C), indicating the N-terminal is essential for bL21 function. This short but essential extension has only two interactions with mL58 at residue N53 (Figure S2). Both, the wild-type bL21–48 allele, and the bL21–75 allele with the replaced MTS complemented the null mutant respiratory growth defect (Figure 6B).

2.1.6. Steady-state level of the truncated alleles and mitoribosome assembly

To evaluate the expression of the MRPs truncated alleles, we analyzed the steady-state level of the uL3, uL16, uL24, uL29, and uL31 proteins, a subset selected considering the availability of working antibodies. Mitochondria from yeast strains harbouring the indicated constructs were isolated from cells grown on rich galactose media as described in the methods section. The replacement of the endoMTS by NcATP9 MTS showed to be functional. Indeed, the replacement by the GPD (glyceraldehyde-3-phosphate dehydrogenase) promoter resulted in a higher steady-state level of truncated uL3 products compared to the wild-type strain or the transformant strain harbouring the wild-type gene with its endogenous promoter (uL3–58) (Figure 7A). uL16–8 and uL16–9 constructs have the endogenous MTS replaced by NcATP9 MTS, while uL16–9 complemented the respiratory deficiency of the null mutant, uL16–8 did not, but this last one also had 55 residues of the C-terminus removed. The uL16–9 construct has its N-terminal shorten in 9 residues, a clear band corresponding to WT uL16 size could be observed; cross-reacting bands make it difficult to interpret the uL16–8 product (Figure 7A). uL24–79 contains the wild-type gene, and a band with a similar intensity was observed in the wild-type. uL24–80 is a construct devoid of 114 residues at the C-terminus, and its product could not be detected. In the uL24–81 construct, we substituted the endogenous MTS and its 60 residues long specific N-terminal by the NcATP9 MTS, which resulted in a product with a higher content than the wild-type and an apparent lower molecular weight (Figure 7A). It is plausible to speculate that the complementation of the null mutant by uL24–81 was achieved because of its overexpression. The allele uL29–31 is shortened in 13 residues at the C-terminus, and it is able to complement the respiratory deficiency of the null mutant; again, its product was not clearly detected, perhaps due to the presence of cross-reacting bands. The same problems with cross-reacting bands make it difficult the detection of uL29 protein in uL29–32 and uL29–33 strains. uL29–32 rescues the respiratory capacity since it essentially expresses the wild-type gene with the MTS replaced by the NcATP9 MTS. uL29–33 is 102 residues shortened with the N and C extensions removed. At least, the uL29 product was clearly detected in uL29–34 (89 residues of the N-extension removed) and in uL29–54, which express the wild-type allele (Figure 7A). Finally, for bL31, the steady-state level analyses were more informative than the previous proteins. bL31 was not detected in the construct bL31–39, which misses the N and C extensions – a shorter bL31 product was detected in the respiratory competent harbouring the bL31–40 construct that has 20 residues shorter than the wild-type at the N-extension. Removal of the bL31 C-extension produced a stable protein observed in bL31–37 with a migration a little faster than the wild-type (Figure 7A).

Figure 7 -.

A) Steady-state levels of uL3, uL16, uL24, uL29, and bL31 were evaluated from mitochondrial protein extracts obtained from galactose grown cultures. The respective antibodies used are indicated on the left of each panel, and protein size standard markers are positioned on the right. Membranes were also probed against anti-porin antibody for loading control. B) Steady-state level of mitochondrial proteins from cell extracts obtained from galactose grown cultures of the indicated respiratory competent constructs. The used antibodies are on the right side of the panel and the position of the protein standard markers on the right.

The fact that the uL16 and uL24 proteins are not detected by immunoblotting in cells harboring the uL16–8 and ul24–80 truncated alleles raises the concern about their fate. The negative results could be due to several intercurrences, from limited expression to protein degradation. The most plausible scenario, however, is that the truncated proteins fail to incorporate into the mitoribosome and are subsequently degraded.

The absence of functional antibodies turned more difficult the analyses of the truncated alleles that did not rescue the respiratory phenotype of the wild-type mitochondrial translation properties. It is important to mention that the intercurrences cited above may be occurring in other non-complementing constructs whose steady-state level was not investigated. Therefore, we further investigated those that are respiratory competent for the steady-state levels of uL3, uL4, uL16, bL19, uL23, uL24, uL29, and bL31 using the aforementioned antibodies (Figure 7B). Cells were grown in rich galactose media, and proteins were extracted, and TCA precipitated after overnight growth. The expression of the truncated versions of uL3–83, uL16–9, uL24–81, uL29–31, and bL31–40 was confirmed once again, with stronger bands with lower molecular weight observed because of the extension’s truncation and GPD-NcATP9 fusions. The steady-state level of OXPHOS complexes components was also evaluated, and all mutants present the same level of the tested proteins Atp6p (complex V), Cor1p (complex III), and Cox1p (complex IV). Prx1p level was investigated considering its reported accumulation in uL23 truncated mutants, but its level was not altered in any of the tested truncated mutant proteins (Kaur and Stuart, 2011). Prx1p is a mitochondrial 1-Cys peroxiredoxin that catalyzes the reduction of endogenously generated H₂O₂; the two observed bands (Figure 7B) result from its process state and consequent distribution to the mitochondrial intermembrane space and matrix (Gomes et al., 2017).

2.1.7. Assembly of mature mitoribosome is modified in protein extension mutants

To determine whether the respiratory growth deficiency of a given mutant is due to an assembly problem or to a non-functional assembled mitoribosome, we assessed the mitoribosome assembly status in strains expressing uL16, bL17, uL22, uL23, uL24, and bL31 mutant constructs. Mitochondria were isolated and mitoribosomes extracted in the presence of 5 mM EDTA to dissociate the two subunits (De Silva et al., 2013) and then subjected to sedimentation in a sucrose gradient as previously described (De Silva et al., 2013). We used the uL29 antibody to follow the sedimentation of the 54S subunit, mS37 antibody to follow the 37S subunit, and Prx1 antibody as a sedimentation control. In the process of the 54S biogenesis, uL29 is proposed to assemble during intermediate stages (Zeng et al., 2018). In respiratory-competent strains (WT, uL16–9, uL24–81, and bL31–40) uL29 sediments at the bottom of the sucrose gradient, this corresponds to the 54S subunit that is enriched in fractions 2 and 3 (Figure 8). The mS37, and therefore 37S subunit, is enriched in fractions 4 and 5 in all tested strains. In the respiratory deficient uL22–19 and uL23–22, uL29 did not sediment in the heavier fractions of the gradient, indicating that 54S is not being formed. Unlike uL23–22, which has the C-extension removed, uL23–21 had removed its N-extension, but although respiratory and translationally deficient, mitochondrial extracts from this strain accumulated uL29 in the heavier fractions indicating 54S assembly. bL17–12 has a broad distribution of uL29 from fraction 1 to 8, suggesting that the assembly of the mature 54S is being hampered and intermediates are being accumulated. uL29 was also detected in lighter fractions (10–14), together with a cross reacting product of higher molecular weight. Prx1 migrates on the lighter fractions of the gradient as expected. Triton was used as the detergent in the extraction process and may have disrupted Prx1p association with uL23 and the mitoribosome (Kaur and Stuart, 2011); nevertheless, Prx1p is soluble in the intermembrane space and has a weak association with the mitochondria inner membrane at the matrix side (Gomes et al. 2017).

Figure 8 -.

Assembly of the mitoribosome 54S and 37S subunits were assessed from uL16–9, bL17–12, uL22–19, uL23–21, uL23–22, uL24–81, and bL31–40. Mitochondrial proteins were extracted in the presence of EDTA and sedimented in a linear sucrose gradient as described in the Methods section and depicted in the top of the panel. Fractions were collected from the bottom (1) to top (14). The western blots were assayed for anti-uL29 (54S component), anti-mS37 (37S component) and anti-Prx1p (soluble – sedimentation control). The asterisk in the anti-uL29 panel indicates the predicted uL29 product. Prx1p is shown in two bands, which are dependent on its processing state (Gomes et al., 2017).

2.1.8. Mitochondrial translation capacity in strains expressing constructs that restored respiratory competency

To better characterize the functionality of the mitochondrial translation machinery of the respiratory competent strains, we performed additional growth tests and translation assays. First, the transition from fermentative growth to respiratory growth was assessed through the spot of cells at early logarithmic growth (A₆₀₀ = 0.5) on rich ethanol-glycerol media, and growth was photographed 24 hours of the spot, and therefore the adaptation to the respiratory media can be qualitatively evaluated and distinguished from those that have a slow growth on ethanol-glycerol. Confirming several growth tests, uL23–23, uL24–81 and uL29–31 present a general slow growth on ethanol-glycerol, while uL4–78 and bL31–40 growth were particularly delayed when spotted from a fermentatively growing culture, indicating a delay in the adaptation to the diauxic shift (Figure 9A). The respiratory competent mutants were also spotted on rich ethanol-glycerol media and incubated at different temperatures. At a higher temperature of 37°C, the growth of uL23–23 was impacted negatively, while at 18°C its growth is rescued, while uL24–81 is more severely affected (Figure 9B). The identification of more restrictive growth conditions will be useful for screening genetic suppressors of these genes and, therefore, the identification of new participants of the mitoribosome biogenesis.

Figure 9 -.

A) Respiratory growth comparison of the indicated truncated mutants spotted from glucose fermentative growing cultures at early exponential phase and photographed after one day or two of incubation. B) Respiratory growth comparison of the indicated truncated mutants from overnight galactose cultures and spotted at different temperatures. Growth was registered after 48h of incubation. C) Mitochondrial newly synthesized products from mutants grown on the indicated carbon sources were analysed as in Figure 2, and the frequency of rho⁺ cells was scored and indicated in the bottom of the Glucose and Galactose panels.

Mitochondrial newly synthesized products were also analysed from these cells at glucose fermentative logarithmic growth, galactose stationary late-log growth, and ethanol-glycerol late-log phase (Figure 9C). In glucose, the accumulation of mitochondrial products is diminished compared to the wild-type strain in uL4–78, bL19–16, uL24–81, uL29–31, and bL31–40, indicating that lower translation activity can explain the delay in the diauxic shift adaptation. In galactose, uL4–78 newly synthesized products bands are more intense, perhaps triggered by a higher rate of petite formation, estimated at 45% in this experiment (Figure 9C). Finally, in ethanol-glycerol, the newly synthesized products are more evenly distributed without any noticeable change in the accumulation of the mitochondria-encoded polypeptides.

To evaluate the quality of the mitochondrial translation process, we incubated the yeast cells in the presence of paromomycin. This aminoglycoside antibiotic is known to act on the mitoribosome, inducing translational misreading and proteotoxic stress in S. cerevisiae mitochondria (Suhm et al., 2018; Vargas Möller-Hergert et al. 2018). All tested respiratory-competent mutants showed an increased resistance of paromomycin compared to the wild-type strain, with uL23–23, bL31–40, and bL19–16, showing the higher rates of mitochondrial products accumulated compared to the untreated cells (Figure 10, top and bottom panels). The elevated resistance to paromomycin observed in these strains indicated an elevated adaptation of the mitoribosome to the drug, perhaps changing the bridges with the SSU and its rotational capacity during translation or due to accommodation in its structure flexibility by the loss of non-covalent interactions (Table S3).

Figure 10 -.

Analyses of paromomycin resistance of the truncated mutants in the syntheses of mitochondrial products in the presence (+) and absence (−) of the antibiotic. Mitochondrial products were analysed as in Figure 2, and the band intensity of each labelling experiment analysed using the histogram for pixels densitometry of the Adobe Photoshop program. The obtained values were compared in the bar graph shown on the bottom.

3-. Discussion

Mitoribosome have greatly diverged from bacterial ancestors with a clear increment of the protein/RNA ratio. New protein sequences have appeared in different organisms, with the most dramatic example being the mitoribosome of the unicellular parasite Trypanosoma brucei, which contains the smallest known rRNAs and 126 proteins. (Ramrath et al., 2018). Yeast mitoribosome evolution moved toward rRNA expansion, RNA moieties in the mtLSU are substantially larger than in bacteria; consequently, new proteins and protein segments were required for the stabilization of rRNA expansions (Ott et al., 2016). Here, we have evaluated the relevance of MRPs mitochondria-specific extensions present in yeast; detailed non-covalent interaction analyses showed an intricate network of interactions with the 54S proteins and the 21S rRNA (Table S3, Figure S2).

Mitochondrial protein synthesis is highly dependent on functional mitoribosomes. In yeast, impairments of mitochondrial protein synthesis lead to pleiotropic consequences on mtDNA stability (Myers et al., 1985), a condition that has historically hampered functional studies of mitoribosome components. The expressions of a nuclear version of VAR1 and a multicopy plasmid containing RNR1 or YMC2 have overcome this difficulty (Zeng et al., 2018). Here, we took advantage of these suppressors to study the complementation of mitoribosome null mutants by truncated allele variants. The detection of partial complementation in some of the constructs opens the opportunity for further studies of factors involved in the modular and hierarchical assembly of the mitoribosome large subunit (Zeng et al., 2018).

Previous studies of ribosome protein extensions have focused on one or two proteins (Woogeng and Kitakawa, 2018; Tutuncuoglu et al., 2016). For instance, it was showed that deleting extensions of uL30 and uL29 from yeast cytosolic 60S ribosome impairs early and middle steps of 60S biogenesis (Tutuncuoglu et al., 2016).

In this study, we constructed and evaluated 70 allele variants of 19 different components of the 54S mtLSU mitoribosome that have bacterial counterparts but also have mitochondria-specific extensions at the C or N-terminus. Most of these extensions turned out to be necessary for respiratory growth and mitochondrial protein synthesis.

Altogether, 20 constructs with one mitochondrion-specific extension removed from the studied gene failed to complement the respiratory capacity of the respective null mutant (Table 1). Except for uL29-C, these extensions have an intricate network of non-covalent interactions with other components of the mitoribosome, which suggests an essential function for such structures as also previously indicated (Amunts et al., 2014; Desai et al., 2017). On the other hand, uL29 histidine-tagged in the C-terminus cross-linked with the mitoribosome receptor Mba1 and, therefore, a role of uL29-C in facilitating the insertion of newly synthesized products in the inner-membrane can be speculated (Gruschke et al., 2010).

Table 1 –

Summary of the MRPs truncated alleles. The size of the removed extension (Ext/R), the percentage of the sequence remained (% Prot), the respiratory growth (YPEG) and the annotated non-covalent interactions missed in the allele construct are indicated.

Allele	Ext/R	% Prot	YPEG Growth	Annotated non-covalent interactions
uL1–1	N/48	83.1	Normal	Not described previously
uL3–3	N/60	75.9	Normal	21S rRNA(10) bL19(3) mL44(1) uL13(8) uL14(1)
uL4–77	C/93	63.9	Absent	21S rRNA(2) mL49(1) mL50(2)
uL5–7	N/115	57.8	Absent	21S rRNA(15) bL27(2) bL31(1) mL38(2) mL40(2) mL46(4) uL16(3)
bL9–71	N/16	81.5	Normal	21S rRNA (6)
uL13–14	C/27	81.8	Absent	21S rRNA (1) mL43(1) mL44(3) uL3(3)
uL13–73	N/13	91.2	Absent	mL43 (2) uL22 (4)
uL16–8	C/44	77.4	Absent	21S rRNA(4) bL27(5) uL5(3)
uL16–9	N/6	96.9	Normal	21S rRNA(14)
bL17–13	N/6	97.4	Absent	21S rRNA(4)
bL17–83	C/120	48	Absent	21S rRNA(8) bL19(3) bL32(3) uL3 (1)
bL19–16	C/13	91.5	Normal	21S rRNA (1)
bL21–24	N/19	86.6	Absent	mL58(2)
uL22–19	N/77	69.4	Absent	21S rRNA(7) mL44(7) uL23(1) uL24(3)
uL22–84	C/39	84.5	Absent	21S rRNA(4) mL43(4) mL44(2) mL58(2) uL13(5)
uL23–21	N/27	87.2	Absent	21S rRNA(2) mL41(2) mL67(1)
uL23–22	C/89	58	Absent	21S rRNA(4) uL22(6) uL29(3)
uL24–80	C/113	61.8	Absent	21S rRNA(5) bL28(7) mL41(10) mL59(5) uL29(8)
uL24–81	N/60	79.7	Slow	21S rRNA(11) bL34(1) uL23(2) uL29(2)
bL27–25	C	31.7	Absent	21S rRNA (26) mL38(21) mL46(3) uL16(2) uL5(1)
bL28–27	C/120	49.3	Absent	21S rRNA(9) uL24(2)
bL28–30	N/61	74.2	Absent	21S rRNA(18) uL24(4)
uL29–31	C/93	69.4	Normal	None
uL29–34	N/89	70.7	Absent	21S rRNA(15) bL32(1) mL41(3) mL67(7) uL23(5) uL24(4)
uL30–35	C/30	63.8	Absent	21S rRNA(7) bL21(6) mL60(1)
bL31–37	C/69	57.4	Absent	uS19(1)
bL31–40	N/26	83.9	Normal	21S rRNA(10) bL33(1) uL5(4)
bL32–42	C/44	61	Absent	21S rRNA (8) mL67 (5) uL22(1) uL23 (1) uL29(1)

Several explanations can be raised for the negative complementation by the truncated alleles. From expression to incorporation into the mitoribosome, several events can affect the fate of the truncated protein, including its folding, stability, and import into mitochondria. However, their truncated versions can still block the process of mitoribosome biogenesis and therefore promote the degradation of MRPs. Loss of proper mitochondrial localization of the truncated alleles could also explain the absence of respiratory complementation; however, a functional MTS was fused to them and worked well in all positive controls except for bL32–41. The need for the endogenous bL32 mitochondrial targeting sequence to fold mature bL32 explains this negative result (Bonn et al., 2011). Due to the absence of antibodies for all studied proteins, epitope-tagged proteins might be a solution to evaluate their steady-state level. However, tag additions at the N or C-terminus in truncated proteins at the same ends would again constraint the conclusion about its functionality and fate. Although, steady-state analyses of the truncated constructs were limited, some expected results were confirmed, such as the higher expression level derived from GPD promoter fusion (Zampol et al., 2010); changes in size due to truncations, and indeed the absence of some proteins.

The formation of mature 54S subunit (Figure 8) was assessed in a subset of mutants. While the respiratory competent mutants presented a sedimentation pattern like the wild-type, respiratory deficient mutants did not, except for uL23–21(N) that showed the sedimentation of uL29 in heavier fractions and bL17(C), which presented the mature sedimentation of 54S subunit but did not synthesize detectable amounts of mitochondrial gene products.

In the tested truncated mutants, we did not observe a specific deficiency in the translation of a specific mitochondrial gene; i.e., in the respiratory growth deficient mutants, the translation of newly synthesized products was severely impacted, whereas the mutants with positive respiratory growth or even slow growth present normal accumulation of newly synthesized mitochondrial products. In a previous study of bL34 temperature-sensitive mutants, we observed a specific impairment of Cox1p and Cox3p translation (Guedes-Monteiro et al., 2018). bL34 does not contain mitochondria-specific segments and interacts with the tunnel exit components uL23 and uL24. Curiously, temperature-sensitive mutants of mL38 (MRPL35), a component of the central protuberance, showed elevated translation of COX1 but a profound defect in cytochrome c oxidase assembly, while uL29 point mutants along with mam33 deletion do not display an overall inhibition in mitochondrial protein synthesis but demonstrate a problem in COX assembly, particularly in Cox1p translation (Box et al., 2017; Hillman and Henry, 2019). Possibly, these point mutants affect the coordination of the translational process with the chaperoned events in the biogenesis of the respiratory complexes, which would explain the specific changes rather than an overall deficit of all mitochondrial products.

As expected, the long and highly complexed extensions present in uL4(C), uL5(N), uL16(C), bL17(C), uL22(N), uL22(C), uL23(N), uL23(C), uL24(C), bL27(C), uL29(N), uL30(C), bL31(C), and bL32(C) are all essential for mitoribosome assembly and function (Table 1). uL5(N), uL16(C) extensions enter a region occupied by 5S rRNA in bacteria (Amunts et al., 2014), while uL22(N-C), bL27(C), and uL29(N) also compensate for rRNA diminishment and the consequent absence of helix 24 and helix 63. Finally, extensions of uL22, uL23, and uL24 contribute to block the bacterial exit tunnel (Amunts et al., 2014). Unlike the C-terminus, removing the N-terminus extension of uL24 and uL16 allowed partial respiratory competence for the transformed mutants. In bacteria, the deletions of uL23 and uL24 loops lead to significant protein folding changes of newly synthesized polypeptides (Kudva et al., 2018).

Smaller extensions with no more than 30 residues present in bL17, uL13, bL21, and bL31 also showed to be essential (Table 1). bL17 extensions are located close to the bacterial peptide exit tunnel (Amunts et al., 2014) and present several interactions with other components of the mitoribosome (Table S3), particularly to 21S rRNA branches. Curiously, the unstructured N-terminal extension of bL31 shows non-covalent interactions with 21SrRNA and other MRPs, but it is not essential for its function, while the C-terminal α-helix extension reaches the 37S subunit, interacts with uS19, and it is essential for bL31 function. Finally, bL32 C-terminus extension removal leads to a destitute growth on respiratory media (Figure S3), but the same was true for the construct that maintained the C-terminus but had replaced the endogenous MTS by NcATP9. At any rate, bL32 presents a Zn²⁺ coordination domain, which is affected in the bL32–42 construct. The observed bL32–42 poor growth should be further evaluated in future studies to determine the effect of the Zn²⁺ binding domain in protein folding. Oxidative stress can alter the redox state of the cysteines coordinating the Zn²⁺ binding domain and therefore affect bL32 maturation and function (Bonn et al., 2011).

The removal of mitochondria-specific sequences from uL1, uL3, bL9, bL19, or uL29(C) complemented the respiratory deficiencies of the respective null mutant (Table 1); therefore, a regulatory function for these accessory structures can be speculated. uL29 (C) sticks out the mitoribosome and interacts with Mba1p (Gruschke et al., 2010), suggesting a role in the coordination of the synthesis and inner membrane insertion of newly synthesized protein. Both uL3 and bL19 N-terminus extensions interact with 21S rRNA helix 0-ES2 (Amunts et al., 2014). The assembled mitoribosome in these mutants could be further investigated to identify new assembly factors that are being recruited and may somehow compensate and bypass the observed growth deficit, providing new insights into its assembly pathway. The identification of such assembly factors can be pursued using classic suppressor search or cutting-edge cryoelectron microscopy structure of new types of mitoribosome intermediates.

Likewise, the partial growth observed in the transformants uL4–78, uL23–23, uL24–81, bL32–42 provide new means to study and evaluate defective assembly modules in the biogenetic process (Zeng et al., 2018). The mitoribosome truncated proteins obtained in this study may stabilize assembly modules of the mitoribosome that are unstable in the null mutant and, therefore, allow the identification through mass spectrometry putative mitoribosome intermediates (Zeng et al., 2018).

We also observed an elevated resistance to paromomycin in the respiratory competent MRPs mutants. Paromomycin binds to the SSU and induces translational misreading in bacteria and mitoribosome (Wilson et al., 2014; Suhm et al., 2018). The observed resistance to this antibiotic (Figure 10) suggests an increased translational fidelity. Therefore, one could speculate that acquisition of these mitochondria-specific extensions during evolution seems to have resulted in a diminishment in mitochondria translation fidelity. However, a role for ES27L rRNA expansion segment in translation fidelity has been proposed for cytosolic ribosomes, in which the expansion acts as a scaffold for proteins involved in translation accuracy (Fujii et al., 2018).

Finally, the study of mitoribosome proteins provides a better understanding of the development of mitochondria translation deficiencies in human patients. Defects in mitochondrial protein synthesis account for an important subgroup of mitochondrial diseases (Ferrari et al., 2021). Mitochondrial translation impairments lead to a diverse clinical presentation, most severe in newborns. The following mitoribosome proteins have been linked to mitochondrial diseases: bS1, bS16, uS2, uS7, uS9, uS11, mS14, mS22, mS23, mS25, mS34, mS39, ms39, uL3, bL12 uL24 and, mL44 (Pearce et al., 2013; Bugiardini et al., 2019; Ferrari et al., 2021).

In conclusion, the MRPs variants described here provide new perspectives for the study of patients with mitoribosome mutations and offer new tools for the study of their function on the assembly of the mitoribosome.