Mitochondrial ribosomal protein 20 (Mrp20) is a component of the yeast mitochondrial large (54S) ribosomal subunit and is homologous to the bacterial L23 protein, located at the ribosomal tunnel exit site. The carboxy‐terminal mitochondrial‐specific domain of Mrp20 was found to have a crucial role in the assembly of the ribosomes. A new, membrane‐bound, ribosomal‐assembly subcomplex composed of known tunnel‐exit‐site proteins, an uncharacterized ribosomal protein, MrpL25, and the mitochondrial peroxiredoxin (Prx), Prx1, accumulates in an mrp20ΔC yeast mutant. Finally, data supporting the idea that the inner mitochondrial membrane acts as a platform for the ribosome assembly process are discussed.
Mitochondrial ribosomes have several similarities with their bacterial ancestors, in particular with respect to their protein composition and antibiotic sensitivities (Ban et al, 2000; Sharma et al, 2003; Smits et al, 2007). The mitochondrial ribosomes, however, synthesize only a small subset of proteins (eight in Saccharomyces cerevisiae). These are mainly integral membrane proteins, including essential proteins of the mitochondrial oxidative phosphorylation (OXPHOS) system (Borst & Grivell, 1978). Understanding of the assembly and function of mitochondrial ribosomes is thus important, because defects in mitochondrial‐translation abilities underlie many human diseases that result in impaired OXPHOS capacities (O'Brien, 2002; Chrzanowska‐Lightowlers et al, 2011).
Data indicate that all protein synthesis within the mitochondria is performed by membrane‐anchored ribosomes, to ensure a close coupling between the protein‐synthesis and membrane‐insertion events (Spithill et al, 1978; Liu & Spremulli, 2000; Stuart, 2002; Fiori et al, 2003; Gruschke et al, 2010). By contrast, the bacterial ribosomes synthesize many diverse proteins—both soluble and integral membrane—and consequently both cytosolic and membrane‐bound populations of ribosomes exist, and presumably cycle, within bacteria. How, and at what stage during the formation of translationally active ribosomes, mitochondrial ribosomal proteins become tethered to the inner membrane is unclear.
The molecular details of the mitochondrial ribosome‐assembly pathway and the assembly factors involved are mostly uncharacterized. One report indicates that a late‐stage assembly intermediate of the large (54S) ribosomal subunit, termed the pre54S particle, completes its assembly at the inner membrane, where it incorporates the mitochondrial ribosomal protein (Mrp)L32, proteolytically matured by the Yta10/Yta12 membrane protease (Nolden et al, 2005). Although it is thought to be a matrix‐soluble assembly intermediate, it has not yet been shown whether the assembly of the pre54S complex actually occurs in the matrix. It is also possible that the inner membrane, the ultimate site of ribosomal activity, functions as a platform to localize these assembly events.
In this study, we have focused on a large ribosomal subunit component termed the Mrp20 protein. Previous work using chemical crosslinking approaches has identified a network of proteins—Mrp20, MrpL4, MrpL40, MrpL27, MrpL22, MrpL13 and MrpL3—that are located near to the exit site of the polypeptide tunnel of the 54S particle in yeast (Jia et al, 2009; Gruschke et al, 2010). This network includes contact points for the association of the ribosomes with components of the inner membrane, Oxa1 and Mba1, that are involved with the membrane insertion of nascent chains (Jia et al, 2003, 2009; Ott et al, 2006; Gruschke et al, 2010). Mrp20 is homologous to the bacterial ribosomal protein L23 and it contains an extended carboxy‐terminal region unique to its mitochondrial relatives, termed mitochondrial‐specific or ‘mitospecific’ domain. Indeed, several mitochondrial ribosomal proteins differ from their bacterial ancestors by containing such mitospecific extensions, and the function of these diverse additional domains remains uncharacterized (Sharma et al, 2003; Smits et al, 2007).
Here, we demonstrate that the truncation of the mitospecific domain of Mrp20 has severe consequences for the assembly of translationally active ribosomes, and hence for the capacity of the cell to support OXPHOS activity and aerobic growth. Mitochondria harbouring the mutant Mrp20 protein accumulate a new membrane‐bound ribosome‐assembly subcomplex, enriched in proteins located near the exit site of the ribosomal polypeptide tunnel. On the basis of our findings, we propose that the inner mitochondrial membrane acts as a stage for ribosome‐assembly events.
Results and Discussion
The Mrp20ΔC mutant has an OXPHOS defect
In addition to its L23 homology domain (residues 84–176), Mrp20 contains an extended C‐terminal mitospecific domain (residues 176–263; Fig 1A). To investigate the functional relevance of this mitospecific domain, a yeast mutant (mrp20ΔC) harbouring a C‐terminally truncated derivative of Mrp20 was created. The Mrp20ΔC protein has the final 23 residues removed, a deletion that was selected because it compromises a predicted α‐helical domain at the C‐terminus of Mrp20 (Fig 1A). Haploid mrp20ΔC yeast cells had a severe growth defect on glycerol‐containing media, a non‐fermentable carbon source, demonstrating their impaired ability to support aerobic respiration (Fig 1B). In organello translation in the presence of [35S]methionine indicated that the mrp20ΔC mitochondria had a reduced ability to synthesize the mitochondrially encoded proteins, which include essential components of the OXPHOS complexes (Fig 1C). Failure to efficiently synthesize these proteins hinders the assembly, and thus the levels, of the OXPHOS complexes. Consistently, a strong reduction in the steady‐state levels of OXPHOS‐complex subunits was also observed in the mrp20ΔC mitochondria (Fig 1D). The levels of subunit 1 of the cytochrome c oxidase (Cox1), the Rieske FeS protein (of the cytochrome bc1 complex) and ATP synthase subunit e—all OXPHOS‐complex subunits, known to be susceptible to proteolytic turnover when the assembly of their respective complexes is prevented—were found to be drastically reduced in the mrp20ΔC mitochondria, similarly to that observed in the wild‐type ρ0 mitochondria (Fig 1D). The ρ0 yeast strains are deficient in mitochondrial ribosomes and protein synthesis because they lack mitochondrial DNA (mtDNA) and thus the coding capacity for ribosomal RNA (rRNA).
The mrp20ΔC mutant has a ribosome‐assembly defect
The steady‐state levels of the Mrp20ΔC protein in the mrp20ΔC mitochondria were reduced relative to the wild‐type protein, indicating that the C‐terminal truncation might have caused instability of the Mrp20 protein (Fig 2A). When the expression of the Mrp20ΔC protein in the mrp20ΔC mutant was placed under the control of the strong GAL10 promoter (GAL10‐Mrp20ΔC), the steady‐state level of the Mrp20ΔC protein was comparable with that of the full‐length Mrp20 in wild‐type mitochondria (Fig 2A). Thus, elevated production of the mutant Mrp20ΔC protein compensated for its proposed proteolytic instability. Overexpression of the Mrp20ΔC protein in this manner did not overcome the defective translational activity of the mrp20ΔC mitochondria, and appeared to exacerbate it (Fig 1C). Consistently, the steady‐state levels of the OXPHOS subunits remained severely compromised after overexpression of the Mrp20ΔC protein, and aerobic growth of the mutant remained severely affected (Fig 1B,D). Thus, taking the GAL10‐Mrp20ΔC results together, we conclude that the translational defects observed in the mrp20ΔC mutant are not caused by the limiting levels of the Mrp20ΔC protein per se, but are instead directly caused by the consequences of the C‐terminal truncation of the Mrp20 protein. It remains unclear why the GAL10‐Mrp20ΔC mutant was more compromised in growth and translational activity than the mrp20ΔC mutant, despite the non‐limiting levels of the Mrp20 protein.
Due to the severely impaired mitochondrial translation, the assembly state of the ribosomes in the GAL10‐Mrp20ΔC mitochondria was analysed (Fig 2B). In the wild‐type, the assembly of the large ribosomal (54S) subunit was indicated by co‐sedimentation of its components, MrpL40, Mrp20, MrpL36 and Mrp49, towards the bottom of the gradient (Fig 2B). The small ribosomal (37S) subunit, as indicated by Mrp10, was recovered in fractions immediately before those containing the 54S proteins. In the GAL10‐Mrp20ΔC mutant, the ribosomal proteins were all recovered towards the top of the gradient, indicating that they failed to assemble into a stable large ribosomal subunit (Fig 2B). Monitoring Mrp10 indicated a defect in the assembly of the small ribosomal subunit as well, suggesting that assembly of the large and small ribosomal subunits might not be independent processes. We conclude that truncation of the Mrp20 protein causes a defect in ribosome assembly, which in turn might account for the impaired translational abilities of these mitochondria.
Analysis of the levels of other ribosomal proteins in the mrp20ΔC and GAL10‐Mrp20ΔC mitochondria indicated that, with the exception of MrpL40, they were similar, or elevated, in comparison with those of the wild‐type mitochondria. This was in contrast to ρ0 mitochondria, in which the levels of these ribosomal proteins are reduced because they are highly susceptible to proteolytic turnover in the absence of rRNA (Fig 2A). Yeast mutants defective in mitochondrial protein synthesis have been reported to lose their mtDNA with high frequency—that is, to change from ρ+ to ρ0/ρ− status (Myers et al, 1985)—and thus the normal steady‐state levels of the ribosomal proteins in the mrp20ΔC mutant were unexpected, given their lack of translational activity. When tested by genetic crossing, the mrp20ΔC mitochondria were, however, demonstrated to contain their mtDNA; they were ρ+ (Fig 2C). Furthermore, both 21S and 15S rRNA were present in the mrp20ΔC mitochondria, although the level of 21S rRNA was slightly reduced when compared with the wild‐type (supplementary Fig S1 online). Thus, the observed proteolytic stability of ribosomal proteins in the mrp20ΔC mitochondria might reflect that they have achieved at least a partial assembly state in the presence of the rRNA.
The levels of MrpL40 were decreased in the mrp20ΔC mitochondria, but elevated in the GAL10‐Mrp20ΔC mitochondria (Fig 2A). MrpL40 is closely associated with the Mrp20 protein in the assembled ribosome (Jia et al, 2009; Gruschke et al, 2010), and these data suggest a dependence of MrpL40 on the Mrp20 protein—but not its extreme C‐terminal region—for its stability.
Characterization of the Mrp20ΔC subcomplex
Taken together, our data indicate that truncation of the mitospecific domain of Mrp20 leads to a defect in ribosome assembly and hence translation. Due to the observed stability of the mtDNA, rRNA and ribosomal proteins, but the lack of stable 54S (or pre54S) particles, we propose that the defective stage of ribosome assembly in the mrp20ΔC mutant is downstream from that in ρ0 mitochondria—in which ribosomal protein instability is observed—and upstream from the Δyta10 (or Δyta12) mutant—in which the accumulation of stable pre54S particles is observed (Nolden et al, 2005). Thus, we conclude that the mrp20ΔC mutant represents a new type of ribosome‐assembly mutant.
Sucrose‐gradient centrifugation indicated that the detergent‐solubilized Mrp20ΔC protein was present in a complex of approximately 150 kDa, which co‐migrated with MrpL40 but separately from MrpL36 (Fig 2D). To identify further components of this Mrp20ΔC subcomplex, we expressed a C‐terminally His‐tagged derivative, Mrp20ΔCHis, in the mrp20ΔC strain (Fig 3A) and performed affinity purification using Ni‐NTA (Ni‐nitriloacetic acid) beads (Fig 3B). Several of other proteins were observed to specifically copurify with Mrp20ΔCHis (Fig 3C). The indicated silver‐stained bands were excised from the gel, and subjected to tryptic digestion and mass spectrometry analysis (supplementary Table S1 online). The analysis indicated these proteins to be MrpL40, MrpL4, MrpL25 and MrpL27, all large ribosomal proteins. In addition, one minor, but reproducible, band at approximately 28 kDa, was identified as peroxiredoxin 1 (Prx1), a mitochondrial peroxiredoxin protein. By using available antibodies, the specific recovery of MrpL40, MrpL4 and Prx1 with the Mrp20ΔCHis protein was verified (Fig 3D). Recovery of members of this complex, such as MrpL40, with the purified Mrp20ΔCHis, reflected their physical association in the mitochondria before the lysis step, and was not due to their nonspecific association following detergent extraction (supplementary Fig S2 online). The association of MrpL25 and MrpL27 with the Mrp20ΔC subcomplex was independently tested by expressing His‐tagged derivatives of these proteins in cells harbouring the non‐tagged Mrp20ΔC protein (Fig 3E). The Mrp20ΔC, MrpL4 and MrpL40 proteins specifically copurified with both the MrpL25His and MrpL27His proteins from the mrp20ΔC mitochondria (Fig 3F). The ribosomal proteins MrpL36, Mrp7 and Mrp49 were not found in association with the Mrp20ΔCHis, MrpL25His or MrpL27His proteins, indicating that the Mrp20ΔC subcomplex was limited in its composition and does not contain all of the large ribosomal subunit proteins (Fig 3D,F). The Mrp20ΔC‐subcomplex members Mrp20, MrpL4, MrpL40 and MrpL27 share a location within the 54S particle—the exit site of the ribosomal tunnel (Gruschke et al, 2010)—suggesting that this subcomplex might represent a true 54S assembly intermediate. MrpL25 is a mitochondria‐specific ribosomal protein of unknown location in the ribosome (Smits et al, 2007), but its association with the Mrp20ΔC subcomplex suggests that it might be located near the tunnel exit site. MrpL36 and Mrp7 are homologues of bacterial L31 and L27 ribosomal proteins, respectively, large subunit components located at the small ribosomal subunit interface; far from the tunnel exit site (Smits et al, 2007). Their absence from the Mrp20ΔC subcomplex indicates that different regions of the large ribosomal subunit might independently assemble before the completion of the 54S particle. An intact mitospecific domain of Mrp20 is not required for the formation of the Mrp20ΔC subcomplex. The presence of an intact mitospecific domain of the Mrp20 protein, however, seems to be crucial for the interaction of this Mrp20 subcomplex with an assembly factor and/or with specific components of another ribosome assembly intermediate required to ensure the next step in the 54S assembly pathway.
Finally, the proposed association of Prx1 with the Mrp20ΔC subcomplex was further analysed (Fig 3G). A C‐terminally His‐tagged Prx1 derivative, Prx1His, copurified with the Mrp20ΔC, MrpL4 and MrpL40 proteins, further verifying the specificity of the Mrp20ΔCHis results and the presence of Prx1 in the Mrp20ΔC subcomplex. Interestingly, other large ribosomal proteins—Mrp7, MrpL36 and Mrp49—were also specifically recovered with Prx1His, suggesting that Prx1 can associate with more than one ribosomal subcomplex, and that it is not exclusively associated with the Mrp20ΔC subcomplex. Prx1 and its cytosolic homologue, Tsa1, have antioxidant‐protective roles and Tsa1 can function to prevent the aggregation of ribosomal proteins under stress conditions (Trotter et al, 2008; Greetham & Grant, 2009). The association of Prx1 with diverse ribosomal subcomplexes indicates that it might function as an assembly chaperone or prevent oxidative damage during the assembly of ribosomal subcomplexes.
Mrp20ΔC subcomplex is tethered to the inner membrane
The mrp20ΔC mutant accumulates a new ribosomal‐assembly subcomplex composed of tunnel‐exit‐site proteins. To address whether this complex is membrane tethered in the absence of stable 54S assembly, mrp20ΔC mitochondria were subjected to a mild sonication treatment under low‐salt conditions, followed by low‐speed centrifugation to recover the mitochondrial membranes (Fig 4A). The ribosomal proteins, such as Mrp20ΔC, MrpL40 and MrpL36, all cofractionated with the inner‐membrane marker, the ADP/ATP carrier protein and in a salt‐sensitive manner. The soluble mitochondrial matrix protein Cpr3 was recovered exclusively in the supernatant fraction, indicating that the sonication conditions successfully disturbed the integrity of the mitochondrial membranes. The absence of ribosomal proteins in the supernatant indicated that the non‐assembled ribosomal proteins in the mrp20ΔC mitochondria are not soluble proteins, but that they are membrane tethered (Fig 4A). Recovery of the ribosomal proteins in the pellet fraction was dependent on their association with the membrane, as when detergent was added to solubilize the membrane, the ribosomal proteins were released together with ADP/ATP carrier protein into the supernatant. A similar membrane‐association behaviour of non‐assembled ribosomal proteins was also observed in the ρ0 mitochondrial sample (Fig 4A). In contrast to the full‐length Mrp20 protein in ρ0 mitochondria, a fraction of the Mrp20ΔC protein was recovered in the supernatant. Thus, the membrane association of Mrp20 might be partly compromised by truncation of its mitospecific domain (Fig 4A). When the membrane‐bound fraction of Mrp20ΔCHis was detergent extracted, it was found to be in association with components of the subcomplex, for example, MrpL40 and MrpL4 (supplementary Fig S3 online). We therefore conclude that the Mrp20ΔC protein associated with the membrane is present in its subcomplex form.
In an independent approach, membrane flotation was used to demonstrate the association of ribosomal proteins with membranes of mitochondria isolated from the mrp20ΔC and the late‐stage ribosome assembly mutant, Δyta10 strains (Fig 4B). Following sonication, mitochondrial samples were placed at the bottom of a sucrose gradient and subjected to ultra‐centrifugation (Fig 4B). Soluble proteins—for example, Cpr3—are retained at the bottom of the gradient, whereas membrane‐associated proteins—for example, ADP/ATP carrier protein—float up in the gradient. Flotation of ribosomal proteins with the inner membrane was observed in both the mrp20ΔC and Δyta10 samples, in a manner that was similar to that reported previously for assembled ribosomes from wild‐type mitochondria (Prestele et al, 2009; Fig 4B).
On the basis of the ρ0, mrp20ΔC and Δyta10 data, we conclude that the association of these ribosomal proteins with the inner membrane is not dependent on the assembly of these proteins into a 54S (or pre54S) particle. We therefore propose that the process of mitochondrial ribosome assembly occurs on the surface of the inner membrane, which is the site of ultimate translational activity. Staging the assembly process on the membrane circumvents the need to target an assembled matrix‐localized large ribosomal subunit to the membrane when translation is required. It is possible that certain membrane proteins or lipids might form specialized ribosome‐assembly sites in the membrane, to which ribosomal proteins become targeted after their import into mitochondria.
Subfractionation of mitochondria by sonication. The mitoplasts (400 μg) were resuspended in 600 μl buffer A (20 mM HEPES‐KOH, pH 7.4, 40 mM KCl (or 160 mM KCl, as indicated), 1 mM DTT and 1 mM PMSF), subjected to sonication on ice with a Branson sonifier 250 microtip (45% amplitude, five pulses of 6 s each interrupted by 90 s interval). For differential centrifugation analysis, the resulting sample was centrifuged for 10 min at 9,271g to give rise to a membrane pellet and supernatant fractions. For membrane flotation analysis, the sample was adjusted to 2.2 M sucrose (final volume 1 ml) and layered with (each 1 ml) 2, 1.5 and 1 M sucrose (in buffer A), and centrifuged for 16 h at 162,502 g at 2 °C (Beckman MLS‐50 rotor).
Ni‐NTA purification of His‐tagged proteins. Mitochondria (200 μg protein) were lysed in 200 μl of lysis buffer (20 mM HEPES‐KOH, pH 7.4, 160 mM KCl, 1 mM PMSF and 1% Triton X‐100) and Ni‐NTA purification of His‐tagged protein was performed, essentially as described previously (Jia et al, 2009).
Further methods are available as supplementary information online.
Supplementary information is available at EMBO reports online (http://www.emboreports.org).
Conflict of Interest
The authors declare that they have no conflict of interest.
We thank V. Strogolova and M. Otero for valuable technical assistance. We are grateful to Drs M. Boguta, C. Grant, T. Mason, M. Ott and A. Tzagoloff for gifts of antisera. The research was supported by National Science Foundation grants MCB 0347025 and MCB 0744067 to R.A.S.
- Copyright © 2011 European Molecular Biology Organization