Involvement of Mitochondrial Ribosomal Proteins in Ribosomal RNA-mediated Protein Folding

Abstract

The peptidyl transferase center of the domain V of large ribosomal RNA in the prokaryotic and eukaryotic cytosolic ribosomes acts as general protein folding modulator. We showed earlier that one part of the domain V (RNA1 containing the peptidyl transferase loop) binds unfolded protein and directs it to a folding competent state (FCS) that is released by the other part (RNA2) to attain the folded native state by itself. Here we show that the peptidyl transferase loop of the mitochondrial ribosome releases unfolded proteins in FCS extremely slowly despite its lack of the rRNA segment analogous to RNA2. The release of FCS can be hastened by the equivalent activity of RNA2 or the large subunit proteins of the mitochondrial ribosome. The RNA2 or large subunit proteins probably introduce some allosteric change in the peptidyl transferase loop to enable it to release proteins in FCS.

Introduction

Discovery of the existence of relics of the RNA world such as “ribozymes,” where the RNA component comprised the active principle and could provide catalytic function by itself in vitro, strengthened a RNA-world hypothesis. A corollary of the RNA world model is that with time RNAs would have to be gradually replaced by proteins as catalysts due to their greater flexibility in conformations that gives them the versatility needed for varied catalytic properties (1–6). Protein catalysts generally have much higher turnover numbers (shorter reaction times) but not necessarily lower K_m values than their RNA counterparts (5, 7). So it may be interesting to find relics of such isofunctional activities in biology.

Ribosomes have been identified as a general protein folding modulator on the basis of their ability to successfully fold all the denatured and newly synthesized proteins studied so far both in vitro and in vivo (8–20). For each unfolded protein, the level of recovery in activity with ribosome is around 80–100%. The protein folding activity has been found to reside in the domain V of the 23 S rRNA in 50 S subunit of the bacterial ribosome and its homolog elsewhere, as in those from plant and animal cytosols. This domain has long been known to harbor the peptidyl transferase center (PTC).⁵ The structure-function relationship of this RNA segment became easier to study when we could split bacterial domain V into the more conserved central loop region (RNA1, the PTC) and the highly variable stem-loop region (RNA2) outside the central loop. The RNA1 transforms an unfolded protein to a FCS, and the RNA2 part is required to dissociate the folding-competent state (FCS) from RNA1 (21).

In mitochondrial ribosome (mitoribosome) in a number of organisms, the conserved central loop of the peptidyl transferase center (corresponding to bacterial RNA1 part) is present, but the variable stem loop region (corresponding to the bacterial RNA2 region) is almost completely deleted. Here we report that the mitoribosomal PTC is an efficient protein folding modulator like the prokaryotic and eukaryotic cytosolic ribosomes, and mitoribosomal proteins compensate for the absence of the RNA2 equivalent. Bacterial ribosomal proteins have no such activity. Furthermore, we show that the RNA2 segment from Escherichia coli rRNA domain V could complement for the contributions from the mitoribosomal proteins in an in vitro folding reaction. A comparison of the regions of some mitoribosomal proteins that bind to its PTC with bacterial homologous proteins showed regions where they match. Interestingly, considerable extra amino acid sequences could be marked in the mitoribosomal proteins. This observation suggests structural and functional replacement of rRNA by proteins in the mitoribosome. In this publication we present experimental results associated with these observations.

MATERIALS AND METHODS

Ethics Statement

The liver samples were collected from the local slaughter house, and use of this sample for our experiments was allowed by the animal ethics committee of the University of Calcutta.

Isolation of Mitochondrial Ribosomes

The preparation of bovine liver mitochondria was based on the procedure of Matthews et al. (22). 500 g of bovine liver was ground in a food processor, diluted with Buffer A (350 mm sucrose, 1 mm EDTA, 5 mm Tris·HCl, pH 7.5), and homogenized in a motor-driven homogenizer. All the steps were carried out at 4 °C unless otherwise mentioned. The homogenate was filtered through cheesecloth and centrifuged at 600 × g for 30 min in a Hitachi RPR 9–2 rotor. The pellet containing cellular and extracellular debris was discarded, and the supernatant was spun at 7300 × g for 45 min as above. The supernatant was discarded, the pellet was resuspended in buffer A, and the procedure was repeated twice. The mitochondrial pellet was resuspended in sufficient buffer A, freeze-thawed, and then treated with 100 μg/ml digitonin for 30 min to remove the outer mitochondrial membrane to prepare the mitoplast. The digitonin-treated resuspension was then centrifuged at 14,500 × g in a Hitachi RPR 20-2 rotor for 30 min. The crude mitoplast pellet was resuspended in buffer A and washed twice. Then the supernatant was spun at 14,500 × g in a Hitachi RPR 20–2 rotor for 30 min at 4 °C. The pellet was then resuspended in an appropriate amount of buffer B (260 mm sucrose, 40 mm KCl, 15 mm MgCl₂, 20 mm Tris·HCl, pH 7.5, 5 mm 2-mercaptoethanol, 0.8 mm EDTA, 50 μm spermine, 50 μm spermidine) and kept in −70 °C until further use. Purified bovine mitoplast was then thawed as required. The amount of protein was estimated by the Lowry method and diluted with buffer C (20 mm Tris·HCl (pH 7.6), 20 mm MgCl_2, 100 mm KCl, 6 mm 2-mercaptoethanol) to a concentration of 20 mg/ml protein. This diluted mitoplast suspension was lysed with 0.1% sodium deoxycholate and 1% Brij-35 and centrifuged for 15 min at 14,500 g in a Hitachi RPR 20-2 rotor at 4 °C. The supernatant was discarded, and the pellet was resuspended in buffer D (20 mm Tris·HCl (pH 7.6), 20 mm magnesium acetate, 10 mm 2-mercaptoethanol, 1 m NH₄Cl) followed by centrifugation as above. The supernatant containing the mitochondrial ribosome was taken and layered over 24% sucrose cushion in buffer C and centrifuged at 230,000 × g in a Beckman TL-Optima 100.2 TLA rotor at 4 °C for 2 h to sediment the crude mitochondrial ribosome. The 55 S ribosomes were further purified by centrifugation in 5–20% linear sucrose gradient in buffer C at 140,000 × g for 2 h in a Beckman SW28 rotor. The fractions were collected, and their absorbance was monitored at 260 nm to detect fractions containing ribosomes. The appropriate fractions were pooled and precipitated, dissolved in required amount of buffer C, and stored at −70 °C until use.

To prepare 39 S and 28 S mitochondrial ribosomal particles, the 55 S mitochondrial ribosomes were dialyzed overnight in buffer E (20 mm Tris·HCl (pH 7.6),100 mm KCl, 6 mm 2-mercaptoethanol), and the dialyzed sample was layered over a 5–20% linear sucrose gradient in buffer E and centrifuged at 140,000 × g for 6 h in a Beckman SW28 rotor at 4 °C. The gradient fractions were collected, and their absorbance was monitored at 260 nm to detect separate peaks due to 39 S and 28 S particles. To avoid cross-contamination, fractions from the heavier side of 39 S peak and lighter side of the 28 S peak were pooled separately and precipitated with chilled ethanol then dissolved in required amount of buffer C and stored at −70 °C until use.

Extraction of Proteins from Mitochondrial Ribosomes

Mitochondrial ribosomes were adjusted directly to a concentration of 16 A₂₆₀ units in buffer C containing 4 m LiCl. The extraction mixture was stirred at 4 °C for 16 h (23) and then centrifuged at 200,000 × g in a Beckman TL Optima 100.2 TLA rotor at 4 °C for 2 h. The supernatant, containing the ribosomal proteins, was concentrated in a Heto Maxi dry lyo lyophilizer and then dialyzed against buffer C containing 1 m LiCl. Small aliquots of this preparation were stored on ice until use. The pellet containing the rRNAs was resuspended in buffer (20 mm Tris·HCl (pH 7.5), 10 mm MgAc, 30 mm NH₄Cl) and gently extracted with 1:1 (v/v) mixture of phenol and chloroform to remove traces of ribosomal proteins. The final volumes of the preparations of the RNAs and the proteins were kept the same to make their molar concentrations equal. In all refolding reactions involving the mitochondrial ribosomal proteins, they were diluted at least 100 times from the stock, and the residual LiCl concentration had no effect upon the enzymatic assay of bovine carbonic anhydrase (BCA). The proteins were run in 15% denaturing polyacrylamide gel.

Isolation of Bovine Mitochondrial DNA

An aliquot of mitoplast suspension was taken and washed twice in buffer (10 mm Tris·HCl (pH 8), 1 mm EDTA). To it 0.1% sodium deoxycholate and 1.2% Brij-35 was added, and it was kept in ice for 30 min. To this an equal volume of water-saturated phenol was added, mixed for 5 min, and centrifuged at 5000 rpm for 15 min at room temperature. The upper aqueous phase was taken. The above step was repeated followed by the addition of equal volume chloroform:isoamyl alcohol (49:1) to the aqueous portion. They were mixed and centrifuged as above, and the aqueous phase was taken again. The DNA in the aqueous phase was precipitated with salt and ethanol, washed twice with 70% chilled ethanol, and the pellet was dried, dissolved in 10 mm Tris·HCl (pH 8), and checked by running in 1% agarose gel. A sharp band was seen by ethidium bromide staining. The position of the band was compared with sequencing ladder corresponded with bovine mitochondrial DNA (16 kbp).

Cloning of Mitochondrial Ribosomal PTC

The DNA segments encoding the bovine mitochondrial ribosomal PTC RNA (450 bases) was amplified using bovine mitochondrial DNA as template. The PCR product was used for cloning into vector pTZ57R/T provided with Ins T/A Clone^TM PCR Product Cloning kit (Fermentas Life Sciences). Primer sequences for cloning bovine mitochondrial PTC were as follows: BMR5′, 5′-gac aag ctt g gtg aaa ttg acc ttc c-3′; BMR3′, 5′-g gaa ttc c tac gta ata gat aga aac cg-3′.

For cloning of Leishmania donovani mitoribosomal PTC (300 bases), PCR was done with the appropriate primers from the total cellular DNA of L. donovani. PCR product was cloned into the transcription vector pGEM4Z under the T₇ promoter. The primer sequences were as follows: 5′-primer, 5′-gag aag aga agc tta aaa aat taa aat agg gca agt cc-3′; 3′-primer, 5′-ttc ccc cgg gaa ttc cct tgc gta cta ata-3′.

The sequences were confirmed by thermosequencing using the α-³²P-labeled primer and the Thermosequenase kit (U. S. Biochemical Corp.). The sequence matched exactly with those of Bos taurus and L. donovani reported earlier (24).

In Vitro Synthesis of Ribosomal RNA

Bovine and Leishmania mitochondrial ribosomal PTC RNA, bacterial RNA1, and RNA2 were transcribed from the plasmids containing the respective rDNA inserts under the T₇ promoter. The plasmid was linearized using restriction enzyme EcoRI cutting at the end of rDNA to obtain templates for transcription. In vitro transcription was carried out in a reaction volume of 100 μl containing the recommended transcription buffer (40 mm Tris·HCl, pH 7.9, 6 mm MgCl₂, 2 mm spermidine, and 10 mm DTT), 1 mm each of RNTP, 1 μg of linearized plasmid as the template, and 50 units of T₇ RNA polymerase (Bangalore Genei). The reaction mixture was incubated at 37 °C for 2.5 h, and the synthesized RNA was precipitated with salt and ethanol. The amount of RNA synthesized was estimated.

Denaturation and Refolding of Proteins

BCAII, lactate dehydrogenase (LDH), porcine heart mitochondrial malate dehydrogenase (MDH), and human carbonic anhydrase I (HCAI) were denatured with guanidine hydrochloride (25–27), and the loss of their secondary structures was verified by CD spectra (19, 21). For all the enzymes 6 m guanidine hydrochloride was used for denaturation. The denatured enzymes were diluted 100 times (final concentration 300 nm) in the refolding buffer (50 mm Tris-HCl, pH 7.5, 100 mm NaCl, 10 mm magnesium acetate) in the absence or in the presence of ribosome, large subunit of ribosome, rRNA of the large subunit, or PTC RNA and were incubated for 30 min at 25 °C to allow folding. The molar concentration of RNA was equal to that of the enzymes. The residual amount of guanidine hydrochloride had no effect on the activities of the enzymes. The activity of refolded proteins was determined following published protocols (25–27) and was expressed as percent of activity of the same amount of native protein.

For measurement of CD spectra, 11 μm BCA was denatured. The concentrations of guanidine hydrochloride and other conditions of denaturation were the same as before. For refolding, the denatured BCA was diluted 80-fold in refolding buffer containing equimolar amounts of PTC RNA and incubated as before. After incubation the RNA was removed by treatment of DEAE cellulose in refolding buffer containing 200 mm NaCl; at this salt concentration the RNA will bind, but the protein will not bind to the DEAE cellulose. It was kept in ice for some time, and the process was repeated. This was then centrifuged, and the supernatant was taken along with the native protein and denatured protein for CD spectral analysis in a Biologic Science Instrument MOS450 CD spectrometer.

Synthesis of Mutant Domain V rRNA

Site-directed mutagenesis was done to introduce point mutations in the E. coli domain V rRNA. Eighteen mutations (U2473C, U2473A, U2491C, U2491G, C2551A, U2552C, A2560G, U2561C, A2560U, A2587G, A2587U, A2059G, A2062G, G2251A, G2251U, G2252A, G2253C, and G2253A) were introduced one at a time, and each mutation was confirmed by sequencing.

Time Course of Release of Refolding Protein from Bovine Mitochondrial rRNA

The release of refolding protein (BCA) from bovine mitochondrial PTC RNA was studied by an electrophoretic mobility shift assay. The PTC RNA was synthesized in the presence of [α-³²P]UTP. After the initiation of the refolding reaction, the samples were irradiated on a glass dish at 254 nm with 4 lamps (GS Gene Linker, Bio-Rad) in parallel at a distance of 6 cm for 2 min on ice at different time points. Irradiated samples were precipitated by salt-ethanol and then washed with 70% ethanol. The samples along with the control set (UV-irradiated RNA only) were electrophoresed on 5% native polyacrylamide gel in 1× Tris borate-EDTA buffer at 10 mA for 3.5 h. The gels were dried at 80 °C for 20 min, and the positions of the bands were visualized by autoradiography using x-ray films.

Interaction of Mitochondrial PTC RNA and E. coli RNA2 with Refolding Proteins (BCA, Lactate Dehydrogenase, MDH, Lysozyme); Primer Extension Analysis with Thermoscript Reverse Transcriptase

The unfolded protein interacts with different sites on PTC RNA at various stages of the refolding process. We wanted to see whether the interaction sites on RNA were specific and in that case, which are the bases interacting with the refolding protein(s). For this purpose, RNA-protein interactions were fixed by UV-cross-linking (described below), and the cross-linked products were used for primer extension studies.

UV Cross-linking of PTC RNA Refolding Protein Complex

At the time of the refolding reaction (in this case the RNA:protein molar ratio was 1:5, which should allow all the PTC RNA to interact with the unfolded protein) 300 μl of sample (∼5 μg of RNA) was irradiated on a glass dish at 254 nm with four lamps (GS Gene Linker) in parallel at a distance of 6 cm for 2.5 min on ice. Irradiated samples were precipitated by salt-ethanol and then washed by 70% ethanol, cross-linked products were checked by gel electrophoresis, and these cross-linked protein-bound PTC RNA were used for reverse transcription extension assay (Primer extension assay).

Primer Extension Assay on Cross-linked Protein-bound PTC RNA

10 pmol of primers for bovine PTC, Leishmania PTC, and E. coli RNA2 were mixed with cross-linked protein-bound PTC or RNA1 and RNA2 of domain V (about 3–5 μg) and kept at 90 °C in water in a beaker for 1 min. The beaker was then left at room temperature to cool down slowly to anneal the primer with the RNA. Annealed primers were labeled by [α-³²P]dCTP by the 3-dNTP method using Thermoscript Reverse Transcriptase (Invitrogen) enzyme at 49 °C for bovine and Leishmania PTC and at 55 °C for E. coli RNA2. Labeled primers were extended at 59 °C after the addition of all the four dNTPs in excess by the same enzyme for about 45 min. Reverse transcriptase (RT) product was precipitated and washed with 70% ethanol. Primer extension assay were carried out using four proteins: BCA, lactate dehydrogenase, malate dehydrogenase, and lysozyme. For the control experiment primer extension assay was carried out on a UV-irradiated domain V only (without any protein). All the RT products were analyzed on a 6.5% sequencing gel containing 8 m urea, next to sequencing ladder. Sequencing of PTC rDNA was carried out using the same primers used for primer extension with Thermo Sequenase DNA Polymerase (Thermo Sequenase Cycle Sequencing kit, U. S. Biochemical Corp.).

RESULTS

Refolding of Proteins by Bovine Mitochondrial and E. coli Ribosomes

BCA was unfolded with 6 m guanidine hydrochloride and diluted 100-fold in the refolding buffer in the absence or in the presence of ribosomes. After incubation for 30 min the activities of the proteins were measured and expressed in terms of the activities of the same amounts of native proteins (see “Materials and Methods“ and the references therein). The activities due to folding in the absence of ribosome were around 25–28% of the native level and was attributed to self-folding. The recovery of enzyme activity increased with increasing concentrations of ribosomes up to a 1:1 ratio of ribosomal particles to enzyme but did not increase with further addition of ribosomal particles, as shown for bovine mitochondrial 55 S ribosome (Fig. 1A). Therefore, optimum folding was achieved when the ribosomes were used at equimolar concentration with the enzymes. With bovine mitochondrial 55 S ribosomes, at a 1:1 stoichiometric concentration of unfolded BCA, the activities of the enzyme increased from 25 to 75–80% with time (Fig. 1D). Also, as we showed earlier with complete ribosome and its large subunit from prokaryotic (Fig. 1, B and F) and eukaryotic cells, the intact 55 S mitochondrial ribosomes as well as its 39 S large subunit (Fig. 1, B and C) are efficient protein folding modulators. Because BCA is not a mitochondrial protein, we repeated the refolding experiments using porcine heart malate dehydrogenase of mitochondrial origin and got similar results (Fig. 1E). Incubation with intact ribosome or its subunits did not change the activities of the native forms of any of the proteins mentioned here (data not shown). We preferred BCA for further studies because its monomeric nature and relatively smaller molecular weight made it easier to compare this with our studies on eubacterial and eukaryotic ribosome-mediated protein folding. Apart from measuring the activity of the protein, the recovery of the tryptophan fluorescence and circular dichroic spectra of the folding BCA to the level of the native protein was monitored (Fig. 1G).

FIGURE 1.

Refolding of BCA and mitochondrial MDH in the presence of ribosomes. A, titration of 55 S ribosomes with 200 nm BCA. B, folding of denatured BCA with 55 S, 39 S, 28 S, 70 S, 50 S, and 30S (self-folding 28%). C, refolding of denatured proteins (BCA, human carbonic anhydrase I (HCA), MDH) in the presence of 55 S ribosomes. Bar graphs represent the mean folding values from three independent experiments. D, time course of refolding of denatured BCA as modulated by 55 S bovine mitochondrial ribosome. E, time course of refolding of denatured MDH as modulated by 55 S bovine mitochondrial ribosome. F, time course of refolding of denatured BCA as modulated by E. coli 70S ribosomes. Gi, CD spectra of native BCA (black line), denatured BCA (dashed line), and BCA after refolding in presence of PTC RNA (dotted line). Gii, gradual increase in intensity of tryptophan fluorescence emission peak after the addition of PTC RNA to denatured BCA in refolding buffer. The tryptophan fluorescence for native BCA and denatured BCA are marked in the figure, and the intermediate lines show the gradual increase of fluorescence with time on the addition of PTC RNA. a.u., arbitrary units.

Refolding of Proteins by Mitochondrial and E. coli rRNA

When refolding studies with mitochondrial large subunit rRNA (Fig. 2, A and B) or its PTC were carried out (Fig. 2, C, D, and E), recovery of activity of the denatured protein was found to be much slower than that with the bacterial PTC RNA (Fig. 2F). The cause of such slow recovery might be due to inefficient release of the refolding intermediate of BCA from the RNA refolding protein intermediate complex (see below). Refolding and release was measured for several hours to verify this. The refolding reached its maximum (around 60% of native BCA activity) after 1.5 h with bovine mitochondrial large rRNA or its PTC and about 2.5 h with Leishmania mitochondrial large rRNA or its PTC, unlike bacterial 23 S rRNA PTC, which took only 12 min to complete the folding (Fig. 2F). Similar results were obtained when we used mitochondrial malate dehydrogenase in place of bovine carbonic anhydrase (Fig. 2C). The extent of folding of the denatured protein was found to be maximum, with complete ribosomal particle (around 80%) and less with the rRNA (∼60%). We have consistently observed this in the case of ribosome and ribosomal RNAs from many sources. This difference could be due to RNA conformation in free state compared with that within intact ribosome. Magnesium was required for this reaction, and folding was found to be optimum at 10 mm magnesium.

FIGURE 2.

Time course of refolding of denatured BCA and MDH in the presence of rRNA. The time course of refolding of denatured BCA as modulated by bovine mitochondrial (BM) large subunit (LSU) rRNA (A) and L. donovani mitochondrial (LM) large subunit rRNA (B). C, time course of refolding of denatured MDH as modulated by bovine mitochondrial PTC (BMDV). Time course of refolding of denatured BCA as modulated by bovine mitochondrial PTC (BMDV) (D), L. donovani mitochondrial PTC (LMDV) (E), and E. coli domain V (DV, F).

Release of Refolding Protein from Bovine Mitochondrial PTC RNA

As only unfolded proteins bind to PTC RNA, we could study the release of the refolding protein from the mitochondrial PTC RNA by electrophoretic mobility shift assay after UV-cross-linking the RNA-protein complex at different times of interaction. In Fig. 3, C1 is the control lane where PTC RNA has been UV-irradiated in the presence of the native protein. It clearly shows that there is no interaction between the native protein and the RNA, C2 shows only UV-irradiated RNA, and other lanes represent UV-cross-linked mitochondrial PTC RNA with the refolding protein at times of interaction 1, 15, 30, 45, 60, 90, and 120 min. The upper bands are due to the protein-RNA complex, and the lower bands are due to free, unbound RNA alone. Results indicate that the unfolded protein bound to the rRNA within 1 min after mixing and that refolding was complete (release from the bovine mitochondrial PTC RNA) in about 90 min (Fig. 3). With increasing time, the loss of the RNA-protein complex and gain in freed RNA is clear from Fig. 3. We have shown earlier that the addition of any other RNA (tRNA, small subunit rRNA, domain II rRNA, etc.) to unfolded BCA or other proteins did not show gel mobility shift upon UV irradiation (17, 27, 28).

FIGURE 3.

Time course of release of protein (BCA) from bovine mitochondrial PTC. Lane C1 is the control lane where PTC RNA has been cross linked to the native protein, and C2 shows only UV-irradiated RNA. Lanes 1–7 show UV-cross linked mitochondrial PTC RNA with the refolding protein at different times of interaction: 1, 15, 30,45, 60, 90, and 120 min, respectively. See “Materials and Methods” for details.

The recovery of the folded form is apparently linear for the first 90 min in the BCA-Bov.mt-PTC RNA interaction (see Fig. 2A), whereas the gel shift shows that most of the protein is released from the RNA protein complex in about 90 min. The discrepancy might be attributed to the contribution from self-folded protein in the population. Even without added rRNA, about 30% of the proteins will spontaneously fold; spontaneous folding is completed in about 15 min. These contributed to the early part of rise in activity with time shown in Figs. 2, A and D. The later rise in activity was from RNA-mediated folding after release of the protein. With 30-min time points taken in these experiments, minor deviation from linearity could be averaged out.

Refolding of Denatured BCA by Bovine Mitochondrial PTC RNA Can Be Accelerated by the Addition of Bacterial RNA2; Functional Complementation of Mitochondrial PTC RNA Deficiency by Bacterial RNA2

As mentioned before, in the mitochondrial PTC, the RNA2 part is severely truncated compared with its bacterial counterpart (Fig. 4, compare A with B and C). We have shown earlier that the bacterial PT loop is responsible for folding the protein to a competent state, and the RNA2 region is required to release it (21). Taking the cue from these, we added bacterial RNA2 in the refolding reaction to see whether it could supplement in the folding of the denatured protein by mitochondrial PTC (21). When an equimolar amount of bacterial RNA2 was added to the mitochondrial PTC, the protein recovered 60% of its activity in about 12 min (Fig. 5A). Release of refolding protein from bovine mitochondrial PTC RNA in the presence of bacterial RNA2 was also studied using the gel assay mentioned above. In the autoradiogram (Fig. 5C), C1 is the control lane where PTC RNA was UV-irradiated with the native protein. It clearly shows that there is no interaction between the native protein and the PTC RNA, C2 shows only UV-irradiated RNA, and the other lanes represent UV-cross-linked mitochondrial PTC RNA with the refolding protein at different times of interaction (0.5, 2, 5, and 16 min) in the presence of bacterial RNA2 (unlabeled). Fig. 5C clearly shows that the protein was released before 16 min.

FIGURE 4.

Secondary structures of PTC of large ribosomal RNA of E. coli (A), bovine mitochondria (B) and Leishmania mitochondria (C).

FIGURE 5.

Refolding and release of protein in presence of bovine mitochondrial PTC/RNA1 and bacterial RNA2 or ethanol. Time course of refolding of denatured BCA as modulated by bovine mitochondrial PTC (MDV) complemented with Bacillus subtilis RNA 2 or ethanol (A) and B. subtilis RNA1 with RNA2 or ethanol (B). C, Time course of release of protein (BCA) from bovine mitochondrial PTC in the presence of bacterial RNA2. C1 is the control lane, where PTC RNA has been cross-linked to the native protein; C2 shows only UV-irradiated RNA; lanes 1–5 show UV-cross-linked radiolabeled mitochondrial PTC RNA with the refolding protein at different times of interaction in presence of bacterial RNA2 at 0.5, 2, 5, 10, and 16 min, respectively. It can be clearly seen that the protein was released before 16 min.

Interaction Sites for the Folding Protein with Mitochondrial PTC RNA

Further experiments were designed to find out the interaction sites of the folding protein(s) with the mitochondrial PTC and the RNA2 region of E. coli ribosomal RNA when they complemented in the protein folding activity. The mitochondrial PTC and the E. coli RNA2 region, when folding the protein, were UV-cross-linked to FCS. The sites of interaction of the FCS with the two RNA were deciphered by extending the ³²P-end-labeled appropriate oligo DNA primers annealed to the two RNAs by Thermoscript reverse transcriptase. The reverse transcription products were blocked at the nucleotides that interacted with the proteins. It was seen that, on both the mitochondrial PTC and the E. coli RNA2, the sites of interactions were the same for all four unfolded proteins (BCA, lactate dehydrogenase, MDH, and lysozyme). A comparison of the nucleotides with the sequencing ladder indicates that there were prominent blocks at nucleotides U1413–G1417 (1), C1408 (2), G1404, U1405 (3), U1391–A1393 (4); U1380–U1383 (5), U1320–U1322 (6), U1302–C1304 (7), and U1289–C1291 (8) in bovine mitochondrial PTC (Fig. 6, A and D). For Leishmania mitochondrial PTC, the blocks corresponded to nucleotide positions U1120–A1122 (1), U1104, A1105 (2), U1091–U1094 (3), U1060 (4), and U1040–C1042 (5) (see Fig. 6, B and E). For E. coli RNA2 the binding sites were at G2221, G2252, A2346, A2358, U2387, G2391, and G2400 (Fig. 6, C and F). There were very few reverse transcriptase pause sites in the control lane C. Thus, presence of native protein during UV irradiation in the absence of unfolded proteins did not inhibit reverse transcription activity, and full-length transcript was the major product in each case.

FIGURE 6.

Primer extension analysis on PTC RNA cross linked to proteins. A, primer extension analysis on bovine mitochondrial PTC RNA after UV irradiation in the absence (lane 1) or presence (lanes 2–5) of proteins BCA, MDH, lactate dehydrogenase (LDH), and lysozyme (LYS), respectively. Lanes 6–9 show the corresponding sequencing ladder. D, secondary structure of the bovine mitochondrial PTC RNA. The boxes include the nucleotides that are identified to be cross-linked with proteins. B, primer extension analysis on Leishmania mitochondrial PTC RNA after UV irradiation in the absence (lane 1) or presence (lanes 2–5) of proteins BCA, lactate dehydrogenase, MDH, and lysozyme, respectively. Lanes 6–9 show the corresponding sequencing ladder. E, secondary structure of the Leishmania mitochondrial PTC RNA. The boxes include the nucleotides that are identified to be cross-linked with proteins. C, primer extension analysis of E. coli RNA2 UV-cross-linked to RNA1 refolding protein complex. Lane 1 is the control lane and shows reverse transcription on UV-irradiated RNA in absence of any protein. Lanes 2–5 shows the corresponding sequencing ladder. Lanes 6–9 shows reverse transcriptase road blocks seen on RNA 2 in the RNA1 refolding protein (BCA, lactate dehydrogenase, MDH, and Lys, respectively) RNA2 complex. F, the secondary structure of E. coli RNA2 shows binding sites with refolding proteins.

This highlights the universality of (i) protein folding by the PTC as well as (ii) the nature of cooperation of the two RNAs in protein folding. The proteins not only bound specifically to mitochondrial PTC, but their interactions with the RNA2 regions were also specific.

From studies on the origin and evolution of mitochondria it is quite obvious that mitochondrial PTC, an RNA fragment of eukaryotic origin, could functionally cross-talk with prokaryotic RNA2, ignoring the species barrier to promote folding. When low concentrations of ethanol (3%) or detergents like Triton X-100 (0.1%) were added in the refolding reaction instead of RNA2, then also the process became efficient (Fig. 5, A and B). This agrees with our observation that in case of bacterial ribosome the folding is done by RNA1 and the role of RNA2 in releasing the folding protein from RNA1 could be substituted by ethanol or Triton X-100 (21). But in the absence of bacterial RNA1 or mitochondrial PTC RNA, when such low concentrations of ethanol (3%) or Triton X-100 (0.1%) were added in the refolding reaction there was no increase in the self- folding activity. This also highlights the fact that the PTC RNA is the folding modulator, and the role of RNA2 is to release the folding-competent protein from the PTC.

Effect of Mutation in RNA1 and RNA2 Regions of Bacterial PTC RNA on Protein Folding

If RNA2 has a secondary effect on protein folding by RNA1 (or PTC of mitochondria), then mutations in the RNA2 should probably be less inhibitory than that in RNA1 for the process. To check this, single base mutations were introduced by site-directed mutagenesis in the RNA1 and RNA2 regions of the E. coli domain V rRNA. The mutants were U2473C, U2473A, U2491C, U2491G, C2551A, U2552C, A2560G, U2561C, A2560U, A2587G, A2587U, A2059G, A2062G, G2251A, G2251U, G2252A, G2253C, and G2253A. We found that mutations in the RNA1 severely inhibit protein folding; that is, the activity of the protein increases from 30% (self folding in the absence of any folding modulator) to around 62% in the presence of wild type PTC RNA but is always within 38–48% in all the different RNA1 mutants, but the effect of mutations in RNA2 on the release of protein folding intermediates is not strikingly affected by mutations as shown in Fig. 7.

FIGURE 7.

Refolding of BCA in the presence of wild type and mutant E. coli domain V rRNA; effect of mutation on folding in bases of RNA1 (gray bars) and RNA2 regions (black bars). Folding is expressed as the percentage gain of native BCA activity.

These experiments establish that the RNA1 (PTC) region of ribosome is indispensable for folding. The role of RNA2 region is to facilitate the release of the folding-competent intermediates. The intermediates are released from RNA1 itself but at a very slow rate (21). The mutations introduced into RNA1 and RNA2 did not influence the stability of the two RNAs to very different extents. This experiment further justifies the fact that mitochondrial PTC, a RNA1 analog, although missing the “RNA2” part, contains all the information necessary to guide the unfolded state of the protein to FCS.

Role of Mitochondrial Ribosomal Proteins

The mitoribosome as a whole or its large subunit could fold proteins efficiently and rapidly, but its large ribosomal RNA or its PTC took a longer time. It seemed reasonable to assume that the mitochondrial ribosomal proteins might have some role in the dissociation of the folding-competent proteins.

To verify this assumption, mitochondrial ribosomal proteins were added in the refolding reaction containing bovine mitochondrial PTC or its large subunit rRNA, and the protein recovered 60% of its activity within 15 min (Fig. 8A). The addition of mitochondrial ribosomal proteins alone showed very poor recovery of BCA activity (only 3–5%), much less than even the self-folding of BCA (27%). Results were similar when the order of addition of the rRNA and the mitochondrial ribosomal proteins was reversed in the refolding reaction of BCA, i.e. the RNA was added after 15 min of incubation of the refolding protein with the ribosomal proteins. The refolding BCA and the mitochondrial ribosomal proteins formed some sort of “irreversible” complex, presumably between largely hydrophobic ribosomal proteins and the unfolded protein. We suggest that the role played by mitochondrial ribosomal proteins in protein folding was similar to bacterial RNA2 (Fig. 5B), and it complemented with its naturally truncated rRNA to release the refolding proteins. This unique effect of ribosomal protein to release folding proteins from the PTC in the cellular cytosol was never seen before in any of our experiments when we added ribosomal proteins to its PTC (data not shown). It should be noted here that mitochondrial ribosomal proteins binding to mitochondrial PTC RNA are much larger in size compared with E. coli ribosomal proteins (Fig. 8B) (29). This substitution of RNA2 region by proteins would be an example of RNA activity substituted by proteins.

FIGURE 8.

A, time course of refolding of BCA as modulated by bovine mitochondrial PTC (BMPTC) complemented with mitoribosomal large subunit proteins (MRPs) and control set where only proteins were added. B, mitochondrial ribosomal proteins interacting with PTC RNA and their E. coli homologs.

DISCUSSION

The presence of two overlapping functions of the ribosome-protein synthesis and protein folding on the PTC (18, 30–33) suggests their co-evolution in the RNA world. The protein folding activity of ribosome has added a new perspective to the protein-folding problem in general (8–17, 19, 20, 34, 35). This is not only an in vitro property of the ribosome, but the nascent proteins in vivo interact with the PTC identically. Dependence of post-translational folding in vivo of some proteins, e.g. β-galactosidase and the molecular chaperone DnaK, on the participation of active PTC has also been shown (19, 20, 35).

The information necessary to guide ribosomal RNA-mediated protein folding resides at the central loop of domain V, more specifically RNA1 (PTC) (21), in bacteria. The RNA1 interacts with the protein-folding intermediate quite stably. This interaction is very specific, as only a conserved set of nucleotides bind to polypeptide of any primary sequence (Ref. 28; Fig. 6). The polypeptide is detached from RNA1 after it reaches a “folding competent state” by the action of RNA2 that constitutes the rest of the domain V in bacteria and eukaryotes. The released FCS then achieves its active state by itself (15–17, 19–21, 25, 26, 34, 35). In the absence of RNA2, the RNA1 can still fold a protein to the same extent as with the entire domain V rRNA, but the reaction is extremely slow; it needs hours to complete because spontaneous dissociation of the folded protein is very slow (data not shown). So the RNA1 provides the folding function that needs to be modulated by RNA2. In fact, the secondary role of RNA2 becomes evident when one finds that the folded protein can be released from the RNA1 with 3% ethanol or 0.1% Triton X-100. These are known to disrupt nucleic acid protein interactions by changing the dielectric constant of water and hydrophobic interactions. Also, mutations in the RNA2 region do not have a significant effect on folding, whereas mutations in the RNA1 region of domain V suppress folding significantly (Fig. 7).

The PTC of mitoribosomal large rRNA contains the region homologous to bacterial RNA1, but there are major deletions in the regions corresponding to RNA2 and elsewhere throughout the large rRNA (24). The secondary structures of the domain V RNA segments of bacterial (E. coli) 23 S rRNA and the homologous region of mitochondrial large rRNA from Leishmania and B. taurus are shown in Fig. 4. The RNA1 homolog bears all the conserved features of this region, but the part corresponding to RNA2 is severely truncated. It is thus expected that the mitochondrial peptidyl transferase center, although able to specifically interact with and bind the unfolded protein, would not be able to release it efficiently to lead to a productive folding. We found that the bovine and Leishmania mitoribosomal PTC in the large subunit rRNA could bind the unfolded polypeptide but were very inefficient in releasing it (Figs. 2 and 3). Therefore, the folding was very slow and took hours to complete but attained the level observed with bacterial domain V RNA, and the folded polypeptide could be dissociated by ethanol or bacterial RNA2 to attain active state rapidly (Fig. 5A). The nature of the interaction of mitochondrial PTC with the folding protein was the same as that with RNA1. The ability of bacterial RNA2 to interact productively with mitochondrial PTC RNA crossing the species barrier makes it obvious that the mechanism of this RNA-mediated protein folding was highly conserved in evolution. The entire large mitoribosomal rRNA showed similar results as that with mitochondrial PTC, indicating that there are no other regions outside the PTC that could provide the RNA 2 function. As shown above, this activity was substituted by the large subunit proteins of the mitoribosome (Fig. 8A). The structure and composition of the mammalian mitoribosome shows that the loss of a large portion of the rRNA moiety is in parallel with considerable gain in the ribosomal protein masses (36). For example, the ribosomal RNA helices 77 and 78 (PTC) are absent in the mitoribosome, and they are possibly replaced by proteins L1 and L9, which are much larger. L3 binding sites are absent in the mitochondrial large rRNA, and the mitochondrial L3 protein is much larger. These and other observations suggest that the sequence losses of the rRNAs in the mitochondrial ribosome may be compensated both structurally and functionally by the larger proteins (29, 36). Our future aim will be to identify the proteins that specifically mediate the release of the FCS from the mitochondrial PTC RNA. Another possibility is that the homologous ribosomal proteins stabilize rRNA conformation and thus make the release more efficient. The possibility that the slow release is due to a defect in the folding property of mitochondrial rRNA in the absence of ribosomal proteins can be ruled out, as early dissociation of the folding protein by ethanol and Triton X would lead to efficient folding.

The complementation of the protein folding activity of mitochondrial PTC RNA with mitoribosomal large subunit proteins that are mostly nuclear genome-encoded reminds us of the proposed origin and evolution of the mitochondria. The endosymbiosis theory (37) holds that mitochondria have evolved from ancient eubacteria, specifically Rickettsia-like α proteobacteria (38). The mammalian mitochondrial ribosome has a resemblance to the bacterial ribosome with regard to antibiotic susceptibility, sequence similarity of ribosomal proteins, translation factors, and aminoacyl-tRNA synthetases. The bacterial ribosome, therefore, might well be regarded as a direct predecessor of the mitochondrial ribosome. Mammalian mitochondria contain small mitochondrial DNA (16 kbp) capable of encoding 13 proteins of inner mitochondrial membrane involved in oxidative phosphorylation, 22 tRNAs, and 2 shortened rRNAs. The mammalian mitochondrial ribosome comprises 2 RNAs derived from mtDNA and more than 70 different proteins encoded in the nuclear genome. It has been suggested that during the process of mitochondrial evolution, a large portion of mtDNA was transferred to the nuclear genome (38). Comparison of amino acid sequences of large ribosomal proteins of human cytosol, mitochondria, and corresponding bacterial proteins was done using ClustalW. The amino acid sequence homology between human cytosolic, mitochondrial, and bacterial (E. coli) ribosomal protein L3 is shown in Fig. 9. The sequence homology between human cytosolic and human mitochondrial protein is much less (score 11) compared with the similarity between mitochondrial and bacterial proteins (score 30) Fig. 9A. The mitochondrial and cytosolic ribosomal proteins show a characteristic N- or C-terminal extension that results in the increased molecular mass when compared with corresponding E. coli protein, but there is very little homology in this additional part between mitochondrial and cytosolic counterparts (score 5) Fig. 9B. Therefore, the extra portion of the mitoribosomal protein is significantly different from the corresponding cytosolic protein. It is tempting to speculate that this distinct additional amino acid stretch might have some role in the release of the folding protein from the shortened mitochondrial PTC RNA. Thus, proteins play an important role in further improvements in RNA function and in the overall evolution of an increasingly interdependent protein-RNA system in the case of mitoribosomes.

FIGURE 9.

Multiple sequence alignment of E. coli ribosomal protein RL3, human mitochondrial ribosomal protein RM3, and human cytosolic ribosomal protein RL3 using ClustalW. The homology values between the protein sequences are shown in the tables. A shows sequence alignment and score of the entire proteins, and B shows sequence alignment of only the extended portion of human RM3 and RL3. Similar results were found for other ribosomal proteins also.

Thus, although the mitoribosomal RNA itself performs the actual function of protein folding, the responsibility for the release of the FCS that belongs to RNA2 region can be bestowed upon some of the proteins in the mitochondrial ribosomal large subunit. We saw earlier that the rRNA performs all the basic function of ribosomes whether it is translation or protein folding. This report suggests that a number of ribosomal proteins evolved to substitute the cooperation between RNA1 and RNA2 with a similar cooperation between RNA1 and the proteins themselves, thereby promoting a RNA-mediated process to a RNP-mediated one.