mtRF1a Is a Human Mitochondrial Translation Release Factor Decoding the Major Termination Codons UAA and UAG

Summary Human mitochondria contain their own genome, encoding 13 polypeptides that are synthesized within the organelle. The molecular processes that govern and facilitate this mitochondrial translation remain unclear. Many key factors have yet to be characterized—for example, those required for translation termination. All other systems have two classes of release factors that either promote codon-specific hydrolysis of peptidyl-tRNA (class I) or lack specificity but stimulate the dissociation of class I factors from the ribosome (class II). One human mitochondrial protein has been previously identified in silico as a putative member of the class I release factors. Although we could not confirm the function of this factor, we report the identification of a different mitochondrial protein, mtRF1a, that is capable in vitro and in vivo of terminating translation at UAA/UAG codons. Further, mtRF1a depletion in HeLa cells led to compromised growth in galactose and increased production of reactive oxygen species.


Yeast growth conditions, plasmid and general strain constructions and complementation assays
The S. cerevisiae MRF1 ORF was PCR amplified from genomic DNA and cloned into pFL61-BEB (a derivative of pFL61 containing a BamHI site). The S. cerevisiae MATα ∆mrf1::Kan R haploid strain Y14510 was ordered from Euroscarf, depleted of mtDNA by growth on ethidium bromide containing medium, and crossed to a wild type strain containing an intron-less mitochondrial genome (CW252/A, a MATa version of CW252, (Saint-Georges et al., 2002). The resulting heterozygous ∆mrf1 diploid KV2 and a control wild type diploid were transformed with the control vector or plasmids producing the human or yeast proteins. Diploids were sporulated and tetrads dissected (Sanger micromanipulator) on minimal medium lacking uracil to maintain the plasmids. For biochemical studies isogenic strains were generated by transforming the rho° versions of Y14510 (∆mrf1Sc) or BY4742 (wt) with empty vector or one producing either S. cerevisiae Mrf1 or human mtRF1a, and then introducing the intron-less mtDNA through cytoduction with the karyogamy deficient strain JC8/252.

Transient transfection of HeLa cells, microscopy and image capture
HeLa cells were grown to 50% confluency on coverslips and transfected with vectors expressing the GFP fusion constructs (1 µg) using Superfect (Qiagen) as recommended.
Cells were cultured for a further 24 h prior to incubation with Mitotracker Red CM-H2XRos (1 µM final, Invitrogen). After brief fixation (4% paraformaldehyde in PBS, 15 min at room temperature), cells were mounted in Vectashield (H-1500 Vector Laboratories Inc) and visualised by fluorescence microscopy using a Leica (Nussloch Germany) DMRA. Images were recorded as a Z-series (0.5 µm slices) using a cooled CCD camera and imaging system (Spot -II Diagnostics Instruments, Sterling Heights MI, USA).

Over-expression and purification of human mitochondrial and bacterial proteins
E.coli strain Rosetta(DE3)pLysS (Novagen) was transformed with constructs for the overexpression of the human mitochondrial release factors. IPTG (1mM) was added to bacterial cultures (0.40-0.45 A 600nm ) and induced for 16 h at 16°C. Cells were harvested, resuspended in PBS and sonicated 10 × 10 s, amplitude 18 microns (Soniprep 150). The lysate was centrifuged at 35,000xg for 20 min and the supernatant filtered before application to a Glutathione Sepharose 4B (Amersham Bioscience) column. The column was washed with PBS/150mM NaCl and protein released from GST by cleavage with PreScission Protease (Amersham Bioscience), resulting in the removal of 49 N-terminal residues for purified mtRF1 and 32 residues for mtRF1a. E.coli RF1 was overexpressed and purified from strain BL21 pLysS following the protocol described in (Tate and Caskey, 1990).

In vitro translation termination assay
This was performed essentially as described in (Caskey et al., 1971;Tate and Caskey, 1990) with complexes simulating the state of the ribosome at the terminal stage of protein synthesis being generated to test release activity of proteins with specific RNA oligomers.

Cell preparations and western analysis
Lysate was prepared by pelleting human cells from 75 cm 2 flasks and resuspending in 150 µl of lysis buffer (50mM Tris-HCl pH 7.5, 130mM NaCl, 2mM MgCl 2 , 1% Nonidet P-40, 1 mM PMSF and Roche EDTA protease inhibitor cocktail). This was vortexed briefly, spun at 1500g for 5 min at 4˚C and the supernatant retained.
Human mitochondrial fractions were prepared from cell pellets resuspended in 0.6 M mannitol, 10 mM Tris-HCl pH7.4, 1 mM EGTA, 0.1% BSA and hand homogenised by 15 passes in a Teflon:glass tissue grinder. Following centrifugation at 600g the supernatant was retained and the pellet was re-homogenised and spun as before. The combined supernatants were cleared at 600g and then the crude mitochondria pelleted at 15kg. These were washed in buffer lacking BSA and frozen in liquid nitrogen until required. S. cerevisiae and S. pombe mitochondria were purified from cells grown in minimal galactose (S. cerevisiae) or glucose (S. pombe) media lacking uracil as described in (Wallis et al., 1994) and (Chiron et al., 2007) respectively, except that protease inhibitors tablets (Roche) were added before breaking the protoplast and during all subsequent steps. S. pombe total proteins were extracted as in (Chiron et al. 2005). Blue native gel electrophoresis was performed as described in (Nijtmans et al., 2002) with the modification that 30 µg protein was loaded per lane and wet transfer as above.

siRNA transfection
Both negative control and mtRF1a specific siRNA duplexes were purchased pre-annealed from Eurogentec. The negative control is a unique sequence that does not match any sequence in the genome. Sense strands for siRNA specific to the mRNA encoding mtRF1a were :-1) GACGCUGCAUGAUCUUGAAdTdT, 2) CCAUGACUGUAGCAAUAUUdTdT and 3) CGAUGAGAAUGAAGAUUUAdTdT.
HeLa cells in a 35 mm dishes were grown to 20-30% confluency, incubated in 800 µl DMEM (D6429, Sigma) supplemented with 10% FCS, 50 µg/ml uridine and 1x nonessential amino acid, to which the Oligofectamine (Invitrogen) transfection mix in OptiMEM (Gibco) was added to give a final concentration of 0.2 µM siRNA. Cells were cultured in this medium for 3 days and then either harvested or retransfected for the 6 day time point.

De Novo mitochondrial protein synthesis
HeLa cells were cultured in 35 mm dishes until 70% confluent, then incubated in methionine free medium for 1 h. This medium was replaced with 1 ml labeling medium (methionine/cysteine-free DMEM, 10% dialyzed FBS, emetine 10 µg /ml, and 500 µCi [ 35 S]methionine/cysteine [3,000 Ci/mmol; Amersham]) and cells incubated at 37°C for 2 h. After removal of the incubation medium, the cells were washed in PBS and pelleted. Protein samples (20 μg) were separated on 12% SDS PAGE and products visualized with a PhosphorImager system with ImageQuant software (Molecular Dynamics, GE Heathcare).
S. cerevisiae transformants were grown in minimal galactose medium lacking uracil, harvested at the exponential phase, resuspended in 40 mM KH 2 PO 4 / K 2 HPO 4 , 2% galactose buffer pH 6 and incubated for 1 h at 28°C. After a 10 min pre-incubation with 600 μg/ml cycloheximide to block cytosolic translation, 72 μCi of [ 35 S] methionine/cysteine (Redivue Pro-mix, Amersham) were added and incubation was carried out for 30 min.
After removal of the medium a 10 min chase was performed by adding 12 μl of 0.25 M methionine and 0.25 M cysteine. Equal amount of total proteins samples were loaded on a 16% acrylamide-0,5% bisacrylamide gel and products visualized by autoradiography or PhosphorImaging as above.
Supp Mat Fig 5