Figure 1.
Identification of a mitochondrial respiratory chain deficiency and defective mtDNA translation.
A) Cytochrome c oxidase (COX) histochemical reactivity revealed a mosaic of COX-deficiency in patient skeletal muscle compared to control. B) The assessment of individual respiratory chain enzyme activities identified a combined OXPHOS deficiency affecting complexes I and IV in skeletal muscle from the proband. The mean activity measured in 25 controls was set at 100%. C) Patient fibroblasts (grey) are less capable of responding to stress in comparison to control fibroblasts (black), as measured by the spare respiratory capacity. *: P<0.05. D) The coupling efficiency of ATP synthesis and respiration, and therefore the level of proton leak, is not decreased in patient fibroblasts (grey) compared to controls (black). The error bars displayed on each graph indicate standard deviation. E) In vitro metabolic labelling of mitochondrial translation in patient fibroblasts (P) showed a generalised decrease in translation activity, with the subunits of Complex I (notably ND5) and Complex IV (notably COXI) most substantially decreased. Even loading was confirmed by Coomassie blue staining (CBS). F) Immunoblotting demonstrated a generalised decrease in the level of individual subunits from respiratory chain complexes I–V (normalised to β-actin) in patient fibroblasts, whilst the mitochondrial marker, TOMM20 was unchanged.
Figure 2.
A TRIT1 mutation segregates with disease and disrupts a conserved tRNA-binding basic side-chain.
A) Targeted resequencing of TRIT1 confirmed that the proband (II-3; arrow), and his clinically affected sister (II–1) are homozygous for the c.968G>A TRIT1 mutation, while his unaffected older brother (II–2) and both of his parents (I–1 and I–2) are heterozygous carriers. B) The TRIT1 mutation is located in exon 8, whilst there is a putative mitochondrial targeting sequence in exon 1 and a matrin-type zinc finger domain spanning exons 10 and 11. C) The mitochondrial sub-localisation of TRIT1 is demonstrated by sub-fractionation and immunoblotting, using markers for each sub-fraction to confirm there was no contamination: TOMM20 (mitochondrial outer membrane), AIF (mitochondrial intermembrane space), GDH (mitochondrial matrix), NDUFA9 (mitochondrial inner membrane) and eIF4E (cytosol). TRIT1 localized with eIF4E in the cytosol (lane 2) and showed the same profile as GDH (lanes 3–6), but was undetectable in the inner mitochondrial membrane fraction (lane 7). A total of 40 µg protein was loaded for each sample, and all mitochondrial subfractions were prepared from the same mitochondrial lysate. D) Clustal Omega alignment of the TRIT1 protein and known orthologs revealed that the affected amino acid (p.Arg323) is conserved in each species excluding S. pombe, S. cerevisiae and E. coli, where the equivalent amino acid is lysine, which has similar electrochemical properties (:). Asterisks (*) indicate completely conserved residues. E–G) The co-crystal structure of Mod5 bound to a substrate tRNA (based on [14]) shows the interaction of the tRNA backbone (nucleotides 27–29) with an extended α-helix in which are located multiple basic side chains (indicated in red) of the enzyme including that corresponding to the mutated position (Lys294).
Figure 3.
Mutant TRIT1 has decreased in vitro activity and cannot complement tit1-Δ defects in S. pombe.
A) A Coomassie-blue stained SDS polyacrylamide gel confirmed the parallel purification of His-tagged TRIT1-WT and His-tagged TRIT1-Mut from E. coli. B) Loss of the isopentenyltransferase activity of TRIT1 carrying the p.Arg323Gln mutation was demonstrated by an in vitro assay using varying concentrations (nM) of His-tagged TRIT1-WT and TRIT1-Mut as well as a standard amount of RNA substrate (2.5 µM) [11]. C) Mutant TRIT1 activity was shown to be very low rather than absent by repeating the assay using 10 µM RNA and 400 nM of protein. D) Wild-type (yYH1) and tit1-Δ (yNB5) strains of the fission yeast Schizosaccharomyces pombe were transformed with an empty vector (+V), the S. pombe tRNA isopentenyltransferase (+tit1+), wild-type human TRIT1 (+TRIT1-WT) or mutant human TRIT1 (+TRIT1-Mut) and plated onto media containing limited adenine to assay for the loss of function of tRNASer(UCA) due to lack of isopentenyl modification. tit1-deleted yeast carrying the empty vector or mutant TRIT1 showed no recovery of tRNASer(UCA) function (red colonies), but knock-down yeast carrying wild-type TRIT1 or tit1+ showed recovery of tRNASer(UCA) activity (white colonies) similar to wild-type yeast. E) Transformation of tit1-Δ yeast with wild-type TRIT1 or tit1+, but not the empty vector or mutant TRIT1, could also complement the respiratory deficiency illustrated by slow growth on glycerol compared to growth on glucose; spots reflect 10-fold serial dilutions of the same amounts of cells as determined by OD600.
Figure 4.
The TRIT1 mutation disrupts modification activity on cytosolic and mitochondrial tRNAs but not enzyme abundance.
A) No decrease in the levels of the native TRIT1 protein in patient fibroblasts was observed by immunoblotting (using β-actin as a loading control) B) The isopentenyl modification status of both mitochondrial (mt-) and cytosolic (cy-) tRNAs in patient fibroblasts (lane P) compared to controls (lane C); by this approach a positive signal is due to lack of isopentenyl modification as detected by an anticodon loop (ACL) probe (the bulky modification on the N of adenine blocks base pairing with the probe, such that no signal for cy-tRNASer(UGA) with the ACL probe indicates efficient modification in the control cells [11]); a body probe to a different region of the same tRNA is used as a control for calibration and calculation of steady-state levels. Each panel shows hybridisation of the same blot with a different probe as indicated to the right. The cytosolic tRNASer(UGA) is poorly modified in patient fibroblasts (strong ACL probe signal), but tRNASer(UGA) steady-state levels are unchanged. Mt-tRNASer(UCN) is also poorly modified in patient fibroblasts, although a small pool of mt-tRNASer(UCN) in control fibroblasts lacks the modification. The modification appears to be influential on mt-tRNASer(UCN) stability, as steady-state levels are decreased by 40% in the patient. The non-substrate mt-tRNACys was probed as a control.
Figure 5.
TRIT1 activity in patient fibroblasts is rescued by transduction with wild-type TRIT1.
A) Immunoblotting of the TRIT1 protein in patient and control fibroblasts after transduction with an empty vector, wild-type TRIT1 (TRIT1-WT) or mutant TRIT1 (TRIT1-Mut) and selection with puromycin demonstrated overexpression, whilst Ponceau S staining confirmed even loading. B) The PHA6 assay was performed on patient and control fibroblasts transduced with wild-type TRIT1, mutant TRIT1 or an empty vector, for both mitochondrial (mt-) and cytosolic (ct-) tRNAs serine. Loading controls are provided by U5 RNA as well as two mitochondrial non-TRIT1 substrates, mt-tRNACys and mt-tRNALeu(UUR). Each sample was run in duplicate with 3 µg (1X) and 6 µg (2X) of RNA as indicated above the lanes. Transduction of patient fibroblasts with both wild-type and mutant TRIT1 rescued the i6A modification of cytosolic tRNASer(UGA). C) However, only transfection with wild-type TRIT1 was able to efficiently rescue the i6A modification of mt-tRNASer(UCN). *: p<0.05. D) The steady-state level of cytosolic tRNASer(UGA), calculated using the body probe and the probe to U5 RNA, was found to be increased in all three transduced patient fibroblasts, although this up-regulation was found to be non-specific for cy-tRNASer(UGA) (data not shown). E) The steady-state level of mt-tRNASer(UCN) was shown to be recovered by transduction with wild-type TRIT1 and to a lesser extent, mutant TRIT1 (using the body probe and the probes for U5 RNA (black columns) and mt-tRNACys (grey columns)). The error bars indicate the difference between the duplicate determinations.
Figure 6.
A mt-tRNASer(UCN) point mutation, m.7480A>G, impairs i6A37 modification.
A) The previously reported m.7480A>G mutation [15] is located at position 38 in the anticodon loop (ACL) of mt-tRNASer(UCN), in the A36A37A38 recognition sequence of TRIT1. B) The i6A modification of position 37 is decreased (stronger binding of the mt-tRNASer(UCN) double ACL probe) by ∼86% in patient skeletal muscle relative to control muscle. Binding of the ACL probe to the non-TRIT1 substrate, mt-tRNACys, confirmed even loading. The steady-state level of mt-tRNASer(UCN) in patient skeletal muscle is 30% lower than in control muscle (calculated using the two body probes). Both control and patient samples were run in duplicate using 2 µg (1X) and 4 µg (2X) of RNA as indicated above the lanes. C) In vitro isopentylation assay using purified TRIT1, 14C-DMAPP, and synthetic minihelixes representing the anticodon stem loops (ASLs) of the tRNAs indicated above the lanes. The upper blot is an ethidium bromide-stained gel of the ASLs after in vitro reaction indicating even loading, whilst the lower blot is the autoradiograph obtained after 3 days of exposure. D) The PHA6 assay was validated and the loss of i6A37 modification in m.7480A>G mutant mt-tRNASer(UCN) confirmed by in vitro modification of synthetic templates. For reaction samples (R), synthetic RNA minihelixes were in vitro modified using unlabeled-DMAPP and recombinant His-TRIT1. For mock-treated samples (M), all the components except the His-TRIT1 enzyme were added. After purification, the RNA samples were transferred to a membrane which was repeatedly hybridized, stripped and rehybridized with 5 different 32P-labeled ASL oligo probes as indicated to the right. Ethidium bromide staining of the gel confirmed equal loading of each pair of reactions (mock and reaction samples) for each synthetic ACL interrogated.