A Deafness- and Diabetes-associated tRNA Mutation Causes Deficient Pseudouridinylation at Position 55 in tRNAGlu and Mitochondrial Dysfunction*

Several mitochondrial tRNA mutations have been associated with maternally inherited diabetes and deafness. However, the pathophysiology of these tRNA mutations remains poorly understood. In this report, we identified the novel homoplasmic 14692A→G mutation in the mitochondrial tRNAGlu gene among three Han Chinese families with maternally inherited diabetes and deafness. The m.14692A→G mutation affected a highly conserved uridine at position 55 of the TΨC loop of tRNAGlu. The uridine is modified to pseudouridine (Ψ55), which plays an important role in the structure and function of this tRNA. Using lymphoblastoid cell lines derived from a Chinese family, we demonstrated that the m.14692A→G mutation caused loss of Ψ55 modification and increased angiogenin-mediated endonucleolytic cleavage in mutant tRNAGlu. The destabilization of base-pairing (18A-Ψ55) caused by the m.14692A→G mutation perturbed the conformation and stability of tRNAGlu. An approximately 65% decrease in the steady-state level of tRNAGlu was observed in mutant cells compared with control cells. A failure in tRNAGlu metabolism impaired mitochondrial translation, especially for polypeptides with a high proportion of glutamic acid codons such as ND1, ND6, and CO2 in mutant cells. An impairment of mitochondrial translation caused defective respiratory capacity, especially reducing the activities of complexes I and IV. Furthermore, marked decreases in the levels of mitochondrial ATP and membrane potential were observed in mutant cells. These mitochondrial dysfunctions caused an increasing production of reactive oxygen species in the mutant cells. Our findings may provide new insights into the pathophysiology of maternally inherited diabetes and deafness, which is primarily manifested by the deficient nucleotide modification of mitochondrial tRNAGlu.

Several mitochondrial tRNA mutations have been associated with maternally inherited diabetes and deafness. However, the pathophysiology of these tRNA mutations remains poorly understood. In this report, we identified the novel homoplasmic 14692A3 G mutation in the mitochondrial tRNA Glu  Deafness is one of the major public health problems, affecting 360 million persons worldwide. Deafness can be grouped into non-syndromic deafness (where hearing loss is the only obvious medical problem) and syndromic deafness (where hearing loss is coupled with other medical problems, such as diabetes) (1). Among syndromic manifestation, maternally inherited diabetes and deafness (MIDD) 3 is characterized by the onset of sensorineural hearing loss and diabetes in adulthood (2,3). Mutation(s) in mtDNA is the most important cause of MIDD (4 -7). In fact, the human mtDNA encodes 13 subunits of the oxidative phosphorylation system, two rRNAs, and 22 tRNAs required for mitochondrial protein synthesis (8). Among these tRNAs, eight tRNAs (such as tRNA Glu and tRNA A1a ) reside at the cytosine-rich light (L) strand. The remaining tRNAs, including tRNA Lys and tRNA His , are located at the guanine-rich heavy (H) strand (9,10). These MIDD-associated mtDNA mutations are often present in heteroplasmy (a mixture of wild-type and mutated molecules) (4). Of these, large deletions or duplications in mtDNA have been associated with diabetes and deafness in some pedigrees (5,11). The most prevalent tRNA point mutation associated with MIDD is the heteroplasmic m.3243A3 G mutation in the tRNA Leu(UUR) gene, accounting for over 1% of diabetic patients in some populations (6,(12)(13)(14). The other MIDD-associated mtDNA mutations were the tRNA Leu(UUR) 3264T3 C, tRNA Glu 14709T3 C and tRNA Gly 10003T3 C mutations (15)(16)(17). These mutations may lead to the structural and functional consequences of tRNAs, including processing of the RNA precursor, aminoacylation, and nucleotide modification of tRNAs (15)(16)(17)(18)(19)(20). In particular, the m.3243A3 G mutation caused the loss of m 5 U, m 5 s 2 U, and taurine modifications in the mutant tRNA Leu(UUR) (20 -22). A failure in tRNA Leu(UUR) metabolism impaired mitochondrial protein synthesis and subsequently reduced the activities of oxidative phosphorylation (18,19,23,24). However, the pathophysiology of these mtDNA mutations remains poorly understood.
As part of a genetic screening program for deafness in a cohort of 2651 Han Chinese hearing-impaired subjects (25,26), we identified the U-to-C transition at position 14692 (m.14692A3 G) in the tRNA Glu gene in three genetically unrelated probands whose families exhibited the maternal transmission of deafness and diabetes. As shown in Fig. 1, the m.14692A3 G mutation is localized at a highly conserved nucleotide (U55) of the T-loop of tRNA Glu (27). The uridine at position 55 (U55) of tRNA Glu is modified to pseudouridine (⌿55), which plays an important role in the structure and function of this tRNA (27)(28). Therefore, substitution of U with C at position 55 of tRNA Glu may cause loss of pseudouridine (⌿55) and destabilize the A18-⌿55 base pair between the T-loop and D-loop of tRNA Glu , thereby altering the tertiary structure and function of tRNA Glu . In particular, the mutation may affect the stability of this tRNA, thus impairing mitochondrial protein synthesis. It was also proposed that an impairment of mitochondrial translation caused by the tRNA mutation altered respiration and production of ATP and reactive oxygen species (ROS). To investigate the pathophysiology of this mutation, we generated lymphoblastoid cell lines derived from two affected matrilineal relatives carrying the m.14692A3 G mutation and two control subjects lacking the mtDNA mutation but belonging to the same mtDNA haplogroup. First we examined whether the m.14692A3 G mutation perturbs the ⌿55 modification of tRNA Glu by using CMCT modification/reverse transcription (29) and in vitro angiogenin cleavage assays (30). We then investigated whether the m.14692A3 G mutation alters the conformation and stability of RNA Glu . This m.14692A3 G mutation was further assessed for the effect on mitochondrial translation, respiration, production of ATP, mitochondrial membrane potential, and the generation of ROS through the use of mutant and control lymphoblastoid cell lines.

Identification of the tRNA Glu 14692A3G Mutation in a Large
Cohort of Hearing-impaired Subjects-The m.14692A3 G mutation in the tRNA Glu gene was identified in three genetically unrelated probands among 2651 Chinese hearing-impaired probands but absent in 574 Chinese control subjects. As shown in Fig. 1, the m.14692A3 G mutation is localized at a highly conserved nucleotide (U55) of the T-loop of the tRNA Glu (31,32). The uridine at this position (U55) of tRNA Glu is modified to pseudouridine (⌿55), which forms a tertiary base pair with the A18 in the D-loop and stabilizes the L-shaped tRNA structure (32)(33)(34). Thus, it was anticipated that the U-to-C substitution at position 55 by the m.14692A3 G mutation would perturb the ⌿55 modification and destabilize the base pairing (18A-⌿55) of tRNA Glu . The sequence analysis of the entire mtDNA in these three probands exhibited the identical m.14692A3 G mutation and distinct sets of polymorphisms belonging to the Eastern Asian haplogroups B5 and D4, respectively (supplemental Table S1) (35). However, there were no other functional significant variants in their mtDNAs. Further analysis showed that the m.14692A3 G mutation was present in homoplasmy in all matrilineal relatives but not in other members of three Chinese families (supplemental Fig. S1).
Clinical Presentation of Three Chinese Families-All available members of three Han Chinese families carrying the m.14692A3 G mutation, as shown in supplemental Table S2, underwent comprehensive evaluations of their medical histories and physical examination with the aim to identify any clinical abnormalities, genetic factors related to the deafness, and diabetes. The audiological examination was performed as detailed elsewhere (36). The diagnosis of diabetes was based on the criteria of the American Diabetes Association (37). Of 11 matrilineal relatives of pedigree WZD81, as shown in supplemental Fig. S2, six (two male and four female) individuals suffered from diabetes (three subjects with only diabetes and three subjects with both diabetes and hearing impairment). The average ages at onset of diabetes and deafness were 60 (from 42-72) and 27 years, respectively. In pedigree WZ82, only one matrilineal relative (WZD82-II-3) had diabetes at the age of 50, and one matrilineal relative (WZD82-III-1) exhibited severe hearing loss at the age of 26. Among seven matrilineal relatives of pedigree WZD83, two members (WZD83-III-2, and WZD83-IV-1) suffered from moderate hearing loss, whereas two subjects (WZD83-II-2 and III-3) exhibited both deafness and diabetes. The average ages at onset of diabetes and deafness were 50 (from 40 -60) and 19 (from 16 -22) years, respectively. Moreover, these matrilineal relatives showed no other clinical abnormalities, including cardiac failure, muscular diseases, visual failure, and neurological disorders. On the other hand, none of the other members in these families exhibited diabetic or other clinical abnormalities. Therefore, these familial histories indicate the maternal inheritance of diabetes and deafness.
Deficient Pseudouridinylation at Position 55 of Mitochondrial tRNA Glu -To determine whether the m.14692A3 G mutation alters the pseudouridinylation of tRNA Glu , we subjected mitochondrial RNAs from mutant and control lymphoblastoid cell lines to CMCT/reverse transcription with a DIGlabeled oligonucleotide probe specific for tRNA Glu (Fig. 1). This approach involved carbodiimide (CMCT) adduct formation with U, G, and pseudouridine, followed by mild alkali to remove the adduct from U and G but not from N 3 -[N-cyclohexyl-NЈ-␤-(4-methyl-morpholinium) ethylcarbodiimide]-⌿ (N 3 -CMC-⌿) (29). This results in the attenuation of primer reverse transcription, causing a stop band one residue 3Ј to the pseudouridine on the sequence gel. As shown in Fig. 1B, the ⌿55 modification was not detected in tRNA Glu derived from the mutant cell lines (III-6, III-9), whereas the ⌿55 modification was present in the tRNA Glu derived from the control cell lines (C3, C4). However, ⌿40, ⌿28, and m 1 A9 modifications were detected in tRNA Glu obtained from both control and mutant cell lines.
We further examined whether the m.14692A3 G mutation perturbs the pseudouridinylation of tRNA Glu by analyzing the sensitivity of wild-type and mutant tRNA Glu to digestion with angiogenin (ANG), a tRNA-specific enzyme of the RNase A superfamily (38). In fact, the loss of certain tRNA modifications increases the angiogenin-medicated cleavage of tRNAs. For this purpose, the wild-type (⌿55) and mutant (C55) tRNA Glu obtained from in vitro transcription as well as from control and mutant cell lines were digested by angiogenin and followed by Northern blotting analysis. As shown in Fig. 2, the mutant (C55) tRNA Glu transcript and mutant tRNA Glu obtained from the mutant cells (III-9) were more sensitive to angiogenin-mediated digestion than those in the wild-type (U55) tRNA Glu transcript and tRNA Glu obtained from the control cell line (C3), respectively.
Altered Conformation and Stability of tRNA Glu -It was anticipated that the destabilization of base pairing (18A-⌿55) by the m.14692A3 G mutation leads to structural alterations of tRNA Glu . The stability of the transcripts of wild-type and mutant tRNA Glu (72 nt) as well as tRNA Phe (74 nt), tRNA Asp (71 nt), and tRNA Thr (69 nt) were first examined by the measurement calculating the derivatives of the absorbance against a temperature curve. As shown in supplemental Fig. S3, the T m values for wild-type (U55) and mutant (C55) tRNA Glu transcripts were 50°C and 46°C, respectively. This suggested that the global folding of mutant tRNA Glu was less stable at high temperature than that of wild-type tRNA Glu . These transcripts . Solid lines represent the DIG-labeled oligonucleotide probe specific for tRNA Glu . Broken lines represent the potential stops of reverse transcription reaction caused by base modification, such as CMC-pseudouridine and m 1 A. B, 2 g of mitochondrial RNA from control (C3 and C4) and mutant (III-6 and III-9) cells was incubated with CMCT for CMC modification of ⌿ residues. Reverse transcription was carried out to identify the stops caused by CMC-pseudouridine. The tRNA Glu transcript was used as a control without base modification. Marker, DIG-labeled oligonucleotides of variable length.
were then assessed for conformation change by PAGE analysis under denaturing and native conditions. As shown in Fig. 3, there was no migration difference between wild-type (U55) and mutant (C55) tRNA Glu transcripts under denaturing conditions, whereas the wild-type (U55) tRNA Glu transcript migrated slightly faster than the mutant (C55) tRNA Glu transcript under the native condition. To further test whether the m.14692A3 G mutation affects the conformation of tRNA Glu , total mitochondrial RNA was electrophoresed through a 10% polyacrylamide gel (native condition) in Tris borate-EDTA buffer and then electroblotted onto a positively charged nylon membrane for hybridization analysis with oligodeoxynucleotide probes for tRNA Glu and tRNA His , respectively. As shown in Fig. 3, the electrophoretic patterns showed that the tRNA Glu in two cell lines carrying the m.14692A3 G mutation migrated much slower than those of control cell lines lacking this mutation. These data indicated that the m.14692A3 G mutation changed the conformation of tRNA Glu .
Marked Decrease in the Steady-state Levels of tRNA Glu -To further evaluate whether the m.14692A3 G mutation impairs tRNA metabolism, we subjected mitochondrial RNAs from mutant and control lymphoblastoid cell lines to Northern blotting with probes for tRNA Glu , tRNA Ala as representatives of the whole L-strand transcription unit and tRNA His , tRNA Ile , and tRNA Thr derived from the H-strand transcription unit, as well as a nucleus-encoded mitochondrial 5S RNA (under denaturing conditions) (8,9,39). As shown in Fig. 4A, the amount of tRNA Glu in two mutant cell lines was markedly decreased compared with those in two control cell lines. For comparison, the average levels of each tRNA in control or mutant cell lines were normalized according to the level of the 5S rRNA. As shown in Fig. 4B, the steady-state levels of tRNA Glu in two mutant cell lines were 35.5% (p Ͻ 0.01) of the average values of two control cell lines lacking the mtDNA mutation. However, the average steady-state levels of tRNA His , tRNA Ile , tRNA Thr , and tRNA Ala in two mutant cell lines were comparable with those in control cell lines.
Mitochondrial Protein Synthesis Defect-To assess whether the m.14692A3 G mutation affects mitochondrial translation, a Western blotting analysis was carried out to examine the steady-state levels of seven mtDNA-encoded respiratory complex subunits in mutant and control cells with GAPDH as a loading control. As shown in Fig. 5, the levels of CO 2 (subunit II of cytochrome c oxidase); ND1, ND4, ND5, and ND6 (subunits 1, 4, 5, and 6 of NADH dehydrogenase); ATP6 (subunit 6 of the H ϩ -ATPase); and CYTB (apocytochrome b) exhibited the variable reductions in two mutant cell lines, with an average of 28.8% (p ϭ 0.033), relative to the mean value measured in the control cell lines. In the mutant cells, the average levels of ND1, ND6, and CO2 with a higher proportion of glutamic acid residues were much lower than those of ND4, ND5, ATP6, and CYTB with relatively less proportion of glutamic acid residues with respect to the average values of control cells. However, the levels of polypeptide synthesis in mutant cells relative to those FIGURE 2. In vitro angiogenin cleavage assays. A, angiogenin digestion pattern of in vitro transcripts of wild-type (55U) and 14692A3 G (55C) mutant tRNA Glu . 2 g of purified tRNA transcripts was used for the ANG cleavage reaction at various lengths (from 0 -45 min). Cleavage products of tRNA transcripts were electrophoresed through a denaturing polyacrylamide gel and stained with methylene blue. Broken squares indicate the differences between wild-type (U55) and mutant tRNA Glu (C55). B, angiogenin digestion pattern of mt-tRNA Glu purified from the control cell line (C3) and mutant cell line (III-9). 2 g of human mitochondrial RNAs was used for the ANG cleavage reaction at various lengths (from 0 -36 h). Cleavage products of mitochondrial tRNAs were resolved in 15% denaturating polyacrilamide gels with 8 M urea, electroblotted, and hybridized with a DIG-labeled oligonucleotide probe specific for tRNA Glu . Broken squares indicate the differences between RNAs from control (C3) and mutant (III-9) cells, respectively.
in control cells showed no significant correlation with either the number of codons or the proportion of glutamic acid residues (supplemental Table S3).
Reduced Activities of Complex I and IV-To investigate the effect of the m.14692A3 G mutation on oxidative phosphorylation, we measured the activities of respiratory complexes by isolating mitochondria from two mutant and two control cell lines. Complex I (NADH ubiquinone oxidoreductase) activity was determined by following the oxidation of NADH with ubiquinone as the electron acceptor (40,41,42). Complex III (ubiquinone cytochrome c oxidoreductase) activity was measured as the reduction of cytochrome c (III) using D-ubiquinol-2 as the electron donor. The activity of complex IV (cytochrome c oxidase) was monitored by following the oxidation of cytochrome c (II). As shown in Fig. 6, the activity of complexes I, II, III, and IV in the mutant cells carrying the m.14692A3 G mutation were 53.2%, 94%, 96.7%, and 75.4% of the mean value measured in two control cell lines, respectively.
Respiration Deficiency-To evaluate whether the m.14692A3 G mutation alters cellular bioenergetics, we examined the oxygen consumption rates (OCRs) of two mutant cell lines carrying the m.14692A3 G mutation and two control cell lines. As shown in Fig. 7, the basal OCR in mutant cell lines was ϳ59.5% (p ϭ 0.005) relative to the mean value measured in the control cell lines. To investigate which of the enzyme complexes of the respiratory chain was affected in the mutant cell lines, the OCR was measured after sequential addition of oligomycin (to inhibit ATP synthase), FCCP (to uncouple the mitochondrial inner membrane and allow for maximum electron flux through the electron transfer chain), rotenone (to inhibit complex I), and antimycin A (to inhibit complex III). The difference between the basal OCR and the drug-insensitive OCR yields the amount of ATP-linked OCR, proton leak OCR, maximal OCR, reserve capacity, and non-mitochondrial OCR. As shown in Fig.  7, the ATP-linked OCR, proton leak OCR, maximal OCR, reserve capacity, and non-mitochondrial OCR in mutant cell lines wereϳ60%, 56%, 64%, 67%, and 71.5% relative to the mean value measured in the control cell lines (p ϭ 0.014, 0.058, 0.017, 0.024, and 0.194), respectively.
Reduced Production of Mitochondrial ATP-The capacity of oxidative phosphorylation in mutant and wild-type cells was examined by measuring the levels of cellular and mitochondrial ATP using a luciferin/luciferase assay. Populations of cells were incubated in media in the presence of glucose or 2-deoxy-D- glucose with pyruvate (26). The levels of ATP production in mutant cells in the presence of glucose (total cellular levels of ATP) were comparable with those measured in the control cell lines (Fig. 8A). By contrast, the levels of ATP production in mutant cell lines in the presence of pyruvate and 2-deoxy-Dglucose to inhibit glycolysis (mitochondrial levels of ATP) were  . Western blotting analysis of mitochondrial proteins. A, 20 g of total cellular proteins from various cell lines was electrophoresed through a denaturing polyacrylamide gel, electroblotted, and hybridized with seven respiratory complex subunits in mutant and control cells with GAPDH as a loading control. B, quantification of mitochondrial protein levels. The levels of mitochondrial proteins in two mutant cell lines and two control cell lines were determined as described elsewhere (26). C, quantification of seven respiratory complex subunits. The levels of ND1, ND4, ND5, ND6, CO2, CYTB, and ATP6 in two mutant cell lines and two control cell lines were determined as described elsewhere (26). The graph details and symbols are explained in the legend for Fig. 4. only 61% (p Ͻ 0.001) relative to the mean value measured in the control cell lines (Fig. 8B).
Decrease of Mitochondrial Membrane Potential-The mitochondrial membrane potential (⌬⌿m) is the central bioenergetic parameter that controls respiratory rate, ATP synthesis, and the generation of ROS and is itself controlled by electron transport and proton leaks. The levels of ⌬⌿m were measured in two mutant and two control cell lines using a fluorescence probe JC-10 assay system. The ratios of fluorescence intensity Ex/Em ϭ 490/590 and 490/530 nm (FL590/FL530) were recorded to delineate the⌬⌿m level of each sample. The relative ratios of the FL590/FL530 geometric mean between mutant and control cell lines were calculated to represent the level of ⌬⌿m. As shown in Fig. 9, the levels of ⌬⌿m in the mutant cell lines carrying the m.14692A3 G mutation ranged from 57 % to 73%, with an average of 65% (p Ͻ 0.001) of the  The non-mitochondrial OCR was determined as the OCR after rotenone/antimycin A treatment. The basal OCR was determined as the OCR before oligomycin minus OCR after rotenone/antimycin A. The ATP-lined OCR was determined as the OCR before oligomycin minus OCR after oligomycin. Proton leak was determined as basal OCR minus ATP-linked OCR. The maximal OCR was determined as the OCR after FCCP minus non-mitochondrial OCR. The reserve capacity was defined as the difference between maximal OCR after FCCP minus basal OCR. The average values of five determinations for each cell line are shown. The horizontal dashed lines represent the average value for each group. The graph details and symbols are explained in the legend for Fig. 4. mean value measured in the control cell lines. In contrast, the levels of ⌬⌿m in mutant cells in the presence of FCCP were comparable with those measured in the control cell lines.
The Increase in ROS Production-The levels of ROS generation in viable cells, derived from two matrilineal relatives carrying the m.14692A3 G mutation and two control individuals lacking the mutation, were measured with flow cytometry under normal and H 2 O 2 stimulation (42,43). Geometric mean intensity was recorded to measure the rate of ROS of each sample. The ratio of geometric mean intensity between unstimulated and stimulated with H 2 O 2 in each cell line was calculated to delineate the reaction upon increasing levels of ROS under oxidative stress. As shown in Fig. 10, the levels of ROS generation in mutant cells ranged from 124.7% to 140.9%, with an average of 132.9% (p Ͻ 0.001) of the mean value measured in the control cells.

Discussion
In this study, we investigated the pathogenic mechanism of the deafness and diabetes-associated m.14692A3 G mutation in the mitochondrial tRNA Glu gene. The occurrence of the m.14692A3 G mutation in three genetically unrelated Chinese families affected by deafness and diabetes strongly indicates that this mutation is involved in the pathogenesis of these disorders. The m.14692A3 G mutation affected a highly conserved nucleotide (U55) of the T-loop that is important for the interaction of the tRNA with several components of translation machinery (27,31). The uridine at this position (U55) of tRNA Glu is modified to pseudouridine (⌿55), catalyzed by TRUB2, whereas the adenine at this position (A55) of tRNA Lys is not modified (31,44,45). The ⌿55 forms a tertiary base pair with the A18 in the D-loop and stabilizes the L-shaped tRNA structure (33,34). In this study, we demonstrated that the m.14692A3 G mutation caused the loss of ⌿55 in mutant tRNA Glu . The U55-to-C55 substitution may destabilize the base pairing (18A-⌿55) of tRNA Glu , cause improper folding, and then alter its structure and function, as in the case of the m.12315G3 A mutation in the tRNA Leu(CUN) (46). In particular, the perturbed tertiary structure then makes the mutant tRNA Glu more unstable and subject to degradation. Here the decreased melting temperature and altered conformation were observed in mutant tRNA Glu transcript (C55) compared with  the wild-type transcript (U55). Alternatively, the altered tertiary structure may affect the aminoacylated efficiency of tRNA Glu (27,31). Notably, mutant cell lines carrying the m.14692A3 G mutation exhibited ϳ65% decrease in the steady-state level of tRNA Glu in contrast with a 40% reduction in tRNA Lys in cells carrying the m.8344A3 G mutation at position 55 in tRNA Lys (47,48). However, the reduced level of tRNA Glu in mutant cells carrying the m.14692A3 G mutation was above the proposed threshold level (Ͼ70%) to produce a clinical phenotype associated with a mitochondrial tRNA mutation (9,26,49,50). This suggests that the m.14692A3 G mutation alone is insufficient to produce a clinical phenotype, as in the case of deafness-associated 12S rRNA 1555A3 G and m.1494C3 T mutations (36,51).
The aberrant tRNA metabolism, including shortage of tRNA Glu , contributed to the impairment of mitochondrial protein synthesis. Alternatively, the mutant tRNA Glu may not interact correctly with the translational machinery, thereby altering mitochondrial protein synthesis (47,48). In this study, reduced levels of mitochondrial proteins (an average decrease of ϳ29%) were comparable with the reduced rate of mitochondrial translation observed in lymphoblastoid cell lines carrying the m.4263A3 G and m.4435A3 G mutations (50,52). Variable decreases in the levels of seven mtDNA-encoded polypeptides were observed in the mutant cell lines. As shown in supplemental Table S3, mutant cell lines carrying the m.14692A3 G mutation exhibited marked reductions (34%-66%) in the levels of three polypeptides (ND1, ND6, and CO2) harboring 3.5%-5.7% of glutamic acid codons, relative mild reductions (20%-30%) in the levels of ATP6, ND4, and ND5 carrying 1.3%-2% of glutamic acid codons, but a mildly increased level (21%) in CYTB with only 1.1% of glutamic acid codons. In contrast to what was shown previously in cells carrying the tRNA Lys 8344A3 G mutation (48), polypeptide levels in mutant cell lines, relative to those in control cell lines, did not significantly correlate with the number of glutamic acid codons. Thus, the impaired synthesis of ND1, ND4, ND5, and ND6, the subunits of complex I and CO2, and the subunit of complex IV was responsible for the reduced activities of complex I and complex IV but not those of complex III in these mutant cell lines carrying the m.14692A3 G mutation. Furthermore, the impairment of mitochondrial translation led to reduced rates in the basal OCR or ATP-linked OCR, reserve capacity, and maximal OCR in the mutant cell lines. This observation is clearly consistent with the critical role of tRNA Glu metabolic failure in producing their respiratory phenotypes, as in the cases of cell lines carrying the deafness-associated tRNA Ser(UCN) 7445A3G, 7511T3 C, and tRNA His 12201T3 C mutations (9,26,49).
The respiratory deficiency caused by the m.14692A3 G mutation may result in the uncoupling of the oxidative pathway for ATP synthesis, oxidative stress, and subsequent failure of the cellular energetic process (53). In this study, a 39% drop in mitochondrial ATP production in cell lines carrying the m.14692A3 G mutation, which may be caused by the defective activity of complex I and IV, was much lower than those in cell lines carrying the tRNA Lys 8344A3 G and tRNA Leu(UUR) 3243A3 G mutations (54,55). Furthermore, the deficient activities of respiratory chain complexes caused by tRNA mutations often alter mitochondrial membrane potentials, which is a key indicator of cellular viability (26,56). Indeed, mitochondrial membrane potentials reflect the pumping of hydrogen ions across the inner membrane during the process of electron transport and oxidative phosphorylation (57,58). In this study, a 35% reduction in mitochondrial membrane potential in lymphoblastoid cell lines carrying the m.14692A3 G mutation was much lower than those in cell lines carrying the m.12201T3 C mutation (26). The abnormal oxidative phosphorylation and mitochondrial membrane potential resulted in increased production of reactive oxygen species and the subsequent failure of cellular energetic processes in mutant cells carrying the m.14692A3 G mutation. In particular, the overproduction of ROS can establish a vicious cycle of oxidative stress in the mitochondria, thereby damaging mitochondrial and cellular proteins, lipids, and nucleic acids and increasing apoptotic signaling (59 -60). Hair cells and neurons in the cochlea or ␤ cells in the pancreas may be preferentially affected because they are somehow exquisitely sensitive to subtle imbalances in the cellular redox state or increased level of free radicals (61)(62)(63)(64). This would lead to the dysfunction or death of hair cells in the cochlea or/and ␤ cells in the pancreas carrying the m.14692A3 G mutation.
The variable phenotypes of deafness and diabetes among the maternal lineage indicated the involvement of modifier factors in the phenotypic manifestation. Of 23 matrilineal relatives in three Chinese families, three members exhibited only deafness, four individuals only exhibited diabetes, five subjects suffered from both deafness and diabetes, and 11 other matrilineal relatives did not have any clinical phenotype. In contrast to the low penetrance of hearing loss in families carrying the m.7445A3 G and m.1555A3 G mutations (65, 66), the average penetrances of deafness in these Chinese families were relatively high (30.4%). The striking feature in these families was that matrilineal relatives developed hearing loss at the average age of 22 years but not congenital hearing loss. However, some of the matrilineal relatives in certain families carrying the m.1555A3 G mutation exhibited profound congenital hearing loss (66,67). Furthermore, the average age at onset of diabetes in matrilineal relatives of three families carrying the m.14692A3 G mutation was 57 years, whereas the average ages at onset of diabetes in subjects carrying the heteroplasmic tRNA Leu(UUR) 3243A3 G and homoplasmic tRNA Gly 10003T3 C mutations were 37 and 50 years, respectively (13,14,17). In addition to diabetes and hearing loss, some subjects harboring the m.3243A3 G mutation developed other symptoms, including neurological disorders, cardiac failure, and renal failure (13,68). This discrepancy of clinical features between the m.14692A3 G and m.3243A3 G mutations may be attributed to the nature of the mutations. In fact, the m.14692A3 G mutation only caused a relatively mild decrease (27.8%) in the levels of mitochondrial translation. By contrast, the heteroplasmic tRNA Leu(UUR) 3243A3 G and tRNA Lys 8344A3 G mutations were severe mtDNA mutations causing profound mitochondrial defects (19,48). In this regard, the heteroplasmic levels of these mtDNA mutations in tissues such as the endocrine pancreas or cochlea determined the phenotypic manifestations (69). In this investigation, the homoplasmy, mild mitochondrial dysfunction, late onset, and incomplete penetrance of diabetes and deafness observed in these Chinese families suggested that the m.14692A3 G mutation is an inherited risk factor for the development of deafness and diabetes. Thus, the other genetic, epigenetic, and environmental modifier factors may be involved in the phenotypic manifestation of m.14692A3 G mutation in these Chinese pedigrees (13,60,68,70). Alternatively, tissue-specific differences in tRNA metabolism may also contribute to the development of deafness and diabetes (71,72).
In summary, our findings convincingly demonstrate the pathogenic mechanism leading to impaired oxidative phosphorylation in cells carrying the deafness-and diabetes-associated m.14692A3 G mutation in the tRNA Glu gene. The m.14692A3 G mutation altered the structure and function of tRNA Glu . This included the loss of ⌿55 modification, improper folding of tRNA Glu , and a decrease of the steady-state level in tRNA Glu . A failure in tRNA metabolism impaired mitochondrial translation and respiration. Therefore, the respiratory deficiency impaired mitochondrial ATP production, membrane potentials, and production of oxidative reactive species. Therefore, mitochondrial dysfunction caused by the m.14692A3 G mutation manifests as MIDD. However, the tissue specificity of this pathogenic mtDNA mutation is likely due to the involvement of nuclear modifier genes or tissue-specific differences in tRNA metabolism. Thus, our findings may provide new insights into the pathophysiology of MIDD, which is manifested by the deficient modification of mitochondrial tRNA Glu .

Experimental Procedures
Subjects-A total of 2651 Chinese hearing-impaired probands were recruited from the Otology Clinics, Wenzhou Medical University, China, as detailed previously (25,26). This study was in compliance with the Declaration of Helsinki. Informed consent, blood samples, and clinical evaluations were obtained from all participating family members under protocols approved by the Ethics Committees of Zhejiang University and Wenzhou Medical University. Audiological and neurological examinations of hearing impairment were defined as detailed previously (36). Diagnosis of diabetes was based on the criteria of the American Diabetes Association (37). All available members of three pedigrees carrying the m.14692A3 G mutation were evaluated at length to identify both personal and family medical histories of deafness, diabetes, and other clinical abnormalities. The 574 control subjects were obtained from a panel of unaffected subjects of Han Chinese ancestry from the same region.
Mutational Analysis of Mitochondrial DNA-Genomic DNA was isolated from whole blood of 2651 probands and 574 Chinese control subjects by using the QIAamp DNA Blood Mini Kit (Qiagen, 51104). Subject DNA fragments spanning the mitochondrial tRNA Glu gene were amplified by PCR using oligodeoxynucleotides corresponding to mtDNA at positions 14448 -14747 (8). Each fragment was purified and subsequently analyzed by direct sequencing. These sequence results were compared with the updated consensus Cambridge sequence (GenBank accession number NC_012920) (8). The entire mitochondrial genomes of three probands (WZD81-III-9, WZD82-III-1, and WZD83-III-2) carrying the m.14693A3 G mutation and two control subjects (C3 and C4) were PCR-amplified in 24 overlapping fragments using sets of the light (L) strand and the heavy (H) strand oligonucleotide primers, as described previously (73). These sequence results were compared with the updated consensus Cambridge sequence as described above. To quantify the m.14692A3 G mutation, the PCR segment (300 bp) was amplified using genomic DNA as the template and oligodeoxynucleotides corresponding to mtDNA at positions 14448 -14747. Equal amounts of various digested samples were then analyzed by electrophoresis through 3% agarose gel. After ethidium bromide staining, an Image-Quant program was used to determine the proportions of digested and undigested PCR product to determine whether the m.14692A3 G mutation is in homoplasmy in these subjects.
Cell Lines and Culture Conditions-Lymphoblastoid cell lines were immortalized by transformation with the Epstein-Barr virus as described elsewhere (74). Cell lines derived from two affected subjects (WZD81-III-6 andWZD81-III-9) carrying the m.14692A3 G mutation and two Chinese control subjects (C3 and C4) lacking the mutation were grown in RPMI 1640 medium supplemented with 10% FBS.
Detection of Pseudouridine Residues in tRNA Glu Using CMCT Modification/Reverse Transcription-Total mitochondrial RNA was obtained from mitochondria isolated from lymphoblastoid cell lines (ϳ1.0 ϫ 10 8 cells) (75). 2 g of mitochondrial RNA was incubated with 160 mM CMCT for 20 min at 37°C to allow for CMCT modification of ⌿ residues. The reaction contained 7 M urea, 4 mM EDTA, and 50 mM Bicine (pH 8.5). Then the modified RNAs were precipitated by adding 2 l of pellet paint co-precipitant, 50 l of 3 M sodium acetate (pH 5.5), and a triple volume of ethanol and incubated for at least 2 h at Ϫ20°C before centrifuging at 12,000 rpm for 30 min. Then we dissolved the RNA pellets in 1 M sodium carbonate (pH 10.4), incubated for 4 h at 37°C, and precipitated the RNA again as described above. The Primescript II First Strand cDNA Synthesis Kit (TAKARA) was used for reverse transcription with a digoxigenin (DIG)-labeled oligodeoxynucleotide (5Ј-TGG-TATTCTCGCACG-3Ј) probe specific for tRNA Glu . RNase A was added to the extension reaction to remove the mitochondrial RNA. Then the DNA was precipitated by ethanol at Ϫ20°C overnight after phenol extraction. 2 g of DNA samples was applied onto 15% polyacrylamide, 7 M urea electrophoresis gel and electroblotted onto a positively charged nylon membrane. The quantification of density in each band is described elsewhere (29).
In Vitro Angiogenin Cleavage Assay-In vitro transcriptions of human mitochondrial tRNA Glu (wild-type and mutant) were performed as described previously (76). 2 g of purified tRNA transcripts or human mitochondrial RNAs were used for the cleavage reaction with 2.5 g/ml recombinant angiogenin in the buffer containing 30 mM HEPES (pH 7.4), 30 mM NaCl, and 0.01% bovine serum albumin. Mixtures were incubated at 37°C for the indicated times and quenched by adding 5 l of gel loading buffer. Cleavage products of tRNA transcripts were electrophoresed through a denaturing polyacrylamide gel and stained with methylene blue. Cleavage products of human mitochondrial RNAs were resolved in 15% denaturating polyacrilamide gels with 8 M urea, electroblotted, and hybridized with a DIG-labeled oligonucleotide probe specific for the tRNA Glu (30).
Mitochondrial tRNA Analysis-The tRNA concentrations from in vitro transcription were measured by UV absorbance at 260 nm, and the extinction coefficient was calculated according to the sequence of each tRNA (77). 1 g of each tRNA transcript was electrophoresed through a 10% polyacrylamide gel (native gel) or with 8 M urea (denaturing gel). The UV melting studies for tRNA transcripts were performed as described elsewhere (46).
For the tRNA Northern blotting analysis, 2 g of total mitochondrial RNA was electrophoresed through a 10% polyacrylamide/8 M urea gel in Tris borate-EDTA buffer after heating the sample at 65°C for 10 min and then electroblotted onto a positively charged nylon membrane for the hybridization analysis with oligodeoxynucleotide probes. Oligodeoxynucleosides used for DIG-labeled probes of tRNA Glu , tRNA His , tRNA Ile , tRNA Ala , tRNA Thr , and 5S rRNA were described as elsewhere (26, 42, 78 -80). The hybridization and the quantification of density in each band was made as detailed previously (42,79).
Assays of Activities of Respiratory Complexes-The enzymatic activities of complexes I, II, III, and IV were assayed as detailed elsewhere (40 -42).
Measurements of Oxygen Consumption-The rates of OCR in lymphoblastoid cell lines were measured with a Seahorse Bioscience XF-96 extracellular flux analyzer as detailed previously (26,42,81).
ATP Measurements-The CellTiter-Glo luminescent cell viability assay kit (Promega) was used for the measurement of cellular and mitochondrial ATP levels according to the modified instructions of the manufacturer (26,42).
Assessment of Mitochondrial Membrane Potential-Mitochondrial membrane potential was assessed with the JC-10 Assay Kit-Microplate (Abcam) following the general recommendations of the manufacturer, with some modifications, as detailed elsewhere (26,58).
Author Contributions-M. X. G. designed the experiments and wrote the main manuscript text. M. W., H. L., M. Z., and X. L. performed biochemical assays. J. Z., L. W., and H. W. performed the mutational screening of mitochondrial tRNA genes. B. C. and W. F. collected clinical data and performed the evaluation. X. Z., G. E., and P. J. performed data analysis. All authors reviewed the manuscript.