Human Molecular Genetics, 2008, Vol. 17, No. 6 775–788 doi:10.1093/hmg/ddm349 Advance Access published on November 27, 2007 Transcriptional activators HAP/NF-Y rescue a defect in yeast and human cells

Flavia Fontanesi1, Can Jin3, Alexander Tzagoloff3 and Antoni Barrientos1,2,

1Department of Neurology and 2Department of Biochemistry & Molecular Biology, The John T. MacDonald Foundation Center for Medical Genetics, University of Miami Miller School of Medicine, Miami, FL, USA and Downloaded from https://academic.oup.com/hmg/article/17/6/775/600424 by guest on 29 September 2021 3Department of Biological Sciences, Columbia University, New York, NY, USA

Received October 29, 2007; Revised and Accepted November 26, 2007

Cell survival and energy production requires a functional mitochondrial respiratory chain. Biogenesis of cytochrome c oxidase (COX), the last of the mitochondrial respiratory chain, is a very complicated process and requires the assistance of a large number of accessory factors. Defects in COX assembly alter cellular respiration and produce severe human encephalomyopathies. in SURF1, a COX assembly factor of exact unknown function, produce Leigh’s syndrome (LS), the most frequent cause of COX deficiency in infants. In the yeast Saccharomyces cerevisiae, deletion of the SURF1 homologue SHY1 results in a similar COX deficiency. In order to identify genetic modifiers of the shy1 mutant phenotype, we have explored for genetic interactions involving SHY1. Here we report that overexpression of Hap4p, the catalytic subunit of the CCAAT binding transcriptional activator Hap2/3/4/5p complex, suppresses the respiratory defect of yeast shy1 mutants by increasing the expression of nuclear-encoded COX subunits that interact with the mitochondrially encoded Cox1p. Analogously, overexpression of the Hap complex human homologue NF-YA/B/C transcription complex in SURF1-deficient fibroblasts from an LS patient effi- ciently rescues their COX deficiency.

INTRODUCTION The subunits composing COX are thought to be matured and inserted into the nascent complex in an ordered fashion. Eukaryotic cytochrome c oxidase (COX), the terminal enzyme At least four COX assembly intermediates (or subcomplexes) of the respiratory chain, catalyzes electron transfer from cyto- can be detected in human mitochondria, which represent major chrome c to molecular . COX is located in the mito- steps in the COX assembly process (1,2). Although more dif- chondrial inner membrane and is made up of some dozen ficult to detect, COX subassemblies have also been described different subunits encoded by both mitochondrial and in some yeast mutants (3,4). Shy1p is an important COX nuclear DNA. The catalytic core of the enzyme consists of assembly factor which functions early in the assembly three subunits (Cox1p, 2p and 3p) all of which are mitochon- pathway (5,6). Identifying the precise function of this drial products. Cox1p contains two redox centers, one is of considerable interest because mutations in its formed by a low-spin heme a and the other by a binuclear human homologue, SURF1, are responsible for most diag- center consisting of a high-spin heme a3 and CuB. A third nosed cases of Leigh’s syndrome (LS) presenting a COX redox center is formed by the two copper ions present in the deficiency, a devastating early-onset encephalomyopathy CuA site of Cox2p. The 8 (yeast) or 10 (mammalian) (7,8). Unlike other assembly-defective strains of yeast that nuclear encoded subunits act as a protective shield surround- display a complete absence of COX, shy1 mutants produce ing the catalytic core. 10–15% fully assembled and functional COX (9), similar to

To whom correspondence should be addressed at: Department of Neurology and Biochemistry & Molecular Biology, The John T. Macdonald Center for Medical Genetics, Universtiy of Miami Miller School of Medicine, 1600 NW 10th Ave., RMSB # 2067, Miami, FL 33136, USA. Tel: þ1 3052438683; Fax: þ1 3052433914; Email: [email protected]

# The Author 2007. Published by Oxford University Press. All rights reserved. For Permissions, please email: [email protected] 776 Human Molecular Genetics, 2008, Vol. 17, No. 6

LS patients who also have 10–30% of normal enzyme (7,8). The shy1 null and point mutant strains transformed with an This suggests that the function of Shy1p/SURF1 increases empty plasmid respire at 15 or 20% of the wild-type rate, the efficiency of some step in the assembly process. Several respectively, in whole cell assays performed in minimum functions have been proposed for Shy1p/SURF1 including media containing galactose. The KCN-sensitive respiratory (i) incorporation of subunit II into the nascent intermediates activity is increased to 42–55% of wild-type in the shy1 composed of subunit I plus subunit IV and V (yeast subunits null and point mutants over-expressing HAP4 (Fig. 1B). The 5a and 6) (2,10), (ii) assembly/stability of the heme a3 presence of an additional chromosomally integrated copy of present in the binuclear center of Cox1p (6). HAP4 in the Dshy1 strain increased its respiratory capacity The COX deficiency of yeast shy1 mutants is rescued by consistently from 15 to 20% (Fig. 1B) which is insufficient mutations in Mss51p, a COX1 mRNA specific translational to produce a visible growth phenotype on solid YPEG (Sup- activator, which increase Cox1p expression in the shy1 plementary Material, Fig. S2). mutants (5,11). Suppressor mechanisms may develop also in mammals and they could open a window for therapeutic inter- Downloaded from https://academic.oup.com/hmg/article/17/6/775/600424 by guest on 29 September 2021 ventions in LS patients. In line with this, SURF1 knockout Partial restoration of shy1 respiratory growth defect by mice have a mild mitochondrial disease and COX deficiency HAP4 over-expression results from an increase in COX mainly affecting muscle and liver, and virtual absence of Spectra of mitochondrial cytochromes indicated the HAP4 overt neurological symptoms (12). over-expressors contain more ‘a’-type cytochromes than To detect new genetic interactions of SHY1, we have shy1 mutants (Fig. 1C). This was paralleled by a similar searched for multicopy suppressors of the shy1 respiratory increase in overall NADH oxidase activity, which was and COX defects. In this communication, we present evidence 2.5-fold higher in the transformant and corresponded to 35– that HAP4, which codes for the catalytic subunit of the CCAAT 52% of the wild-type rate (Fig. 1D). The COX activity, which binding transcriptional activator complex Hap2/3/4/5p, can act was measured either polarographically following the oxidation as a multicopy suppressor of shy1 mutants by over-expression of ascorbate in the presence of TMPD (N,N,N’,N’-tetramethyl of Cox5ap and Cox6p, which have been proposed to interact p-phenylenediamine) or spectrophotometrically by following with Cox1p at early step of COX assembly. The mammalian the oxidation of ferrocytochrome c at 550 nm was also 2.5 homologue of the Hap complex is the ubiquitous transcriptional times higher in the HAP4 over-expressors than in the shy1 activator system NF-Y (13). Like Hap4p, over-expression of the mutants and ranged between 35 and 50% of wild-type values catalytic subunit NF-YA rescues the COX activity defect of (Fig. 1D and E). Consistently, the steady-state concentrations SURF1 mutant fibroblasts from an LS patient. of COX subunits in the HAP4 transformants were higher com- pared with the shy1 mutants (Fig. 1F). Interestingly, the NADH cytochrome c reductase activity, RESULTS which was previously reported to be higher in shy1 mutants (9), remained increased in the suppressed strains (Fig. 1E), HAP4 is a high copy suppressor of the shy1 mutants and probably represents a compensatory mechanism of the respiratory growth defect partial COX defect. In order to explore for genetic interactions involving SHY1, the C173/U1 strain carrying a point in the SHY1 Only the Hap4p component of the Hap2/3/4/5p complex region coding for the first transmembrane domain of Shy1p restores the respiratory defect of shy1 mutants (9) was used in a multicopy suppressor screen. Twenty respir- atory competent clones were obtained with a yeast genomic Hap4p is the catalytic subunit of a multimeric transcription multicopy plasmid library constructed in YEp24. With one activator complex formed by three additional subunits, exception all the clones contained plasmids with SHY1. The Hap2p, Hap3p and Hap5p. Interestingly only two of these sub- restriction map of pG91/T5, the single plasmid with an unre- units, Hap4p and Hap2p, have nuclear import sequences. lated nuclear DNA insert, had end points between coordinates Hap3p and 5p form a ternary complex with Hap2p in the cyto- 229 300 and 240 578 on XI (Supplementary plasm and are imported together by a piggy-back mechanism Material, Fig. S1). This plasmid was used to subclone the sup- (15). The ternary complex binds to the consensus sequences pressor gene by transferring different regions of the insert to a present in the promoter of their target and, upon yeast episomal vector and testing the ability of the new con- binding of Hap4p to the complex, transcription is activated. structs to confer respiration in Dshy1 cells. The suppressor HAP2, HAP3 and HAP5 genes are constitutively expressed, activity mapped to a XmaI/PstI fragment containing only whereas HAP4 expression is highly regulated by carbon HAP4 (Supplementary Material, Fig. S1), coding for the cata- source: cells grown in media containing lactate, a respiratory lytic subunit of the Hap complex (14). HAP4 acted as high substrate, have 4–5-fold more HAP4 mRNA than cells copy suppressor of both, shy1 null and point mutant strains grown in media containing glucose (14,16). (Fig. 1A and Supplementary Material, Fig. S2). HAP4 trans- We tested if over-expression of the Hap4p partner subunits formants had a generation time in respiratory complete could suppress shy1 mutants. No clearly detectable effect of ethanol–glycerol containing media of 4.2 h compared with HAP2, 3 and 5 on the ability of shy1 mutants to grow after 4 28 h for the shy1 mutant and 2.4 h for the wild-type. days on YPEG was noted (Fig. 1A). However, the respiratory Expression of HAP4 from an integrative plasmid did not activity of the null mutant and the W125 point mutant expres- rescue the respiratory growth defect (Supplementary Material, sing HAP2 from a high copy plasmid increased from 15 to Fig. S2). 20% and from 18 to 25% of the wild-type values, respectively Human Molecular Genetics, 2008, Vol. 17, No. 6 777 Downloaded from https://academic.oup.com/hmg/article/17/6/775/600424 by guest on 29 September 2021

Figure 1. Characterization of shy1 mutants and HAP transformants. (A) The respiratory competent wild-type strain W303 and a Dshy1 strain transformed with either empty plasmids (ep) or plasmids over-expressing genes for the different subunits of the HAP complex (HAP2, 3, 4 and 5) were grown over-night in liquid selective WO-D media. Ten-fold serial dilutions of the strains were plated on solid YPD or YPEG media and incubated at 308C. Pictures were taken after 2 and 4 days of incubation, respectively. (B) KCN-sensitive endogenous cell respiration was measured polarographically in the presence of galactose in shy1 null (Dshy1) and point mutants (W125 and C173) transformed with plasmids over-expressing the different subunits of the HAP complex. Where indicated, the genes were expressed from either integrative (i-) or episomal (e-) plasmids. (C) Total mitochondrial cytochrome spectra. The absorption bands corresponding to cytochromes a and a3 have maxima at 603 nm (a,a3). The maxima for cytochrome b (b) and for cytochrome c and c1 (c,c1) are 560 and 550 nm, respectively. (D) Oxygen consumption was assayed polarographically in mitochondria prepared from the indicated strains in the presence of either NADH or ascorbate-TMPD as sub- strates as described under Materials and Methods. (E) Mitochondrial respiratory chain enzyme spectrophotometric measurements in isolated mitochondria. Cyto- chrome oxidase (COX) and NADH cytochrome c reductase (NCCR) were measured as described under Materials and Methods. (F) Steady-state levels of COX subunits. Total mitochondrial (30 mg) separated by 12% SDS–PAGE were transferred to nitrocellulose and probed with subunit-specific antibodies to COX subunits 1, 2, 4 and 5. The western blots were also probed with an antibody that recognizes a 45 kDa mitochondrial outer membrane protein (MOM45) to normalize the signals for protein loading. The bars indicate the mean + SD from at least three independent sets of measurements. 778 Human Molecular Genetics, 2008, Vol. 17, No. 6 Downloaded from https://academic.oup.com/hmg/article/17/6/775/600424 by guest on 29 September 2021

Figure 1. Continued.

(Fig. 1B). A possible explanation is that over-expression shy1 null mutants by enhancing synthesis of Cox1p (5). of HAP2 increases the amount of the ternary Hap2/3/5p HAP4 could indirectly affect Cox1p by increased DNA binding complex in the nucleus. expression of Mss51p. Northern blot analyses showed that the steady-state levels of MSS51 mRNA in Dshy1 cells and in mutant cells over-expressing HAP4 were virtually identical Suppression by HAP4 over-expression is not common (data not shown). Additionally, western blot analyses showed among null mutants of COX assembly factors the amount of Mss51p in the mitochondrial membranes was Suppression of the respiratory growth defect of different COX not increased in shy1 mutants over-expressing HAP4 (Sup- assembly defective strains was assessed by transforming plementary Material, Fig. S3). cox10, cox11, , cox15, cox16, , oxa1, mss51 and Suppression by HAP4 could also be the result of a general pet191 null mutants with HAP4 on a multicopy plasmid. A increase in mitochondrial mass resulting in increased null mutant of cox5a was also included in the screening expression of mitochondrial proteins. HAP4 has been shown because, similar to Dshy1 cells, it retains 15% of COX to stimulate mitochondrial biogenesis (17). This is supported activity, in this case resulting from enzyme containing the by a significant increase in mitochondrial DNA nucleoids in isoform Cox5bp. In all cases tested, HAP4 over-expression wild-type and shy1 mutant cells over-expressing this transcrip- failed to induce any suppression of their respiratory growth tion factor although the nucleoids are of smaller size (Sup- defect (data not shown), indicating that the suppression by plementary Material, Fig. S4A and B). There is also a HAP4 was specific to the shy1 null mutant, and suggesting a significant alteration in the morphology of the mitochondrial specific mechanism of HAP4 suppression in this strain. network which appears to be more fragmented, characteristic The specificity of HAP4 suggested that it might act by either of conditions favoring mitochondrial biogenesis (Supplemen- increasing the expression of a protein with an overlapping tary Material, Fig. S4C). Levels of mitochondrial DNA were function with Shy1p or increasing the amount of specific estimated by Southern-blot analyses and were not significantly COX subunits and/or assembly factors favoring a more effi- increased in wild-type or mutant cells overexpressing HAP4 cient assembly of COX in the absence of Shy1p. (data not shown). Also, levels of COX1, and COX2 mRNAs, were not increased in cells overexpressing HAP4 as estimated by northern blot analyses of mitochondrial RNAs (Fig. 2A). Suppression by HAP4 over-expression Accordingly, in vivo translation assays did not show any sig- does not involve MSS51 nificant increase of 35S-methionine incorporation into newly Mutations in the COX1 mRNA specific translational activator synthesized mitochondrial gene products in HAP4 over- MSS51 or over-expression of wild-type MSS51 suppresses expressing strains when cells were grown in galactose Human Molecular Genetics, 2008, Vol. 17, No. 6 779 Downloaded from https://academic.oup.com/hmg/article/17/6/775/600424 by guest on 29 September 2021

Figure 2. Effect of HAP4 overexpression on the steady-state levels of COX-related mRNAs. (A) Northern blots of total RNA probed for COX11, COX10, HEM2, COX4, COX5a, COX6 and ACT1. After processing, the membranes were exposed to X-ray film. In the lower panel, the images were digitalized and densitometry performed using the histogram function of the Adobe Photoshop program. The values were normalized by the signal of ACT1 as the loading control and expressed as the value relative to the control. (B) Northern blots of mitochondrial RNA extracts were hybridized with probes contain- ing the entire coding sequence of cytochrome b (COB), subunits 1 (COX1)and2(COX2) and of 21S-rRNA as a loading control. The precursor (p) and mature (m) forms of COB and COX1 mRNAs are identified in the margin. In the lower panel the intensities of the bands quantified as in (A) were nor- malized by the 21S-rRNA signal and expressed as the value relative to the control. The values measured in two independent assays did not differ by more than 5%.

(Fig. 3A) and only a mild increase in cells submitted to a diauxic mss51F199I revertant transformed with HAP4 was further shift (Fig. 3B). The apparent discrepancy between the observed increased to 85% (Fig. 3C). This strain grew on non- increase in mitochondrial nucleoids and the absence of obvious fermentable carbon sources (either YPEG or the more stringent mitochondrial protein synthesis enhancement in cells over- WO-EG) significantly better than the HAP4 over-expressor expressing HAP4 can be explained by the absence of signifi- or the shy1 revertant (Fig. 3D). These results suggest that cant increase in mitochondrial DNA and RNA, and by HAP- the HAP4 suppression is not related to an increase in Cox1p independent factors such as Mss51p that may be limiting synthesis but to enhancement of downstream events in the translation of Cox1p. These results also argue against an maturation/assembly of this and/or other COX subunits. Mss51p-mediated mechanism of HAP4 suppression. Despite Copper and heme A are both prosthetic groups of Cox1p. some variability intrinsic to in vivo mitochondrial protein syn- Because Shy1p could be involved in maturation of Cox1p by thesis experiments, the amount of newly synthesized Cox1p contributing to the formation and/or stabilization of the was higher in the HAP4 suppressor than in the mutant after CuB-heme a3 center, we tested if supplementation of the 2 h of chase (Fig. 3B) suggesting that a larger amount of growth media with either hemin or copper induced any rescue newly synthesized Cox1p is spared from proteolysis and can of the respiratory defect observed in shy1 mutants. No effect proceed in the assembly pathway. on the growth of the shy1 null mutant strain was observed in Suppression by HAP4 and MSS51 were found to be addi- the presence of hemin (not shown). As shown in Fig. 3D and tive. A shy1 strain that reverted to a respiratory-competent Supplementary Material, Fig. S5A, respiratory media sup- F199I phenotype by a spontaneous mss51 mutation (5) respired plementation with 0.15% CuSO4 allowed a null mutant of shy1 at 55% the wild-type rate. The comparable value for the HAP4 to grow more efficiently although not at wild-type levels. The suppressed mutant was 40%. The respiratory activity of the improved respiratory growth resulted from an increased cell 780 Human Molecular Genetics, 2008, Vol. 17, No. 6 Downloaded from https://academic.oup.com/hmg/article/17/6/775/600424 by guest on 29 September 2021

Figure 3. Additive effects of mss51 mutations and copper supplementation to the media on the shy1 suppression by HAP4 over-expression. (A) In vivo mito- chondrial protein synthesis. Wild-type (WT), the null mutant Dshy1 cells (Dshy1) transformed with either an empty plasmid (2) or a plasmid over-expressing either HAP4 or HAP2, a spontaneous revertant of shy1 carrying a F199I mutation in MSS51 (5), and the revertant strain over-expressing HAP4, growing in minimum media containing 2% galactose and supplemented with the appropriate auxotrophic requirements were labeled with [35S]-methionine at 308C for 15 min in the presence of cycloheximide. Equivalent amounts of total cellular proteins were separated by PAGE on a 17.5% polyacrylamide gel, transferred to a nitrocellulose membrane and exposed to an X-ray film. The mitochondrial translation products are identified on the left. (B) The same strains as in (A) plus a wild-type strain overexpressing HAP4 and a Dshy1 strain overexpressing COX5a and COX6, were grown on minimum glucose media overnight and trans- ferred to minimum ethanol/glycerol media during 4 h. Protein synthesis was performed as in (A). After 15 min pulse, 4 mg/ml puromycin and excess of cold methionine were added to stop the incorporation reaction and chase the synthesized products for the indicated times. To quantify the Cox1p signal images were digitalized and densitometry performed using the histogram function of the Adobe Photoshop program. The values measured in two independent assays did not differ by more than 10%. The values reported are averages of the two assays. (C) Endogenous cell respiration in cells grown in liquid WO-Gal media measured polarographically as described under Materials and Methods. (D) The same strains as in (A) were grown over-night in liquid selective WO-D media. Ten-fold serial dilutions of the different strains were plated on solid YPD, WO-EG, YPEG and YPEG media supplemented with 0.15% CuSO4 and incubated at 308C. Pictures were taken after 2, 5 and 4 days of incubation respectively. (E) Same as in (C) with the exception that the different strains were grown in solid YPEG plates supplemented or not with 0.15% of CuSO4. Immediately before the assay, the cells were scraped from the plates and suspended in YPEG liquid media. The bars indicate the mean + SD from at least three independent sets of measurements. respiratory capacity in YPEG from 20% of wild-type in the (Supplementary Material, Fig. S5B). The suppression by absence of Cu supplementation to almost 40% in the presence copper supplementation was no additive to either the suppres- of the metal, a consequence of an increase in COX activity sion by mutations in mss51 or HAP4 over-expression (Fig. 3E). Human Molecular Genetics, 2008, Vol. 17, No. 6 781

Suppression by HAP4 over-expression does not involve mutants. Over-expression of COX4 in combination with each over-expression of other COX assembly factors of the other genes did not produce any effect. However, over- expression of COX5a with COX6 significantly improved the Data derived from systematic analyses of yeast gene growth on non-fermentable minimal ethanol/glycerol and expression (17–19) summarized in Supplementary Material, increased the respiratory activity of the D Table S1, have revealed that although nuclear-encoded COX shy1 mutant from assembly factors are not co-regulated, some are directly or 15 to 30–32% of wild type. indirectly regulated by the Hap system. The enhanced efficiency of COX assembled in the absence In our system, HAP4 expression was under the control of its of Shy1p can be explained by an increase in a subassembly con- own promoter, which is repressed by glucose. Overexpression taining Cox5ap and Cox6p. In these conditions a larger amount of HAP4 in a Dshy1 mutant was induced in cells undergoing a of newly synthesized Cox1p is spared from proteolysis, as diauxic shift by transferring them from galactose to ethanol– observed in our chase experiments (Fig. 3B), and can be matured to form a ternary complex with Cox5ap–6p. This glycerol containing media for 4 h. Northern blot analyses Downloaded from https://academic.oup.com/hmg/article/17/6/775/600424 by guest on 29 September 2021 allowed us to estimate HAP4 mRNA to be increased 5-fold in result is consistent with results obtained with human fibroblasts cells overexpressing HAP4 (data not shown). In these metabolic from LS patients carrying mutations in SURF1, which were conditions, we assessed the steady-state mRNA levels of a found to accumulate a COX assembly intermediate consisting series of candidate genes involved in COX biogenesis. We of subunit I and subunits CoxIV and CoxVa, the latter two did detect a higher than 2-fold increase in the expression of being the human homologues of yeast Cox5ap and Cox4p, COX10, a gene involved in heme a biosynthesis (Fig. 2B) and respectively (2). Hap4p increases cell respiration to 42% of a consistent 1.5-fold increase in the amount of HEM2 mRNA, wild-type, while concomitant over-expression of Cox5ap and the gene encoding for the d-aminolevulinate dehydratase, Cox6p achieves a rate of 30% of wild-type, suggesting that which catalyzes the conversion of delta-aminolevulinic acid other COX related factors may contribute to suppression to porphobilinogen, the second and rate-limiting step in the by HAP4. In this line, it was recently reported that co-over- heme biosynthetic pathway and known to be regulated by the expression of COX10 enhanced the suppressor effect of MSS51 D Hap system (20). Overexpression of these 2 genes in a shy1 over-expression in a shy1 strain, probably by increasing the mutant strain did not produce any suppression of its respiratory amount of matured heme A-containing Cox1p (22). deficiency. No significant change was detected in the expression To further test if suppression by HAP4 is mediated through of COX11 (Fig. 2B), a gene involved in copper delivery to higher expression of Cox5a and 6p, we changed the promoter Cox1p. Although shy1 mutants are rescued by copper sup- of COX5a by inserting a constitutive HAP4-independent pro- plementation to the media, overexpression of COX11 and moter using a yeast promoter collection comprising 11 other genes involved in mitochondrial copper metabolism mutants of the strong constitutive Saccharomyces cerevisiae TEF1 (translation and elongation factor 1) promoter (23). such as COX17, COX19, SCO1 and COX23 neither produced The promoter activities of these mutants range from 8– any suppression of the shy1 respiratory defect nor enhanced 120% of the normal TEF1 promoter, independent of the the suppression by copper supplementation (data not shown). carbon source. A cassette with promoter TEF1-1, which retain 100% of TEF1 activity, was used to replace the promo- ter of COX5a in wild-type and Dshy1 strains overexpressing The mechanism of suppression by HAP4 over-expression is HAP4 (Fig. 4C). The substitution did not compromise the mediated through the over-expression of nuclear encoded ability of the wild-type to grow in non-fermentable carbon COX structural subunits sources (Fig. 4D), indicating that the amount of COX5a In contrast to COX assembly factors, all the nuclear genes mRNA expressed was enough to sustain a wild-type respir- coding for COX subunits except COX5b, are transcriptionally ation level. In the Dshy1 mutant over-expressing HAP4, the co-regulated by the Hap activation complex, which interacts substitution of the COX5a by the TEF1-1 promoter reduced with the Hap2/3/5p consensus CCAAT target sequence in the levels of COX5a mRNA to that of the Dshy1 strain their promoters (19) (Supplementary Material, Table S1). In (Fig. 4E) and abolished the suppressed phenotype (Fig. 4D). our experiments, northern blot analyses showed that in shy1 mutants undergoing a diauxic shift, overexpression of HAP4 induced a 1.7–1.9-fold increase in the steady-state levels of Effect of over-expression on shy1 mutants of the human COX4, COX5a and COX6 mRNA (Fig. 2B). Overexpression CCAAT binding transcription factor NF-Y subunit A, of either COX4 or COX6 did not produce any significant effect homologue of yeast HAP2 on the ability of shy1 mutants to grow (Fig. 4A) and respire on EG (Fig. 4B). COX5a overexpression, however, increased The Hap complex is conserved from yeast to humans where shy1 cell respiration on minimum ethanol/glycerol media from the complex, termed NF-Y, contains three subunits, NF-YA, 15–21% of wild-type values (Fig. 4B), resulting in a clearly B and C, homologues of Hap2p, 3p and 5p, respectively. observable growth of the mutant on this medium (Fig. 4A). Some of the subunits are interchangeable across species. For Cox5ap is stabilized by Cox6p but not Cox4p (21). In example, human NF-YA, that contains both a DNA binding addition, Cox5ap and Cox6p have been found to physically domain and a Q-rich activation domain, complements the res- interact forming a sub-complex that accumulates in some piratory defect of a hap2 yeast mutant (13). In our system, COX assembly mutants (3). For these reasons, we decided overexpression of human NF-YA was not able to suppress to explore the effects of different combinations of the the Dshy1 respiratory growth defect in YPEG but it did COX nuclear structural genes on respiration of shy1 slightly increase its respiratory capacity from 15 to 20% 782 Human Molecular Genetics, 2008, Vol. 17, No. 6 Downloaded from https://academic.oup.com/hmg/article/17/6/775/600424 by guest on 29 September 2021

Figure 4. Suppresion of shy1 by HAP4 overexpression is mediated by overexpression of Cox5ap and Cox6p. (A) The respiratory competent wild-type strain W303 and a strain carrying a null allele of shy1 (Dshy1) transformed with either empty plasmids (ep) or plasmids over-expressing COX4, COX5a and COX6 alone or in combination were grown over-night in liquid selective WO-D media. Ten-fold serial dilutions of the two strains were plated on solid YPD or WO-EG media and incubated at 308C. Pictures were taken after 2 and 5 days of incubation, respectively. (B) KCN-sensitive endogenous cell respiration was measured polarographically in the same strains than in (A) growing in WO-EG. The bars indicate the mean + SD from at least three independent sets of measurements. (C) Diagram representing the strategy used to replace the promoter of COX5a by several mutated versions of the constitutive HAP4-independent transcription factor 1 (TEF) promoter retaining different levels of expression induction with respect to the unmutated TEF1 promoter (modified from 23). (D) The respiratory competent wild-type strain W303 and a strain carrying a null allele of shy1 (Dshy1) overexpressing HAP4, either maintaining their endogenous pro- moter in COX5a or having it substituted by the mutated version of TEF1, TEF1-1, were grown over-night in liquid selective WO-D media. Ten-fold serial dilutions of the two strains were plated on solid YPD or WO-EG media and incubated at 308C. Pictures were taken after 2 and 5 days of incubation, respectively. (E) Steady-state levels of COX5a mRNA detected by northern blot in the strains mentioned in (D). similar to what was observed by overexpressing yeast HAP2 and oxidative phosphorylation (19). In contrast, the mamma- (Supplementary Material, Fig. S6). lian genes are regulated by a variety of transcription factors, most notably NRF-1 and NRF-2 (24). Recently expression of at least two key proteins of the mammalian mitochondrial Restoration of the COX assembly defect of fibroblasts translational machinery, mitoribosomal protein S12 (Mrps12) from an LS patient carrying a mutation in SURF1 and mitochondrial seryl-tRNA ligase, were shown to be by over-expression of NF-YA regulated by NF-Y (25). Since NF-Y regulates expression of The Hap2/3/4/5 complex is the main activator of transcription fatty acid oxidation genes, its stimulatory effect on mito- of yeast nuclear genes involved in mitochondrial biogenesis chondrial biogenesis could be indirect as reported for other Human Molecular Genetics, 2008, Vol. 17, No. 6 783 transcriptional factors such as the peroxisome proliferator- of Hap4p rescues the COX assembly lesion and the respiratory activated receptor PPARd in skeletal muscle (26). defect of yeast shy1 mutants. We tested the effect NF-Y on a fibroblast cell line from an To identify the targets of Hap4p action responsible for the LS patient carrying a homozygous mutation in SURF1 (27) suppression, we first tested if suppression might be mediated and displaying a severe COX deficiency, a phenotype that through enhanced expression of MSS51, which was previously was also maintained after immortalization as demonstrated shown to suppress shy1 mutants. Expression of MSS51 was by the absence of the characteristic cytochemical staining found to be independent of Hap4p thereby excluding its (Supplementary Material, Fig. S7). Immunostaining for COX product from playing a role in the mechanism of suppression. subunit 1 with an antibody conjugated to Alexa Fluor 488 This was also true of genes coding for proteins involved in showed green staining super-imposable on the mitochondrial heme A biosynthesis, mitochondrial copper homeostasis and network that was intense in normal fibroblasts but significantly other aspects of COX assembly. As an alternative approach, weaker and more diffused in SURF1 mutant fibroblasts (Sup- we over-expressed either singly or in combination candidate plementary Material, Fig. S8). The difference in immunostain- genes known to be regulated by the Hap complex. Downloaded from https://academic.oup.com/hmg/article/17/6/775/600424 by guest on 29 September 2021 ing is even more evident when wild-type and SURF1 mutant This strategy revealed that over-expression of the nuclear- fibroblasts were mixed prior to staining (Supplementary encoded Cox5ap subunit effectively compensated for Material, Fig. S8). absence of Shy1p. Suppression by COX5a was abolished Fibroblasts from the SURF1-LS patient were transformed when the normal promoter of this gene was replaced by a con- with an empty pIRES2 plasmid, and the same plasmid overex- stitutive Hap-independent promoter, indicating that the mech- pressing either SURF1 of NF-YA. Cytochemical staining for anism of suppression depended minimally on Cox5ap COX activity 48 h after transfection showed that while the over-expression. The more efficient rescue of the shy1 empty plasmid transfection failed to show an increase in mutant transformed with COX5a in combination with COX6 COX activity, the fibroblasts over-expressing SURF1 (10– but not COX4 indicated that suppression by Hap4p was also 12% of the culture) recovered COX activity (Fig. 5A). This mediated by increased expression of Cox6p (Fig. 6A). was also true of the fibroblasts that had been transfected COX assembly appears to be a sequential process (1,2) with NF-YA (12–15% of the culture). The cultures from initiated by the synthesis of Cox1p, which during maturation the SURF1 mutant patient transfected with either wild-type of its metal and heme A prosthetic groups forms a subassem- SURF1 or NF-YA showed cells with an intense COX I green bly with Cox5ap and Cox6p (mammalian CoxIVþCoxV) signal that co-localized with the branched mitochondrial (2,3). These subunits are in direct contact with Cox1p in the network stained with Mitotracker red (Fig. 5B). The percen- mature enzyme (28). Cox5ap interacts with Cox1p through tage of COX positive cells is in agreement with a control its single transmembrane domain, while Cox6p caps the transfection with the plasmid pEGF-N1 (Clonthech, Palo matrix side of Cox1p. In S. cerevisiae, Cox1p is extremely Alto, CA) expressing enhanced green fluorescent protein prone to proteolysis in mutants blocked at any assembly (eGFP), which on average yielded 15% cells were transfected step. Over-expression of Cox5ap and Cox6p may restore expressing eGFP. COX assembly by sparing newly synthesized Cox1p from The restoration of COX by SURF1 and NF-YA was con- proteolysis. Suppression of the shy1 mutant by HAP4 and firmed by assays of COX activity in cell homogenates MSS51sup is additive, presumably as a result of increased obtained from the transfected cultures. Only 16% of wild- synthesis and availability of Cox1p for interaction with type activity was detected in the SURF1 mutant fibroblasts. Cox5a and Cox6 and a higher flux through the assembly The COX activity measured in fibroblasts transfected with pathway (Fig. 6A). either wild-type SURF1 or NF-YA increased to 25 or Interestingly, the Cox1p–Cox5ap–Cox6p sub-complex 28% of wild-type, respectively, (Supplementary Material, accumulates in fibroblasts of SURF1 patients (2), although it Fig. S9) consistent with the values expected for full restoration does not accumulate in heme A deficient fibroblasts of patients of COX activity with a 15% transfection efficiency. with lesions in COX10 or COX15 (2,29). These findings suggested that the presence of heme A in Cox1p might stabil- ize its binding to Cox5ap and Cox6p. Because the two heme A groups are deeply buried within the interior of Cox1p (28) their insertion probably occurs either immediately after mem- DISCUSSION brane insertion of the translated but unassembled subunit or Shy1p/SURF1 is a conserved nuclear-encoded COX assembly during formation of the Cox1p–Cox5ap–Cox6p intermediate. factor that plays a role in maturation/assembly of subunit 1 of The presence of SURF1 orthologues in terminal oxidase COX. Here we have reported that overexpression of the cata- operons of several prokaryotes (30) in which COX contains lytic subunits of conserved CCAAT binding transcription acti- the evolutionary conserved core subunits (Cox1p, Cox2p, vators suppress the COX assembly defect of both, yeast shy1 Cox3p) but not the supernumery subunits such as Cox5ap mutant cells and human SURF1 mutant fibroblasts. and Cox6p, suggests a role of Surf1p/Shy1p in forming the The Hap2/3/4/5 transcriptional complex of yeast plays an catalytic core of the enzyme. Studies in Rhodobacter sphaer- important role in globally activating transcription of nuclear oides, suggest that bacterial Surf1p is required either for genes involved in mitochondrial respiration during transition insertion of heme A at the a3 center or stabilization of the from fermentation to respiration. This activation depends on a3-CuB binuclear center in Cox1p (6). A function of Shy1p the synthesis of the Hap4p subunit, which is regulated by in maturation of the a3 center would imply the existence of carbon source (14). Our evidence indicates that over-expression a second chaperone specific for the heme A of the cytochrome 784 Human Molecular Genetics, 2008, Vol. 17, No. 6 Downloaded from https://academic.oup.com/hmg/article/17/6/775/600424 by guest on 29 September 2021

Figure 5. COX expression in SURF1 fibroblasts overexpressing NF-YA. (A) Cytochemical cytochrome c oxidase activity staining. Wild-type (WT) and SURF1 mutant (surf1) fibroblasts were grown in cover-slips. The SURF1 mutant fibroblasts were transformed with an empty plasmid (pIRES2) or plasmids overexpres- sing either wild-type SURF1 or NF-YA. Forty eight hours after transfection, cells were submitted to cytochemical COX activity as described under the Materials and Methods. Pictures were taken at original magnification of 40.(B) Fibroblast cultures were grown on coverslips, stained with Mitotracker red, fixed and immunochemically stained for COX subunit 1. Original magnification: 100. Human Molecular Genetics, 2008, Vol. 17, No. 6 785 Downloaded from https://academic.oup.com/hmg/article/17/6/775/600424 by guest on 29 September 2021

Figure 6. Mechanisms of suppression of the COX assembly defect of yeast shy1 and human SURF1 mutants. (A) Different mechanisms of suppression of the yeast shy1 mutants. After Cox1p synthesis, this subunit is matured by insertion of heme and copper prosthetic groups, before or during the formation of an assembly intermediate containing Cox5ap and Cox6p. This ternary complex will subsequently assemble with metallated Cox2p. Shy1p could play a role either in Cox1p maturation or assembly with Cox2p. Mutant and additional copies of mss51 act as shy1 suppressors by increasing the amount of newly synthesized Cox1p, while overexpression of Hap4p suppresses mostly by increasing the amount of Cox5ap-Cox6p. (B) Model depicting the mechanism of suppression of the respiratory defect in yeast and human cellular models of Leigh’s syndrome by overexpression of CCAAT binding transcription activators HAP4 and NF-YA, respectively.

a center. Smith et al. (6) proposed that Surf1p could perform matured by insertion of copper at the CuA site. The fact that this role in an indirect way by interacting with subunits 1 or 2 SCO1- like SURF1-deficient fibroblasts accumulate the to facilitate formation of the heme a3-CuB center. This would COXI–COXIV–COXVa (yeast Cox1p–Cox5ap–Cox6p) be in agreement with a previous observation suggesting an subassembly (2) strongly supports this possibility as SCO1 is interaction of SURF1 with COXII in human cultured cells required for the formation of CuA (32) and hence maturation (31). of COXII. No interaction of yeast Shy1p was detected with either Cox2p may associate with Cox1p–Cox5ap–Cox6p upon newly synthesized Cox1p or Cox2p (11). However, it is poss- Cox2p metallation which in turn could affect the stability of ible that such an interaction could occur only after the Cox1p– the CuB-hem a3 center by capping the heme-insertion Cox5ap–Cox6p subassembly is formed and Cox2p has been channel formed in Cox1p–Cox5ap-Cox6p subassembly (33). 786 Human Molecular Genetics, 2008, Vol. 17, No. 6

Shy1p could play a role in facilitating this interaction. Our transcriptional activation of the same or a different set of results showing suppression of shy1 by increasing the genes than those proposed for suppression of yeast shy1 amount of Cox1p–5a–6p supports this possibility although a mutants by HAP4 (Fig. 6B). Knowledge of such mechanisms direct role of Shy1p on Cox1p maturation cannot be excluded can be helpful in future attempts to devise therapeutic inter- (Fig. 6A). Supporting a role of Shy1p as assembly factor, it ventions to combat LS as well as other forms of mitochondrial was recently reported that it promotes COX biogenesis COX deficiency. through association with different protein modules, potential In the wider context, our results would suggest that COX assembly intermediates containing Cox1p and Cox5ap differences in the expression of oxidative phosphorylation among other proteins (34). (OXPHOS) structuraland assembly factor genes can modify Surf1p does not play a direct role in copper insertion into disease mechanisms and may provide a basis for the influence COX (6). However, we have observed that copper supplemen- of genetic background and variable tissue involvement in tation to the media is able to partially restore the respiratory mitochondrial disorders. A greater understanding of the influ- defect of yeast shy1 mutants. In yeast shy1 mutants copper ence of genetic factors such as single nucleotide polymorph- Downloaded from https://academic.oup.com/hmg/article/17/6/775/600424 by guest on 29 September 2021 supplementation could facilitate the formation of the CuB isms (SNPs) in OXPHOS-related genes and physiological site in a slightly larger pool of Cox1p that could proceed in changes in may therefore provide a better the assembly process finally resulting in increased levels of understanding of differences in the progression of inherited the Cox1p–Cox5ap–Cox6p intermediate. Alternatively or mitochondrial diseases between individuals. maybe concurrently, copper supplementation could increase the pool of CuA containing Cox2p, which also would increase the pace of the assembly process in the absence of Shy1p. MATERIALS AND METHODS However, the suppressor effect was not enhanced by overex- pression of mitochondrial copper metabolism genes including Yeast strains and media COX17, SCO1 and COX11. It is also intriguing that the sup- Most of the work was carried out with the W303-1A wild-type pressor effect of copper supplementation is not additive to strain (MATa leu2-3,112 trp1-1 can1-100 ura3-1 ade2-1 the effect of neither increased amount of Cox1p by mutations his3-11,15) the isogenic mutant strain Dshy1 in which the in mss51 nor increased amounts of Cox5ap and Cox6p by whole gene had been deleted by insertion of a URA3 marker overexpression of HAP4. In fact, copper supplementation is (5) and the isogenic strain W125 carrying a point mutation some how deleterious to the shy1 revertant strain with in the region of SHY1 coding for the second transmembrane mutated mss51. If the amount of heme A available for domain of Shy1p (9). A shy1 mutant with a point mutation Cox1p maturation in shy1 mutant is limiting, an excess of in the SHY1 region coding for the first transmembrane copper could contribute to build unstable pro-oxidant inter- domain of Shy1p (C173/U) was also used in some studies mediates, like the ones recently proposed (35), when the (9). The genotypes and source of these and other strains of amount of newly synthesized Cox1p is enhanced. S. cerevisiae strains are listed in Supplementary Material, NF-Y regulates the expression of many genes involved in Table S2. The compositions of the growth media have been cell cycle and has also been proposed to play a role in tran- described elsewhere (40) and are detailed in the Supplemen- scriptional regulation of cellular aging through p53 (36). tary Materials and Methods. NF-Y and the transcriptional factor Sp1 (37) cooperate to regulate the expression of several genes involved in lipid metabolism including fatty acid synthase and carnitine palmi- toyltransferase (CPT)-I. These genes are also regulated by Characterization of yeast mitochondrial respiratory chain peroxisome proliferator-activated receptor PPAR, a ligand- Mitochondria were prepared from strains grown in media con- activated nuclear receptor that induces transcription of genes taining 2% galactose, according to the method of Faye et al. encoding fatty acid oxidation as well as the uncou- (41) except that zymolyase 20T (ICN Biochemicals Inc., pling proteins (26). An increase in PPARd activity induces Aurora, OH) instead of Glusulase was used for the conversion increased muscle mitochondrial biogenesis (38). Because of cells to spheroplasts. it has been recently reported that a high-fat diet induces Mitochondria prepared from the different strains were increased mitochondrial biogenesis in muscle from rats assayed polarographically for KCN-sensitive NADH oxidase through activation of the PPARd (39), it is possible that and ascorbate -N-N-N’-N’-tetramathyl-p-phenilenediamine over-expression of NF-Y also produces a similar effect on (TMPD) oxidation using a Clark type polarographic oxygen mitochondrial biogenesis. Although mammalian NF-Y complex electrode from Hansatech Instruments (Norfolk, UK) at 248C does not play a general role in expression of genes involved as described (5). Endogenous cell respiration was also in mitochondrial biogenesis, it was recently shown to regu- assayed in the presence of galactose or ethanol–glycerol. late the expression of two genes required for mitochondrial The specific activities reported were corrected for translation (25). KCN-insensitive respiration. These reports prompted us to examine if over-expression of Mitochondria were also used for spectrophotometric assays NF-Y in human SURF mutants fibroblasts could mimic the carried at 248C to measure KCN-sensitive cytochrome c suppressor effect of HAP4 in yeast shy1 mutants. We oxidase (COX) activity and antimycin A-sensitive NADH showed that NF-YA over-expression is able to increase the cytochrome c reductase (NCCR) activity as described (5). mitochondrial COX in SURF1 deficient fibroblast. The mech- Total mitochondrial cytochromes spectra was obtained as anism of COX-deficiency suppression by NF-Y may occur by reported (42). Human Molecular Genetics, 2008, Vol. 17, No. 6 787

For in vivo analysis of mitochondrial protein synthesis, SUPPLEMENTARY MATERIAL mitochondrial gene products were labeled with 35S-methionine Supplementary Material is available at HMG Online. in whole cells at 308C in the presence of cycloheximide (5). Details of these and other methods are described in the Sup- plementary Materials and Methods. ACKNOWLEDGEMENTS We thank Dr C. Moraes, Dr S. Williams, D. Horn and I.C. Soto for critically reading the manuscript. We thank Dr Human cell lines, culture conditions, transfections, enzyme A. Ro¨tig (Hopital des Enfants Malades, Paris, France) for pro- assays and immunodetections viding with the SURF1 deficient fibroblast cell line, Dr L. Pon A previously reported primary fibroblast line from an LS (Columbia University, NY) for the anti MOM45 Ab, Dr C. Moraes (University of Miami, FL) for the HPV 16 E6/E7

patient was obtained from Agnes Ro¨tig (Necker Hospital, Downloaded from https://academic.oup.com/hmg/article/17/6/775/600424 by guest on 29 September 2021 Paris, France) (27). The patient carried a homozygous 1-bp oncogenes, Dr G. Stephanopoulos (Massachusetts Institute of deletion within exon 6 of the SURF1 gene that causes a fra- Technology, Cambridge, MA) for the TEF1-based promoter meshift mutation resulting in a premature stop-codon at replacement cassettes and Dr B. Grimpe (University of residue 187 of the polypeptide. The fibroblast Miami, FL) for the pIRES2 plasmid. cell lines were immortalized by transducing human papil- loma virus (HPV) 16 E6/E7 oncogenes. Human neonatal Conflict of Interest statement. None declared. fibroblasts (CCD-10645 k) were obtained from ATCC (Manassas, VA) and used as a control cell line. Cells were FUNDING cultured in high glucose Dulbecco’s modified Eagle’s medium supplemented with 10% fetal bovine serum and This research was supported by National Institutes of Health 100 mg/ml sodium pyruvate and 0.25 mg/ml uridine. For Research Grant GM071775A (to A.B.), GM50187 (to A.T.) transient expression studies, the genes of interest were and a Research Grant from the Muscular Dystrophy Associ- cloned into the plasmid pIRES2 (Clontech, Mountain View, ation (to A.B.). F.F. is supported by Telethon-Italy, Fellowship CA) and transfected into human cell lines by using Lipofec- no. GFP05008. tamine-2000 (Invitrogen, Carlsbad, CA). For immunocytochemistry, cells were grown on coverslips. Forty eight hours after transfection, living cells were incubated REFERENCES 20 min with 50 nM mitochondrial dye Mitotracker red 1. Nijtmans, L.G., Taanman, J.W., Muijsers, A.O., Speijer, D. and Van den (CMXRos; Molecular Probes, Invitrogen). Coverslips were Bogert, C. (1998) Assembly of cytochrome-c oxidase in cultured human washed with PBS and fixed with 2% paraformaldehyde in cells. Eur. J. Biochem., 254, 389–394. PBS, permeabilized with cold methanol and then incubated 2. Williams, S.L., Valnot, I., Rustin, P. and Taanman, J.W. (2004) Cytochrome c oxidase subassemblies in fibroblast cultures from patients with an anti-COXI monoclonal antibody conjugated to Alexa carrying mutations in COX10, SCO1,orSURF1. J. Biol. Chem., 279, Fluor 488 (Molecular Probes, Invitrogen) at a concentration 7462–7469. of 2 mg/ml in 5% BSA in PBS for 2 h. For cytochemistry, 3. Church, C., Goehring, B., Forsha, D., Wazny, P. and Poyton, R.O. (2005) cells were grown on coverslips. Forty eight hours after trans- A role for Pet100p in the assembly of yeast cytochrome c oxidase: fection, cells were stained for succinate dehydrogenase and interaction with a subassembly that accumulates in a pet100 mutant. J. Biol. Chem., 280, 1854–1863. cytochrome c oxidase activities as described (8). 4. Horan, S., Bourges, I., Taanman, J.W. and Meunier, B. (2005) Analysis of To measure COX activity in cells homogenates, cells were COX2 mutants reveals cytochrome oxidase subassemblies in yeast. collected by trypsinization 48 h after transfection, and sub- Biochem J., 390, 703–708. mitted to three freezing–thawing cycles. COX activity was 5. Barrientos, A., Korr, D. and Tzagoloff, A. (2002) Shy1p is necessary for full expression of mitochondrial COX1 in the yeast model of Leigh’s measure in the presence of lauryl maltoside as described (43). syndrome. EMBO J., 21, 43–52. 6. Smith, D., Gray, J., Mitchell, L., Antholine, W.E. and Hosler, J.P. (2005) Assembly of cytochrome-c oxidase in the absence of assembly protein Surf1p leads to loss of the active site heme. J. Biol. Chem., 280, 17652– Miscellaneous procedures 17656. 7. Zhu, Z., Yao, J., Johns, T., Fu, K., De Bie, I., Macmillan, C., Cuthbert, Standard procedures used for DNA cloning, bacterial and A.P., Newbold, R.F., Wang, J., Chevrette, M. et al. (1998) SURF1, yeast transformation, northern and western blot analyses are encoding a factor involved in the biogenesis of cytochrome c oxidase, is mutated in Leigh syndrome. Nat. Genet., 20, 337–343. described in the Supplementary Materials and Methods. 8. Tiranti, V., Hoertnagel, K., Carrozzo, R., Galimberti, C., Munaro, M., Granatiero, M., Zelante, L., Gasparini, P., Marzella, R., Rocchi, M. et al. (1998) Mutations of SURF-1 in Leigh disease associated with cytochrome c oxidase deficiency. Am. J. Hum. Genet., 63, 1609–1621. Statistical analysis 9. Mashkevich, G., Repetto, B., Glerum, D.M., Jin, C. and Tzagoloff, A. (1997) SHY1, the yeast homolog of the mammalian SURF-1 gene, Most experiments were done at least in triplicate. Data are pre- encodes a mitochondrial protein required for respiration. J. Biol. Chem., sented as means + SD of absolute values or percent of control. 272, 14356–14364. 10. Tiranti, V., Galimberti, C., Nijtmans, L., Bovolenta, S., Perini, M.P. and The values obtained for wild-type, mutant and over-expressor Zeviani, M. (1999) Characterization of SURF-1 expression and Surf-1p strains for the different parameters studied were compared by function in normal and disease conditions. Hum. Mol. Genet., 8, 2533– Student’ t-test. P , 0.05 was considered significant. 2540. 788 Human Molecular Genetics, 2008, Vol. 17, No. 6

11. Barrientos, A., Zambrano, A. and Tzagoloff, A. (2004) Mss51p and The whole structure of the 13-subunit oxidized cytochrome c oxidase at Cox14p jointly regulate mitochondrial Cox1p expression in 2.8 A. Science, 272, 1136–1144. Saccharomyces cerevisiae. EMBO J., 23, 3472–3482. 29. Antonicka, H., Mattman, A., Carlson, C.G., Glerum, D.M., Hoffbuhr, 12. Dell’agnello, C., Leo, S., Agostino, A., Szabadkai, G., Tiveron, C., Zulian, K.C., Leary, S.C., Kennaway, N.G. and Shoubridge, E.A. (2003) A., Prelle, A., Roubertoux, P., Rizzuto, R. and Zeviani, M. (2007) Mutations in COX15 produce a defect in the mitochondrial heme Increased longevity and refractoriness to Ca2þ-dependent biosynthetic pathway, causing early-onset fatal hypertrophic neurodegeneration in Surf1 knockout mice. Hum. Mol. Genet., 16, 431– cardiomyopathy. Am. J. Hum. Genet., 72, 101–114. 444. 30. Poyau, A., Buchet, K. and Godinot, C. (1999) Sequence conservation 13. Becker, D.M., Fikes, J.D. and Guarente, L. (1991) A cDNA encoding a from human to prokaryotes of Surf1, a protein involved in cytochrome human CCAAT-binding protein cloned by functional complementation in c oxidase assembly, deficient in Leigh syndrome. FEBS Lett., 462, yeast. Proc. Natl Acad. Sci. USA, 88, 1968–1972. 416–420. 14. Forsburg, S.L. and Guarente, L. (1989) Identification and characterization 31. Nijtmans, L.G., Artal Sanz, M., Bucko, M., Farhoud, M.H., Feenstra, M., of HAP4: a third component of the CCAAT-bound HAP2/HAP3 Hakkaart, G.A., Zeviani, M. and Grivell, L.A. (2001) Shy1p occurs in a heteromer. Genes Dev., 3, 1166–1178. high molecular weight complex and is required for efficient assembly of cytochrome c oxidase in yeast. FEBS Lett., 498, 46–51. 15. Steidl, S., Tuncher, A., Goda, H., Guder, C., Papadopoulou, N., Downloaded from https://academic.oup.com/hmg/article/17/6/775/600424 by guest on 29 September 2021 Kobayashi, T., Tsukagoshi, N., Kato, M. and Brakhage, A.A. (2004) A 32. Leary, S.C., Kaufman, B.A., Pellecchia, G., Guercin, G.H., Mattman, A., single subunit of a heterotrimeric CCAAT-binding complex carries a Jaksch, M. and Shoubridge, E.A. (2004) Human SCO1 and SCO2 have nuclear localization signal: piggy back transport of the pre-assembled independent, cooperative functions in copper delivery to cytochrome complex to the nucleus. J. Mol. Biol., 342, 515–524. c oxidase. Hum. Mol. Genet., 13, 1839–1848. 16. McNabb, D.S. and Pinto, I. (2005) Assembly of the Hap2p/Hap3p/Hap4p/ 33. Cobine, P.A., Pierrel, F. and Winge, D.R. (2006) Copper trafficking to the Hap5p-DNA complex in Saccharomyces cerevisiae. Eukaryot. Cell, 4, and assembly of copper metalloenzymes. Biochim. 1829–1839. Biophys. Acta, 1763, 759–772. 17. Lascaris, R., Bussemaker, H.J., Boorsma, A., Piper, M., van der Spek, H., 34. Mick, D.U., Wagner, K., van der Laan, M., Frazier, A.E., Perschil, I., Grivell, L. and Blom, J. (2003) Hap4p overexpression in glucose-grown Pawlas, M., Meyer, H.E., Warscheid, B. and Rehling, P. (2007) Shy1 Saccharomyces cerevisiae induces cells to enter a novel metabolic state. couples Cox1 translational regulation to cytochrome c oxidase assembly. Genome Biol., 4, R3. EMBO J., 26, 4347–4358. 18. DeRisi, J.L., Iyer, V.R. and Brown, P.O. (1997) Exploring the metabolic 35. Khalimonchuk, O., Bird, A. and Winge, D.R. (2007) Evidence for a and genetic control of gene expression on a genomic scale. Science, 278, pro-oxidant intermediate in the assembly of cytochrome oxidase. J. Biol. Chem., 282, 17442–17449. 680–686. 36. Matuoka, K. and Chen, K.Y. (2002) Transcriptional regulation of cellular 19. Buschlen, l., Amillet, J.-M., Guiard, B., Fournier, A., Marcireau, C. and ageing by the CCAAT box-binding factor CBF/NF-Y. Ageing Res. Rev., Bolotin-Fukuhara, M. (2003) The S. cerevisiae HAP complex, a key 1, 639–651. regulator of mitochondrial function, cooordinates nuclear and 37. Roder, K., Wolf, S.S., Larkin, K.J. and Schweizer, M. (1999) Interaction mitochondrial gene expression. Comp. Funct. Genom., 4, 37–46. between the two ubiquitously expressed transcription factors NF-Y and 20. Schlaepfer, I.R., Mattoon, J.R. and Bajszar, G. (1994) The sequence and Sp1. Gene, 234, 61–69. potential regulatory elements of the HEM2 promoter of Saccharomyces 38. Tanaka, T., Yamamoto, J., Iwasaki, S., Asaba, H., Hamura, H., Ikeda, Y., cerevisiae. Yeast, 10, 227–229. Watanabe, M., Magoori, K., Ioka, R.X., Tachibana, K. et al. (2003) 21. Glerum, D.M. and Tzagoloff, A. (1997) Submitochondrial distributions Activation of peroxisome proliferator-activated receptor delta induces and stabilities of subunits 4, 5 and 6 of yeast cytochrome oxidase in fatty acid beta-oxidation in skeletal muscle and attenuates metabolic assembly defective mutants. FEBS Lett., 412, 410–414. syndrome. Proc. Natl Acad. Sci. USA, 100, 15924–15929. 22. Pierrel, F., Bestwick, M.L., Cobine, P.A., Khalimonchuk, O., Cricco, J.A. 39. Garcia-Roves, P., Huss, J.M., Han, D.H., Hancock, C.R., and Winge, D.R. (2007) Coa1 links the Mss51 post-translational function Iglesias-Gutierrez, E., Chen, M. and Holloszy, J.O. (2007) Raising plasma to Cox1 cofactor insertion in cytochrome c oxidase assembly. EMBO J., fatty acid concentration induces increased biogenesis of mitochondria in 26, 4335–4346. skeletal muscle. Proc. Natl Acad. Sci. USA, 104, 10709–10713. 23. Nevoigt, E., Kohnke, J., Fischer, C.R., Alper, H., Stahl, U. and 40. Myers, A.M., Pape, L.K. and Tzagoloff, A. (1985) Mitochondrial protein Stephanopoulos, G. (2006) Engineering of promoter replacement cassettes synthesis is required for maintenance of intact mitochondrial genomes in for fine-tuning of gene expression in Saccharomyces cerevisiae. Appl. Saccharomyces cerevisiae. EMBO J., 4, 2087–2092. Environ. Microbiol., 72, 5266–5273. 41. Faye, G., Kujawa, C. and Fukuhara, H. (1974) Physical and genetic 24. Scarpulla, R.C. (2006) Nuclear control of respiratory gene expression in organization of petite and grande yeast mitochondrial DNA. IV. In vivo mammalian cells. J. Cell Biochem., 97, 673–683. transcription products of mitochondrial DNA and localization of 23 S 25. Zanotto, E., Shah, Z.H. and Jacobs, H.T. (2007) The bidirectional ribosomal RNA in petite mutants of Saccharomyces cerevisiae. J. Mol. promoter of two genes for the mitochondrial translational apparatus in Biol., 88, 185–203. mouse is regulated by an array of CCAAT boxes interacting with the 42. Tzagoloff, A., Akai, A. and Needleman, R.B. (1975) Assembly of the transcription factor NF-Y. Nucleic Acids Res., 35, 664–677. mitochondrial membrane system. Characterization of nuclear mutants of 26. Evans, R.M., Barish, G.D. and Wang, Y.X. (2004) PPARs and the Saccharomyces cerevisiae with defects in mitochondrial ATPase and complex journey to obesity. Nat. Med., 10, 355–361. respiratory enzymes. J. Biol. Chem., 250, 8228–8235. 27. von Kleist-Retzow, J.C., Vial, E., Chantrel-Groussard, K., Rotig, A., 43. Barrientos, A., Kenyon, L. and Moraes, C.T. (1998) Human Munnich, A., Rustin, P. and Taanman, J.W. (1999) Biochemical, genetic xenomitochondrial cybrids. Cellular models of mitochondrial complex I and immunoblot analyses of 17 patients with an isolated cytochrome deficiency. J. Biol. Chem., 273, 14210–14217. c oxidase deficiency. Biochim. Biophys. Acta, 1455, 35–44. 44. Hlavacek, O., Bourens, M., Salone, V., Lachacinski, N., Lemaire, C. and 28. Tsukihara, T., Aoyama, H., Yamashita, E., Tomizaki, T., Yamaguchi, H., Dujardin, G. (2005) The transcriptional activator HAP4 is a high copy Shinzawa-Itoh, K., Nakashima, R., Yaono, R. and Yoshikawa, S. (1996) suppressor of an oxal yeast mutation. Gene, 354, 53–57.