| Human Molecular Genetics | Pages |
©1999 Oxford University Press |
Expression and functional analysis of SURF1 in Leigh syndrome patients with cytochrome c oxidase deficiency
Introduction
Results
Expression of SURF1
hSurf1 is not detected in COX-deficient LS patient cells
hSurf1 localizes to mitochondria
hSurf1 is an integral inner membrane protein
Functional analysis of mutant hSurf1 proteins
Discussion
Materials And Methods
Cell lines
hSurf1-FLAG expression construct and immunofluorescence
Mitochondrial import
Preparation of antibodies to hSurf1 protein
Immunoblotting
Human SURF1 mutant cDNA constructs
Virus production and infection
Acknowledgements
References
Expression and functional analysis of SURF1 in Leigh syndrome patients with cytochrome c oxidase deficiency
Received August 6, 1999; Revised and Accepted September 14, 1999
Leigh syndrome (LS) associated with cytochrome c oxidase (COX) deficiency is an autosomal recessive neurodegenerative disorder caused by mutations in SURF1. Although SURF1 is ubiquitously expressed, its expression is lower in brain than in other highly aerobic tissues. All reported SURF1 mutations are loss of function, predicting a truncated protein (hSurf1) product. Western blot analysis with anti-hSurf1 antibodies demonstrated a specific 30 kDa protein in control fibroblasts, but no protein in LS patient cells. Steady-state levels of both nuclear- and mitochondrial-encoded COX subunits were also markedly reduced in patient cells, consistent with a failure to assemble or maintain a normal amount of the enzyme complex. An epitope (FLAG)-tagged hSurf1 was targeted to mitochondria in COS7 cells and a mitochondrial import assay showed that the hSurf1 precursor protein (35 kDa) was imported and processed to its mature form (30 kDa) in a membrane potential-dependent fashion. The protein was resistant to alkaline carbonate extraction and susceptible to proteinase K digestion in mitoplasts. Mutant proteins in which the N-terminal transmembrane domain or central loop were deleted, or the C-terminal transmembrane domain disrupted, did not accumulate and could not rescue COX activity in patient cells. Co-expression of the N- and C-terminal transmembrane domains as independent entities also failed to rescue the enzyme deficiency. These data demonstrate that hSurf1 is an integral inner membrane protein with an essential role in the assembly or maintenance of the COX complex and that insertion of both transmembrane domains in the intact protein is necessary for function.
INTRODUCTION
The biogenesis of the cytochrome c oxidase (COX) complex requires many nuclear-encoded assembly factors in addition to the 13 structural subunits of the enzyme complex itself. In yeast, >30 different genetic complementation groups for COX assembly have been identified (1,2) and a number of genes required for assembly of the yeast COX complex have been documented (3-7). In humans, COX assembly genes such as OXA1 (8), COX10 (9) and COX17 (10) have been cloned by functional complementation of their corresponding mutants in yeast. Recently, additional human genes involved in the biogenesis of the COX complex including PET112, SCO1, COX15 and COX11 were identified by homology search of the expressed sequence tag database against yeast protein sequences (11). The SURF1 gene was identified more than a decade ago (12) but its function remained unknown until recently when pathogenic mutations in the gene were identified in COX-deficient LS (13,14). LS is a fatal neurological disorder characterized by specific subcortical lesions in the brain. It is associated with a number of different enzyme deficiencies, all of which impair aerobic energy metabolism. The COX deficiency in LS patients apparently results from a failure to assemble a normal amount of active enzyme (15,16), suggesting a role for the SURF1 gene product in the biogenesis of the COX complex.
The SURF1 gene is located in a cluster of housekeeping genes known as the surfeit locus (17). The locus consists of at least six genes (SURF1-6), none of which is related by sequence homology, clustered in ~36 kb of genomic DNA. Adjacent surfeit genes are separated by only a small number of base pairs, some share bidirectional promoters and others share overlapping reading frames (18-20). This unique structure has been conserved over 250 million years of divergent evolution between birds and mammals, suggesting that the gene cluster is of functional significance (21).
The human SURF1 cDNA is predicted to code for a protein of 300 amino acids (19). The protein (hSurf1) has a characteristic mitochondrial targeting sequence at its N-terminus. Hydropathy plots of hSurf1 predict two transmembrane domains, one at amino acid position 61-79 and the other at position 275-293, very close to the C-terminus (14). The reported mutations in SURF1 all predict a truncated protein, some very near the C-terminus (13,14); however, it is not known whether any of the mutant proteins accumulate.
A homologue of SURF1, SHY1, has been identified in a yeast pet (respiratory chain) mutant (22). Mutations in SHY1 produce a partial pleiotropic respiratory chain deficiency (which involves COX) and an inability to grow on non-fermentable substrates. Surprisingly, disruption of SHY1 between the N- and C-terminal transmembrane domains does not result in a pet phenotype and the expression of the C-terminal half of the protein is sufficient to rescue the growth phenotype of point mutations, but not SHY1 deletions. These data suggest that the two domains of Shy1p can form a functional protein when expressed separately (22). It is not known whether hSurf1 can function in a similar fashion.
Here we have investigated the expression, targeting and structure-function relationships of hSurf1 in control and LS patient fibroblast cell lines. We show that hSurf1 is an integral inner mitochondrial membrane protein and that all truncating mutations in patient cells result in an unstable protein that does not accumulate. The steady-state levels of subunits in patient cells are drastically reduced, consistent with an essential role for hSurf1 in the assembly or maintenance of the COX complex. This phenotype is reversed when hSurf1 is restored with both transmembrane domains and the central loop intact.
RESULTS
Expression of SURF1
The structural features of the surfeit locus suggest that all of the SURF genes are housekeeping genes. As LS is primarily a central nervous system disorder, we investigated the tissue-specific expression of SURF1 in a number of human tissues. Northern blot analysis showed that SURF1 is ubiquitously expressed as expected, but that there are significant variations in steady-state SURF1 mRNA levels among tissues. In particular, the expression in the brain appears low compared with other highly aerobic tissues such as heart, skeletal muscle and kidney (Fig. 1).
Figure 1. Northern blot analysis of SURF1 expression in different human tissues. A commercial multiple tissue blot (Clontech, Palo Alto, CA) containing ~2 µg of poly(A)+ RNA per lane was probed with a human SURF1 cDNA according to the manufacturer's protocol. The blot was probed with a [beta]-actin probe as a loading control.
hSurf1 is not detected in COX-deficient LS patient cells
It has been demonstrated that the steady-state levels of SURF1 mRNA are reduced in LS patient fibroblasts (14) and all mutations in the SURF1 gene reported so far are predicted to produce truncated proteins (13,14). To test whether any of the predicted truncated proteins accumulate in the patient cells, we measured hSurf1 by western blot analysis. Antibodies prepared against a bacterially expressed hSurf1 revealed a major protein of ~30 kDa in control cells but failed to detect any hSurf1 in fibroblasts from three different patients (Fig. 2A). Our antibody would not be expected to detect proteins with large C-terminal truncations. However, the steady-state levels of mRNAs from such alleles are drastically reduced and very little translation of hSurf1 would be expected.
Figure 2. Western blot analysis of hSurf1 and COX subunit expression in patient fibroblasts. (A) Mitochondrial protein (8 µg) from each cell line was separated by SDS-PAGE, transferred to nitrocellulose and immunoblotted with polyclonal anti-hSurf1 antibodies and a monoclonal anti-porin antibody. (B) As for (A) except that monoclonal antibodies specific to COX-I, COX-II, COX-IV, COX-Vb and porin were used.
The expression of both mitochondrial- and nuclear-encoded COX subunits in the patient cells was also examined using monoclonal antibodies specific to the enzyme subunits COX-I, COX-II, COX-IV and COX-Vb. As shown in Figure 2B, all four subunits studied were synthesized in patient cells, but the steady-state concentrations of all subunits were reduced compared with control cells.
hSurf1 localizes to mitochondria
The predicted hSurf1 protein contains a typical mitochondrial targeting sequence (rich in hydrophobic and basic residues) at its N-terminus. MitoProtII (23) predicts that hSurf1 is very likely to be a mitochondrial protein (P = 1.0). To demonstrate mitochondrial targeting of the hSurf1 protein, a chimeric construct expressing hSurf1 with a C-terminal FLAG epitope was constructed and expressed transiently in COS7 cells. Figure 3 shows the immunofluorescence pattern of transfected COS7 cells double-labelled with the M2 monoclonal antibody specific for the FLAG epitope and polyclonal antibodies to COX-II, a protein of the inner mitochondrial membrane in the COX complex. A typical mitochondrial staining pattern was observed with both antibodies (Fig. 3a and b) and superimposition of the two panels showed complete overlap of the two patterns (Fig. 3c). No other subcellular structures exhibited immunoreactivity with the anti-FLAG antibody, demonstrating that hSurf1 localizes exclusively to mitochondria, even when overexpressed.
Figure 3. Immunofluorescence micrographs showing co-localization of a FLAG-tagged hSurf1 protein with a mitochondrial protein (COX-II) in COS7 cells. COS7 cells were transiently transfected with pcDNA3/HS1-FLAG and the intracellular distribution of the hSurf1-FLAG fusion protein was visualized using the M2 anti-FLAG monoclonal antibody followed by Alexa 488 goat anti-mouse IgG (a). The expression of the COX-II subunit in the same cells was assessed by the successive application of polyclonal antibodies against COX-II, biotinylated goat anti-rabbit IgG and Cy3-streptavidin (b). Superimposition of (a) and (b) shows complete overlap of the two staining patterns (c).
To analyse further the targeting and proteolytic processing of the hSurf1 protein, an in vitro mitochondrial import assay was performed. In vitro transcription and translation of a human SURF1 cDNA using a reticulocyte lysate preparation directed the synthesis of a major protein of ~35 kDa (Fig. 4, lane 1). Incubation of the translation products with freshly prepared mitochondria resulted in the appearance of a second protein band (~30 kDa) in addition to the band corresponding to the precursor protein (Fig. 4, lane 2). This band corresponds to the processed mature protein in which the N-terminal mitochondrial targeting sequence (44 amino acids as predicted by MitoProtII) has been removed. The uptake of the mature protein into mitochondria was tested by adding either trypsin or proteinase K to the post-import mixture. Imported mature hSurf1 was protected from enzymatic digestion, while the surface-bound precursor protein was completely degraded (Fig. 4, lanes 3 and 4), indicating that the mature protein is internalized within the mitochondria and inaccessible to protease. To determine whether the protein import and processing were membrane potential-dependent, the proton electrochemical gradient was collapsed with the uncoupler carbonyl cyanide m-chlorophenylhydrazone (CCCP). In the presence of 1 µM CCCP, the processing of the precursor protein was completely blocked (Fig. 4, lane 5), indicating that mitochondrial import of hSurf1 is membrane potential-dependent.
Figure 4. Mitochondrial import of hSurf1 protein. Radiolabelled hSurf1 protein was synthesized by in vitro transcription and translation using a rabbit reticulocyte system (lane 1). The mitochondrial import assay was performed by incubating the translation product with freshly isolated mitochondria (lane 2). The retention of imported product was assessed by adding either trypsin (lane 3) or proteinase K (lane 4) to the post-import reaction mixtures. The import reaction was also performed in the presence of 1 µM CCCP (lane 5) or both CCCP and proteinase K (lane 6). The positions of the precursor (p) and mature proteolytically processed protein (m) are indicated.
hSurf1 is an integral inner membrane protein
hSurf1 is predicted to be membrane-associated based on computer analysis using a program that predicts transmembrane domains (TMpred). To assess this property of hSurf1, mitochondria were sonically disrupted and fractionated into soluble matrix proteins and membrane vesicles. Western blotting using anti-hSurf1 antisera showed that hSurf1 was present in the membrane fraction (Fig. 5A, lane 3), indicating that hSurf1 is a component of the mitochondrial membrane. To determine whether hSurf1 is a peripheral or integral membrane protein, its susceptibility to alkaline carbonate extraction was assessed. Mitochondria were treated with sodium carbonate (pH 11.5) and the soluble and insoluble fractions were subjected to western blot analysis. As shown in Figure 5A, hSurf1 is found in the insoluble fraction (lane 5), indicating that it is an integral membrane protein that cannot be extracted by alkaline carbonate. This result was confirmed by extraction and phase separation of mitochondrial proteins using the non-ionic detergent Triton X-114. The partition of hydrophilic proteins into the aqueous phase and integral membrane proteins into the detergent phase during separation in Triton X-114 has been demonstrated previously (24). Figure 5A shows that hSurf1 was found exclusively in the detergent phase (lane 7) after separation, supporting the conclusion that hSurf1 is an amphiphilic integral membrane protein.
Figure 5. Localization of hSurf1 to the inner mitochondrial membrane. (A) Isolated mitochondria from cells expressing a wild-type hSurf1 protein were either sonically disrupted or treated by sodium carbonate as described (42). Membrane pellet and supernatant fractions were obtained by centrifugation at 150 000 g for 60 min. In addition, proteins from mitochondria were extracted by Triton X-114 followed by phase separation as described (24). All fractions were subjected to western blot analysis using anti-hSurf1 antibodies. Lane 1, untreated mitochondria; lanes 2 and 3, supernatant and pellet fractions, respectively, from sonically disrupted mitochondria; lanes 4 and 5, supernatant and pellet fractions, respectively, from sodium carbonate-treated mitochondria; lanes 6 and 7, aqueous and detergent phases, respectively, after Triton X-114 extraction. (B) Mitochondria from mouse liver and heart were converted to mitoplasts by hypotonic swelling as described (22). Both mitochondria and mitoplasts were incubated for 30 min on ice in the presence or absence of 40 µg/ml proteinase K. The treated mitochondria and mitoplasts were recovered by centrifugation and subjected to western blot analysis using anti-hSurf1 antibodies. The 70 kDa subunit of complex II was used as a control matrix protein.
To test whether hSurf1 is a component of the inner membrane, its susceptibility to protease digestion in mitoplasts was evaluated. Mitochondria isolated from mouse liver and heart were converted to mitoplasts by hypotonic swelling, or were left untreated, and both intact mitochondria and mitoplasts were treated with proteinase K. hSurf1 was protected against proteolysis in intact mitochondria but was degraded by proteinase K in mitoplasts, indicating that the protease has access to hSurf1 from the intermembrane space when the outer membrane is ruptured (Fig. 5B). The concentration of the 70 kDa subunit of complex II, which is located on the matrix side of the inner mitochondrial membrane, was not altered significantly in this experiment, showing that the protease did not have access to the matrix compartment. Based on these results, we conclude that hSurf1 is probably an integral protein of the inner mitochondrial membrane.
Functional analysis of mutant hSurf1 proteins
To evaluate the structure-function relationships of the hSurf1 protein, a series of mutant cDNA constructs was prepared and expressed in patient W cells via retroviral infection. The function of the mutated proteins was determined by measuring their ability to restore COX activities in the patient cells. The mutant constructs are depicted in Figure 6. Constructs [Delta]55-138 (N-terminal transmembrane domain deleted), A238E-A284D (C-terminal transmembrane domain disrupted) and [Delta]147-235 (central loop region deleted) were prepared to assess the importance of the two transmembrane domains and central loop for function. Constructs [Delta]295-300 (last six amino acids deleted) and K291E-K292E-R295D (positively charged amino acids replaced by negatively charged ones at the C-terminus) were prepared to investigate the role of the C-terminal tail of the protein. All of the constructs were tested in an in vitro transcription/translation system and all directed the synthesis of a protein of the predicted size (data not shown). Table 1 shows the COX activity measured in patient W cells transduced with the SURF1 mutant cDNA constructs. Expression of the wild-type hSurf1 protein restored the COX activity in the patient cells to near control levels. The mutant protein with the last six amino acids deleted ([Delta]295-300) also restored full COX activity in the patient cells. However, none of the other mutant proteins was able to rescue the COX deficiency in the patient cells. Immunoblot analysis of mitochondria isolated from the transduced patient cells demonstrated that hSurf1 was expressed abundantly in cells transduced with the wild-type cDNA and the protein with its last six amino acids deleted (Fig. 7). All of the other mutated proteins, except for the construct in which the central loop was deleted (Fig. 7, lane 9), were detected on western blot analysis of isolated mitochondria (Fig. 7, lanes 6-8 and 10-11) at levels similar to or slightly less than that of hSurf1 in control cells (Fig. 7, lane 1). This demonstrates that the mutant proteins were targeted correctly to mitochondria, but were probably unstable and rapidly degraded. The fact that the steady-state levels of the mutant proteins were similar to that of hSurf1 expressed from the native promoter suggests that failure to rescue the biochemical defect was the result of a non-functional protein and not of inadequate expression. Further analysis of the expression of the COX subunits in these transduced cells clearly showed that the levels of a mitochondrial- and a nuclear-encoded subunit returned to control levels in patient cells expressing either the wild-type or the C-terminal truncated protein, but not in cells transduced with the other mutant constructs (Fig. 7). This experiment indicates that a functional hSurf1 requires both transmembrane domains and the central loop for proper insertion, but that the C-terminal tail of the protein is dispensable.
Figure 6. Schematic representation of the mutant hSurf1 constructs. [Delta]295-300, deletion of the C-terminal tail; [Delta]55-138, deletion of the N-terminal transmembrane domain; [Delta]147-235, deletion of the central loop; A283E-A284D, disruption of the C-terminal transmembrane domain; K291E-K292E-R295D, amino acid substitution at the C-terminal tail. The cross-hatched areas denote the mitochondrial targeting sequence (TS). The black areas represent the two transmembrane domains (TM).
Figure 7. Functional analysis of mutant hSurf1 proteins. Equal amounts of mitochondrial protein (8 µg) from a control cell line and patient cells transduced with wild-type or mutant SURF1 cDNA constructs were separated by SDS-PAGE, transferred to nitrocellulose and immunoblotted with either polyclonal anti-hSurf1 antibodies or monoclonal antibodies specific to COX-II, COX-IV and porin. All constructs were tested in patient W cells. [Delta]55-138 was also tested in patient C cells (last lane). The arrow in the top panel points to the mutant hSurf1 protein in lanes 6, 10 and 11.
Table 1. COX activity in patient fibroblasts transduced with SURF1 mutant cDNA constructs
| Samples | COX/citrate synthase |
| Control (C75) | 0.640 |
| Patient W | 0.096 |
| pLXSH | 0.081 |
| Wild-type | 0.541 |
| [Delta]295-300 | 0.545 |
| [Delta]55-138 | 0.071 |
| K291E-K292E-R295D | 0.121 |
| A283E-A284D | 0.056 |
| [Delta]147-235 | 0.072 |
| [Delta]55-138 + A283E-A284D | 0.067 |
| [Delta]55-138 (Patient C) | 0.029 |
To test whether separately expressed domains of hSurf1 protein could cooperate to form a functional protein, constructs [Delta]55-138 (expressing a protein with the N-terminal transmembrane domain deleted) and A283E-A284D (encoding a protein with the C-terminal transmembrane domain disrupted) were used to co-infect patient W cells. In addition, the [Delta]55-138 construct alone was used to infect patient C cells, which are predicted to produce a near full-length hSurf1 protein with an intact N-terminal transmembrane domain from one allele. Mutant hSurf1 did not accumulate significantly in the transduced cells, which had low residual COX activities and concentrations of the COX structural subunits (Table 1, Fig. 7). This result suggests that separately expressed N- and C-terminal transmembrane domains of hSurf1 are not able to form a functional protein.
DISCUSSION
This study represents the first characterization of a human protein involved in the assembly of the COX complex. Several lines of evidence demonstrate that the hSurf1 protein localizes to mitochondria. First, the protein contains a characteristic mitochondrial targeting sequence at its N-terminus. Second, immunofluorescence analysis of COS7 cells expressing a C-terminal FLAG-tagged hSurf1 protein demonstrated co-localization of the hSurf1 protein with a mitochondrial protein (COX-II). Finally, mitochondrial import assay of in vitro translated hSurf1 protein revealed that the precursor protein was imported and proteolytically processed to a mature protein, and that its translocation was membrane potential-dependent. Based on the observations that hSurf1 protein could not be extracted by alkaline sodium carbonate, yet was susceptible to proteinase K digestion in mitoplasts, we further conclude that hSurf1 is probably an integral protein that is embedded in the inner mitochondrial membrane with its central loop facing the intermembrane space. This membrane topology is similar to that proposed for Shy1p in yeast (22).
The expression of endogenous hSurf1 protein in normal human fibroblasts was relatively low in isolated mitochondria, consistent with a role as an assembly factor rather than a structural subunit of the COX complex. Variations in the steady-state levels of SURF1 mRNA among tissues did not provide any real insight in the tissue-specific nature of the clinical phenotype. Despite the relatively low expression of SURF1 in the brain, it is clear that the protein serves a critical function in this tissue as central nervous system pathology is a prominent feature of LS. None of the predicted truncated proteins, caused by mutations leading to premature stop codons in the SURF1 gene, was detected in the patient cells even though some of the mutations did not significantly affect the steady-state levels of SURF1 mRNA (14). In particular, one of the disease alleles in patient C is predicted to produce a near full-length protein with only the last 10 amino acids truncated (and the C-terminal transmembrane domain disrupted), but no immunoreactive protein was detected on western blot analysis. These data predict that most, if not all, truncated hSurf1 proteins will be unstable and that the LS patients will be functionally null for hSurf1. If so, the molecular diagnosis of LS caused by SURF1 mutations could be performed by immunocytochemical or western blot analyses on patient cells. So far, disease alleles with missense mutations in the SURF1 gene have not been reported (13,14), suggesting that such mutations, if they exist, are rare.
The transcription (25) and translation (16) of the structural subunits of the COX complex appear to be normal in LS COX-deficient patients. Thus, the decrease in the steady-state concentrations of COX subunits in LS patient cells indicates that either the assembly or the maintenance of the COX complex is impaired without functional hSurf1, leading to the degradation of the unassembled subunits. We observed that both the mitochondrial- and nuclear-encoded subunits were affected to a similar extent. This pattern is similar to that observed for COX assembly mutants in yeast (3,5,26). However, in many yeast COX assembly mutants, the steady-state concentrations of the mitochondrial-encoded subunits, in particular COX-I, show greater reductions than nuclear-encoded subunits. The turnover of unassembled respiratory subunits is thought to be mediated by mitochondrial ATP-dependent proteases; however, the specific proteases responsible for the turnover of COX subunits are not known (27) and the differences between human and yeast cells could be a reflection of different pathways of degradation.
We have demonstrated previously that expression of a wild-type hSurf1 protein in patient fibroblasts restores COX activity (14). In this study, we have used these COX-deficient cells for functional analysis of hSurf1 mutants to investigate the structure-function relationships of the protein. We chose patient W cells as an expression system for the mutant constructs since the steady-state level of SURF1 mRNA is reduced to ~15% of control levels in these cells. Thus, even if a small amount of a truncated protein were produced, it is unlikely to interfere with the mutant proteins expressed from the retroviral vector.
The results obtained from the functional expression analysis demonstrate that the C-terminal tail of the hSurf1 protein, at least the last six amino acids, is not essential for its function. It is likely that this truncated protein is targeted to and integrated into the inner mitochondrial membrane in a manner identical to that of wild-type protein. This is based on the observations that this protein was able to restore COX activity in patient cells and was expressed at a level similar to that of the wild-type protein from the retroviral construct. Mutant proteins with either the N-terminal transmembrane domain or central loop deleted, or the C-terminal transmembrane domain disrupted, failed to rescue the COX deficiency, suggesting that the protein requires both domains to anchor itself in the inner membrane in order to function properly. The fact that most of the mutant proteins were detectable at much reduced levels compared with the wild-type protein overexpressed from the retroviral vector suggests that they are highly unstable and are degraded rapidly. Membrane association as an essential requirement for function has also been observed in yeast Sco1p (28), a protein thought to function in mitochondrial copper recruitment (6).
The mutant hSurf1 protein in which the positively charged Lys291, Lys292 and Arg295 residues were substituted by negatively charged amino acids (either Glu or Asp) also did not restore COX activity in the patient cells. Substitution of these three amino acids is not predicted (by TMpred) to affect the C-terminal transmembrane domain, and Arg295 is within the last six amino acids that were deleted in another construct without affecting the function of the protein. Despite this, the level of the mutant protein was greatly reduced in the transduced patient cells, implying a critical role for the two lysine residues. It is known that ionic interactions between membrane lipids and positively charged amino acids bordering the transmembrane domain are important in stabilizing integral membrane proteins (29,30). Lys291 and Lys292 are adjacent to the C-terminal transmembrane domain and could be involved in the interactions with negatively charged phospholipid head groups of the inner mitochondrial membrane to hold the transmembrane domain in place. Substitution of these two amino acids by negatively charged residues would disrupt this ionic interaction, destabilizing the C-terminal transmembrane domain and therefore the entire protein. It is also possible that the introduction of negatively charged residues immediately C-terminal to the transmembrane domain disrupts an internal mitochondrial targeting sequence necessary for the insertion of hSurf1 into the inner membrane. Such internal targeting sequences are essential for proper sorting of other proteins with N-terminal cleavable targeting sequences such as cytochrome c1 (31).
Many membrane proteins have been shown to be able to form a functional protein when N- and C-terminal halves of a protein are expressed separately (32). Most of the examples are proteins with multiple transmembrane domains or duplicated structures. Although Shy1p appears to possess this ability, based on the lack of an obvious phenotype when the gene was disrupted between the two transmembrane domains, hSurf1 does not. Co-expression of two hSurf1 mutants, one with an intact N-terminal transmembrane domain and the other with an intact C-terminal transmembrane domain, failed to restore COX activity in patient cells. Likewise, expression of an intact hSurf1 C-terminal transmembrane domain in a patient cell line (patient C) in which one allele is predicted to produce a protein with a truncation of only the last 10 amino acids, failed to restore COX activity. This suggests that, unlike Shy1p in yeast, the two domains of hSurf1 when expressed separately cannot stabilize or interact with each other to form a functional protein. The biochemical phenotype of the SHY1 deletion in yeast is apparently more complex than the isolated COX deficiency seen in human cells null for hSurf1. We have also observed that expression of SURF1 is not able to rescue the pet phenotype of SHY1 deletion mutants in yeast (unpublished data). The different behaviours of Shy1p and hSurf1 may reflect different cellular functions of the two proteins.
MATERIALS AND METHODS
Cell lines
Primary fibroblast lines were established from three patients (W, Wi and C) with COX-deficient LS. All patients had a severe deficiency in COX activity and a typical LS phenotype carrying nonsense mutations in both SURF1 alleles. The sites of the mutations were as follows: patient W, C765T (nonsense codon in exon 7), T337 + 2C (mutation in the donor splice site of intron 4); patient Wi, G88A (nonsense codon in exon 2), 588insCTGC (insertion in exon 6); patient C, 326insATdelTCTGCCAGCC (insertion/deletion in exon 4), 855delCT (deletion in exon 9). The primary fibroblast lines were transduced with retroviral vectors expressing the E6E7 region of type 16 human papilloma virus (a kind gift of Denise Galloway, Fred Hutchinson Cancer Research Centre, Seattle, WA) (33) and telomerase (a kind gift of Geron Corp., Menlo Park, CA) (34) to extend their life span, and grown in high glucose Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum.
hSurf1-FLAG expression construct and immunofluorescence
A human SURF1 cDNA lacking its natural stop codon was generated by PCR using the human SURF1 cDNA clone (pcDNA3/HS1) (14). The same sense primer (containing a BamHI site and consensus Kozak sequence: 5[prime]-GGATCCGCCACCATGGCGGCGGTGGCTGCGTTGCA-3[prime]) used for generation of the cDNA clone and an antisense primer (5[prime]-AAATCTCGAGCACACCAGGTGTCCCACGTA-3[prime]) with an engineered XhoI site immediately upstream of the natural stop codon were used in the PCR. After PCR cloning, a BamHI-XhoI fragment was excised and subcloned into the BamHI-XhoI sites of a modified pcDNA3 vector containing a FLAG epitope followed by a stop codon immediately downstream of the XhoI site. The resulting construct is designated pcDNA3/HS1-FLAG.
COS7 cells were transiently transfected with the pcDNA3/HS1-FLAG construct using SuperFect (Qiagen, Mississauga, Ontario, Canada) according to the manufacturer's instructions. After 24 h, the cells were replated on coverslips and grown for 24 h. The coverslips were washed in phosphate-buffered saline (PBS), fixed in 4% paraformaldehyde/PBS for 10 min and permeabilized in 0.1% Triton X-100/PBS for 3 min. The cells were then incubated with anti-FLAG M2 monoclonal antibody (6 µg/ml; Sigma, Oakville, Ontario, Canada) and rabbit anti-COX-II polyclonal antibodies (1:250 dilution; a gift of Dr Anne Lombes, Institut de Myologie, Paris) in 1% bovine serum albumin (BSA)/PBS for 2 h, followed by another incubation with Alexa 488 goat anti-mouse IgG (20 µg/ml; Molecular Probes, Eugene, OR) and biotinylated goat anti-rabbit IgG (1:100 dilution; Cedarlane, Hornby, Ontario, Canada) for 1 h. Finally, the cells were incubated with Cy3-conjugated streptavidin (1:1000 dilution; Jackson ImmunoResearch, West Grove, PA) for 30 min and viewed under a Leitz fluorescence microscope.
Mitochondrial import
hSurf1 was synthesized by in vitro transcription and translation using the TNT-coupled reticulocyte lysate system (Promega, Madison, WI) and the pcDNA3/HS1 construct in the presence of [35S]methionine. The import assay was performed in a 50 µl reaction mixture containing mitochondria (0.5 mg protein/ml) freshly isolated from rat heart and 10% (v/v) in vitro translation products in an import buffer as described (35). The import reaction mixture was incubated at 37°C for 1 h in the absence or presence of 1 µM CCCP. To assess the protease resistance of the imported protein, the post-import reaction mixture was incubated with either 40 µg/ml proteinase K or 125 µg/ml of trypsin at 4°C for 20 min, followed by another incubation with either 2 mM phenylmethylsulfonyl fluoride or 1.25 mg/ml trypsin inhibitor at 4°C for 20 min. The mitochondria were recovered by centrifugation and analyzed by SDS-PAGE.
Preparation of antibodies to hSurf1 protein
The QIAexpress system (Qiagen) was used for expression and purification of recombinant hSurf1. A SmaI-HindIII fragment of the human SURF1 cDNA encoding the C-terminal half of the protein (amino acids 147-300) was cloned in the bacterial expression vector, pQE-31. The resulting plasmid expresses a 19 kDa protein consisting of an N-terminal His6-tag fused in-frame to the Surf1 sequence. The plasmid was transformed into Escherichia coli M15 for protein expression and the fusion protein was purified under denaturing conditions by Ni-NTA chromatography (36). The purified protein was mixed with Titermax Gold (Cedarlane) and used to raise antibodies in rabbits.
Immunoblotting
Mitochondria from confluent cultures of fibroblasts were isolated by differential centrifugation as described previously (37), omitting the use of protease inhibitors in the isolation buffer. An 8 µg aliquot of the mitochondrial protein, as determined with the Pierce (Rockland, IL) Micro BCA protein kit, was separated on 12.5% SDS-polyacrylamide gels and transferred electrophoretically to 0.2 µm nitrocellulose membranes. The membranes were blocked with 5% milk powder in Tris-buffered saline containing 0.1% Tween-20 for at least 1 h, followed by incubation at 4°C overnight with the following primary antibodies: anti-hSurf1 polyclonal antibodies raised in rabbit (1:200 dilution) and monoclonal antibodies against human COX-I (1 µg/ml; Molecular Probes), COX-II (4 µg/ml; a gift from Dr Anne Lombes), bovine COX-IV (0.5 µg/ml; Molecular Probes), bovine COX-Vb (3 µg/ml; Molecular Probes), human porin (1.4 µg/ml; Calbiochem, La Jolla, CA) and the 70 kDa subunit of complex II (0.1 µg/ml; a kind gift of Dr R. Capaldi, University of Oregon, Portland, OR). For specific hSurf1 detection, the blots were incubated in goat anti-rabbit IgG conjugated to horseradish peroxidase (HRP; 1:20 000 dilution; Pierce) and developed using SuperSignal West Femto substrate (Pierce). The specific bands for the COX subunits, 70 kDa complex II subunit and porin were detected using 1:1000 diluted goat anti-mouse IgG coupled to HRP and LumiGLO chemiluminescent reagent (New England BioLabs, Mississauga, Ontario, Canada).
Human SURF1 mutant cDNA constructs
Human SURF1 cDNAs coding for proteins mutated at the C-terminus of the protein were generated by PCR using antisense primers incorporating desired mutations and the cDNA clone, pcDNA3/HS1, as a template. The antisense primers (all containing a BamHI site for subsequent cloning) were as follows: 5[prime]-GGATCCTCATAGGATTTCTTAAACCACAGTAGG-3[prime] (a stop codon introduced at amino acid position 295) for truncation of the C-terminus ([Delta]295-300); 5[prime]-GGATCCTCACACACCAGGTGTCCCACGTAGGAATTTCTT-
AAACCCAGGTAGGATGTATCTTCAGAGAGTCC-3[prime] (containing mutations, bold and underlined, corresponding to Ala283->Glu and Ala284->Asp) for disruption of the C-terminal transmembrane domain (A283E-A284D); and 5[prime]-GGATCCTCACACACCAGGTGTCCATCTAGGAATTCCTCAAACCACAG-3[prime] (containing mutations, bold and underlined, corresponding to Lys291->Glu, Lys292->Glu and Arg295->Asp) for amino acid substitution of the C-terminus (K291E-K292E-R295D). In each case, the sense primer was the same as that used for the pcDNA3/HS1-FLAG construct. All amplified mutant cDNAs were first cloned using a PCR cloning kit (Stratagene, La Jolla, CA) and then subcloned into the BamHI site of a retroviral expression vector, pLXSH (38). Construction of the mutant cDNA expressing a protein with its N-terminal transmembrane domain deleted ([Delta]55-138) was achieved by PCR amplifying a cDNA fragment encoding the first 54 amino acids using the same sense primer (containing an NcoI site at the initiator ATG) as above and the antisense primer with an engineered NcoI site (5[prime]-CCATGGTGGCAGATGCTTCTGCTGCAGA-3[prime]). After PCR cloning, the NcoI fragment was used to replace an NcoI fragment (encoding amino acids 1-138) from the wild-type cDNA. The resulting cDNA encodes a protein with amino acids 55-138 deleted. The construct coding for a protein with the central loop deleted ([Delta]147-235) was prepared by removing an SmaI-MscI fragment (coding for amino acids 147-235) from the cDNA and religating (blunt ends) the remaining sequences in-frame. The presence of the specific mutations as well as the fidelity of the PCR-amplified cDNAs were confirmed by automated DNA sequencing.
Virus production and infection
Stable virus-producing cell lines were generated using procedures described previously (39). Briefly, the retroviral constructs were used to transfect GP + E86 ecotropic packaging cells (40) and virus thus produced was used to infect the amphotropic packaging cell line PA317 (38). Selection was performed in 200 U/ml hygromycin B 48 h after infection and continued until colonies were visible. The colonies were pooled and expanded, thus establishing the virus-producing cell lines. Primary patient cells were infected by exposure in virus-containing medium ovenight in the presence of polybrene (4 µg/ml) as described (41).
ACKNOWLEDGEMENTS
We thank Tim Johns and Katherine Fu for excellent technical assistance and Isla Ogilvie for comments on the manuscript. This research was supported by a grant from the Sick Childrens Hospital Foundation, Toronto, Canada, and the Muscular Dystrophy Association. J.Y. is the recipient of a Jeanne Timmins postdoctoral fellowship. E.A.S. is an MNI Killam Scholar.
REFERENCES
+To whom correspondence should be addressed. Tel: +1 514 398 1997; Fax: +1 514 398 1509; Email: eric{at}ericpc.mni.mcgill.ca
This page is run by Oxford University Press, Great Clarendon Street, Oxford OX2 6DP, as part of the OUP Journals
Comments and feedback: jnl.info{at}oup.co.uk
Last modification:
Copyright© Oxford University Press, 1999.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  F. A. Bundschuh, A. Hannappel, O. Anderka, and B. Ludwig Surf1, Associated with Leigh Syndrome in Humans, Is a Heme-binding Protein in Bacterial Oxidase Biogenesis J. Biol. Chem., September 18, 2009; 284(38): 25735 - 25741. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Stiburek, K. Vesela, H. Hansikova, H. Hulkova, and J. Zeman Loss of function of Sco1 and its interaction with cytochrome c oxidase Am J Physiol Cell Physiol, May 1, 2009; 296(5): C1218 - C1226. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. N. Baden, J. Murray, R. A. Capaldi, and K. Guillemin Early Developmental Pathology Due to Cytochrome c Oxidase Deficiency Is Revealed by a New Zebrafish Model J. Biol. Chem., November 30, 2007; 282(48): 34839 - 34849. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. A. Zordan, P. Cisotto, C. Benna, A. Agostino, G. Rizzo, A. Piccin, M. Pegoraro, F. Sandrelli, G. Perini, G. Tognon, et al. Post-transcriptional Silencing and Functional Characterization of the Drosophila melanogaster Homolog of Human Surf1 Genetics, January 1, 2006; 172(1): 229 - 241. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D M Kirby, R McFarland, A Ohtake, C Dunning, M T Ryan, C Wilson, D Ketteridge, D M Turnbull, D R Thorburn, and R W Taylor Mutations of the mitochondrial ND1 gene as a cause of MELAS J. Med. Genet., October 1, 2004; 41(10): 784 - 789. [Full Text] [PDF]
|
|
|
|
 |
|
 M. Zeviani and S. Di Donato Mitochondrial disorders Brain, October 1, 2004; 127(10): 2153 - 2172. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C E Oquendo, H Antonicka, E A Shoubridge, W Reardon, and G K Brown Functional and genetic studies demonstrate that mutation in the COX15 gene can cause Leigh syndrome J. Med. Genet., July 1, 2004; 41(7): 540 - 544. [Full Text] [PDF]
|
|
|
|
|
|
S. L. Williams, I. Valnot, P. Rustin, and J.-W. Taanman Cytochrome c Oxidase Subassemblies in Fibroblast Cultures from Patients Carrying Mutations in COX10, SCO1, or SURF1 J. Biol. Chem., February 27, 2004; 279(9): 7462 - 7469. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Antonicka, S. C. Leary, G.-H. Guercin, J. N. Agar, R. Horvath, N. G. Kennaway, C. O. Harding, M. Jaksch, and E. A. Shoubridge Mutations in COX10 result in a defect in mitochondrial heme A biosynthesis and account for multiple, early-onset clinical phenotypes associated with isolated COX deficiency Hum. Mol. Genet., October 16, 2003; 12(20): 2693 - 2702. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A.-R. Moslemi, M. Tulinius, N. Darin, P. Aman, E. Holme, and A. Oldfors SURF1 gene mutations in three cases with Leigh syndrome and cytochrome c oxidase deficiency Neurology, October 14, 2003; 61(7): 991 - 993. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. C. Goldenberg, R. D. Steiner, L. S. Merkens, T. Dunaway, R. A. Egan, E. A. Zimmerman, G. Nesbit, B. Robinson, and N. G. Kennaway Remarkable improvement in adult Leigh syndrome with partial cytochrome c oxidase deficiency Neurology, March 11, 2003; 60(5): 865 - 868. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Jazayeri, A. Andreyev, Y. Will, M. Ward, C. M. Anderson, and W. Clevenger Inducible Expression of a Dominant Negative DNA Polymerase-gamma Depletes Mitochondrial DNA and Produces a rho 0 Phenotype J. Biol. Chem., March 7, 2003; 278(11): 9823 - 9830. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. A. Shoubridge Nuclear genetic defects of oxidative phosphorylation Hum. Mol. Genet., October 1, 2001; 10(20): 2277 - 2284. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J.-C. VON KLEIST-RETZOW, J. YAO, J.-W. TAANMAN, K. CHANTREL, D. CHRETIEN, V. CORMIER-DAIRE, A. RÖTIG, A. MUNNICH, P. RUSTIN, and E. A SHOUBRIDGE Mutations in SURF1 are not specifically associated with Leigh syndrome J. Med. Genet., February 1, 2001; 38(2): 109 - 113. [Full Text]
|
|
|
|
|
|
V. Tiranti, P. Corona, M. Greco, J.-W. Taanman, F. Carrara, E. Lamantea, L. Nijtmans, G. Uziel, and M. Zeviani A novel frameshift mutation of the mtDNA COIII gene leads to impaired assembly of cytochrome c oxidase in a patient affected by Leigh-like syndrome Hum. Mol. Genet., November 1, 2000; 9(18): 2733 - 2742. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Jaksch, I. Ogilvie, J. Yao, G. Kortenhaus, H.-G. Bresser, K.-D. Gerbitz, and E. A. Shoubridge Mutations in SCO2 are associated with a distinct form of hypertrophic cardiomyopathy and cytochrome c oxidase deficiency Hum. Mol. Genet., March 22, 2000; 9(5): 795 - 801. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||