Nature Genetics
20, 337 - 343 (1998)
doi:10.1038/3804
SURF1, encoding a factor involved in the biogenesis of cytochrome
c oxidase, is mutated in Leigh syndromeZhiqing Zhu1, 6, Jianbo Yao1, 6, Timothy Johns1, Katherine Fu1, Isabelle De Bie1, Carol Macmillan1, Andrew P. Cuthbert3, Robert F. Newbold3, Jia-chi Wang4, Mario Chevrette4, Garry K. Brown5, Ruth M. Brown5
& Eric A. Shoubridge.1, 21 Montreal Neurological Institute, 3801
University Street, Montreal, Quebec, Canada
H3A 2B4. 2 Department of Human Genetics, 1205 avenue
Dr. Penfield, McGill University, Montreal, Quebec,
Canada H3A 1B1. 3 Department of Biology and Biochemistry, Brunel University,
Uxbridge, UK. 4 Montreal General Hospital Research Institute, Department
of Surgery, Urology Division, McGill University, Montreal,
Canada. 5 Genetics Unit, Department of Biochemistry, Oxford University,
South Parks Road, Oxford, U.K. 6 These authors contributed equally to this work.
Correspondence should be addressed to Eric A. Shoubridge. eric@ericpc.mni.mcgill.caLeigh Syndrome (LS) is a severe neurological disorder characterized by
bilaterally symmetrical necrotic lesions in subcortical brain regions that
is commonly associated with systemic cytochrome c oxidase (COX) deficiency.
COX deficiency is an autosomal recessive trait and most patients belong to
a single genetic complementation group. DNA sequence analysis of the genes
encoding the structural subunits of the COX complex has failed to identify
a pathogenic mutation. Using microcell-mediated chromosome transfer, we mapped
the gene defect in this disorder to chromosome 9q34 by complementation of
the respiratory chain deficiency in patient fibroblasts. Analysis of a candidate
gene (SURF1) of unknown function revealed several mutations, all of
which predict a truncated protein. These data suggest a role for SURF1 in
the biogenesis of the COX complex and define a new class of gene defects causing
human neurodegenerative disease.Introduction
LS is a subacute neurodegenerative condition characterized by bilaterally
symmetrical necrotic lesions in the brainstem, basal ganglia, thalamus and
spinal cord1,
2,
3,
4. Microscopically, these lesions are associated
with vascular proliferation, gliosis, neuronal loss, demyelination and cystic
cavitation. LS onset usually occurs in infancy, but adult cases have been
reported5,
6. LS is a genetically heterogeneous disease caused
by defects in enzymes involved in aerobic energy metabolism. These include
the X-linked E1 subunit of pyruvate dehydrogenase7,
8,
the mtDNA-encoded ATP6 subunit of ATP synthase8,
9,
10, respiratory
chain complex I (refs 8,12) and COX (refs 3,13). It has also been found in association with point mutations
in mitochondrial tRNA genes6,
8,
14,
15 and a mutation in the
Fp subunit of succinate dehydrogenase16.
Systemic COX deficiency presenting as LS is inherited as an autosomal recessive
trait and is one of the most common causes of LS; however, the underlying
molecular defect remains unknown. COX activity in these patients is reduced
in all tissues of the body, often to very low residual levels, with little
or no tissue specificity in the severity of the defect17,
18.
A biochemically distinct form of LS with COX deficiency exists in the French-Canadian
population in which the brain and liver are severely affected and fibroblasts
and skeletal muscle are relatively spared19. Somatic cell genetic
studies have demonstrated that the majority of patients with the classic form
of COX-deficient LS belong to a single genetic complementation group20,
21. DNA sequence analysis of cDNA for all 13 structural subunits
of the COX complex, in both the classical and French-Canadian forms of the
disease, have not revealed any pathogenic mutations22,
23. This
is consistent with earlier biochemical and molecular genetic studies that
suggested a failure to assemble an active enzyme complex as the basis of the
lack of enzyme activity17,
18,
24.
The assembly of COX requires the expression of a much larger number of
genes than those encoding the structural subunits of the complex. More than
30 different genetic complementation groups for COX assembly have been identified
in yeast25,
26. Many of these, such as factors that modulate
translational efficiency27,
28 by binding to the 5´ or 3´
UTR of mtDNA-encoded COX subunits, probably do not have mammalian homologues,
as mammalian mitochondrial mRNAs do not have UTRs. Although human COX assembly
genes have been identified that could be considered potential candidate genes
for LS (refs 29,30),
the number and identity of the genes involved in COX biogenesis in mammals
remains largely unknown.
Identifying the genetic defect in COX-deficient LS by sequencing candidate
genes as they are identified is, therefore, an uncertain prospect. Likewise,
the small family size in most cases precludes gene mapping by conventional
linkage analysis. To circumvent these problems, we have attempted to map and
identify the defective gene by functional complementation of the enzyme defect
in patient cells, using microcell-mediated chromosome transfer31.
We show here that transfer of a normal human chromosome 9 into patient fibroblasts
rescues the classical COX deficiency associated with LS. We were able to localize
the genetic defect to a small region of approximately 4.5 cM in 9q34 using
deletion mapping. DNA sequence analysis of a candidate gene (SURF1)
in the region revealed several different pathogenic mutations, all of which
predict a truncated protein.
Results
Microcell-mediated chromosome transfer
Fibroblasts
from LS patients with COX deficiency have some residual COX activity but,
similar to yeast pet mutants, are unable to grow in medium that lacks
a fermentable substrate. To determine the chromosomal location of the gene
defect in these patients, we rescued this phenotype by functional complementation
with a normal human chromosome incorporated into the genome of patient fibroblasts
by microcell-mediated chromosome transfer. A panel of stable monochromosomal
human:mouse hybrid cell lines containing human chromosomes tagged with a dual
selectable marker (HyTK) was used as the source of donor human chromosomes32. All 22 autosomes and the X chromosome were transferred, one at
a time, into a patient fibroblast line (patient W) and the cells were selected
in hygromycin B. Surviving colonies were picked 3-4 weeks after fusion and
COX activity was measured on pooled colonies derived from each chromosome
transfer using a spectrophotometric assay in whole cell extracts or a cytochemical
assay in intact cells (Table 1). Transfer of chromosome
9 from the K1-9 monochromosomal hybrid line restored specific COX activity
to near control levels (Table 1, Fig.
1). When normalized to citrate synthase, a mitochondrial matrix marker,
the COX activity was approximately 50% that of control cells. (Full restoration
of COX activity may require two normal chromosomes.) COX activity was not
restored by transfer of any other autosome or the X chromosome, averaging
approximately 10% of control values when normalized to citrate synthase (Table 1). Transfer of chromosome 9 also restored the ability
of patient fibroblasts to grow in medium lacking a fermentable substrate (medium
in which galactose is substituted for glucose). In fact, complemented clones
could be recovered in this medium without the use of hygromycin B after transfer
of chromosome 9.
 | | |
| | |
To test the specificity of this result, chromosome 9 was transferred into
fibroblasts from a second patient (patient C) in the same genetic complementation
group, and additional clones were isolated from patient W. COX activity was
assessed in a large number of clones, each representing an independent transfer
of chromosome 9. This was done by cytochemical assay in most clones (Table 1). In two clones (9-1 and 9-4, Table
1), sufficient cells were available for the spectrophotometric assay.
Enzyme activity was restored in 21 of 23 independent clones derived from patient
W cells and 6 of 7 clones in cells derived from patient C (
Table 1, and data not shown).
Transfer of deleted human chromosome 9 maps the gene defect to 9q
To refine the map position of the defective gene on chromosome
9, we first introduced deleted versions of human chromosome 9 from two different
monochromosomal hybrid lines: A9-9del(p) and A9-9HyTK-1.5. The deleted chromosomes
in these lines have been mapped with a panel of microsatellite markers and,
in the case of clone 1.5, with additional markers in 9p21 (33). We further characterized the chromosome 9 deletions in
these lines by Alu-PCR FISH in normal human lymphoblasts34
using a reverse painting probe generated from donor hybrid lines with human-specific
Alu PCR primers (Fig. 2). The A9-9del(p) line lacks
most of the p arm of human chromosome 9 (Fig. 2a),
and the A9-9HyTK-1.5 line has a complex pattern of large deletions in both
the p and q arms (Fig. 2b). We detected one major
deletion in the p arm and three in the q arm, (encompassing q21.2-q22.1, q22.3-q32
and q34). Microcell hybrids constructed from the A9-9HyTK-1.5 line failed
to rescue the COX defect in all clones derived from patient W (8/8) and patient
C (8/8). On the other hand, all clones from patient W (5/5) and some clones
from patient C (16/21) were rescued after transferring human chromosome 9
del(p). (COX activity was determined cytochemically in all clones.) From these
data, we concluded that the gene defect maps to one of the deleted regions
on 9q in the A9-9HyTK-1.5 cell line.
| | |
Deletion mapping localizes the gene defect to 9q34
Chromosomes incorporated into the genome of recipient cells by microcell transfer
often delete and rearrange31,
35. To further narrow the region
containing the gene of interest, we used polymorphic microsatellite markers
to map the regions of human chromosome 9 that were incorporated into the genomes
of complementing and non-complementing clones isolated after selection in
hygromycin B. Markers that were informative for the presence of the donor
chromosome and missing in complemented clones could be excluded from the region
of interest, as could informative markers present in non-complementing clones.
Analysis of 6 complementing clones and 1 non-complementing clone from patient
C, and 20 complementing and 2 non-complementing clones from patient W (all
recovered following transfer of the entire chromosome 9 from the K1-9 cell
line), allowed us to map the gene defect telomeric to D9S149 (Fig. 3). Most of this region is missing in human chromosome
9 in the A9-9HyTK-1.5 cell line, in which the first detectable markers telomeric
to D9S149 are D9S158 and D9S1838. These data map the
candidate gene to a small region of approximately 4.5 cM in 9q34.
| | |
Exclusion mapping
To test whether the region of interest
could be narrowed further by exclusion mapping, we genotyped two small families
with a dense set of markers telomeric to D9S260. Affected family members
should share common alleles at the locus of the gene defect, whereas affected
and unaffected individuals should be discordant. This analysis allowed us
to exclude regions centromeric to D9S159 and telomeric to D9S1818
, but did not permit any further narrowing of the region of interest (Fig. 3).
Mutations in SURF1 result in reduced transcript stability
A large part (approximately 1.4 Mb) of the chromosomal region
surrounding 9q34 has been completely sequenced as part of a successful search
for TSC1, a gene associated with tuberous sclerosis36.
This gene-dense region contains at least 30 genes between the markers
D9S149 and D9S114. SURF1 maps to the non-excluded region
in our patients. The function of SURF1 is yet to be determined but a yeast
homologue of SURF1 (SHY1) rescues a yeast nuclear pet
mutant with a partial decrease in COX activity37, making
SURF1 a candidate gene for LS with COX deficiency.
We amplified all nine exons of SURF1 by PCR using genomic DNA from
patients W and C and an affected member from family D, and sequenced them
to screen for mutations. Two heterozygous mutations were found in patient
W: a C765T mutation that produces a nonsense codon in exon 7 and a 337+2 T C
mutation in the donor splice site of intron 4 (Fig. 4a,
b). Northern-blot (Fig. 5a) and quantitative
RT-PCR (Fig. 5b) analyses of total RNA derived
from patient W fibroblasts showed a large overall reduction (80-85%) in the
level of SURF1 mRNA, and RT-PCR revealed two PCR products (
Fig. 5c). Sequencing of the 771-bp minor RT-PCR product revealed
that it lacked exon 4. This appears to result from the use a cryptic donor
sequence (GT) at the 5´ end of exon 4, which causes the removal of exon
and intron 4, if the wild-type donor consensus sequence in intron 4 is mutated.
Deletion of exon 4 produces a frameshift in the mRNA, resulting in the appearance
of several nonsense codons, the first of which is encountered in exon 5, 28
bases downstream from the 3´ end of exon 3. An unaffected sibling of
patient W was heterozygous for the wild-type allele and the C765T mutation.
| | |
| | |
Genomic DNA sequence analysis in patient C also revealed two heterozygous
mutations (Fig. 4c,d): an insertion/deletion
mutation in exon 4 (326insATdelTCTGCCAGCC), which creates a nonsense codon
at the site of the mutation, and a 2-bp deletion in exon 9 which removes one
of three CT repeats between positions 855 and 860. As we do not know which
of these repeats is removed, we have arbitrarily assigned the mutation to
the first repeat (855delCT). This results in the appearance of a nonsense
codon at position 884, very near the carboxy-terminal coding region of the
gene. Northern-blot (Fig. 5a) and RT-PCR analysis
(Fig. 5b) of patient C fibroblasts demonstrated
a reduction of 50-60% in steady-state level of SURF1 mRNA. DNA sequence
analysis of the RT-PCR products from patient C, however, showed that the mRNA
transcribed from the allele containing the exon 4 insertion/deletion was a
minority species that did not interfere with interpretation of the sequence
of the allele with the exon 9 mutation. The decrease in the steady-state mRNA
level in patient C appears, therefore, to be due largely to a decrease in
the message from the allele with the nonsense codon in exon 4.
Finally, a homozygous insertion mutation was found in exon 9 in an affected
case from family D, which creates a nonsense codon at position 884, as in
patient C (Fig. 4e). The mutation inserts a T
into a string of Ts; thus, the exact position of the mutation cannot be determined.
We refer to this mutation as 882insT. Thus, all five pathogenic alleles identified
in SURF1 in three pedigrees (Fig. 4f)
result in the appearance of nonsense codons in the gene that predict a truncated
protein product.
Rescue of the COX deficiency by SURF1 cDNA
To test directly whether the lack of a functional SURF1 protein was responsible
for the COX deficiency in LS, patient fibroblasts were transiently transfected
with SURF1 cDNA in an expression vector using a biolistic device, and
COX activity was determined cytochemically 48 hours later. As a control, patient
cells were transfected with cDNA encoding green fluorescent protein (GFP).
COX activity was restored in some cells transfected with SURF1 cDNA,
but not with GFP (Fig. 6). The proportion of either
COX-positive or GFP-positive cells was similar in the two experiments (less
than 2% of surviving cells).
Discussion
Several lines of evidence indicate that the classical COX deficiency associated
with LS is caused by mutations in SURF1. First, functional complementation
of the respiratory chain deficiency in patient fibroblasts occurred only when
the region of chromosome 9 between markers D9S1830 and D9S1818,
which includes the SURF locus, was stably transferred. Second, exclusion
mapping in two small pedigrees using polymorphic microsatellite markers mapped
the gene defect to a small region on 9q encompassing the SURF locus.
Third, pathogenic mutations predicting a truncated SURF1 protein were found
in five different alleles in three independent pedigrees, in two cases as
compound heterozygotes and in one case as a homozygote. Fourth, northern-blot
and RT-PCR analyses showed a reduction in the steady-state levels of SURF1
mRNA from patient fibroblasts, the magnitude of which correlated with
the predicted position of the first nonsense codon in the gene. Decreased
message stability is a characteristic feature of premature translation termination
signals38. Finally, transient expression of SURF1 cDNA
is able to restore COX activity in patient fibroblasts.
SURF1 is embedded in a cluster of housekeeping genes (the surfeit
locus); the structure of this cluster is unique in the mammalian genome39. The locus consists of six housekeeping genes (SURF1-
6), unrelated by sequence in approximately 36 kb of genomic DNA. In the
mouse, the largest distance between any two genes is 73 bp; two genes,
SURF2 and SURF4, share overlapping reading frames. This basic structure
has been conserved over 250 million years of divergent evolution between birds
and mammals, suggesting it has functional importance in coordinate gene regulation40. Adjacent genes are transcribed in opposite orientations, and some
(such as SURF1 and SURF2) share bidirectional promoters41,
42. Except for SURF3, which encodes the ribosomal protein
L7a (43), the function of SURF genes
is largely unknown.
SURF1 was evaluated as a candidate gene in LS patients with COX
deficiency because it mapped to the non-excluded region of chromosome 9 and
because a yeast homologue, SHY1, has been shown to restore the ability
of a yeast pet mutant in the G91 complementation group to grow on a
non-fermentable substrate37. The biochemical defect rescued
by SURF1 in LS patient fibroblasts is a specific and severe deficiency
in the activity of COX, whereas the biochemical abnormality in the yeast G91
complementation group is pleiotropic37 and cannot be explained
by the activities of individual components of the respiratory chain. It has
been suggested that the phenotype may result from inefficient transfer of
electrons between the bc1 (ubiquinol cytochrome c reductase) and COX
complexes37. Biochemical studies of LS patient fibroblasts demonstrate
that all of the structural subunits of the COX complex are present, albeit
at reduced levels17,
18,
24, and that the synthesis of the COX
subunits is normal. These data suggest that SURF1 has a role in the assembly
or maintenance of an active holoenzyme COX complex, and point to different,
but perhaps overlapping, functions of the human and yeast proteins.
SHY1 encodes a protein of 389 aa and SURF1, a protein of
300 aa (Fig. 7). The overall amino acid homology between
the yeast and human genes is 25.6%. The region of greatest homology is between
aa 37 and 248 of the human sequence, where the sequences are 33% identical.
Expression studies have localized Shy1 to the mitochondrial inner membrane37 and computer analysis with a program that evaluates mitochondrial
targeting (MitoProt II) supports a mitochondrial localization for human SURF1
(44). Both human and yeast proteins have
an amino terminus with a characteristic mitochondrial targeting sequence.
Expression of human SURF1 cDNA with a C-terminal FLAG tag in COS-7
cells co-localizes SURF1 with Mitotracker, a mitochondrial marker (unpublished
data). Expression of the C-terminal half of SHY1 is able to rescue
point mutations in SHY1 in yeast, but not a targeted gene deletion37. Hydropathy plots of Shy1p predict two membrane-spanning domains
close to the N and C termini, and it has been suggested that the C-terminal
domain alone is sufficient to interact with a mutant protein and rescue the
pet phenotype37. These transmembrane domains are highly
conserved in human, mouse and Fugu rubripes proteins (
Fig. 7). The MitoProt II programme44 predicts an additional
putative hydrophobic domain in the middle of human SURF1.
| | |
The mutations that we have so far identified in SURF1 all predict
truncated proteins. Some, like 855delCT and 882insT, predict a truncation
very near the C terminus of the protein, with an intact N-terminal domain.
These alleles, in which the last 19 aa are affected by the mutation (the last
10 are truncated), are predicted to severely disrupt the C-terminal transmembrane
domain (Fig. 7) and are unable to rescue growth of patient
cells in glucose-free medium. This suggests an important function for the
C terminus of human SURF1 or a decreased stability of SURF1 with a truncated
C terminus.
Nuclear genes involved in respiratory chain defects have been difficult
to identify. The first such reported gene defect was found in a structural
subunit of succinate dehydrogenase (Fp) in a family presenting
with LS (16), and a mutation in a nuclear
gene encoding structural subunit of complex I has recently been reported45. Most LS patients with COX deficiency probably have mutations in
SURF1, as nearly all belong to the same major genetic complementation
group20,
21. Very recently, mutations in a mitochondrial ATPase,
paraplegin, were reported in hereditary spastic paraplegia46,
a neurodegenerative disease characterized by degeneration of the cortico-spinal
and dorsal column axons. Paraplegin is homologous to a family of yeast mitochondrial
genes, the AAA proteases, which possess both proteolytic and chaperone-like
activities47. These proteins are thought to be important in
the assembly and turnover of the subunits of the complexes of the mitochondrial
respiratory chain47. Our data indicate that SURF1 encodes
a putative assembly or maintenance factor which, in humans, appears to be
specific for the COX complex. Together, these studies define a new class of
genes responsible for human neurodegenerative disorders. The biochemical evidence
suggests that at least one more such biogenesis factor will be found in the
French Canadian form of COX-deficient LS (19).
As in other neurodegenerative diseases associated with ubiquitously expressed
genes, the basis for selective cellular vulnerability in the classical form
of COX-deficient LS remains a mystery.
The functional complementation approach we have taken to identify the gene
defect in COX-deficient LS could serve as a paradigm to map and clone other
nuclear genes associated with respiratory chain disorders, such as mtDNA depletion
syndrome48 or complex I-deficient LS. Family size is usually
small in these disorders, and the extent of genetic heterogeneity is unknown.
We have demonstrated that deletion mapping in cells derived from a single
patient can be used to narrow the chromosomal location of a disease gene to
a region small enough to identify and test candidate genes. The ultimate identification
of the gene defect in this study was aided by previous studies of yeast
pet mutants, illustrating the value of using model organisms for clues
to the genetic basis of human disease.
Methods
Cell lines.
Primary fibroblast lines were established
from two patients (W and C) with COX-deficient LS, both of whom belong to
the same major genetic complementation group20 (unpublished
data). Both patients had a severe deficiency in COX activity and a typical
LS phenotype. The primary fibroblast lines were transduced with a retroviral
vector expressing the E6E7 region of type 16 human papilloma virus to extend
their life span49 and grown in high glucose DMEM supplemented
with 10% fetal bovine serum.
Microcell-mediated chromosome transfer.
A panel of
human:mouse monochromosomal hybrids32 was used as the source
of normal human donor chromosomes. All 22 autosomes and the X chromosome were
transferred into one of the patient cell lines by microcell-mediated chromosome
transfer50. Briefly, donor cells were plated in DMEM containing
10% fetal bovine serum and hygromycin B (400 U; Calbiochem) 3 d before fusion.
The medium was then changed to DMEM plus 20% fetal bovine serum and the cells
were exposed to colchicine (0.03-0.06  g/ml) for 48 h to induce micronucleation.
The cells were then collected by trypsinization and plated on plastic 'bullets'
(custom-made from tissue culture plates to fit into 50-ml centifuge tubes)
coated with crosslinked concanavalin A (Sigma). Microcells were prepared by
centrifugation at 34-37 °C and 15,000 r.p.m. (SS34 rotor) in media containing
cytochalasin B (10 g/ml; Sigma), filtered through 8 m and 5 m
filters, pelleted at 3000 r.p.m. on a benchtop centrifuge (Beckman) and resuspended
in serum-free medium. The microcell suspension was added to the plate containing
the recipient cells along with phytohaemagglutinin (100 g/ml) and incubated
for 15-20 min. Microcells were fused with 45% PEG plus 10% DMSO for 60 s,
washed with serum-free medium and incubated in DMEM plus 10% FBS. After 48-72
h, fused cells were selected in hygromycin B (100 units/ml). Colonies were
picked, expanded and analysed 3-4 weeks later. In one experiment, cells were
selected in media in which galactose was substituted for glucose (glucose-free
media).
COX cytochemistry and enzymology.
COX activity was
measured spectrophotometrically51 and cytochemically52
as described.
FISH analysis.
Normal human lymphoblasts were collected
in accordance with the standard procedure for metaphase preparations for high
resolution FISH analysis. Human-specific Alu primers (primers 153,
154, 450, 451) were used to generate probes from the A9(del)p and A9-9HYTK-1.5
cell lines53. Alu-PCR products were pooled and biotin-labelled
with the BRL BioPrime DNA labelling kit (8094SA). Cot-1 human DNA (2 g)
was added to the final probe mixture containing biotin-labelled DNA (100 ng).
After hybridization in a humid chamber containing 50% formamide/2 SSC
at 37 °C overnight with the denatured probe, the slides were washed in
50% formamide/2SSC and 2SSC, 10 min each and in 4SSC/5%
Triton 100 for 5 min. Hybridization signals were detected with rabbit
anti-biotin (1:100; Enzo Diagnostics), goat anti-rabbit IgG (1:100; Gibco
BRL) and amplified with streptavidin FITC (1:100; Gibco-BRL). Chromosomes
were counterstained with propidium iodide (1.25 g/l) and visualized
by epifluorescence.
Deletion and exclusion mapping.
Oligonucleotide primers
for polymorphic microsatellite markers on chromosome 9 were obtained (Research
Genetics). Deletion mapping was done on DNA isolated from hygromycin B-resistant
clones obtained following microcell fusion from the K1-9 monochromosomal hybrid
line using spin columns (Qiagen Genomic DNA). Exclusion mapping was carried
out in two small families: family W, which includes patient W, one other affected
and one unaffected member; and family D with two affected individuals and
one unaffected20. The primers were 5´-end labelled with 
33P-dATP and PCR was performed using the Research Genetics protocol
using AmpliTaq gold (PE Applied Biosystems). PCR products were run on 6% sequencing
gels and analysed on a Storm phosphorimager (Molecular Dynamics).
Mutation detection.
All nine exons of SURF1
were amplified using Pfu polymerase in five separate PCR reactions
with intronic primer pairs designed using the OLIGO primer design program.
PCR products were gel purified using the Qiaex II gel extraction kit (Qiagen)
and sequenced by automated sequencing. PCR products containing exons 1 and
2 could only be amplified using 7-deaza guanine54, as this region
is very GC rich due to the proximity of the CpG island at the 5´ end
of the gene. To verify the allelic nature of the mutations in patient C, PCR
products were cloned into a PCR cloning vector (Stratagene) and sequenced.
RT-PCR products, generated as outlined below, were also sequenced in patients
W and C. This allowed us to verify the allelic nature of the pathogenic mutations
in patient W because of the size difference of the PCR products. The positions
of mutations are numbered using the cDNA sequence for SURF1 (42).
Quantitative RT-PCR of SURF1 mRNA and northern
-blot analysis. Total RNA from control and patient fibroblasts was
isolated using an RNA isolation kit (Qiagen). SURF1 cDNA was amplified
by a single step RT-PCR procedure, the Titan One Tube RT-PCR kit (Boehringer),
using total RNA (1 g) as starting material for reverse transcription.
Two pairs of primers were included in the RT-PCR mixture: one that would specifically
amplify a 854-bp SURF1 product, and another that would amplify a 187-bp
human ribosomal RPL27 product, a representative low expression housekeeping
protein55. Reverse transcription was carried out for 30 min
at 50 °C, followed by 26 (RPL27) or 30 (SURF1) PCR cycles,
at which point amplification of the PCR products was still in the linear range.
For quantitative analysis, the PCR was carried out with 32P-dCTP;
the reaction products were separated on a 10% polyacrylamide gel and analysed
on a Storm phosphorimager using ImageQuant (Molecular Dynamics). To visualize
the two spliced forms of the SURF1 message in patient W, PCR was carried
out for 35 cycles, the products separated on 1% agarose gels and stained with
ethidium bromide.
For northern-blot analysis, poly(A)+ mRNA (1 g) was
purified from total RNA using the Oligotex mRNA kit (Qiagen) and electrophoresed
through a 1% agarose gel containing formaldehyde. After transfer to a Zetaprobe
membrane, the blot was hybridized overnight with a random-prime labelled
SURF1 cDNA probe. The blot was washed at RT for 20 min in 2SSC
0.1% SDS and then at 55 °C for 30 min in 0.1SSC 0.1% SDS. SURF1
mRNA was quantitated on a Storm phosphorimager (Molecular Dynamics) using
ImageQuant and normalized to the level of actin mRNA .
Expression of human SURF1 cDNA in patient cells.
Human SURF1 cDNA was generated by RT-PCR from total RNA isolated
from normal human fibroblasts using primers corresponding to the published
human SURF1 cDNA sequence42. A BamHI site and
consensus Kozak sequence (GCCACC) were engineered into the cDNA immediately
preceding the initiator ATG, and another BamHI site was introduced
after the stop codon. The amplified cDNA was first cloned using PCR-script
amp cloning kit (Stratagene) and then subcloned into the BamHI site
of pcDNA3 (Invitrogen).
Transfer of SURF1 cDNA into patient fibroblasts was performed by
particle bombardment. Plasmid DNA was precipitated onto gold particles as
described56. Cells were grown on glass coverslips and bombarded
with DNA-coated particles using a biolistic device similar to PDS-1000/He
(Bio-Rad). Approximately 0.5 g of DNA on 0.3 mg of particles were used
per coverslip. Cells were stained for COX activity after 48 h. As a control
for the efficiency of the biolistic process, cells were bombarded with a GFP
expression vector (pEGFP-N1; Clontech).
Received 15 July 1998; Accepted 5 October 1998
REFERENCES
- Leigh, D. Subacute necrotizing encephalomyelopathy in an infant. J. Neurol. Neurosurg. Psychiatry 14, 216−221 (1951). | ISI | ChemPort |
- van Erven, P.M. et al. Leigh syndrome, a mitochondrial encephalo(myo)pathy. A review of the literature. Clin. Neuro. Neurosurg. 89, 217−230 (1987). | ChemPort |
- Van Coster, R. et al. Cytochrome c oxidase-associated Leigh syndrome: phenotypic features and pathogenetic speculations. J. Neurol. Sci. 104, 97−111 (1991). | Article | PubMed | ChemPort |
- Brown, G.K. & Squier, M.V. Neuropathology and pathogenesis of mitochondrial diseases. J. Inherit. Metab. Dis. 19, 553−572 (1996). | PubMed | ISI | ChemPort |
- Kalimo, H., Lundberg, P.O. & Olsson, Y. Familial subacute necrotizing encephalomyelopathy of the adult form (adult Leigh syndrome). Ann. Neurol. 6, 200−206 (1979). | Article | PubMed | ISI | ChemPort |
- Chalmers, R.M. et al. A mitochondrial DNA tRNA(Val) point mutation associated with adult-onset Leigh syndrome. Neurol. 49, 589−592 (1997). | ISI | ChemPort |
- Dahl, H.H. et al. Mutations and polymorphisms in the pyruvate dehydrogenase E1 gene. Hum. Mutat. 1, 97−102 (1992). | PubMed | ChemPort |
- Rahman, S. et al. Leigh syndrome: clinical features and biochemical and DNA abnormalities. Ann. Neurol. 39, 343−351 (1996). | Article | PubMed | ISI | ChemPort |
- Tatuch, Y. et al. Heteroplasmic mtDNA mutation (T to G) at 8993 can cause Leigh disease when the percentage of abnormal mtDNA is high. Am. J. Hum. Genet. 50, 852−858 (1992). | PubMed | ISI | ChemPort |
- Santorelli, F.M., Shanske, S., Macaya, A., DeVivo, D.C. & DiMauro, S. The mutation at nt 8993 of mitochondrial DNA is a common cause of Leigh's syndrome. Ann. Neurol. 34, 827−834 (1993). | Article | PubMed | ISI | ChemPort |
- Robinson, B.H., De Meirleir, L., Glerum, M., Sherwood, G. & Becker, L. Clinical presentation of mitochondrial respiratory chain defects in NADH-coenzyme Q reductase and cytochrome oxidase: clues to pathogenesis of Leigh disease. J. Pediatr. 110, 216−222 (1987). | PubMed | ISI | ChemPort |
- Morris, A.A. et al. Deficiency of respiratory chain complex I is a common cause of Leigh disease. Ann. Neurol. 40, 25−30 (1996). | Article | PubMed | ISI | ChemPort |
- DiMauro, S. et al. Cytochrome c oxidase deficiency in Leigh syndrome. Ann. Neurol. 22, 498−506 (1987). | Article | PubMed | ISI | ChemPort |
- Berkovic, S.F. et al. Myoclonus epilepsy and ragged-red fibres (MERRF). 1. A clinical, pathological, biochemical, magnetic resonance spectrographic and positron emission tomographic study. Brain 112, 1231−1260 (1989). | PubMed | ISI |
- Santorelli, F.M. et al. Maternally inherited encephalopathy associated with a single-base insertion in the mitochondrial tRNATrp gene. Ann. Neurol. 42, 256−260 (1997). | Article | PubMed | ISI | ChemPort |
- Bourgeron, T. et al. Mutation of a nuclear succinate dehydrogenase gene results in mitochondrial respiratory chain deficiency. Nature Genet. 11, 144−149 (1995). | Article | PubMed | ISI | ChemPort |
- Glerum, D.M., Yanamura, W., Capaldi, R.A. & Robinson, B.H. Characterization of cytochrome-c oxidase mutants in human fibroblasts. FEBS Lett. 236, 100−104 (1988). | Article | PubMed | ISI | ChemPort |
- Lombes, A. et al. Biochemical and molecular analysis of cytochrome c oxidase deficiency in Leigh's syndrome. Neurol. 41, 491−498 (1991). | ISI | ChemPort |
- Merante, F. et al. A biochemically distinct form of cytochrome oxidase (COX) deficiency in the Saguenay-Lac-Saint-Jean region of Quebec. Am. J. Hum. Genet. 53, 481−487 (1993). | PubMed | ISI | ChemPort |
- Brown, R.M. & Brown, G.K. Complementation analysis of systemic cytochrome oxidase deficiency presenting as Leigh syndrome. J. Inherit. Metab. Dis. 19, 752−760 (1996). | PubMed | ISI | ChemPort |
- Munaro, M. et al. A single cell complementation class is common to several cases of cytochrome c oxidase-defective Leigh's syndrome. Hum. Mol. Genet. 6, 221−228 (1997). | Article | PubMed | ISI | ChemPort |
- Adams, P.L., Lightowlers, R.N. & Turnbull, D.M. Molecular analysis of cytochrome c oxidase deficiency in Leigh's syndrome. Ann. Neurol. 41, 268−270 (1997). | Article | PubMed | ISI | ChemPort |
- Lee, N., Morin, C., Mitchell, G. & Robinson, B.H. Saguenay Lac Saint Jean cytochrome oxidase deficiency: sequence analysis of nuclear encoded COX subunits, chromosomal localization and a sequence anomaly in subunit VIc. Biochim. Biophys. Acta 1406, 1−4 (1998). | Article | PubMed | ISI | ChemPort |
- Hayasaka, K., Brown, G.K., Danks, D.M., Droste, M. & Kadenbach, B. Cytochrome c oxidase deficiency in subacute necrotizing encephalopathy (Leigh syndrome). J. Inherit. Metab. Dis. 12, 247−256 (1989). | PubMed | ISI | ChemPort |
- Tzagoloff, A. & Dieckmann, C.L. PET genes of Saccharomyces cerevisiae. Microbiol. Rev. 54, 211−225 (1990). | PubMed | ISI | ChemPort |
- McEwen, J.E., Ko, C., Kloeckner-Gruissem, B. & Poyton, R.O. Nuclear functions required for cytochrome c oxidase biogenesis in Saccharomyces cerevisiae. Characterization of mutants in 34 complementation groups. J. Biol. Chem. 261, 11872−11879 (1986). | PubMed | ChemPort |
- Mulero, J.J. & Fox, T.D. PET111 acts in the 5'-leader of the Saccharomyces cerevisiae mitochondrial COX2 mRNA to promote its translation. Genetics 133, 509−516 (1993). | PubMed | ISI | ChemPort |
- Brown, N.G., Costanzo, M.C. & Fox, T.D. Interactions among three proteins that specifically activate translation of the mitochondrial COX3 mRNA in Saccharomyces cerevisiae. Mol. Cell. Biol. 14, 1045−1053 (1994). | PubMed | ISI | ChemPort |
- Glerum, D.M. & Tzagoloff, A. Isolation of a human cDNA for heme A:farnesyltransferase by functional complementation of a yeast cox10 mutant. Proc. Natl Acad. Sci. USA 91, 8452−8456 (1994). | PubMed | ChemPort |
- Bonnefoy, N. et al. Cloning of a human gene involved in cytochrome oxidase assembly by functional complementation of an oxa1-mutation in Saccharomyces cerevisiae. Proc. Natl Acad. Sci. USA 91, 11978−11982 (1994). | PubMed | ChemPort |
- Newbold, R.F. & Cuthbert, A.P. Mapping human senescence genes using interspecific monochromosome transfer. in Culture of Immortalized Cells (eds Freshney, R.I. & Freshney, M.G.) 53−75 (Wiley-Liss, New York, 1996). | ChemPort |
- Cuthbert, A.P. et al. Construction and characterization of a highly stable human: rodent monochromosomal hybrid panel for genetic complementation and genome mapping studies. Cytogenet. Cell Genet. 71, 68−76 (1995). | PubMed | ISI | ChemPort |
- England, N.L. et al. Identification of human tumour suppressor genes by monochromosome transfer: rapid growth-arrest response mapped to 9p21 is mediated solely by the cyclin-D-dependent kinase inhibitor gene, CDKN2A (p16INK4A). Carcinogenesis 17, 1567−1575 (1996). | PubMed | ISI | ChemPort |
- Lichter, P., Ledbetter, S.A., Ledbetter, D.H. & Ward, D.C. Fluorescence in situ hybridization with Alu and L1 polymerase chain reaction probes for rapid characterization of human chromosomes in hybrid cell lines. Proc. Natl Acad. Sci. USA 87, 6634−6638 (1990). | PubMed | ChemPort |
- Leach, R.J., Thayer, M.J., Schafer, A.J. & Fournier, R.E. Physical mapping of human chromosome 17 using fragment-containing microcell hybrids. Genomics 5, 167−176 (1989). | Article |
|
