Advertisement
Complements of the house
Eric A. Schon
Department of Neurology, Columbia University, New York, New York, USA.
Address correspondence to: Eric A. Schon, Department of Neurology, Columbia University, Room 4-431, 630 West 168th Street, New York, New York 10032, USA. Phone: (212) 305-1665; Fax: (212) 305-3986; E-mail: eas3@columbia.edu.
Find articles by Schon, E. in: PubMed | Google Scholar
Published September 15, 2004 - More info
J Clin Invest. 2004;114(6):760–762. https://doi.org/10.1172/JCI22942.
© 2004 The American Society for Clinical Investigation
Forty-six in the first, 4 in the second, 11 in the third, and 13 in the fourth, for a total of 74. Sounds like the score, by quarters, for an erratic college basketball team. Of course, it’s not — it’s the number of subunits in each of the 4 complexes of the human mitochondrial respiratory chain (Figure 1). And that number does not even include the 16 subunits of the fifth complex, ATP synthase, which uses the protons produced by the respiratory chain to fuel oxidative ATP synthesis.
Figure 1The OxPhos system, consisting of NADH dehydrogenase-CoQ oxidoreductase (Complex I), succinate dehydrogenase-CoQ oxidoreductase (Complex II), CoQ-cytochrome c oxidoreductase (Complex III), cytochrome c oxidase (Complex IV), and ATP synthase (Complex V). Nuclear DNA_encoded subunits are shown in blue; mtDNA-encoded subunits are shown in red; mutated OxPhos complex subunit polypeptides that cause mitochondrial disease are shown in bold; mutated assembly polypeptides are shown in plain text; Cyt c, cytochrome c.
And therein lies a problem. In the last 15 years, we have begun to recognize that defects in the mitochondrial respiratory chain/oxidative phosphorylation system (OxPhos) are responsible for a panoply of human disorders, ranging from sporadic myopathies to fatal encephalomyopathies. Those disorders can be maternally inherited, as a result of mutations in any one of the 13 polypeptides encoded by mitochondrial DNA (mtDNA), or they can be Mendelian, as a result of mutations in the 77 polypeptides encoded by nuclear DNA (nDNA). To make matters worse, there are at least 30 other proteins — all nDNA encoded — required for the proper assembly and functioning of these 5 complexes.
With more than 120 potential culprit genes responsible for OxPhos disorders, what’s a geneticist to do? In the old days — say, 5 years ago — the problem was simple. We didn’t have the human genome available online, so we spent lots of time looking for mutations in mtDNA. After all, how hard is it to find mutations in a genome that’s only 16.6 kb in size? In relatively short order, mutations were found in all 13 polypeptides (as well as in both mtDNA-encoded ribosomal RNAs and in 21 of the 22 transfer RNAs) (1). On the other hand, if anyone wanted to find the cause of a Mendelian OxPhos disorder — and there are plenty of them, almost all lethal in early infancy or childhood and almost all recessively inherited (for example, Leigh syndrome) — the task was daunting, for at least two reasons: first, there were, at a minimum, 120 candidate genes; where do you start? Second, there are few large pedigrees, so linkage analysis is usually not an option.
In spite of these obstacles, progress came, albeit slowly. Today, we know of mutations in at least 22 nDNA-encoded polypeptides, of which 8 are assembly proteins (Figure 1). Of the 22, nine are structural subunits of complex I, the largest by far of the 5 OxPhos complexes, which consists of 46 subunits (7 encoded by mtDNA and 39 by nDNA). There is no reason to believe that there could not be mutations in any of the 46 subunits or in complex I assembly protein. In fact, it appears that more patients with respiratory chain disorders have deficits in complex I than in all the other complexes combined (2). What would be the most expeditious way of finding these culprit genes, other than by brute force DNA sequencing?
One path to the solution of this problem was revealed by Eric Shoubridge in Montreal and then by Massimo Zeviani in Milan, both of whom identified mutations in SURF1, a complex IV assembly protein that was responsible for fatal infantile cytochrome c oxidase (COX) deficiency. Both groups had the unusual advantage of access to large pedigrees, which enabled them to map the culprit locus to chromosome 9. They narrowed down the locus using monochromosomal hybrids and microcell-mediated chromosome transfer, in which human chromosomes are “inserted” into patient COX-deficient fibroblasts in order to find the one chromosome (or piece of a chromosome) that could complement the defect. Once the region had been narrowed down to a manageable size (actually a few megabases!), some brute force, plus intelligent choices of candidate genes for sequencing, yielded the responsible gene, namely, SURF1 (3, 4).
A different approach was taken by Eric Lander in Boston and Brian Robinson in Toronto, who collaborated to look for a gene responsible for another recessive COX deficiency prevalent in individuals from the Saguenay Lac St. Jean region of Quebec. Using a combination of genomics, proteomics, and bioinformatics, they were able to identify mutations in another COX assembly gene called LRPPRC (5).
In this issue of the JCI, David Thorburn, Denise Kirby, and colleagues have combined elements of both of these approaches to determine, first, the prevalence of genes mutated in a series of patients with complex I deficiency of unknown origin, and second, the identity of one such new gene, using what might be called a “Venn diagram” approach (Figure 2) (6). They fused complex I–deficient cells from individual patients in pairwise fashion and asked whether fused pairs could rescue the respiratory-deficient phenotype; in other words, they performed a classical complementation experiment. In a second round of fusions, Kirby et al. asked whether patient cells could rescue function in ρ0 cells, which are cells containing mitochondria that are devoid of mtDNA (and hence devoid of respiratory function). Fusions that were rescued did so because the ρ0 nucleus complemented the mutation in the patient cell’s nucleus, which implied that the patient’s defect resided in a nuclear gene. On the other hand, fusions that failed to rescue the complex I deficiency did so because the mtDNA in the patient’s mitochondria could not rescue the respiratory deficiency arising from the ρ0 cell’s lack of mtDNA, which suggests that the patient’s genetic error resided in the mtDNA, not the nDNA.
Figure 2Venn diagram approach used by Kirby et al. (6). Smiling and frowning faces indicate elevated complex I activity (indicating rescue of the phenotype) and reduced complex I activity, respectively. The diagrams show only some of the possible combinations of outcomes.
Using this Venn diagram approach, Kirby et al. (6) were able to pigeonhole cells from 10 patients into 8 complementation groups, including 8 patients with mutations at 7 chromosomal loci and 2 patients with mutations in mtDNA-encoded genes (which were identified rapidly, by sequencing the mtDNA). The power of the approach was demonstrated in the 2 patients who shared the same complementation group. Following analysis by homozygosity mapping, microcell-mediated chromosome transfer, and transcriptome analysis, both patients were found to have mutations in NDUFS6, a complex I subunit gene residing on chromosome 5 to which no pathogenic mutation had been assigned previously.
As demonstrated by Kirby et al. (6), the integrated approach of 3-way complementation, biochemistry, genetics, microarrays, and bioinformatics, although cumbersome, offers a relatively straightforward path to deducing the molecular basis of recessive respiratory chain disorders (or any other recessive disorder, for that matter). As for dominant disorders, or those due to multifactorial causes (uncommon, but not unknown in the field of mitochondrial diseases), however, this approach will fail. Tennis, anyone?
Copyright © 2025 American Society for Clinical Investigation
ISSN: 0021-9738 (print), 1558-8238 (online)