Until the mid-20th century, pentosuria — a benign condition identified by the presence of five carbon sugars in a person's urine — was often misdiagnosed as diabetes, causing many people to needlessly take insulin. A study published in PNAS in October by the University of Washington's Mary-Claire King, Arno Motulsky, Sarah Pierce, and their collaborators describes two mutations in DCXR that cause pentosuria. Genome Technology's Christie Rizk spoke with King about the study. What follows is an excerpt of their conversation, edited for space.
Genome Technology: Why did you do this study?
Mary-Claire King: I was minding my own business, sequencing exomes when we first acquired our [Illumina] GAIIx, back in early 2010. I thought, 'Well, if we're going to sequence exomes, I better see what the data looks like.' So I picked eight people who had given us DNA, and ran the sequences, and studied them carefully to see what the exomes looked like. Lo and behold, there was a frameshift in one of these individuals in a gene I had never heard of, called DCXR.
I looked it up, and it was the gene predicted to encode cellulose reductase, and it is predicted that mutations in this gene should lead to the phenotype pentosuria as described by Garrod in his Inborn Errors of Metabolism in 1908. And I thought, 'That's odd. I would have thought that had been solved.' So I did two things: I looked up the ancestry of this particular participant, and she had indicated her ancestry was Jewish, which made sense as Garrod had described pentosuria in people of Jewish origin; and then we checked it with the Sanger sequences, and the frameshift was indeed there.
Of the four errors described by Garrod, the other three had long since been solved. But this one had not been, in the sense that no one had ever figured out what the mutations were. The reason was that the phenotype has no clinical effect now, and the reason the phenotype originally mattered — up until the middle of the 20th century — is that it was confused with diabetes.
I thought it would be interesting to solve this, even though there is a differential diagnosis for diabetes now, no one pays attention to pentosuria anymore.
GT: How did you find the second mutation?
MCK: When I saw my colleague Arno Motulsky, I told him all this, and he told me he had records and contact information for many families with pentosuria [given to him by Margaret Lasker, who studied pentosuria from the 1930s to the 1960s]. So Arno and Sarah Pierce and I started going through these records and … we found one family in Switzerland and 14 other families in the US. They said they wanted to get involved in the study again, so they sent us blood samples, and off we go extracting DNA, and lo and behold, we found that very mutation that I had found in my control person. It popped up as a homozygote in some, but not all, of the people who had sent us their blood and were alleged to be pentosuric. So Arno said, 'Maybe there's another mutation.' So we sequenced the entire gene in everybody, and we found it.
With the two mutations, all the 15 families were explained. People were either homozygous for one mutation, homozygous for the other mutation, or had one allele of each.
GT: Even though there are no clinical implications for this specific phenotype, does your method have implications for the discovery of other disease genes?
MCK: Yes, absolutely. I think what this does is vindicate the genomics-based gene discovery paradigm and what it demonstrates is that if you find a mutation and you have the capacity to identify people who have a phenotype, you can make the proof. It gives us a sense of confidence in the discovery of new genes.