As genes are turned into proteins, every three base pairs of DNA code for one amino acid. And because there are 64 possible strings of DNA triplets, and only 20 amino acids, some changes in DNA sequence, also known as single-nucleotide polymorphisms (SNP), do not lead to changes in protein sequence. Such mutations are known as silent, because in theory, they should not have an effect on protein function if the amino acid sequence is unchanged.

But as the old joke goes, apparently the proteins didn't read the textbook. In the Dec. 21, 2006 issue of Science, researchers report on a so-called silent mutation in the MDR1 gene that can affect cells' interaction with drugs as well as drug pump inhibitors.

"We were looking for reasons why cancer cells might be resistant to chemotherapy," senior author Michael Gottesman explained to BioWorld Today. The MDR1 gene codes for p-glycoprotein, which is something like the body's sump pump system: it lines the gastrointestinal tract, where it prevents multiple kinds of toxins from entering cells.

P-glycoprotein "probably evolved to protect the organism from toxic compounds in the diet," said Gottesman, who is head of the Laboratory of Cell Biology within the National Cancer Institute's Center for Cancer Research. An unfortunate consequence is that it also can protect cancer cells from chemotherapy toxins, contributing to their resistance to multiple chemotherapy drugs.

Gottesman and his team, who are at the National Cancer Institute in Bethesda, Md. and Seoul National University in South Korea, genetically engineered cultured cells to contain either normal MDR1 or a copy of the MDR1 with one or more SNPs. They then measured pump function with fluorescent dyes attached to several different molecules, including the cancer drugs paclitaxel, vinblastine and verapamil.

One particular SNP resulted in altered interactions with both cancer drugs and pump inhibitors - but only when it occurred together with one or two additional SNPs. Antibody binding studies suggested that the mutation combinations led to changes in the structure of P-glycoprotein.

How the change affects protein structure without affecting amino acid sequence still is a bit of a mystery. "The phenomenon is clear," Gottesman said. "The explanation for it is forthcoming."

He said he thinks the most likely reason is that the protein folding is altered by the change. "For big proteins, the protein is being folded as it is being translated," he said. Rare codons use rare transfer RNAs, and those rarer transfer RNAs take longer to find, essentially altering the rhythm of protein folding.

If his idea is correct, it would also explain why the silent SNP only had functional consequences when it was paired with one or two additional mutations that frequently occur along with it. "A single change is probably not enough, but once you've got two, it may slow down the process," Gottesman said.

The findings could be used to improve drug selection for cancer patients. "Only some drugs are affected by changes in the folding of this protein," Gottesman said. But he cautioned that "we would need a lot more information on which drugs are affected" for the approach to be practical.

Beyond the particular SNP, the findings affect the basic ideas of how SNPs relate to health and disease. Geneticists, Gottesman said, tend to ignore silent mutations when they are assessing the relative risks of disease conferred by SNPs. But silent mutations can have a profound effect on protein function: "To some extent, we do that at our peril."