To date, the clearest examples of genetically determined diseases are point mutations, in which a rare alteration produces a nonfunctioning or malfunctioning protein - cystic fibrosis and Huntington's disease are two well-known examples.

In the December 2004 issue of the American Journal of Human Genetics, researchers from the University of Chicago and the University of Oklahoma at Norman report data suggesting the opposite also can be true. In certain members of the cytochrome P450 gene family, it may be the nonfunctional protein that is advantageous and thus selected for.

The research focused on genetic variation in that gene family. At the most general level, "enzymes that act at the interface between organism and the environment are likely to be targets of natural selection," making them interesting to study from an evolutionary point of view, said Anna DiRienzo, associate professor of human genetics at the University of Chicago and senior author of the study.

The cytochrome P3A subfamily, in particular, is important in drug metabolism, since two family members, CYP3A4 and A5, metabolize more than half of today's marketed drugs. Like nearly all genes, those for cytochrome P450 come in different versions, or alleles. CYP3A5 alleles are unevenly distributed in the population - one particular allele that codes for a nonfunctional protein is found with much greater frequency among Caucasian and Asian-Americans than among African-Americans.

The scientists began with a resequencing study of the CYP genes. As opposed to a database study, in which researchers rely on already-collected sequence information, scientists conducting a resequencing study collect genetic samples from the population of interest and sequence each individual. It's labor-intensive, but it allows researchers to gain a more complete picture of genetic variation.

"In the tests we perform often all the information is in the rare variants," DiRienzo told BioWorld Today. It already was clear that the variance in coding sequence for the cytochrome genes did not explain the observed phenotypic variance. So the researchers decided to search for conserved noncoding regions, which are "the site of likely regulatory elements," DiRienzo said.

Comparison of rat, dog, monkey and human genomes revealed several conserved noncoding regions of the gene, and the researchers also used a bioinformatics approach to identify likely transcription factor-binding sites.

The researchers sequenced the CYP genes in about 75 individuals from three ethnic groups: European-Americans, African-Americans and ethnic Han (Chinese) people from Los Angeles. They then analyzed the number and distribution of haplotypes, which are blocks of linked genetic markers that tend to be inherited together. They found a group of haplotypes, all very similar, that extended over a long range of the chromosome. DiRienzo said that because they are indicative of DNA stretches that have been inherited by virtue of being near a beneficial gene, such long-range haplotypes are "a basic signature of positive selection."

The data show strong racial variation in haplotype frequency, a fact that squares with previously known racial differences in responses to a number of drugs that are metabolized by the cytochrome P450 family. But DiRienzo said that race is something of a red herring, and the more important correlation is with climate.

"The truth is that when we start looking carefully, there also is variation within ethnic groups," she said. "And it is correlated with latitude."

DiRienzo said that the differences might have arisen as a consequence of migration to colder and salt-richer climates. Under those conditions, a gene version that produces a nonfunctional protein protects its carriers from salt-sensitive hypertension. Selective pressures might have led to the rise of that nonfunctional variant in colder climates in which excessive salt loss through sweating is not a major concern.

Proxy' Analysis Suggests Climate Effect

The researchers demonstrated the correlation with climate by studying the distribution of a specific cytochrome allele in a larger population, and its correlation with what DiRienzo called a "proxy" for climate: distance from the equator. When they investigated distribution of the CYP3A5*3 allele, which codes for a nonfunctional protein, in about 1,000 individuals, there was a strong correlation between allele frequency and geographic latitude. The nonfunctional allele was present at frequencies of less than 10 percent in populations of sub-Saharan Africa, and more than 90 percent in parts of Europe and China. Allele frequency and latitude correlated significantly within Asian populations, as well, though not within European populations, which span less latitude overall.

In a final experiment aimed at demonstrating that the differences were not due to nonspecific migration effects, the researchers compared the distribution of the cytochrome P gene with two other types of genes: microsatellite DNA, which are short repeating DNA sequences that do not code for anything and thus are not subject to selective pressures, and the angiotensin gene, which also is known to influence salt and water retention. The hypertension-promoting version of the angiotensin gene also decreased with distance from the equator, suggesting the two genes were exposed to the same type of selective pressure. Microsatellite alleles did not show a similar geographic distribution, controlling for unspecific effects of migration.