As evidenced by the
featured story on the frontpage of my university's website, as well as a seminar I attended earlier today by Dr. Sara Sawyer, who's been in our Molecular Genetics and Microbiology section for all of 5 weeks. She did her PhD at Cornell and post-doc at the Hutchinson Cancer Research Center. Her faculty page reads:
Evolutionary change driven by historical viral epidemics has left a molecular “fossil record” in our DNA sequence. Our goal is to learn about natural strategies that have been successful at beating viruses in the past, and how these might be exploited in the fight against modern viral attacks. We are using a broad array of techniques from molecular evolution, virology, experimental evolution, and comparative genomics to look at human and primate genes that encode inhibitors of viral infection. We are also interested in a systems biology approach to explore how human genes change enough to avoid susceptibility to new viruses, yet still maintain their ability to perform other important cellular functions. We are interested in the inverse effects that evolutionary change can have on multiple, intertwined biological systems.
The title of her talk was "Tracking the Evolutionary Footprints of Viruses". Her talk was structured around 3 main points. The first was the origin of retroviruses in the human genome, the second was how this evolutionary record can be deduced from genomic analysis, and the third was how this inquiry can inform our treatment and possible prevention of retroviral disease (e.g. AIDS).
She began by saying that the human genome is ~8% endogenous retroviruses, which total about 506,000
fragments. She briefly explained how they integrate into the genome (reverse transcription), but since that's all probably old hat to most of you I'll leave it at that.
The model she works with is a simple evolutionary arms race between virus and antivirus genes, such that protein innovation is the selectively beneficial trait. In order to tell fast from slow innovation, she takes the ratio of synonymous to nonsynonymous substitutions, where values over 1 indicate positive selection (assuming synonymous changes are invisible to selection, which
isn't always the case).
From analysis of whole genome sequences of human and chimp we know that signature of adaptive evolution in our lineage are often found in genes involved in reproduction, immunity and environmental perception (e.g. olfaction). With respect to immunity and the topic of Dr. Sawyer's research, anti-HIV factors show high (greater than 1) dN/dS ratios. For example, TRIM5a targets retroviral capsids and halts the viral lifecycle (though it's inactive in humans). Obviously the protein (TRIM5a) and the accompanying retroviral capsule it latches to is a prime target for an evolutionary arms race.
So Sara decided to sequence TRIM5a from 30 primate species (apes, old and new world monkeys) and from these sequences and phyolgenetic analysis found ancient signatures of selection on TRIM5a which were decoupled from HIV/SIV (lentiviruses) infection. She then set out to investigate individual codons for evidence of positive selection, and out of 500, found five such amino acid positions, all of which clustered into a patch of 13 amino acids. Predicting that this region is the interacting domain of the TRIM5a protein, she deleted this region and found an increased susceptibility to HIV in Rhesus macaques. Plugging the Rhesus patch into human cells increased resistance to HIV 2 fold. It follows then that this domain in TRIM5a, poetically named B30.2, is the culprit. This is where viral recognition takes place (the V1 loop).
She said other papers published results soon after hers which supported her conclusion. One of those papers found that 1 amino acid site causes the majority of the susceptibility.
Deciding to see what this region looks like across human populations, she sequenced this region from 37 individuals located across the globe. For this gene, she found almost no variation in the last 3 exons including B30.2, which is the specificity domain. She said she was disappointed with this, but fixed changes do cluster in B30.2, which suggest multiple selective sweeps occurred here. No beneficial SNPs were found, and actually a deleterious (increased HIV susceptibility) allele was discovered. This H43Y genotype is actually common(found in 19% of the population), which seems counterintuitive. Surprisingly, she found 4% of Old World people possess this genotype, but an incredible 43% here in the Americas. She posits three explanations to account for this huge discrepancy. The first is that balancing selection is maintaining this because H43Y is superior at restricting other viruses. To her mind this seems unlikely. The second possibility is random genetic drift. Since only 2 exogenous retroviruses effect human populations currently (HIV and HTLV), perhaps there was no contemporary fitness cost? Obviously HIV itself would change that. She leans toward this explanation out of the three. The final explanation she offered was something about positive selection, where TRIM5a is costly due to "collateral damage". I didn't follow her very well here so I think I'm missing part of her explanation for that.
She closed with her Dual Selection model in which a host protein experiences selective pressure in 2 directions, both in combating the viral protein and in maintaining its original housekeeping function. She says this may explain the as yet unexplained correlation between genetic diseases (housekeeping end of things) and positive selection (virus end of things).
Further research for her concerns systematic evaluation of other candidate genes involved in HIV interaction (i.e., ones exhibiting high dN/dS ratios). There are 3 to 5 she wants to look at next, although none presently look as strongly selected as TRIM5.
Hopefully this restores some of your faith in Texas' science prowess.
UPDATE:
Pat Hardy has successfully
kept her seat. Hurray.
