Research Briefs

By Philip Cohen, PhD

Viral Defense: Use It or Lose It

Human proteins with the innate ability to fight off retroviruses are a hot topic since they could form the basis of novel therapeutic approaches. A recent report by Harmit Malik and Michael Emerman at the Fred Hutchinson Cancer Research Center and their colleagues suggests that at least one of these virus fighters is less powerful now than it may have been in the past (Curr. Biol. 16, 95, 2006).

These researchers were looking at naturally-occurring variation in the human gene for TRIM5a, a protein recently implicated in the species-specificity of retroviral resistance in primates (see Making a monkey out of HIV, IAVI Report 9, 3, 2005). Previously, this group analyzed the primate lineage and found that TRIM5a had undergone sporadic episodes of rapid evolution for at least 33 million years, consistent with selection imposed by the emergence of new retroviruses.

The new work focused on whether naturally-occurring variants of human TRIM5a contribute to a range of susceptibility to different retroviruses. The researchers analyzed genes from people representing 37 different regions around the world.

The analysis yielded 20 single nucleotide polymorphisms (SNPs) in the human population. Based on sequence conservation and structural prediction, three of these were predicted to possibly affect the function of TRIM5a. These candidates were then expressed in cell lines to test their ability to confer resistance to retroviruses. Only one, which changed a histidine to tyrosine at position 43 (H43Y), detectably altered protein function-for the worse.

In one assay, the researchers looked directly at human B-lymphocytes taken from four individuals with two, one, or no copies of H43Y TRIM5a. Cells from the person with no copies of H43Y were able to potently restrict N-MLV, a mouse retrovirus susceptible to the human protein. The two people with two copies of this variant restricted the virus about 100 times less potently. And a single copy of H43Y rendered B cells from the last person about 10-fold less potent.

This defective version of TRIM5a is not rare. The researchers found H43Y in 20% of chromosomes. They suggest that H43Y TRIM5a may have been able to flourish because it hasn't been necessary in recent human history to fight retroviruses, either exogenous ones or those endogenous in our chromosomes. They point out that the human genome sequence contains thousands of examples of endogenous retroviruses, all of which appear to be defective, and that humans only appear to be currently infected by two retroviruses, HTLV and HIV. It's also clear that humans have at least one other layer of innate protection against retroviruses, a protein called APOBEC3G (see Guardian of the genome, IAVI Report 9, 2, 2005), perhaps rendering TRIM5a redundant.

The authors also suggest that diminished pressure from retroviruses could make TRIM5a as much a risk as a benefit. The H43Y SNP lies in the protein's RING domain, which bestows on some proteins the ability to tag target proteins with ubiquitin that are then shuttled to the cell's proteasomal disposal pathway. So, occasionally, TRIM5a may inappropriately destroy the cell's own proteins, perhaps giving people who have eliminated this function a slight biochemical edge. The work suggests that any attempt to redirect innate antiviral machinery against HIV will need to account for the fact that these systems may not be optimized in a significant portion of the human population.

How Cells Get Hooked on HIV

HIV's gp120 core protein participates in many important stages of the viral replication cycle and host response including binding co-receptor, viral fusion with target cells and sensitivity to antibody neutralization. Now, for the first time, researchers have been able to capture a snapshot of one domain of gp120 that is important in all those roles: the third variable region, or V3 (Science 310, 1025, 2005).

The HIV envelope spike is composed of six protein molecules, three of gp120 and three of gp41. Through gp120 the virus attaches to the CD4 protein on a target cell's surface, shifts its structure to grab a coreceptor (either CCR5 or CXCR4), the binding of which then initiates a series of structural transformations in gp41 ending with the fusion of the virus and cell. Biochemical and genetic evidence suggests that V3 plays an important role in coreceptor binding. But presumably due to the flexibility of this region, V3-containing gp120 has proven difficult to crystallize, which has left its detailed structure unknown.

Peter Kwong at the Vaccine Research Center at the National Institute of Allergy and Infectious Diseases and his colleagues began their hunt for V3 crystals by using robots to screen many different combinations of proteins, chemistry, and time points. They screened gp120 from 3 different clade B HIV-1 isolates mixed with a shortened version of CD4 and 1 of 6 anti-gp120 antibodies to create 13 different complexes. Each of those complexes was dissolved in 576 different solutions and left to form crystals. Pictures of each crystallization reaction were taken at time points out to 21 days and visually inspected. Promising candidates for crystal formation were then optimized for growth of larger crystals and good X-ray diffraction characteristics.

The structure derived from those crystals shows the V3 protein chain forming a highly extended "hook" 50 angstroms long, 15 angstroms deep and only 5 angstroms wide. In free virus, the authors suggest, this hook could reach across to associate with the other proteins of the viral spikes, potentially playing a role in driving their interaction or structure. But it also appears to be flexible enough for every surface of it to be exposed to antibodies, explaining why V3 is a major focus of antibody response to HIV.

In gp120's CD4-bound conformation, V3 extends towards the target cell 30 angstroms; from the gp120 core, ready to grasp the appropriate coreceptor molecules. Once bound, the authors also suggest V3 could act as a "rip cord," linking coreceptor binding to deployment of the viral fusion program of gp41.

Microbicide Interferes in Viral Affairs

Vaginal microbicides could be powerful new HIV prevention tools—monoclonal antibodies, CCR5 inhibitors, and other less specific antiviral agents have been studied and clinical trials of some compounds are ongoing. Now researchers at Harvard Medical School led by Judy Lieberman and David Knipe have extended the microbicide concept to harness a hot new technology: RNA interference (Nature 439, 89, 2006).

RNA interference is mediated by small interfering RNA (siRNA) molecules, short (21-23 bp) double stranded RNA molecules that complex with proteins to form the RNA-induced silencing complex (RISC). When the siRNA within the RISC pairs precisely with its target RNA strand, cleavage occurs and the target RNA is destroyed. The acute specificity of siRNA sequences has encouraged researchers to evaluate their efficacy in silencing viral gene expression and thereby mitigating infection and disease. The limiting factor, as with most gene therapy approaches, has been the delivery of the siRNAs to their intended site of action.

To investigate delivery and uptake of siRNAs, the researchers used the transgenic green fluorescent protein (GFP) mouse that constitutively expresses this protein in all cells. SiRNAs were complexed with a transfection lipid to facilitate crossing cell membranes and instilled into the GFP mouse vagina. The siRNAs were efficiently taken up by the vaginal and ectocervical epithelium, as well as the underlying lamina propria and stroma, and genetic silencing of GFP persisted for at least nine days.

They then looked to see if topical siRNA application could protect against herpes simplex virus 2 (HSV-2), a sexually-transmitted infection that is lethal in the mouse. SiRNAs complementary to the UL27 and UL29 genes—which encode an envelope glycoprotein and a DNA-binding protein, respectively—were instilled intravaginally into mice 2 hours before and 4 hours after vaginal challenge with 2LD50 of HSV-2. Only 25% of mice given an irrelevant siRNA survived to day 15, whereas 75% of those given UL29-specific siRNA survived.

This protection was not due to inflammation or induction of interferon-responsive genes. Also, the researchers did not find any indication of escape from the siRNA sequences, but acknowledge that this could be a bigger concern for RNA viruses like HIV that have a higher mutation rate than with DNA viruses like HSV-2.

One of the most striking features of the study was the efficient uptake of and lasting silencing by siRNAs in the vaginal mucosal layer, an important consideration for any practical microbicide. Whether a similarly efficient uptake and longevity of effect will be seen in primates, and more specifically humans, remains to be determined. Another crucial question is whether siRNAs can be effective against HIV in infection models; previous HIV/siRNA research has only been done in cell culture.

The research team now intends to evaluate the concept on HIV infection in primates. They plan to target highly conserved viral genes and HIV's cellular coreceptor, CCR5, to see if down regulation of its expression augments any benefit of targeting HIV's own genes.

What Drives HIV Envelope Evolution?

One of the most frustrating problems that HIV researchers are confronting is the virus' ability to escape from neutralizing antibodies (NAbs). HIV has an intrinsically high mutation rate and the env gene encoding the envelope glycoprotein, gp160, that is the primary target for NAbs evolves at an especially high rate. Immediately following infection, env genetic diversity is low and then narrows further, increasing to a peak after several years. As a result env is extremely genetically diverse and poses one of the trickiest challenges to vaccine development.

A number of factors contribute to this genetic diversity but the relative importance of each and the precise mechanism remain to be determined. Selection for CXCR4 coreceptor usage can generate env diversity, but this doesn't happen until late in infection and involves only a limited number of amino acid residues. Escape from cellular immune responses also drives diversity in env, but cellular responses are usually stronger towards other HIV genes. Selection by NAbs, however, results in rapid, continuous evolution of viral escape at the phenotypic level, and so may be the most significant driving force.

Three mechanisms can contribute to this escape from neutralizing antibodies: point mutations, changes in glycosylation patterns, and insertions and deletions in the envelope. To explore the relative contribution of each, Douglas Richman at University of California at San Diego and his colleagues compared env genetic variation in a cohort of 13 recently HIV-infected men with different rates of escape from NAb responses (Proc. Natl . Acad. Sci. USA 102, 18514, 2005). They measured NAb responses in a virus assay that used recombinant virus particles containing patient virus envelope proteins plus an HIV genomic vector with a firefly luciferase indicator gene insert.

As in previous studies, HIV-specific NAb responses varied greatly within the cohort. The researchers used the rate at which patients' NAb responses decreased against successive autologous virus isolates as a measure of their rate of viral escape. After fitting data to a statistical model, they classified individuals as having high and low rates of viral escape. They then compared the patterns of env genetic variation in the virus from the two groups. There was no correlation between the rate of escape from NAbs and the rate of evolution of glycosylation sites, nor insertions and deletions. However there was good correlation between the rate of NAb escape and the rate of evolution of amino acid substitutions, consistent with previous observations in SIV-infected macaques.

This is contrary to previous in vitro work suggesting that a variable "glycan shield" due to mutations in glycosylation sites provides protection from NAb in recent infection (Nature 422, 307, 2003). The authors note that such NAb escape may require several substitutions at glycosylation sites and, since this will be reliant on the accumulation of single point mutations, this mechanism may be secondary.