The human immunodeficiency virus, HIV-1, can infect, via entry, a variety of cells. The specific cell-type infected by a viral strain is referred to as HIV-1 tropism and can affect treatment and overall patient outcome. Cell-based methods for determining HIV-1 tropism are costly with a slow turnaround, and as a result there has been significant interest in developing computational tools for predicting HIV-1 tropism based on isolated DNA sequences encoding the virus.
As described in an article recently published in PLOS ONE, researchers led by Christodoulos A. Floudas, director of the Texas A&M Energy Institute and Erle Nye '59 Chair Professor for Engineering Excellence, have made a significant breakthrough in the computational prediction of HIV-1 tropism [1]. The research team responsible for this work includes Assistant Professor Phanourios Tamamis, postdoctoral research associate Dr. Chris A. Kieslich, and doctoral students Yannis Guzman and Melis Onel.
HIV-1 remains difficult to treat due to its high genetic variability. One approach that has been adopted to circumvent the difficulties of thwarting the quickly mutating virus is to target a specific class of human cell-bound proteins, referred to as coreceptors, which are hijacked by HIV-1 for cellular entry. The coreceptor used for cellular entry typically drives HIV-1 tropism, and is based on an interaction with a short segment of a HIV-1 envelope protein. Unlike existing tropism prediction methods, the method developed by Professor Floudas and his team accounts for the specific amino acid interactions between the human and viral proteins that drive coreceptor usage. As a result of this development, the new method provides significant improvements in accuracy in comparison to existing methods. This advance was enabled by computationally-derived complex structures for the interactions between the segment of HIV-1 envelope protein gp120 and human coreceptors, CCR5 [2] and CXCR4 [3], which were also developed by the Floudas group.
The team has also developed the web tool CRUSH (CoReceptor Usage prediction for HIV-1). The use of features based on the computationally-derived coreceptor complex structures allows CRUSH to be much simpler than existing methods in terms of computation and complexity. CRUSH accuracies are not anchored to a genetic reference sequence, and as a result, CRUSH provides highly accurate predictions across the major geographic and genotypic HIV-1 subtypes. The improved accuracy and generality of CRUSH makes it an ideal candidate for genotypic determination of HIV-1 tropism in clinical settings, which will be the focus of future work.
1. Kieslich CA, Tamamis P, Guzman YA, Onel M, Floudas CA. Highly Accurate Structure-Based Prediction of HIV-1 Coreceptor Usage Suggests Intermolecular Interactions Driving Tropism. PLoS ONE. Public Library of Science; 2016;11: e0148974. doi:10.1371/journal.pone.0148974
2. Tamamis P, Floudas CA. Molecular Recognition of CCR5 by an HIV-1 gp120 V3 Loop. PLoS ONE. 2014;9. doi:10.1371/journal.pone.0095767
3. Tamamis P, Floudas CA. Molecular Recognition of CXCR4 by a Dual Tropic HIV-1 gp120 V3 Loop. Biophysical Journal. 2013;105: 1502–1514. doi:10.1016/j.bpj.2013.07.049