Structural studies of the deaminase domain of the human HIV-1 restriction protein APOBEC3G

Abstract

According to a report from the UNAIDS on the global AIDS epidemic in 2009, about 33.4 million people were living with the HIV virus, with a 2.7 million increase each year. HIV can lead to AIDS, which is a set of symptoms and infections resulting from the damage to the human immune system. However, due to the rapid mutability and productivity of the HIV, identifying treatments and therapeutic intervention remains challenging and has seen limited progress. In the past year, one novel innate defense against HIV infection was discovered, in which the human protein APOBEC3G (A3G) plays an important role. A3G was identified as a single-strand DNA deaminase that potently inhibited the replication of the HIV-1ΔVif virus. It produces the nonfunctional provirus by deaminating the cytosines to uracils on the minus-strand viral cDNA. Consequently, A3G can genetically inactivate HIV and recent studies have demonstrated that this activity is as potent as any current anti-retroviral drug. However, the exact model of this mechanism at atomic level has not yet been elucidated due to the low solubility of A3G which presents an obstacle for biochemical and structural studies. The research in this thesis took a structural approach to screen for a catalytically active and more soluble C-terminal deaminase of APOBEC3G (A3G-ctd) derivative, determine the NMR solution structure of this variant (A3G-2K3A) and characterize its interaction with single-strand DNA. Due to intrinsically low solubility of A3G-ctd, a strategy to design more soluble derivatives of the catalytic domain was performed prior to NMR structure determination. Two key methods: a solubility test and a E. coli-based mutation assay were used to test the solubility and catalytic activity of APOBEC3G variants. Deletion mutant analyses of APOBEC3G found the minimal catalytic region consisted of amino acids 198-384 (A3G198-384). Various alanine and lysine substitution variants based on this fragment were constructed and examined to screen for improved solubility and enhanced activity. One variant A3G198-384-2K3A (L234K-C243A-F310K-C321A-C356A) showed a significant improvement in both assays, and was purified as a monomer. The three-dimensional structure of A3G198-384-2K3A was then determined by triple resonance NMR spectroscopy. It consists of five β-strands that form a hydrophobic platform surrounded by five α-helices. Summarizing the DNA titration data, E. coli-based catalytic activity, conserved residues and computational modeling, the DNA binding mechanism of A3G was proposed in which a canyon formed by positively charged residues guides single-strand DNA binding and positions the target cytidine for deamination. Subsequently, a longer catalytic domain, A3G191-384-2K3A, was found to have higher activity than that of the A3G198-384-2K3A derivative. The longer domain has an additional α1-helix (residues 201-206) that was not observed in the shorter variant and part of the last α-helix (residues 191-194) of the N-terminal domain. The truncated model of the N-terminal domain was generated from the C-terminal NMR structures based on the sequence homology. Finally, a novel full-length A3G model was constructed by physically overlapping the α-helix (residues 191-194) of the N-terminal domain model and the C-terminal domain structure.