A new HIV-1 Rev structure optimizes interaction with target RNA (RRE) for nuclear export

Abstract

HIV-1 Rev mediates the nuclear export of unspliced and partially-spliced viral transcripts for the production of progeny genomes and structural proteins. In this process, four (or more) copies of Rev assemble onto a highly-structured 351-nt region in such viral transcripts, the Rev response element (RRE). How this occurs is not known. The Rev assembly domain has a helical-hairpin structure which associates through three (A-A, B-B and C-C) interfaces. The RRE has the topology of an upper-case letter A, with the two known Rev binding sites mapping onto the legs of the A. We have determined a crystal structure for the Rev assembly domain at 2.25 Å resolution, without resort to either mutations or chaperones. It shows that B-B dimers adopt an arrangement reversed relative to that previously reported, and join through a C-C interface to form tetramers. The new subunit arrangement shows how four Rev molecules can assemble on the two sites on the RRE to form the specificity checkpoint, and how further copies add through A-A interactions. Residues at the C-C interface, specifically the Pro31-Trp45 axis, are a potential target for intervention.

Highly Mutable Linker Regions Regulate HIV-1 Rev Function and Stability

“…Recent structural studies implicate the turn residues as part of an additional Rev-Rev interaction surface where residues W45, P28, P29 and P31 “hook” with the same residues on an adjacent Rev molecule. This interface has been proposed to act as a molecular bridge between independent oligomerization events along the RRE [24, 25]. Most residues in the turn do not have strong preferences in the CDMS study (Figure 1B) [9], consistent with our understanding of the region as a simple linker.…”

“…Interestingly, the most common allele in patient sequences is Y28, although P28 is also substantially represented and is found in many lab-adapted HIV-1 strains [36]. It is possible that a P28Y substitution could provide stacking interactions to stabilize this aromatic interface within the turn [25], but we found no significant difference in activity between the proline and tyrosine variants (Figure S3). Even an alanine substitution at position 28 is only marginally weaker in replication assays, indicating substantial sequence plasticity, but leaving the manner in which Y28 confers a fitness advantage in competition experiments and in patients still unclear.…”

“…One very provocative feature of the Rev OD is that variation in the packing of its hydrophobic residues leads to multiple configurations of Rev dimers and oligomers [12, 24, 25]. This structural plasticity underlies the ability of Rev to assemble into a oligomeric complex on the RRE and raises the question of how constrained these side chains must be.…”

“…The ability of the turn-8GS mutant to export well in the reporter assay yet fail to replicate suggests that this linker region may perform additional roles in viral replication. Moreover, structural studies have determined that the turn can form a homotypic interaction interface[24, 25], and mutations expected to stabilize this interface (P28Y) are enriched in both our CDMS data and patient data[9], yet do not have a significant effect in either reporter or replication assays. Further characterization will be required to determine the exact, and potentially novel, role of this interface and, more generally the turn itself, in Rev function.…”

“…Given the established plasticity in RRE structure and the hydrophobic Rev interaction surfaces, we proposed earlier that these features allow Rev-RRE complexes with varying sequences to adopt configurations favorable for Crm1 recruitment and nuclear export [12]. Adding the requirement that the OD-NES linker needs to be structured (helical with the limitations imposed by two prolines in the NL4-3 sequence), we built a model of a Rev monomer containing residues 5-83 (Figure 8A), where 83 is the last residue of the NES, from the structures of Rev (5DHV) [25] and a Rev NES-Crm1 complex (3NBZ) [14] using Chimera [46]. Previous studies established a hexamer of Rev on a minimal, functional 240 nucleotide RRE [16].…”