|Home | About | Journals | Submit | Contact Us | Français|
The HIV-1 Rev response element (RRE) is a ~350 nucleotide, highly structured, cis-acting RNA element essential for viral replication. It is located in the env coding region of the viral genome and is extremely well conserved across different HIV-1 isolates. It is present on all partially spliced and unspliced viral mRNA transcripts, and serves as an RNA framework onto which multiple molecules of the viral protein Rev assemble. The Rev-RRE oligomeric complex mediates the export of these messages from the nucleus to the cytoplasm, where they are translated to produce essential viral proteins and/or packaged as genomes for new virions.
The RRE serves as a specific RNA scaffold that coordinates the assembly of a unique homo-oligomeric ribonucleoprotein (RNP) complex to mediate the nuclear export of essential, intron-containing, viral messages.
The small size of the HIV-1 genome necessitates the use of multiple reading frames and alternative splicing to encode all its viral proteins. Transcription from an integrated HIV-1 provirus generates a single 9 kilobase (kb) pre-mRNA containing multiple splice sites. In the early phase of the viral life cycle, this pre-mRNA is completely spliced to RRE-free, 2 kb messages that are exported to the cytoplasm via standard mRNA nuclear export pathways (Fig. 1A). One of these messages encodes the Rev protein, which can translocate to the nucleus.
The late phase of the viral life cycle is characterized by the expression of viral proteins from RRE-containing unspliced (9 kb) and partially spliced (4 kb) messages. Normally, these intron-containing RNAs are retained in the nucleus for splicing/degradation. Nuclear export of such RNAs is achieved by a specific, cooperative assembly of multiple Rev molecules on the RRE, followed by recruitment of the host Crm1/Ran-GTP nuclear export machinery via Rev’s nuclear export sequence (NES). In the cytoplasm, these messages are translated to produce the remaining viral proteins or packaged as genomes for newly budding virions (Fig. 1A) (for a detailed review, see ref. 52).
The RRE was initially identified as a cis-acting element in the env region (nucleotides 7709–8063 of the HXB2 reference strain) of the HIV-1 genome (Fig. 1B), required along with Rev for cytoplasmic accumulation of unspliced HIV-1 mRNAs, viral protein expression, and hence, successful viral replication.2-8 The minimal size of a functional RRE is ~250 nucleotides, although the entire structured region is ~350 nucleotides.9 The full RRE (nt 7709–8063) (Fig. 2) is extremely well conserved across subtypes of the major group M, with conservation pressure from both the RNA structure and the overlapping env protein reading frame.
Early computational predictions of RRE secondary structure indicated a complex, highly branched structure characterized by multiple stem-loops and bulges5,10 (Fig. 3). Chemical and enzymatic probing studies confirmed the secondary structure1,9,11 and in combination with mobility shift assays, established its importance in sequence-specific Rev-binding.1,12-16
Multiple molecules of Rev were shown to bind the RRE,1,17,18 forming a homo-oligomeric complex driven by a combination of hydrophobic Rev-Rev oligomerization interactions and Rev-RRE interactions.17,19
Truncation studies and footprinting experiments established stem IIB as the high-affinity site for Rev18,20-22 (Fig. 3) with binding mediated by a 17-residue arginine-rich motif (ARM) in Rev.23,24 Structural studies indicate an RNA A-form major groove widened by purine-purine base pairs to cradle the Rev-ARM α-helix with both base-specific contacts and electrostatic phosphate-backbone contacts cementing the interaction.25
Although the Rev-stem IIB interaction is an important determinant of Rev-RRE affinity and function, oligomerization of Rev on the RRE is critical for a functional complex that can mediate nuclear export. Studies from our lab have shown that the RRE structural scaffold dictates the formation of a highly cooperative oligomeric Rev-RRE complex with a 500-fold higher affinity than the Rev-IIB interaction.26 Reports from multiple groups have also demonstrated a direct correlation between cooperative Rev assembly on the RRE and export function.26-28
Recent crystal structures of Rev dimers29,30 reveal a hydrophobic core that mediates Rev-Rev interactions, with the ARMs pointing away from the core and available to contact the RNA. The structures, in combination with results from detailed biochemistry that indicate a stoichiometry of 6 Rev monomers on a ~250 nucleotide RRE,31 have led to the proposal of a “jellyfish” model for the Rev-RRE complex29 (Fig. 4). Formation of the complex presumably involves nucleation of Rev assembly at stem IIB followed by sequential addition of Rev monomers,9,27 as supported by kinetic studies of Rev-RRE assembly.32
Rev oligomerization and Rev-RNA interactions contribute synergistically toward formation of a high-affinity cooperative assembly that is directly correlated with its export competence. Studies with chimeric Rev and RRE (Rev fused to MS2 coat-protein and RRE stem IIB replaced with the MS2-RNA)33-35 and with polyvalent Rev-binding elements (e.g., stem IIB concatemers)36,37 have shown that although high affinity correlates with function, affinity alone is unable to recapitulate full export activity of the Rev-RRE complex.
These data suggest a model in which the RRE serves as an architectural scaffold that orchestrates the specific assembly of a Rev oligomer using a combination of affinities derived from RNA binding and oligomeric interactions. This complex in turns influences the structure, positioning, and stoichiometry of the larger Rev-RRE/Crm1/RanGTP export complex.
This model is reminiscent of other biological scaffolding strategies: scaffold proteins are central to organizing and coordinating signal transduction cascades in cells.38 Integral to the functioning of macromolecular machines such as the ribosome,39 spliceosome,40 telomerase41 and signal recognition particle42 are scaffold RNAs that provide the structural framework to position the different components and in some cases to organize their catalytic abilities.
The RRE is an example of a unique RNA scaffold, providing the framework for assembling a homo-oligomeric complex. The protein-binding sites it presents recruit multiple Rev molecules through diverse sets of interactions with specific positional and orientation requirements. Viral evolution has thus served as a selection experiment—identifying RNA-binding partners for Rev and arranging them structurally to derive maximal functional efficiency from such a complex, even under additional constraints imposed by an overlapping protein-coding reading frame. By deriving specificity from three dimensional restraints imposed by oligomer formation, the virus is able to maintain enhanced specificity of Rev for RRE over the pool of cellular RNAs without relying solely on high-affinity sequence recognition.
At least two significant observations support this model of viral evolution selection. Biochemical studies from our lab identified stem IA to be a secondary Rev-binding site26 with very different structural and thermodynamic modes of Rev recognition compared with stem IIB (Fig. 3), as might have been identified by an in vitro selection experiment. More evidence for this model comes from an intriguing study with a dominant negative Rev mutant, RevM10, that harbors a mutation in its nuclear export sequence (NES).43 RevM10 assembles on the RRE like wild-type Rev but fails to trigger nuclear export of RRE-containing RNAs. Viral resistance selection experiments resulted in two silent mutations in the RRE that altered the structure of RRE in stems III/IV/V with no detectable change in binding to Rev or RevM10.44
All complex retroviruses face the problem of exporting unspliced and partially spliced mRNAs and have employed the use of related RNA systems, although the scaffolding abilities of their RNAs is not well understood. Lentiviruses such as HIV-1 or SIV (simian immunodeficiency virus) use the Rev-RRE system although sequence conservation of Rev and RRE across different lentiviral species varies significantly.45 Other complex retroviruses such as some β retroviruses, and all δ retroviruses, encode similar systems: Rem/RmRE and Rex/RxRE systems, respectively.46,47 Although Rev proteins from many of these viruses have been shown to oligomerize along their respective RREs, the role of these oligomeric assemblies in coordinating larger complex formation (by comparison to the Rev-RRE-Crm1/RanGTP complex of HIV-1) has not yet been elucidated.45
Many simple retroviruses, most notably Mason-Pfizer monkey virus (MPMV), do not encode a Rev-like protein, but instead have evolved a cis-acting RNA element, the constitutive transport element (CTE),48 that directly binds to components of the host mRNA export machinery.49 The MPMV CTE is ~220 nucleotides and consists of two identical binding sites for the Nxf1 protein.48,50 Although the CTE can substitute for Rev-RRE in HIV-1, export is significantly diminished in these viruses. Wild-type like export activity can be restored by engineering tandem repeats of the CTE,50,51 suggesting that affinity in this case is enhanced through additional sequence-specific interactions rather than through a specific structural arrangement of dissimilar but cooperative binding sites, as for the Rev-RRE complex.
The RRE is an essential viral RNA element carefully selected by many evolutionary cycles to balance specificity and adaptability. Although stem IIB serves as a high affinity anchor point, binding to the secondary binding sites on the RRE need not be driven by high affinity for the nucleotide sequence, but rather by the architecture of the entire complex. From this viewpoint, the RRE can be seen as a scaffold directing the orientation and contacts of the similarly pliable RNA binding surface of Rev. In this way, the virus is able to encode a complex and specific mechanism to achieve nuclear export without the large evolutionary penalties of maintaining the sequence of multiple high affinity binding sites.
Sequences were retrieved from the HIV sequence database (www.hiv.lanl.gov) for all isolates aligned to nucleotides 7709–8063 of the HIV-1 HXB2 sequence. Non HIV-1 sequences were removed, and sequences were realigned over the RRE using Clustal X 2.1. Analyses were performed using Consensus Maker (www.hiv.lanl.gov) and JalView.
Previously published online: www.landesbioscience.com/journals/rnabiology/article/18178