Retroviral RNA Dimerization: From Structure to Functions

Abstract

The genome of the retroviruses is a dimer composed by two homologous copies of genomic RNA (gRNA) molecules of positive polarity. The dimerization process allows two gRNA molecules to be non-covalently linked together through intermolecular base-pairing. This step is critical for the viral life cycle and is highly conserved among retroviruses with the exception of spumaretroviruses. Furthermore, packaging of two gRNA copies into viral particles presents an important evolutionary advantage for immune system evasion and drug resistance. Recent studies reported RNA switches models regulating not only gRNA dimerization, but also translation and packaging, and a spatio-temporal characterization of viral gRNA dimerization within cells are now at hand. This review summarizes our current understanding on the structural features of the dimerization signals for a variety of retroviruses (HIVs, MLV, RSV, BLV, MMTV, MPMV…), the mechanisms of RNA dimer formation and functional implications in the retroviral cycle.

Keywords: HIV; MuLV; RNA; dimerization; function; retrovirus; structure.

FIGURE 1

Schematic overview of the role of RNA dimerization in the retroviral life cycle. The cycle begins with the entry of the retrovirus within the target cell, followed by reverse transcription of the RNA genome into cDNA. During this step, gRNA dimerization plays an important role since RT may switch between strands, thus allowing genome repair and/or shuffling. The pre-integration complex (PIC) is then translocated into the nucleus where it is integrated in the genome of the target cell. The unspliced mRNAs are transcribed by the host machinery from the integrated provirus and transported to the cytoplasm. There, the single 5′ capped mRNAs serve as genomic RNAs that dimerize and are subsequently selected and packaged into the nascent virions, while mRNAs beginning with two or three guanosine are translated by the host machinery (Kharytonchyk et al., 2016). After budding, immature particles follow a maturation step initiated by the viral protease to produce infectious virions.

FIGURE 2

Motifs involved in the RNA dimerization of alpha-retrovirus. (A) Schematic representation of the 5′-end of alpha retrovirus gRNA. The functional domains and their positions are represented: R, repeated region; U5, unique sequence in 5′; PBS, primer binding site; AUG, gag translation initiation codon; SD, splice donor site; L3 and DLS, dimerization motifs for ALV and RSV, respectively. (B) Predicted secondary structure of the ALV L3 stem-loop. The consensus nucleotides are represented, with the palindromic hexanucleotide sequence highlighted in red. (C) Proposed kissing-loop complex and extended duplex conformations of ALV L3 element.

FIGURE 3

Secondary structure of the 5′-end of MMTV and MPMV genomic RNAs and dimerization models. (A) MMTV gRNA secondary structure, the different palindromic sequence (pal I, II, III, and PBS-pal) and the long-range interaction (LRI) between U5 and the beginning of gag are indicated. The different stem-loops (SL) are numbered as proposed by (Aktar et al., 2014). SL4 comprises both pal II which is the proposed DIS, and the major SD site. The R and PBS regions are also represented. (B) MMTV pal II switch models from kissing-loop complex to extended duplex. The nucleotide positions are represented and the hexanucleotide palindrome is highlighted in red. (C) MPMV gRNA secondary structure, the different stem-loops (SLs) are numbered as proposed by (Aktar et al., 2013), the LRI between U5 and the beginning of gag is also represented. The palindromic sequence folded in a short hairpin (Pal SL) which is the proposed DIS is highlighted in red. (D) MPMV Pal SL switch models from the kissing-loop complex to the extended duplex. The nucleotide positions are represented and the hexanucleotide palindromic sequence is highlighted in red.

FIGURE 4

Secondary structures and dimerization of MuLV genomic RNA. (A) Secondary structure of the 5′-end of MuLV gRNA containing the four stem-loops SL-A to SL-D. The nucleotide positions are indicated. (B) In vitro model of the kissing-loop/loose dimer complex. SL-A and SL-B contain a palindromic sequence of 10 and 16-nts long, respectively (in red), that promote the initiation of RNA dimerization. SL-C (purple) and SL-D (blue) both present a GACG tetraloop involved in heterologous non-canonical loop-loop interactions that stabilize the duplex. In this conformation, several UCUG quartets (green), which constitute high affinity binding sites for the viral NC protein, are trapped within SL-A and SL-B. (C) Secondary structure model of gRNA dimer maturation steps derived from SHAPE data (Grohman et al., 2014): the extracted ex viro immature form (left) is converted by NC into the mature 5′-end RNA dimer (right). In the immature conformation, only SL-A and SL-A′ are paired and adopt an extended conformation while the SL-B elements are unpaired. This conformation exposes the UCUG quartets. In the mature conformation, both SL-A/SL-A′ and SL-B/SL-B′ loop–loop interactions adopt an extended duplex structure, thus increasing dimer stability (Gherghe et al., 2010). This model also exposes the UCUG quartets. The mature dimer is thought to be to be similar to in vitro 5′-end tight dimer structure.

FIGURE 5

HTLV-I DIS model. Contrary to what is observed for other retroviruses, the DIS of HTLV-I presents a peculiar secondary structure harboring a single A residue in the apical loop (Monie et al., 2004). The palindrome is highlighted in red and the nucleotide positions are indicated. The proposed dimer interaction is also represented (Monie et al., 2001).

FIGURE 6

Mechanism of HIV-1 RNA dimerization. (A) Schematic secondary structure model of the 5′-end region of the HIV-1 gRNA. TAR, trans-activation response element; Poly-A, stem-loop containing the 5′-copy of the polyadenylation signal in the apical loop; U5, unique in 5′; PBS, primer binding site; DIS, dimerization initiation site; SL1–4: stem-loops 1–4 containing the Dimerization Initiation Site (DIS), the major Splice Donor (SD) site, the historical packaging signal, and the gag AUG initiation codon, respectively. The packaging signal (Psi) region, spanning SL1-SL4 elements, is indicated. (B) Model of the SL1 switch from the kissing-loop complex to the extended duplex conformation. The nucleotide positions are indicated. (C) Solution structures of SL1 23-mer kissing-loop complex (KC) (up) (Kieken et al., 2006) and of SL1 35-mer extended duplex (ED) (down) (Ulyanov et al., 2006) as determined by NMR. The DIS palindromic sequences are highlighted in red, and the purines flanking the DIS are in orange. Structures were drawn using the coordinates deposited on the PDB (PDB ID: 2F4X – KC – and 2GM0 – ED). (D) X-ray crystal structures of SL1 23-mer both in KC (up) (Ennifar et al., 2001) and ED (down) (Ennifar et al., 1999) conformations. The DIS palindromic sequences are highlighted in red. One Mg²⁺ ion found in the ED crystal structure exposing the highly conserved purines flanking the DIS (in orange) is drawn in purple. Structures were drawn using the coordinates deposited on the PDB (PDB ID: 2B8R – KC – and 2F4X – ED).

FIGURE 7

Secondary structure models of HIV-1 genomic RNA. (A) HIV-1 long-distance interaction model (LDI) as proposed by Huthoff and Berkhout (2001), in which SL1 DIS is base-paired together with the poly-A element. (B) HIV-1 U5:DIS model proposed by Lu et al. (2011), in which SL1 DIS is base-paired with the U5 region. This model was proposed to promote the translation of unspliced gRNAs by repressing dimerization (Lu et al., 2011). (C,D) Dimerization-competent structural models of HIV-1 5′-end. (C) The branched multiple hairpin (BMH)/U5:AUG models (Huthoff and Berkhout, 2001; Lu et al., 2011) in which SL1 DIS is exposed while U5 base-pairs with the region overlapping gag translation initiation codon, which promotes dimerization and was proposed to repress translation (Lu et al., 2011). (D) Keane et al. (2015) recently proposed a putative three-way junction structure of the extended duplex conformation in which the whole region downstream of SL1 is exchanged. This conformation was proposed to be achieved through NC chaperone activity.

FIGURE 8

Secondary structure models of HIV-2 genomic RNA. (A) Schematic representation of the secondary structure model of the 5′- region of the HIV-2 gRNA. The red line delimits the packaging signal (Psi) region containing the five stem-loops SL1, Ψ1 to Ψ3 and SD. Similar to HIV-1, SL1 contains in its apical loop a hexanucleotide palindromic sequence constituting the DIS. SL1 also contains in its basal part a 10-nts palindromic sequence partially entrapped (PAL). The gag translation initiation codon is located within a G-rich region (G-box). Upstream of Psi, are found the TAR, the poly-A, a C-rich region important for RNA dimerization (C-box) and the PBS. (B) In the CGI dimer structural conformation, the C-box and G-Box are base-paired. This conformation restricts HIV-2 RNA dimerization and is adopted by loose dimers when the DIS and TAR hairpin III are involved in kissing-loop interactions (Lanchy et al., 2003a,b). (C) The RNA switch model from loose to tight dimer, as proposed by Purzycka et al. (2011). HIV-2 SL1 stem B is melted through Gag/NC chaperone activity, freeing the PAL region that can thus forms additional intermolecular base-pairings to stabilize the RNA dimer.