Functional Equivalence of Retroviral MA Domains in Facilitating Psi RNA Binding Specificity by Gag

Abstract

Retroviruses specifically package full-length, dimeric genomic RNA (gRNA) even in the presence of a vast excess of cellular RNA. The "psi" (Ψ) element within the 5'-untranslated region (5'UTR) of gRNA is critical for packaging through interaction with the nucleocapsid (NC) domain of Gag. However, in vitro Gag binding affinity for Ψ versus non-Ψ RNAs is not significantly different. Previous salt-titration binding assays revealed that human immunodeficiency virus type 1 (HIV-1) Gag bound to Ψ RNA with high specificity and relatively few charge interactions, whereas binding to non-Ψ RNA was less specific and involved more electrostatic interactions. The NC domain was critical for specific Ψ binding, but surprisingly, a Gag mutant lacking the matrix (MA) domain was less effective at discriminating Ψ from non-Ψ RNA. We now find that Rous sarcoma virus (RSV) Gag also effectively discriminates RSV Ψ from non-Ψ RNA in a MA-dependent manner. Interestingly, Gag chimeras, wherein the HIV-1 and RSV MA domains were swapped, maintained high binding specificity to cognate Ψ RNAs. Using Ψ RNA mutant constructs, determinants responsible for promoting high Gag binding specificity were identified in both systems. Taken together, these studies reveal the functional equivalence of HIV-1 and RSV MA domains in facilitating Ψ RNA selectivity by Gag, as well as Ψ elements that promote this selectivity.

Keywords: Gag; Psi RNA (Ψ RNA); Rous sarcoma virus (RSV); human immunodeficiency virus type 1 (HIV-1); matrix (MA); nucleocapsid (NC).

Figure 1

Rous sarcoma virus (RSV) RNA, and RSV and human immunodeficiency virus type 1 (HIV-1) protein constructs used in this work. (A) Predicted secondary structures of RSV MΨ (top) and RSV 167 (bottom) derived from nucleotides (nt) 156–315 and 1249–1409 of the RSV Prague C genome, respectively. Four mutant MΨ constructs are indicated by the boxed nt and arrows: UGCG to GAGA, GG to CC, A197G, and a triple mutation AGC to UCA. Additional nt at the 5′ and 3′ ends of each RNA that are not present in the RSV genome are shown in gray; (B) RSV and HIV-1 Gag constructs used in this work, with full-length wild-type Gag shown at the top of each set for comparison. Matrix (MA), capsid (CA) and nucleocapsid (NC), as well as various spacer peptides (SP, SP1, and SP2), the HIV-1 p6 domain, and the RSV p2, p10, and protease domain (PR) comprise these constructs. In the case of the two chimeric constructs, H132R and R155H, the residues at the junctions are explicitly shown. H132R contains the 132-residue HIV-1 MA in place of RSV MA in the context of RSV Gag, and R155H contains the 155-residue RSV MA in place of HIV-1 MA in the context of HIV-1 Gag.

Figure 2

(A) Salt dependence of RSV Gag∆PR, RSV CANC (construct lacking RSV MA, p2, p10, and PR), and RSV MA binding to RSV MΨ (solid curves) and RSV 167 (dashed curves). Data from panel A are re-graphed in log-log plots of the apparent binding affinity (K_d) as a function of NaCl concentration for RSV Gag∆PR (B); RSV CANC (C); and RSV MA (D); Bar graphs of K_d(1M) values (E) and Z_eff values (F) were determined by salt-titration assays using the indicated RSV proteins and RNAs. K_d(1M) values describe the nonelectrostatic component of binding and Z_eff values describe the electrostatic component of the protein–nucleic acid interactions [46]. Values of three trials performed in each case are shown with the height of the bar indicating the mean value.

Figure 3

Bar graphs showing K_d(1M) values (A) and Z_eff values (B) determined from salt-titration assays with RSV Gag∆PR, H132R, R155H, and HIV-1 Gag∆p6 with RSV MΨ, RSV 167, HIV-1 Ψ, and HIV-1 TARpolyA. Values of three or four trials performed in each case are shown with the height of the bar indicating the mean value.

Figure 4

Bar graphs of K_d(1M) values (A) and Z_eff values (B) determined from salt-titration assays with RSV Gag∆PR and RSV 167, WT MΨ, and MΨ RNA mutants. Values of three trials performed in each case are shown with the height of the bar indicating the mean value.

Figure 5

HIV-1 RNA constructs used in this work and salt-titration binding results with HIV-1 Gag∆p6. (A) Predicted secondary structures of HIV-1 Ψ (left) and HIV-1 TARpolyA (right) are shown. Eight mutant Ψ RNAs are indicated by the boxed nt and arrows (Mut1-Mut8). Additional 5’ G nt not present in the HIV-1 genome were added to facilitate in vitro transcription and are shown in gray; (B) The location of Mut 1, 2, and 6 mapped onto the all-atom model of HIV-1 Ψ previously determined from small angle x-ray scattering data and computational modeling [61]. Colors in (A) and (B) indicate mutations that had no effect on specificity (blue), and significant effects (red) based on salt-titration binding assays. Bar graphs of K_d(1M) values (C) and Z_eff values (D) determined from salt-titration assays with HIV-1 Gag∆p6 and HIV-1 TARpolyA, Ψ RNA, and Ψ RNA mutants. Values of three or four trials performed in each case are shown with the height of the bar indicating the mean value.