Solution Conformation of Bovine Leukemia Virus Gag Suggests an Elongated Structure

Abstract

Bovine leukemia virus (BLV) is a deltaretrovirus that infects domestic cattle. The structural protein Gag, found in all retroviruses, is a polyprotein comprising three major functional domains: matrix (MA), capsid (CA), and nucleocapsid (NC). Previous studies have shown that both mature BLV MA and NC are able to bind to nucleic acids; however, the viral assembly process and packaging of viral genomic RNA requires full-length Gag to produce infectious particles. Compared to lentiviruses, little is known about the structure of the Gag polyprotein of deltaretroviruses. In this work, structural models of full-length BLV Gag and Gag lacking the MA domain were generated based on previous structural data of individual domains, homology modeling, and flexible fitting to SAXS data using molecular dynamics. The models were used in molecular dynamic simulations to determine the relative mobility of the protein backbone. Functional annealing assays revealed the role of MA in the nucleic acid chaperone activity of BLV Gag. Our results show that full-length BLV Gag has an elongated rod-shaped structure that is relatively rigid, with the exception of the linker between the MA and CA domains. Deletion of the MA domain maintains the elongated structure but alters the rate of BLV Gag-facilitated annealing of two complementary nucleic acids. These data are consistent with a role for the MA domain of retroviral Gag proteins in modulating nucleic acid binding and chaperone activity. IMPORTANCE: BLV is a retrovirus that is found worldwide in domestic cattle. Since BLV infection has serious implications for agriculture, and given its similarities to human retroviruses such as HTLV-1, the development of an effective treatment would have numerous benefits. The Gag polyprotein exists in all retroviruses and is a key player in viral assembly. However, the full-length structure of Gag from any virus has yet to be elucidated at high resolution. This study provides structural data for BLV Gag and could be a starting point for modeling Gag-small molecule interactions with the ultimate goal of developing of a new class of pharmaceuticals.

Keywords: Gag; matrix; molecular dynamics; retrovirus; small-angle X-ray scattering; viral assembly.

Figure 1.

a. Domain structure of BLV Gag and b. segments used to construct the full-length Gag model. The origin of each segment is listed in Table 1.

Figure 2.

Kratky plots for BLV Gag and CA-NC. Plots were generated from scattering curves using PRIMUS [76].

Figure 3.

P(r) distribution of BLV Gag and CA-NC. The P(r) data were generated using GNOM analysis software [77].

Figure 4.

SAXS-derived molecular envelopes for BLV Gag (a) and CA-NC (b). In Figure 4b, the CA-NC envelope is colored blue and fit to the Gag envelope using the volume fit function of UCSF Chimera [95].

Figure 5.

BLV Gag (a) and CA-NC (b) structures after undergoing flexible fitting to molecular envelopes calculated from SAXS data. Envelopes are transparent gray, and the models are shown as surface representations. The Gag structure is color-coded by domain: MA (blue), CA (orange), and NC (purple).

Figure 6.

BLV Gag (a) and CA-NC (b) represented as density maps docked to their respective SAXS envelopes. The density maps (blue for Gag and purple for CA-NC) were generated using the Fit in Map function of Chimera assuming a resolution of 15 Å, which is the resolution of a SAXS envelope generated in SITUS.

Figure 7.

Flexibility of BLV Gag based on 50 ns of molecular dynamics simulation. Trajectories were evaluated using the Turning function of the TimeScapes software package [92, 93]. More flexible regions are red and less flexible regions are blue. Flexibility is expressed using RMS fluctuations correlated with slow, global motion as described in the cited literature.

Figure 8.

Time-based annealing of mini-TAR RNA and mini-TAR DNA. Gag data were fit to a singleexponential function, and CA-NC data were fit to a double-exponential function. Error bars represent the standard error of the mean of at least three independent experiments.