Arenavirus Z-glycoprotein association requires Z myristoylation but not functional RING or late domains

Abstract

Generation of infectious arenavirus-like particles requires the virus RING finger Z protein and surface glycoprotein precursor (GPC) and the correct processing of GPC into GP1, GP2, and a stable signal peptide (SSP). Z is the driving force of arenavirus budding, whereas the GP complex (GPc), consisting of hetero-oligomers of SSP, GP1, and GP2, forms the viral envelope spikes that mediate receptor recognition and cell entry. Based on the roles played by Z and GP in the arenavirus life cycle, we hypothesized that Z and the GPc should interact in a manner required for virion formation. Here, using confocal microscopy and coimmunoprecipitation assays, we provide evidence for subcellular colocalization and biochemical interaction, respectively, of Z and the GPc. Our results from mutation-function analysis reveal that Z myristoylation, but not the Z late (L) or RING domain, is required for Z-GPc interaction. Moreover, Z interacted directly with SSP in the absence of other components of the GPc. We obtained similar results with Z and GPC from the prototypical arenavirus lymphocytic choriomeningitis virus and the hemorrhagic fever arenavirus Lassa fever virus.

FIG. 1.

Functional characterization of GP-Flag. (A) Surface localization of GP is unaltered by the presence of the C-terminal Flag tag in GP. 293T cells were transfected with plasmids containing LCMV (ARM) GP, LCMV (clone 13) GP, LFV (Josiah) GP, or their Flag-tagged counterparts, and surface GP was determined using FACS analysis. Fluorescence intensity is shown on the x axis, and relative cell counts on the y axis. (B) Murine leukemia virus (MLV) virions pseudotyped with arenavirus GP-Flag show infection efficiencies similar to those with untagged GP pseudotypes. The arenavirus GP constructs from panel A were used to prepare MLV pseudotypes (AVGP) and infect A549 cells following the schematic. Luciferase activity of the various pseudotypes is shown in the bar graph. Lanes: 1, LCMV ARM GP-Flag; 2, LCMV GP cl-13-Flag; 3, LFV GP-Flag; 4, LCMV ARM GP; 5, LFV GP; 6, VSV GP; 7, no GP.

FIG. 2.

Subcellular localization and biochemical association of Z and GP. (A) LFV Z and GP colocalize in transfected cells. 293T cells were transiently transfected with LFV Z-HA, GP-Flag, or both and subjected to indirect immunofluorescence. Panels A A, A B, and A C correspond to cells transfected with LFV GP-Flag only; while panels D, E, and F correspond to cells transfected with LFV Z-HA only. Z-HA or GP-Flag was detected using anti-HA (mouse) and anti-Flag (rabbit) antibodies simultaneously to assess cross-reactivity and background. Panels A G to A I and B G to I correspond to cells expressing both Z-HA and GP-Flag, which show colocalization of both proteins. (B) LCMV Z and GP colocalize in transfected cells. Panel B is arranged as in panel A. Results are representative of several independent experiments. (C) Expression of GP and Z proteins. 293T cells were transiently transfected to express Z-HA and GP-Flag, as indicated above the lanes, and total cell lysates were subjected to Western analysis. As a negative control, cells were transfected with empty vector. A section of the membrane stained with Ponceau-S is shown to reflect equal loading. It was noted that expression of LCMV Z-HA was less than that of LFV Z-HA despite the use of the same plasmid to express these proteins. (D) Z and GP biochemically associate. Cell lysates shown in panel C were immunoprecipitated using an anti-Flag antibody, and precipitated proteins were subjected to Western analysis. We could readily detect Z-HA from both LCMV and LFV in Flag immunoprecipitates containing the homologous GPs. (E) Z association with GP is specific. As a control to assess nonspecific interactions, cells were transfected with Z-HA and untagged GP and then subjected to anti-Flag immunoprecipitation. Lanes unrelated to this experiment were cropped. Z-HA was undetectable in immunoprecipitates from cells coexpressing untagged GP, while it was enriched in immunoprecipitates after coexpression with GP-Flag.

FIG. 3.

Biochemical and genetic interactions of Z and GP. (Ai) Expression of the LCMV Z-HA, the GP WT, the GP-D1 mutant, and Flag-tagged GPs in total cell lysates. The various GPs were detected using an anti-GP2 antibody, and Z was detected using an anti-HA antibody. A section of the membrane stained with Ponceau-S is shown to reflect equal loading. (Aii) The GP-D1 mutant associates biochemically with Z. Cell lysates shown in panel Ai were immunoprecipitated using an anti-Flag antibody, and the precipitated proteins were analyzed by Western blotting using an antibody to GP2. Z was visualized by Western blotting with anti-HA as shown in the panel, and bands of interest are indicated with arrows. As controls, cells were transfected with empty vector and with the WT GP or the GP-D1 mutant in the presence of Z-HA and then subjected to anti-Flag IP, which allows detection of nonspecific IP with the anti-Flag resin. To show equal loading, the IgG(L) fragment that stains clearly with Ponceau-S on the membrane is shown. (B) Z-mediated inhibition of the LCMV MG is prevented by the WT GP but not by the GP-D1 mutant. Cells (293T) were transfected with the indicated combination of plasmids, and 72-h cell lysates were prepared for CAT assays. CAT activities were normalized by assigning 100% to the activity determined in lysates from cells transfected with L + NP + T7RP + MG and after subtracting background levels of CAT activity detected in lysates of transfected cells that did not receive the plasmid encoding the L polymerase.

FIG. 4.

Myristoylation greatly impacts the level of Z-GP association. (A) The Z-G2A mutant shows altered subcellular localization. LCMV Z-G2A HA and GP-Flag localization was assessed in transfected cells by using indirect immunofluorescence. As a positive control, indirect immunofluorescence of cells transfected with Z-HA WT and GP-Flag is shown. Panels A A and A D correspond to signals for Z only, panels A B and A E correspond to signals for GP only, while panels A C and A F show a merged image of the Z, GP, and nuclear signals. (B) Expression of GP-Flag and Z-G2A HA proteins. Cells were transfected with empty vector as a negative control and with GP-Flag and Z WT as a positive control, and total cell lysates were subjected to Western analysis. The lanes are labeled above panel B according to the Z-+-GP combination expressed in the transfectants. Staining of the membrane with Ponceau-S and probing for beta-actin were done to assess loading. It was noted that more Z-G2A was detectable than Z WT in cells. (C) The Z-G2A myristoylation mutant is decreased in biochemical association to GP. Cell lysates shown in panel B were subjected to anti-Flag immunoprecipitation. As controls to assess nonspecific associations, cells were cotransfected with the Z WT or Z-G2A and untagged GP. The amount of Z-G2A detectable in immunoprecipitates containing GP-Flag was similar to that which was detectable in the presence of untagged GP. (D) Densitometry analysis of the coimmunoprecipitated Z-G2A shown in panel B. The amount of Z-G2A was normalized to both the amount of the WT Z and the amount of actin in total cell protein. Results are representative of at least two independent experiments.

FIG. 5.

Late and RING domains are not required for Z-GP association. (A) Subcellular localization of Z containing mutations in the late and RING domains. Cells were cotransfected with GP-Flag and various Z mutants, each containing a C-terminal HA tag, and subjected to indirect immunofluorescence of GP and Z. Images of a cell cotransfected to express Z WT and GP are shown as a control. We observed no change in the localization of the Z L domain mutant (Z-AAPA) or the Z-A36 mutant and their colocalization with GP; however, we observed a vesicular pattern in cells expressing the Z F32G35 mutant. (B) Expression of Z mutants in total cell protein. Transiently transfected 293T cells with the GP-Z combinations described in the legend to panel A were lysed, and proteins were detected by Western analysis using anti-Flag and anti-HA antibodies. To assess loading, the membrane was stained with Ponceau-S before probing for GP and Z and also probed for actin specifically. The amounts of GP expression appeared equal among the transfectants, while LCMV Z-AAPA and LCMV Z-A36 showed slightly more expression than LCMV Z WT. There were no differences in expression between LFV Z WT and LFV Z-AAPA-LTAL, but we consistently observed less Z-F32G35 in the βOG-soluble fraction of cells coexpressing GP. Densitometry analysis was used to normalize the expression of Z mutants to that of WT. (C) Late and RING domains are not required to mediate interaction of Z with GP. The cell lysates shown in panel B were subjected to Flag immunoprecipitation and Western blotting to detect Z WT strains and Z mutants. No differences were observed between LFV Z WT and LFV Z-AAPA-LTAL in Z-GP association, and the amounts of LCMV Z WT, LCMV Z-AAPA, and LCMV Z-A36 were consistent with levels in total cell protein. We observed more LCMV Z-F32G35 than LCMV Z WT in immunoprecipitates, however. (D) Densitometry analysis of Z L and RING domain mutants following Flag immunoprecipitation. A separate experiment containing duplicate transfections of the Z L and RING domain mutants was performed to assess the differences in association between the various Z mutants and GP shown in panel C. Following normalization of Z expression in whole-cell lysates, the amount of Z mutants present in Flag immunoprecipitates was compared to that of Z WT. This analysis revealed a modest increase in coimmunoprecipitation of Z-AAPA and Z-A36 with GP and a large increase in Z-F32G35 present in GP immunoprecipitates.

FIG. 6.

Subcellular localization and biochemical association between LCMV Z and SSP. (A) LCMV Z and SSP minimally colocalize. 293T cells expressing LCMV Z-Flag and SSP-HA in the indicated combinations were subjected to indirect immunofluorescence and confocal laser scanning microscopy. As described in the legend to Fig. 2, cells expressing Z-Flag or SSP-HA alone were probed simultaneously with anti-Flag or anti-HA antibodies to assess cross-reactivity and background. We observed modest levels of colocalization of Z and SSP in cotransfected cells, which were 25% and 23% of Z with SSP and the reverse, respectively, as determined with the LSM Image Examiner program to calculate the Pearson product-moment correlation coefficient for the respective signals of Z and SSP. (B) Expression of LCMV Z-Flag and SSP-HA in total cell protein. 293T cells were transiently transfected with LCMV Z-Flag and LCMV SSP-HA, and both proteins were detected in cells by Western analysis of whole-cell lysates. To observe loading, the membrane was stained with Ponceau-S before probing with antibodies, and the length of exposure time following enhanced chemiluminescence to detect SSP-HA is noted. (C) LCMV Z and SSP biochemically associate. Western analysis of anti-Flag immunoprecipitates was done for SSP-HA following coexpression with LCMV Z-Flag. As a control for specificity, anti-Flag immunoprecipitation was done following expression of SSP-HA alone in cells. Densitometry analysis of these exposures indicated that the specific co-IP signal from SSP-HA with LCMV Z-Flag was 9.7-fold higher than the condition where SSP-HA was expressed alone.