YRKL sequence of influenza virus M1 functions as the L domain motif and interacts with VPS28 and Cdc42

Abstract

Earlier studies have shown that the C-terminal half of helix 6 (H6) of the influenza A virus matrix protein (M1) containing the YRKL sequence is involved in virus budding (E. K.-W. Hui, S. Barman, T. Y. Yang, and D. P. Nayak, J. Virol. 77:7078-7092, 2003). In this report, we show that the YRKL sequence is the L domain motif of influenza virus. Like other L domains, YRKL can be inserted at different locations on the mutant M1 protein and can restore virus budding in a position-independent manner. Although YRKL is a part of the nuclear localization signal (NLS), the function of YRKL was independent of the NLS activity and the NLS function of M1 was not required for influenza virus replication. Some mutations in YRKL and the adjacent region caused a reduction in the virus titer by blocking virus release, and some affected virus morphology, producing elongated particles. Coimmunoprecipitation and Western blotting analyses showed that VPS28, a component of the ESCRT-I complex, and Cdc42, a member of the Rho family GTP-binding proteins, interacted with the M1 protein via the YRKL motif. In addition, depletion of VPS28 and Cdc42 by small interfering RNA resulted in reduction of influenza virus production. Moreover, overexpression of dominant-negative Cdc42 inhibited influenza virus replication, whereas a constitutively active Cdc42 mutant enhanced virus production in infected cells. These results indicated that VPS28, a component of ESCRT-I, and Cdc42, a small G protein, are associated with the M1 protein and involved in the influenza virus life cycle.

FIG. 1.

Schematic representation of the influenza A virus M1 protein structure with functional domains. The X-ray crystallographic studies of the M1 protein showed the α-helices (marked as H1 to H9) and loops (marked as L1 to L8). The protein is divided into the N domain (H1 to H4), M domain (H6 to H9), and C domain. The sequence of the H6 domain (residues 91 to 105) is shown. The positively charged amino acids are shown in gray. The structure of the C domain is still unknown (shown as an open bar). The N terminus (residues 1 to 164) possesses the membrane interaction domain; residues 91 to 116 and 165 to 252 are the RNP binding domains; residues 177 to 202 contain the RNA binding domain; residues 1 to 20, 45 to 70, and 107 to 150 are hydrophobic domains (HD); residues 90 to 108 and 129 to 164 are transcription inhibition domains (TID); 100-YRKL-103 is the proposed L domain; 101-RKLKR-105 is the NLS and NEP binding motif; and 148-CATCEQIADSQHRSH-162 is the putative CCHH zinc finger motif (Zn). The potential casein kinase 2 and protein kinase C phosphorylation sites (circled Ps) are Thr5, Thr37, Ser70, Ser161, and Thr185. PE is the protease cleavage site. See the references in our previous paper (35). The ovals represent helices; —, loop.

FIG. 2.

Expression of M1 mutant proteins. cDNA-transfected (Trans) 293T cells were pulse-labeled at 18 h p.t. for 2 h. The cells were then lysed in RIPA buffer, immunoprecipitated with anti-M1 antibody, and resolved by SDS-PAGE.

FIG. 3.

Budding of virus particles by thin-section electron microscopy. MDCK cells grown on a polycarbonate filter were infected with either WT or mutant viruses at an MOI of 3.0. At 12 h p.i., the infected cell monolayers on filters were cross-linked, postfixed, and embedded (8). Sixty-nanometer-thick sections were stained and examined. WT (A), R105E (B), R101A (C), 4A:YRKL^160-161 (D), and mock infection (E). →, elongated virus particles; ⇒, villi.

FIG. 4.

Effect of insertion of YRKL and adjacent sequences on virus rescue. Larger peptides containing YRKL and adjacent sequences were inserted in the H6 (A) or L3 (B) region of M1. The insertion mutants are named for the amino acid residues inserted followed by the position of the insertion site. The positively charged amino acids are shown in gray. The virus titer (PFU/ml) indicates virus collected at 48 h p.i. A minus sign indicates that recombinant virus could not be rescued. On the left is shown the expression of WT and mutant M1 in cDNA-transfected cells (for details, see the legend to Fig. 2).

FIG. 5.

M1 and NP contents of infected cell lysates and purified virions. MDCK cells infected with WT or mutant viruses were labeled from 4 to 16 h p.i. with 250 μCi of ³⁵S protein-labeling mixture. WT, 4A, and R101A/K102A cDNAs (plus seven-plasmid)-transfected 293T cells were labeled at 4 h p.t. for 56 h with 250 μCi/ml of ³⁵S protein-labeling mixture. The labeled cell lysates and purified viruses from the media were analyzed as described in Materials and Methods. Loading in gels reflected 1× cell lysate and 10× medium. The positions of the ¹⁴C molecular mass standards are indicated on the left, and lane numbers are shown below.

FIG. 6.

Intracellular distribution of M1 in cDNA-transfected 293T cells. 293T cells were transfected individually with WT or mutant M1 cDNAs. At 9 h p.t., the cells were fixed and examined for M1 expression by immunofluorescence staining with anti-M1 antibodies. Representative fields from analysis of 20 fields in triplicate experiments are shown. Magnification, ×558.

FIG. 7.

VPS28 interacts with M1 in influenza virus-infected cells. (A) Coimmunoprecipitation of influenza virus-infected cell extracts using antibodies against a number of L-domain-interacting cellular proteins. 549A cells were infected with WT virus (MOI = 5), pulse-labeled at 4 h p.i. (2 mCi in 15 ml for 3 h), and lysed. The cell lysates were coimmunoprecipitated in CoIP lysis buffer as described in Material and Methods with different antibodies (shown above each lane) and resolved by SDS-PAGE (12%). Representative data from four independent experiments are shown. The positions of the ¹⁴C molecular mass standards are indicated on the left, and lane numbers are shown below. No antibody (Mock) and Ig represent negative controls. (B) Immunoprecipitation-Western blot analysis of interaction between M1 and VPS28. A549 cells were infected with WT virus and cell lysates at 7 h p.i., clarified, coimmunoprecipitated in CoIP lysis buffer with different antibodies (shown above each lane), and resolved by SDS-PAGE (15%). The proteins were then transferred to an NC membrane, probed with anti-VPS28 antibody (WB: αVPS28), and detected by enhanced chemiluminescence. Representative data from two independent experiments are shown. The positions of the prestained molecular mass standards are indicated on the right, and lane numbers are shown below. (C) The YRKL motif of M1 interacts with VPS28 but not with TSG101. A549 cells were infected with WT (lanes 2 to 5), PTAP (35) (lanes 6 to 9), and 4A:YRKL^160-161 influenza virus (lanes 14 to 17) and labeled as in panel A. 293T cells were transfected with 4A cDNA plus seven plasmids (lanes 10 to 13) and labeled for 2 h at 12 h p.t. The cells were then lysed in CoIP lysis buffer and used for CoIP assays using different antibodies as described above (panel A). Ig represents the negative control. Representative data from two separate experiments are shown. (D) 293T cells were transfected with siRNAs for Luc, TSG101, and VPS28 (25 nM) twice at an interval of 24 h. A Western blot of the VPS28 protein is shown (inset). At 48 h p.t., siRNA-transfected or mock-transfected cells were infected with WT virus (MOI = 0.1). At 16 h p.i., the media were collected and assayed for PFU titers. The data represent the mean titer ± standard deviation (n = 5). *, P < 0.05 (versus Mock). (E) Immunoprecipitation of M1 and NP of cell lysates and released virus particles from siRNA-transfected 293T cells infected with WT virus. Luc, TSG101, or VPS28 siRNA-transfected 293T cells (see the legend to panel D) were infected (MOI = 0.1) and labeled at 4 h p.t. for 18 h with 250 μCi of ³⁵S protein-labeling mixture. Cell lysates and virus particles from the extracellular medium were immunoprecipitated by using anti-NP and anti-M1 antibodies as described in Materials and Methods. The loading of medium was 10-fold more. The virus release efficiency (percent release relative to the control) from two experiments was calculated as for Table 3. *, P < 0.05 (versus Luc).

FIG. 7.

VPS28 interacts with M1 in influenza virus-infected cells. (A) Coimmunoprecipitation of influenza virus-infected cell extracts using antibodies against a number of L-domain-interacting cellular proteins. 549A cells were infected with WT virus (MOI = 5), pulse-labeled at 4 h p.i. (2 mCi in 15 ml for 3 h), and lysed. The cell lysates were coimmunoprecipitated in CoIP lysis buffer as described in Material and Methods with different antibodies (shown above each lane) and resolved by SDS-PAGE (12%). Representative data from four independent experiments are shown. The positions of the ¹⁴C molecular mass standards are indicated on the left, and lane numbers are shown below. No antibody (Mock) and Ig represent negative controls. (B) Immunoprecipitation-Western blot analysis of interaction between M1 and VPS28. A549 cells were infected with WT virus and cell lysates at 7 h p.i., clarified, coimmunoprecipitated in CoIP lysis buffer with different antibodies (shown above each lane), and resolved by SDS-PAGE (15%). The proteins were then transferred to an NC membrane, probed with anti-VPS28 antibody (WB: αVPS28), and detected by enhanced chemiluminescence. Representative data from two independent experiments are shown. The positions of the prestained molecular mass standards are indicated on the right, and lane numbers are shown below. (C) The YRKL motif of M1 interacts with VPS28 but not with TSG101. A549 cells were infected with WT (lanes 2 to 5), PTAP (35) (lanes 6 to 9), and 4A:YRKL^160-161 influenza virus (lanes 14 to 17) and labeled as in panel A. 293T cells were transfected with 4A cDNA plus seven plasmids (lanes 10 to 13) and labeled for 2 h at 12 h p.t. The cells were then lysed in CoIP lysis buffer and used for CoIP assays using different antibodies as described above (panel A). Ig represents the negative control. Representative data from two separate experiments are shown. (D) 293T cells were transfected with siRNAs for Luc, TSG101, and VPS28 (25 nM) twice at an interval of 24 h. A Western blot of the VPS28 protein is shown (inset). At 48 h p.t., siRNA-transfected or mock-transfected cells were infected with WT virus (MOI = 0.1). At 16 h p.i., the media were collected and assayed for PFU titers. The data represent the mean titer ± standard deviation (n = 5). *, P < 0.05 (versus Mock). (E) Immunoprecipitation of M1 and NP of cell lysates and released virus particles from siRNA-transfected 293T cells infected with WT virus. Luc, TSG101, or VPS28 siRNA-transfected 293T cells (see the legend to panel D) were infected (MOI = 0.1) and labeled at 4 h p.t. for 18 h with 250 μCi of ³⁵S protein-labeling mixture. Cell lysates and virus particles from the extracellular medium were immunoprecipitated by using anti-NP and anti-M1 antibodies as described in Materials and Methods. The loading of medium was 10-fold more. The virus release efficiency (percent release relative to the control) from two experiments was calculated as for Table 3. *, P < 0.05 (versus Luc).

FIG. 8.

Coimmunoprecipitation of M1 and Cdc42. (A) ³⁵S-labeled lysates of influenza virus-infected cells were coimmunoprecipitated in CoIP lysis buffer using different G-protein antibodies and analyzed by SDS-PAGE. For details, see the legend to Fig. 7A. (B) Immunoprecipitation-Western blot analysis of interaction between M1 and Cdc42. An NC membrane was probed with anti-Cdc42 antibody (WB: αCdc42). For details, see the legend to Fig. 7B. (C) A549 cells were infected with WT virus (lanes 2 to 4) and 4A:YRKL^160-161 (lanes 8 to 10) and then labeled for 3 h at 4 h p.i. (see the legend to Fig. 7A). 293T cells were transfected with 4A cDNA (lanes 5 to 7) and labeled for 2 h at 12 h p.t. The cells were then lysed in CoIP lysis buffer, and the cell lysates were coimmunoprecipitated with Ig (negative control) or anti-M1 or anti-Cdc42 antibody as indicated. Representative data from two experiments are shown.

FIG. 9.

Cdc42 is involved in influenza virus replication. (A) The siRNAs for Cdc42 and Luc (25 nM) were transfected into 293T cells twice at an interval of 24 h. A Western blot of the Cdc42 protein is shown (inset). At 48 h p.t., the siRNA-transfected or mock-transfected cells were infected with WT virus (MOI = 0.1). At 16 h p.i., the media were collected and assayed for virus titer by plaque assay. The data represent the mean titer ± standard deviation (n = 5). *, P < 0.05 (versus Luc). (B) Immunoprecipitation of M1 and NP from 293T cell lysates and released virus particles after siRNA transfection and virus infection. Mock, Luc, and Cdc42 siRNA-transfected 293T cells (see the legend to panel A) were infected with WT virus (MOI = 0.1) and labeled at 4 h p.i. for 18 h with 250 μCi/ml of ³⁵S protein-labeling mixture. The cell lysates and released virus particles were immunoprecipitated for NP and M1 (for details, see the legend to Fig. 5). The virus release efficiency relative to the control is shown (bottom). *, P < 0.001 (versus Luc). (C) 293T cells were transfected with plasmids (4 μg) encoding HA-tagged WT Cdc42, dominant-negative HA-tagged Cdc42 (Cdc42-T17N), or constitutively active HA-tagged Cdc42 (Cdc42-G12V). The proteins were radiolabeled for 48 h and immunoprecipitated with anti-HA epitope antibody (inset). These Cdc42-expressing 293T cells were infected with WT virus (MOI = 0.1) at 48 h p.t., and the media were then collected at 16 h p.i. and assayed for PFU titers. The data represent the mean titer ± standard deviation (n = 3). *, P < 0.05; **, P < 0.001 (versus Cdc42). (D) Immunoprecipitation of M1 and NP from 293T cell lysates and released virions after cDNA transfection and virus infection. Mock, Cdc42, Cdc42-T17N, and Cdc42-G12V cDNA-transfected 293T cells (see the legend to panel C) were infected with WT virus (MOI = 0.1) and labeled at 4 h p.i. for 18 h with 250 μCi/ml of ³⁵S protein-labeling mixture. The cell lysates and extracellular medium were used for immunoprecipitation (for details, see the legend to Fig. 5). Representative data from two experiments are shown. Relative virus releases were normalized to intracellular expression levels (bottom). *, P < 0.05; **, P < 0.001 (versus Luc).

FIG. 9.

Cdc42 is involved in influenza virus replication. (A) The siRNAs for Cdc42 and Luc (25 nM) were transfected into 293T cells twice at an interval of 24 h. A Western blot of the Cdc42 protein is shown (inset). At 48 h p.t., the siRNA-transfected or mock-transfected cells were infected with WT virus (MOI = 0.1). At 16 h p.i., the media were collected and assayed for virus titer by plaque assay. The data represent the mean titer ± standard deviation (n = 5). *, P < 0.05 (versus Luc). (B) Immunoprecipitation of M1 and NP from 293T cell lysates and released virus particles after siRNA transfection and virus infection. Mock, Luc, and Cdc42 siRNA-transfected 293T cells (see the legend to panel A) were infected with WT virus (MOI = 0.1) and labeled at 4 h p.i. for 18 h with 250 μCi/ml of ³⁵S protein-labeling mixture. The cell lysates and released virus particles were immunoprecipitated for NP and M1 (for details, see the legend to Fig. 5). The virus release efficiency relative to the control is shown (bottom). *, P < 0.001 (versus Luc). (C) 293T cells were transfected with plasmids (4 μg) encoding HA-tagged WT Cdc42, dominant-negative HA-tagged Cdc42 (Cdc42-T17N), or constitutively active HA-tagged Cdc42 (Cdc42-G12V). The proteins were radiolabeled for 48 h and immunoprecipitated with anti-HA epitope antibody (inset). These Cdc42-expressing 293T cells were infected with WT virus (MOI = 0.1) at 48 h p.t., and the media were then collected at 16 h p.i. and assayed for PFU titers. The data represent the mean titer ± standard deviation (n = 3). *, P < 0.05; **, P < 0.001 (versus Cdc42). (D) Immunoprecipitation of M1 and NP from 293T cell lysates and released virions after cDNA transfection and virus infection. Mock, Cdc42, Cdc42-T17N, and Cdc42-G12V cDNA-transfected 293T cells (see the legend to panel C) were infected with WT virus (MOI = 0.1) and labeled at 4 h p.i. for 18 h with 250 μCi/ml of ³⁵S protein-labeling mixture. The cell lysates and extracellular medium were used for immunoprecipitation (for details, see the legend to Fig. 5). Representative data from two experiments are shown. Relative virus releases were normalized to intracellular expression levels (bottom). *, P < 0.05; **, P < 0.001 (versus Luc).