Influenza A Virus M1 Protein Non-Specifically Deforms Charged Lipid Membranes and Specifically Interacts with the Raft Boundary

Abstract

Topological rearrangements of biological membranes, such as fusion and fission, often require a sophisticated interplay between different proteins and cellular membranes. However, in the case of fusion proteins of enveloped viruses, even one molecule can execute membrane restructurings. Growing evidence indicates that matrix proteins of enveloped viruses can solely trigger the membrane bending required for another crucial step in virogenesis, the budding of progeny virions. For the case of the influenza A virus matrix protein M1, different studies report both in favor and against M1 being able to produce virus-like particles without other viral proteins. Here, we investigated the physicochemical mechanisms of M1 membrane activity on giant unilamellar vesicles of different lipid compositions using fluorescent confocal microscopy. We confirmed that M1 predominantly interacts electrostatically with the membrane, and its ability to deform the lipid bilayer is non-specific and typical for membrane-binding proteins and polypeptides. However, in the case of phase-separating membranes, M1 demonstrates a unique ability to induce macro-phase separation, probably due to the high affinity of M1's amphipathic helices to the raft boundary. Thus, we suggest that M1 is tailored to deform charged membranes with a specific activity in the case of phase-separating membranes.

Keywords: M1 matrix protein; amphipathic helices; fluorescent confocal microscopy; giant unilamellar vesicle (GUV); hemagglutinin; influenza A virus; lipid raft; lipoprotein envelope; membrane deformation; viral budding.

Figure 1

Scheme of the experiment. GUVs in the swelling buffer (SB) were added to the droplet of the working buffer (WB) and equilibrated for 15 min. M1 protein, as well as control proteins, BSA and PLL, were added in the pipette buffer (PB) to the selected GUV through a glass patch micropipette placed next to the selected GUV.

Figure 2

Monitoring the uncharged GUVs from the lipid Mixture 1 after 10 s perfusion with 1 μM M1 protein in the working buffer. The time count is synchronized with the perfusion start. The scale bar is 10 μm.

Figure 3

Monitoring the charged GUVs from the lipid Mixture 2 after 10 s perfusion with 1 μM of the M1 protein (A), BSA (B), or PLL (C) in the working buffer. Scale bars are 20 μm for (A,B) and 10 μm for (C).

Figure 4

Monitoring the charged GUVs from the lipid Mixture 2 after 10 s perfusion with 1 μM of M1 (A), 1 μM of BSA, (B), or 1 μM of PLL (C) in the static hyperosmotic mode. The scale bar is 10 μm.

Figure 5

Monitoring the uncharged GUVs from the lipid Mixture 1 after 10 s perfusion with 1 μM of M1 (A), 1 μM of BSA (B), or 1 μM of PLL (C) in the dynamic hyperosmotic mode. The scale bar is 10 μm.

Figure 6

Monitoring the charged GUVs from the lipid Mixture 2 after 10 s perfusion with 1 μM of M1 (A), 1 μM of BSA (B), or 1 μM of PLL (C) in the dynamic hyperosmotic mode. The scale bar is 20 μm.

Figure 7

Monitoring the charged GUVs from the lipid Mixture 2 after 10 s perfusion with a protein-free solution in dynamic hyperosmotic mode. The scale bar is 20 μm.

Figure 8

Monitoring the “raft” GUVs after 10 s perfusion with M1. The L_d marker Rho-PE is in the red channel, and the L_o marker BODIPY-GM1 is in the green channel. The scale bar is 20 μm.

Figure 9

Monitoring the “raft” GUVs after 10 s perfusion with BSA. The color code is the same as in Figure 8. The scale bar is 20 μm.

Figure 10

Monitoring the “raft” GUVs after 10 s perfusion with PLL. The color code is the same as in Figure 8. The scale bar is 20 μm.