Background
Measles virus (MV) is a prototype member of the Morbillivirus genus in the family
Paramyxoviridae. In virus particles, the negative-stranded RNA genome is encapsidated by the N, P and L proteins, and this ribonucleocapsid (RNP) is surrounded by a lipid bilayer. The two surface glycoproteins, the hemagglutinin H and the fusion protein F, protrude from the viral envelope. The matrix protein (M) is located at the inner surface of the lipid bilayer tethering the RNP to the envelope. Due to its interaction with the glycoproteins and the RNPs, the M protein is essential for MV assembly and particle formation. M binding to the cytoplasmic tails of the glycoproteins at the surface of infected cells is furthermore required to downregulate H/F-mediated cell-to-cell fusion of infected and neighboring uninfected cells [
1‐
5].
The actin network is primarily associated with mechanical stability, cell motility and cell contraction. It is also important for chromosome movement during mitosis and for internal transport, particularly near the plasma membrane. Cargos can be transported either by riding on myosin motors along actin filaments or by pushing forces exerted by actin as it undergoes polymerization [
6]. Cytoskeletal actin not only has a central function in cell physiology but is also an essential component involved in the replication of many RNA and DNA viruses. The molecular mechanisms underlying this important host-virus interaction, however, are extremely diverse [
7]. For MV, several reports have shown that actin is involved in virus maturation at the plasma membrane. This idea was initially based on the findings that actin was identified as an internal component of MV particles [
8,
9] and co-caps with MV H on infected cells [
10]. There is further ultrastructural evidence that actin filaments take part in the process of budding and protrude into viral buds [
7,
8]. Very recently, it was furthermore proposed that F-actin associates with the MV M protein altering the interaction between M and H, hereby modulating MV cell-cell fusion and assembly [
11].
Though there is conclusive evidence that intact actin filaments are important for MV replication, it is not yet defined if a stable actin cytoskeleton is sufficient, or if actin dynamics are required. Aim of this study was thus to analyze the effects of actin-disrupting and actin-stabilizing drugs to define if actin filaments as structural components or rather actin dynamics and treadmilling are essential for MV maturation. Actin treadmilling is a process in which actin filament length remains approximately constant but actin monomers preferentially join with the barbed ends and dissociate from the pointed ends of filaments. This oriented renewal of actin within microfilaments causes a treadmilling involving both, actin monomers and actin-binding proteins. Jasplakinolide (Jaspla) is a cyclic peptide isolated from a marine sponge that binds to and stabilizes filamentous actin, inducing a blockade of actin treadmilling [
12,
13]. In contrast to Jaspla, Cytochalasin D (CD), a fungal metabolite, serves as an actin capping compound that binds to the barbed (+) end of actin filaments and shifts the polymerization-depolymerization equilibrium towards depolymerization of F-actin [
14].
With our studies on MV replication in the presence of CD and Jaspla, we show that defects in actin polymerisation and defects in actin depolymerisation can both interfere with late virus assembly and budding steps without impairing overall viral protein synthesis, cell-associated infectivity or the surface transport of the MV glycoproteins. Most interestingly, the underlying mechanism of interference with late MV maturation steps by CD and Jaspla differs principally. While intact actin filaments that can be disrupted by CD treatment are required for M-RNP cotransport from the cytoplasm to the plasma membrane, actin dynamics at the cytocortex blocked by Jaspla are necessary for later steps in virus maturation at the plasma membrane. Supporting our finding that actin disruption blocks M-RNP transport to the plasma membrane, cell-to-cell spread of MV infection was enhanced upon CD treatment. Due to the lack of M-glycoprotein-interaction at the cell surface, M-mediated fusion downregulation is hindered and a more rapid syncytia formation is observed in CD-treated cells.
Discussion
Consistent with an earlier report using Cytochalasin B [
18], we found that CD inhibited MV replication. Actin inhibitors specifically seem to block late virus assembly and budding steps, because inhibition of earlier replication steps such as RNA synthesis would be reflected in decreased overall virus protein expression levels and reduced titers of both, released virus and cell-associated infectivity. After actin disruption by CD, M-RNP complexes were retained in the cytoplasm implicating a role for intact actin filaments in M-RNP transport. Consistent with M acting as a fusion downregulator, we found an increased cell-to-cell fusion after actin disruption due to the lack of M-glycoprotein interaction at the plasma membrane. Similar to actin disruption, actin stabilization by Jaspla led to decreased cell-free virus titers. However, the underlying molecular mechanism is clearly different. Block of actin dynamics by Jaspla did not result in the intracellular retention of M-RNP complexes. Immunofluorescence and electron microscopic observations rather indicate a defect in bud formation and subsequent pinching-off.
Actin filaments (disrupted by CD) but not actin treadmilling (blocked by Jaspla) are needed for transport of M and MV nucleocapsids from the cytoplasm to the cell surface. Yet, actin dynamics seem to play a role in budding of mature virions at the plasma membrane. Even if an early report in chronically MV-infected cells had suggested that RNP but not the M transport to the cell surface depends on intact actin filaments [
19], we and others have shown in more recent studies that M interacts with RNPs in viral inclusions and movement of viral RNPs to the plasma membrane occurs as co-transport with the M protein [
5,
20,
21]. In agreement with the idea that actin filaments serve as tracks for movement of M-coated RNPs to the cell surface, disruption of the actin filaments resulted in an intracellular retention of both, M and RNPs. A very recent report proposes that interaction with the cytoskeleton is mediated by direct binding of F-actin to phenylalanine 50 in the M protein [
11]. The direct interaction of MV M and actin is in line with findings for the M proteins of Newcastle disease virus and Sendai virus [
22] that also serve as the recognition site for actin. However, it remains to be elucidated if M binds directly to actin, or if interaction is mediated via motor proteins. The latter might be supported by our finding that Jaspla treatment did not affect the M-RNP trafficking to the cell periphery. While movement along non-dynamic actin filaments must be assumed to be less efficient, motor proteins would still be able to transport cargo (M-RNPs) along stabilized actin filaments.
In contrast to many other viral matrix proteins [
23], MV M protein does not possess any known late domain motif. Therefore budding does not depend on the cellular endosomal sorting complex required for transport (ESCRT) machinery [
24]. Given that MV M-RNPs interact with the actin cytoskeleton and that the actin motor protein Myosin Vb interacts with members of the Rab11 family [
25], a mechanism of late domain independent budding might be used by MV. The idea of a M-Rab11 interaction is indeed supported by the very recent finding that apical release from polarized epithelial cells depends on the Rab11A-dependent apical recycling endosomal pathway [
26].
Multifunctional involvement of actin microfilaments during viral infection has been documented in many studies. Dependence of viruses on actin, however, differs drastically not only between different DNA and RNA virus families, but even between closely related virus family members. Thus, it must be concluded that each virus has evolved its own mechanism to interact with the cellular cytoskeleton machinery ensuring optimal replication. In contrast to our findings, HPIV-3 needs intact actin filaments for viral RNA synthesis [
27]. Consequently, actin disruption led to a reduction of HPIV-3 release due to lack of viral proteins. For the budding process of HPIV-3, microtubules rather than actin filaments are important [
28]. Since Cytochalasin B had no negative effect on VSV release, it was supposed that actin is not involved in VSV replication [
29]. However, interaction of VSV M with dynamin was recently shown to be required for assembly [
30]. Treatment of Rotavirus infected polarized epithelial cells with Jaspla did not reduce overall virus release but altered the budding polarity from apical to bipolar [
13]. MV is also released apically from polarized epithelia [
31‐
33], but actin treadmilling does not seem to play a role in polarized MV release since treatment of MV-infected polarized MDCK cells with Jaspla did not alter budding polarity (Dietzel, unpublished observation).
As in MV infection, disruption of the actin cytoskeleton reduced release and viral infectivity of HIV [
34]. Recent cryo electron tomography studies have shown that HIV budding at the plasma membrane can be divided into different categories with respect to their actin context [
35]. Even if most of the budding sites were found adjacent to filamentous actin, only half of them were associated with filopodia-like structures characterized by a parallel actin organization. The rest of the buds were found with cortical actin parallel to the plasma membrane, or with cortical actin directed towards or protruding into the budding site. Even if actin filaments have been shown to protrude into budding MV particles [
8,
18], one might speculate that highly pleomorphic MV particles also bud in different forms. Though we cannot rule out that the only partial block of virus release observed upon Jaspla treatment is due to an incomplete stabilization of actin filaments at the used concentrations, it might be speculated that only budding forms that have a distinctive requirement for actin dynamics or treadmilling are affected by cytocortical actin stabilization.
Recent observations have shown that the interaction of the MV glycoprotein complex with receptors on lymphocytes and dendritic cells (DCs) initiate cytoskeletal dynamics. In DCs, MV binding initiates host cell cytoskeletal dynamics needed for viral uptake and the establishment of functional synapses with T cells. Furthermore, MV binding to T cells causes a loss of polarization, adhesion and motility by actin cytoskeletal paralysis [
36]. It is therefore highly likely that integrity and dynamics of actin filaments not only play an important role in virus maturation at the plasma membrane, but are also involved in MV-mediated immunosuppression.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
ED conceived the study, carried out all cell infections and inhibitor studies, analysed the samples and drafted the manuscript. LK performed the EM work, and edited the manuscript. AM conceived the study and drafted the manuscript. All authors read and approved the final manuscript.