In this report, we show that herpesvirus infection induces the formation of actin filaments in the nucleus. The timing of filament formation (), the association of capsids with filaments ( and ) and myosin V (), and the dependence of VP26 organization on filaments (), suggests that F-actin plays a key role in the formation and organization of viral assembly centers in the nucleus. Our results support the recent finding showing that directed movement of HSV-1 capsids in the nucleus is actin-, and likely, myosin-dependent [34
]. We propose that nuclear actin filaments provide tracks for myosin—rather than actin polymerization-based movement, because elimination of actin dynamics by jasp did not disrupt VP26 localization (). A diverse group of intracellular microorganisms, including Listeria monocytogenes, Shigella
spp., and vaccinia virus, utilize the host actin cytoskeleton to move within and spread between mammalian host cells. We believe that herpesviruses provide another distinct example of a pathogen appropriating the host actin cytoskeleton.
We first observed nuclear actin filaments in infected mouse SMG tissue in volume electron-microscopic images obtained using SBFSEM (). In those images, we detected a network of filaments in the nucleus surrounding the PRV capsids present only in infected cells. Most of the filaments associated with individual genome-filled capsids, showing both end-on and side-on associations. Given the association of filaments exclusively with genome-filled capsids, we propose that the filaments play a role in capsid assembly and/or the transport of capsids in preparation for egress. Similar filaments described as “interwoven fine fibrils” have been found to associate with HSV-1 and HSV-2 nucleocapsids in infected CNS, PNS, and glial cell nuclei [39
]. Thus, the formation of such filaments may be a conserved feature in the alpha-herpesvirus life cycle.
We tested the hypothesis that infection induces filament formation using dissociated, cultured peripheral SCG neurons. Neurons infected with PRV or HSV-1 contained a network of nuclear actin filaments, visualized by fluorescently labeled phalloidin ( and ). However, the extent to which actin filaments were present in the nucleus was variable: some cells contained a single large bundle running along one face of the inner nuclear membrane whereas others possessed larger, more complex structures. We do not yet understand the factors responsible for these differences, but amount of viral production or the size of the actin monomer pool could play a role. The average length of nuclear actin filaments seen by confocal microscopy (4.5 ± 1.9 μm) was similar to that seen with SBFSEM (3.0 ± 1.2 μm) and in both cases filaments specifically associated with capsids and were present only in infected cells. Thus, they may be the same structures. In PK15 cells, nuclear actin filaments appeared finer and less organized than those in SCG neurons, perhaps reflecting differences between polarized and non-polarized cells or primary and transformed cells.
Actin filaments associate with the nuclear lamina that faces the Golgi () which implies that, if actin filaments are indeed utilized for nuclear egress, capsids traveling along the filaments towards the nuclear membrane would emerge from the nuclear envelope ready to engage the secretory pathway. Consistent with this speculation, live imaging of HSV-1 capsids in the nucleus showed directed movements towards the nuclear envelope [34
]. This result also suggests that if actin nucleators are involved in the formation of these filaments, then one would expect to find them asymmetrically localized on the nuclear lamina.
Since actin filaments are required for capsid foci formation, GFP-VP26 foci presumably reflect viral assembly sites as they have been observed and characterized in both PRV and HSV-1-infected cells [31
]. Experiments with latA () suggest that actin filaments are important for establishment and/or maintenance of GFP-VP26 foci. A recent paper has shown that nuclear expansion in epithelial cells infected with HSV-1 is dependent on actin [10
]. LatA also inhibits replication, compartment maturation, and chromatin dispersal [10
], providing support for the notion that actin filaments provide a scaffold for viral assembly. Testing whether such a scaffold is also used for viral DNA replication will require further study.
Jasp, an inhibitor of actin treadmilling dynamics that stabilizes actin filaments, increased the number of cells containing GFP-VP26 foci by 30% (). This finding suggests that capsid foci may be stabilized as a result of filament stabilization. At present, we do not know the degree to which the filaments are dynamic or if they are undergoing rapid turnover; Fluorescence Recovery after Photobleaching (FRAP) experiments with neurons expressing GFP-actin should answer this question. The effects of jasp imply that individual capsids use the actin filaments as tracks for directed movements rather than by using actin polymerization-based movement, which is distinct from the way that Listeria
and vaccinia virus use the actin cytoskeleton. The result that capsids co-localize with the actin motor myosin Va () supports this idea. Class V myosins are among the most thoroughly studied forms of unconventional myosins and considerable evidence supports a role in transport of organelles and vesicles [40
]. Myosin Va is a two-headed myosin that shows processive movement along actin filaments, similar to that of two-headed kinesins and dynein along microtubules [40
]. Myosin Va is one of the fastest myosins, moving along actin filaments at a speed of 300–400 nm/s [42
], which is comparable with the speed reported for directed movements of HSV-1 capsids [34
We do not know whether actin filaments form as a result of rearrangement of the available nuclear monomer pool or if monomeric actin is recruited from the cytoplasm. Several recent publications [43
] have demonstrated that actin is present in the nucleus and is critical for transcription, chromatin remodeling, mRNA export, and nuclear structure and integrity. However, the actin present in the nucleus may not be filamentous, since it is not recognized by phalloidin, which binds only to filaments more than seven monomers long [46
]. By analogy, actin may also play a role in transcription of viral genes. β-actin has been shown to co-purify with RNA polymerase (I, II, and III), is a component of pre-initiation complexes, and appears to be recruited to promoters of genes about to be transcribed [47
]. However, recent work [10
] demonstrates that treatment with inhibitors of actin polymerization does not
affect HSV-1 viral replication (cytochalasin D treatment even increases viral titer about 15 times). These observations taken together with our findings that actin filaments still form when viral DNA replication is prevented () suggest that actin plays an ancillary rather than an essential role in the virus life cycle.
Recent work also shows that host-derived actin is incorporated into the PRV virion and becomes an integral part of the outer tegument layer [50
]. The amount of actin incorporated increased in the absence of VP22, one of the major tegument proteins, providing support for the view of the outer tegument layer as dynamic outer shell [50
]. Virion-associated actin has been reported in other herpesviruses [1
] and other enveloped viruses including paramyxovirus, retrovirus, and rhabdovirus [57
]. Although actin incorporation may be required as a structural element of the virion, actin may also serve an additional function later in infection, such as nuclear egress or envelopment.
We do not know how actin filament formation occurs after infection, but we do know that new protein synthesis is required. We also know that at least one viral immediate-early or early protein is required (). This protein may promote the formation of actin filaments directly, perhaps as an actin nucleator. Studies of the mammalian stress response have revealed that formation of nuclear actin filaments, increased nuclear invaginations, and decondensation of nucleoli, events all associated with viral infection, occur in response to heat shock [62
]. During the heat-shock response and during HSV-1 infection, cellular chaperones Hsp70 and Hsp40 are redistributed to the nucleus during infection and co-localize with ICP0, adjacent to replication complexes, thereby promoting sequestration and compartmentalization of the nucleus [21
]. Herpesvirus infection induces cellular stress responses, which may be exploited to concentrate viral and host proteins required for viral assembly and packaging. Consistent with this speculation, baculovirus-infected cells form nuclear actin filaments at the time of virus assembly; these filaments co-localize with nucleocapsids and the baculovirus major capsid protein has been shown to bind F-actin [63
]. Whether actin filament formation is a viral-induced response or a general stress response, viruses that replicate in the nucleus may utilize these actin filaments as a scaffold for assembly and genome packaging. As a first step toward testing these ideas, it will be important to understand how viral capsids interact with nuclear actin filaments by identifying viral proteins that interact with actin and/or myosin.