As intracellular pathogens, enveloped viruses must usurp the host cell machinery for many stages of the viral life cycle in order to produce a new generation of infectious virions. In one of the less understood steps of viral assembly, viral components including the transmembrane glycoproteins, structural proteins and the viral genome must be targeted to the site of viral budding, where they assemble and are incorporated into a newly formed virion that gains a lipid envelope from a cellular membrane. Recent work has revealed that the cellular recycling endosome pathway, in particular Rab11, plays an important role in the assembly of negative-strand RNA viruses such as respiratory syncytial virus, influenza A virus, Andes virus and Sendai virus. The present mini-review discusses this emerging field and explores the potential roles of the Rab11 pathway in the trafficking, assembly and budding steps of these viruses.
Introduction: Rab11 and the recycling pathway
The Rab family of GTPases play crucial roles in directing cellular vesicular transport throughout the endosomal system, as individual Rab proteins serve as identity markers for a variety of membrane compartments and also help to mediate transport of vesicles throughout the pathway. Rab11 is one of the primary Rabs involved in the recycling endosome pathway, which transports cargo endocytosed at the apical plasma membrane to either the ARE (apical recycling endosome) or the perinuclear recycling endosome compartment, before redirecting vesicular cargo back to the apical plasma membrane (Figure 1A).
Rab11 associates with membranes in a cyclical GTP-dependent manner. After binding GTP, a conformational change allows the isoprenylated C-terminus of Rab11 to associate with a vesicular membrane . Compartmental identity in the endosomal trafficking pathways is determined partially by the identity of the Rabs on a given vesicle, although the exact mechanism that guides a particular Rab to a particular membrane remains unknown. Considerably more is known about the downstream effectors of Rab11, which link Rab11 to various cytoskeletal components and are responsible for directing vesicular movement throughout the recycling pathway. The most well-studied members of the Rab11 effector family include the Rab11-FIPs (family interacting partners). Currently identified FIPs include FIP1/RCP (Rab-coupling protein), FIP2, FIP3/Arfophilin/Eferin, FIP4/Arfophilin-2 and FIP5/Rip11/Gaf-1/pp75, all of which share a highly conserved C-terminal motif termed the RBD (Rab11-binding domain). In addition to the FIPs, other Rab11 effectors include myosin Vb, Rabphilin-1/RabBP (Rab-binding protein), phosphoinositide 4-kinase and sec15 (reviewed in ). In complex with Rab11, the Rab11 effectors then direct vesicles trafficking throughout the recycling pathway, primarily by mediating associations with specific motor proteins. For example, FIP3 is known to bind dynein light intermediate chain 1 and 2 and to be responsible for retrograde trafficking from peripheral sorting endosomes to the perinuclear recycling endosome [3,4]. In contrast, FIP5 binds Kif3b, a component of kinesin II, and is responsible for trafficking vesicles from the perinuclear recycling endosome to the cell periphery . Myosin Vb has also been shown to be important for traffic from the perinuclear recycling endosome to peripheral endosomes, either directly or via FIP2, in addition to tethering vesicles in the cortical actin network [6–8]. Overall, interactions with the various effectors allow Rab11-decorated vesicles access to actin-based cytoskeletal transport in addition to the microtubule network travelled by kinesins and dynein, and are thus capable of mediating active transport throughout the entire recycling pathway.
Viruses face a set of unique challenges in their life cycle of replication, as they must manufacture new viral proteins and genomic DNA in a crowded and complex cellular environment, and target each component to the specific site of budding in a co-ordinated fashion. Some viruses position their replication machinery on a cellular membrane as a way of concentrating viral components in two dimensions, whereas others such as IAV (influenza A virus) replicate their RNA in the nucleus, and must ensure that the viral genome and proteins ultimately meet at the correct location. Enveloped viruses then gain a lipid envelope by budding through a cellular membrane, often at the plasma membrane or sometimes internally, into (for example) the Golgi lumen. This variation in assembly site requires that viruses usurp specific trafficking pathways, in order to deliver their proteins and genome to the correct membrane in a targeted fashion. The cytoskeletal and vesicular trafficking networks play important roles in transporting viral components (during entry as well as assembly) and ensuring the release of enveloped virions from the right location. Finally, a membrane scission event separates the viral and cellular membranes, releasing the virus particle from the host cell. Many, but not all, viruses co-opt the ESCRT (endosomal sorting complex required for transport) machinery to achieve this scission step, although notably IAV and RSV (respiratory syncytial virus) are ESCRT-independent. As discussed in the present mini-review, the Rab11 recycling pathway is involved in the processes of genome trafficking and egress of several enveloped negative-strand RNA viruses.
Rab11 and transport of viral genomes
Work supporting a function for Rab11 in virus trafficking has revealed a role for Rab11 in intracellular transport of negative-strand virus genomes. The RNP (ribonucleoprotein) complexes that contain the genomic RNAs of SeV (Sendai virus) and IAV appear to ‘hitchhike’ across the cytoplasm by binding to Rab11-positive vesicles in transit from the perinuclear recycling endosome to the plasma membrane [9–12] (Figure 1B). Evidence for this comes from altered RNP localization when the Rab11 pathway is disrupted by siRNA (small interfering RNA) depletion or by overexpression of DN (dominant-negative) FIPs, as well as from evident punctate co-localization of Rab11 and viral RNP components in the cytoplasm (Figure 2). Visually striking evidence can also be gained from live-cell microscopy of fluorescently tagged components, where (taking the data in aggregate) RNPs could be seen to co-localize with Rab11 while exhibiting rapid saltatory movement along microtubules [9–13].
The association between Rab11 and viral RNPs probably takes place preferentially on a vesicular membrane. Whereas IAV RNPs bind endogenous Rab11 in pull-down assays, studies conducted with GTP-bound/CA (constitutively active) or GDP-bound/DN Rab11 mutants revealed that this binding occurs only with the CA form of Rab11 [10,12]. As the CA form of Rab11 is expected to be membrane bound, whereas the DN form is generally thought to be cytosolic, it suggests that viral RNPs associate with Rab11 attached to a vesicle. Furthermore, immunogold labelling of SeV RNPs in thin-section electron micrographs showed them to be in direct proximity to vesicular structures . The exact details of how the viral RNPs are targeted to Rab11 await clarification. In the case of IAV, evidence suggests that the PB2 subunit of the viral polymerase mediates the interaction [10,12], whereas the biochemistry of SeV RNP interactions has not yet been investigated. From the cellular side, it is not known whether the RNPs interact directly with Rab11, or indirectly by (for example) a FIP. For IAV, however, it is possible that the cellular protein HRB (HIV Rev-binding protein) plays a role in guiding RNPs to the perinuclear recycling endosome .
The role of Rab11 in viral budding and release
A requirement for Rab11 in negative-strand virus budding was first suggested for RSV, following the observation that expression of DN forms of the Rab11 effectors myosin Vb, FIP1 or FIP2 inhibited overall virus release and disrupted its normal apical polarity [15,16]. A direct role for Rab11 itself was next indicated for the hantavirus ANDV (Andes virus): viral structural proteins were observed to co-localize with Rab11 in the perinuclear region and siRNA-mediated depletion of Rab11 decreased virus release 10-fold . A subsequent report showed that the Rab11 pathway was also needed for efficient release of IAV, as siRNA depletion of Rab11 decreased the number of infectious virions released (again approximately 10–30-fold) and also resulted in aberrant particle morphologies consistent with a budding defect . In the case of IAV, laboratory-adapted strains of virus produce profuse numbers of elongated particles that project away from the plasma membrane, sometimes still apparently attached by a narrow stalk of membrane (Figure 3A). However, after depletion of Rab11, many ‘stumpy’ particles fail to complete membrane scission, and remain attached to the plasma membrane by wide necks of cytoplasm , whereas (less commonly) ‘daisy chains’ of segmented structures that plausibly represented abortive attempts to pinch off the viral membrane are formed (Figure 3A). For RSV and IAV, these findings initially suggested a possible function for the Rab11 pathway akin to that provided by the ESCRT machinery.
The role of the ESCRT pathway in the release of many enveloped viruses, notably HIV, is well characterized, and thought to directly provide the energy and structural components needed to drive scission of viral and cellular membranes . Interestingly, some of the budding defects seen for IAV after Rab11 depletion (Figure 3A) resembled phenotypes exhibited by mutated HIV that were unable to engage the ESCRT pathway. However, both influenza and RSV bud via ESCRT-independent mechanisms [16,20–22], which raised the possibility that Rab11 or one of its interacting partners could be providing an alternative means of membrane scission.
Further supporting a potential role in budding, certain Rab11-FIPs have been shown to influence the shape of virus particles produced by IAV and RSV. Ordinarily, both viruses form long filamentous virions (see Figure 3A for IAV) in addition to smaller spherical particles [23,24], and, although the function of these filamentous forms remains enigmatic, they may well have clinical relevance since the majority of human IAV isolates as well as influenza B, C and other non-segmented negative-sense respiratory viruses produce them [25–28]. Both Rab11 and Rab11-FIP3 are required to support the formation of filamentous IAV particles, as depletion of either protein abrogates this phenotype  (Figure 3A). However, loss of FIP3 has no effect on spherical virion formation, implying that the function of Rab11 during influenza replication is unlikely to be mediated solely via Rab11-FIP3 . As Rab11-FIP3 is involved in trafficking membrane to the cleavage furrow during cytokinesis [29,30], it is possible that its involvement in filament morphogenesis is simply related to the increased membrane surface area required to produce a particle of up to 30 μm in length. Alternatively, Rab11-FIP3 has been implicated in regulating actin dynamics , and influenza filament formation is known to require an intact cortical actin network [32,33].
In contrast with IAV, RSV filament formation is controlled by Rab11-FIP2 . Expression of a DN form of Rab11-FIP2 interfered with normal RSV particle formation, although it led to an increase in filament length, rather than a decrease as seen with IAV and FIP3. Thus, whereas IAV, ANDV and RSV differ in the specifics of engagement with Rab11 and its effectors, in each case, the final steps of the viral life cycles appear to be dependent on a fully functioning Rab11 pathway. In the case of ANDV, the virus probably buds intracellularly into the lumen of an ill-defined perinuclear membrane compartment , so the Rab11 pathway is plausibly required for the outward movement of virus-containing vesicles (Figure 1C). However, for IAV and RSV, where budding unquestionably takes place at the plasma membrane, it is not clear to what extent this requirement for Rab11 is direct, or whether instead it reflects defective transport of essential viral (e.g. the genomic RNPs) and/or cellular factors to the sites of virus budding.
Why co-opt the Rab11 machinery?
As the Rab11 pathway is responsible for directing vesicular transport from the trans-Golgi network and perinuclear recycling endosome to the ARE and plasma membrane, this trafficking route is capable of providing polarized delivery of virion components to the site of assembly. Since many viruses (including the respiratory pathogens from the ortho- and para-myxovirus families which are the main focus of the present review) exhibit polarized budding from the apical surface of epithelial cells [34–36], this is a potentially useful characteristic. Using the cellular vesicular transport system to access the microtubule network also provides a solution to the problem of how to move megadalton-sized RNPs through a crowded and viscous cytoplasm . Furthermore, as Rab11-positive vesicles can traffic outwards on both microtubules and actin filaments, via FIP5–kinesin II and myosin Vb respectively, this system theoretically provides a relatively simple means of navigating transport through both the cytoplasm and cortical actin barrier without having to engage multiple host proteins .
Nevertheless, despite these real and/or hypothetical advantages, questions can be asked as to what degree of reliance any of the viruses under discussion place on the Rab11 pathway. On one hand, DN FIP proteins exert very substantial inhibitory effects on RSV replication, decreasing titres of replicated virus ~1000-fold [15,16], but siRNA depletion of Rab11 itself had more moderate (~10-fold) inhibitory effects on ANDV or IAV replication [17,18]. Do these differential effects reflect genuine differences in the behaviour of the viruses or merely the experimental approaches used? Given that IAV buds from the plasma membrane, whereas ANDV buds into an internal compartment, Rab11 is likely to be playing multiple mechanistic roles depending on the virus in question, which may complicate analysis. Plausible arguments can be made regarding inefficiencies of siRNA depletion or the pleiotropic effects of overexpressing DN mutants, but, nevertheless, it would be interesting to know the effects of Rab11 depletion on RSV replication or to test IAV growth in cell lines expressing DN FIPs. However, whatever the outcome, the conundrum would remain of why drugs that target the microtubule network directly (thus blocking the tracks that Rab11 vesicles run on) have rather lesser effects (post-viral entry) on RSV, SeV or IAV replication [9,10,38]. One likely explanation is that the viruses have evolved redundant strategies to achieve the same end. For instance, whereas live-cell imaging of RNP cytoplasmic transport provides unambiguous evidence for microtubule-mediated transport, movement with the characteristics of actin-based motility or even simple diffusion has also been observed for IAV and RSV [10,13,39,40]. Further work is needed to resolve these questions.
To end on a final correlative speculation, it is interesting to note that, so far, the negative-sense RNA viruses that show a dependency on the Rab11 pathway for genome trafficking are either known (RSV and IAV) or have been postulated (SeV) to use a mechanism other than the ESCRT pathway to achieve the final membrane scission step of virion release [16,20–22,41]. We came to study the Rab11 pathway in search of an alternative mechanism to explain the end stage of IAV budding  before realizing that there were upstream effects on RNP trafficking . Although it is still possible that the virus uses the Rab11 pathway directly as an alternative means of achieving scission via cellular machinery, recent work has suggested that the viral M2 protein may play a role . We therefore now prefer the hypothesis that the budding defects in IAV (and perhaps RSV) observed in the absence of a functioning Rab11 system indicate that virion formation and morphology of these viruses is particularly dependent on correct incorporation of the genome.
E.A.B. was supported by a Ph.D. studentship from the Gates Foundation. A.S. and P.D. acknowledge support from the U.K. Medical Research Council [grant number G0801931] and Institute Strategic Programme Grant Funding from the U.K. Biotechnology and Biological Sciences Research Council [grant number BB/J004324/1].
We thank Dr Robert W. Doms for helpful comments and a critical reading of the paper.
Rab GTPases and Their Interacting Proteins in Health and Disease: A Biochemical Society Focused Meeting held at University College Cork, Cork, Ireland, 11–13 June 2012. Organized and Edited by Mary McCaffrey (University College Cork, Ireland).
Abbreviations: ANDV, Andes virus; ARE, apical recycling endosome; CA, constitutively active; DN, dominant-negative; ESCRT, endosomal sorting complex required for transport; FIP, family interacting partner; IAV, influenza A virus; RNP, ribonucleoprotein; RSV, respiratory syncytial virus; SeV, Sendai virus; siRNA, small interfering RNA
- © 2012 The Authors Journal