SEARCH

SEARCH BY CITATION

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Membrane interactions during FV target cell entry
  5. Glycoprotein-dependent release of infectious FV virions
  6. Glycoprotein-mediated release of subviral particles
  7. Manipulation of FV capsid membrane interactions
  8. Retargeting of PFV Gag to different cellular membranes
  9. Pseudotyping of FV capsids
  10. Conclusions
  11. Acknowledgements
  12. References

Foamy viruses (FVs), a unique type of retroviruses, are characterized by several unusual features in their replication strategy. FVs, common to all non-human primates and several other species, display an extremely broad tropism in vitro. Basically, all mammalian cells and species examined, but also cells of amphibian or bird origin, are permissive to FV glycoprotein (Env)-mediated capsid release into the cytoplasm. The nature of the broadly expressed, and potentially evolutionary conserved, FV entry receptor molecule(s) is poorly characterized. Although recent data indicate that proteoglycans serve as an important factor for FV Env-mediated target cell attachment, additional uncharacterized molecules appear to be essential for the pH-dependent fusion of viral and cellular lipid membranes after endocytic uptake of virions. Furthermore, FVs show a very special assembly strategy. Unlike other retroviruses, the FV capsid precursor protein (Gag) undergoes only very limited proteolytic processing during assembly. This results in an immature morphology of capsids found in released FV virions. In addition, the FV Gag protein appears to lack a functional membrane-targeting signal. As a consequence, FVs utilize a specific interaction between capsid and cognate viral glycoprotein for initiation of thebudding process. Genetic fusion of heterologous targeting domains for plasma but not endosomal membranes to FV Gag enables glycoprotein-independent particle egress. However, this is at the expense of normal capsid morphogenesis and infectivity. The low-level Gag precursor processing and the requirement for a reversible, artificial Gag membrane association for effective pseudotyping of FV capsids by heterologous glycoproteins strongly suggest that FVs require a transient interaction of capsids with cellular membranes for viral replication. Under natural condition, this appears to be achieved by the lack of a membrane-targeting function of the FV Gag protein and the accomplishment of capsid membrane attachment through an unusual specific interaction with the cognate glycoprotein.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Membrane interactions during FV target cell entry
  5. Glycoprotein-dependent release of infectious FV virions
  6. Glycoprotein-mediated release of subviral particles
  7. Manipulation of FV capsid membrane interactions
  8. Retargeting of PFV Gag to different cellular membranes
  9. Pseudotyping of FV capsids
  10. Conclusions
  11. Acknowledgements
  12. References

Spuma or foamy viruses (FVs) are the only genus of the retrovirus subfamily Spumaretrovirinae (Linial et al., 2005). FVs that have coevolved with their natural hosts over millions of years appear to be non-pathogenic in natural hosts, which are most non-human primates, cattle, horses and cats, as well as zoonotically infected humans (reviewed in Rethwilm, 2010). They are promising candidates as retroviral gene transfer systems, partially due to some of their general but mainly as a result of their special molecular and cell biological features. FVs display a unique replication strategy (Fig. 1) with features common to both, other retroviruses (Orthoretrovirinae) as well as hepadnaviruses (reviewed in Lindemann and Rethwilm, 2011). The best-studied and characterized isolate to date is the so-called prototype FV (PFV, formerly known as human FV, HFV). PFV was originally isolated from an African patient who presumably was zoonotically infected by a chimpanzee FV (Achong et al., 1971; Herchenröder et al., 1995; Epstein, 2004).

figure

Figure 1. Schematic overview and comparison of the replication cycles of orthoretroviruses and foamy viruses. MTOC, microtubule organizing centre; ER, endoplasmic reticulum.

Download figure to PowerPoint

Like many other viruses, replication of FVs requires multiple trafficking events in the host cell and a multitude of interactions with host cell machineries and cellular membranes. Unlike other viruses, only few FV host cell interactions have been identified and characterized in detail to date.

Membrane interactions during FV target cell entry

  1. Top of page
  2. Summary
  3. Introduction
  4. Membrane interactions during FV target cell entry
  5. Glycoprotein-dependent release of infectious FV virions
  6. Glycoprotein-mediated release of subviral particles
  7. Manipulation of FV capsid membrane interactions
  8. Retargeting of PFV Gag to different cellular membranes
  9. Pseudotyping of FV capsids
  10. Conclusions
  11. Acknowledgements
  12. References

An extremely broad host range characterizes FVs and only recently few apparently non-permissive cell lines were identified (Hill et al., 1999; Mergia and Heinkelein, 2003; Stirnnagel et al., 2010). The FV glycoproteins mediate specific attachment to target cells by a receptor binding domain located in the glycoprotein surface (SU) subunit constituted by a bipartite sequence motif (Herchenröder et al., 1999; Duda et al., 2006). FV Env precursor cleavage between SU and transmembrane (TM) subunits is not required for target cell attachment and uptake. However, for the release of capsids by Env-directed lipid membrane fusion into the cytoplasm of the target cell the FV Env precursor cleavage is needed (Stirnnagel et al., 2012).

The nature of the cellular receptor(s) of FV attachment and entry by glycoprotein-mediated fusion of viral and cellular lipid membranes remains poorly characterized. Several lines of evidence from recent publications suggest that proteoglycans and heparan sulfate function particularly as attachment factors for FV Env-mediated host cell infection (Stirnnagel et al., 2010; Nasimuzzaman and Persons, 2012; Plochmann et al., 2012). Heparan sulfate expression levels seem to correlate with susceptibility towards FV infection. Furthermore, glycosaminoglycan- or heparan sulfate-deficient cell lines are characterized by a strong reduction in FV susceptibility (Stirnnagel et al., 2010; Nasimuzzaman and Persons, 2012; Plochmann et al., 2012). However, these molecules do not appear to be essential as these cells are still susceptible towards FV Env-mediated transduction (Stirnnagel et al., 2010; Nasimuzzaman and Persons, 2012; Plochmann et al., 2012). Therefore, heparan sulfate appears to function as an attachment factor for FV Env containing virions, potentially resulting in a concentration of viruses on the cell surface. A subsequent engagement of additional, yet unidentified molecules seems to be required for FV Env-mediated lipid membrane fusion releasing the capsid into the cytoplasm of the target cell.

Previous work from our lab demonstrated that fusion activity of all FV glycoproteins examined is pH-dependent with higher activity at low pH (Picard-Maureau et al., 2003). PFV Env was somewhat special as it demonstrated a significantly higher fusion activity at neutral pH than other FV glycoproteins (Picard-Maureau et al., 2003). Therefore, it is thought that FV uptake follows endocytic routes (Fig. 1). Very recently, we developed infectious, double-tagged FV particles composed of EGFP-tagged PFV capsids and mCherry-labelled envelope proteins of PFV or macaque simian foamy virus (SFVmac) (Stirnnagel et al., 2012). This tool was used to investigate uptake and trafficking processes of FV virions upon target cell entry by biochemical methods in bulk populations as well as time-lapsed imaging analysis in individual living cells. Both virus types showed trafficking of double-tagged virions towards the cell centre (Stirnnagel et al., 2012). Upon fusion and subsequent capsid release into the cytosol, accumulation of naked capsid proteins was observed within 4 h in the perinuclear region, presumably representing the centrosomes as reported previously (Lehmann-Che et al., 2005; 2007; Stirnnagel et al., 2012). As mentioned above, virions harbouring fusion-defective glycoproteins still promoted virus attachment and uptake but failed to show perinuclear capsid accumulation. Overall, the analysis indicated that productive FV Env-mediated fusion events occur predominantly within 4–6 h after the attachment of the virus. Non-fused or non-fusogenic viruses are rapidly cleared from the cells by putative lysosomal degradation. Quantitative monitoring of individual FV virions in target cells demonstrated that PFV Env induced fusion within the first few minutes whereas fusion SFVmac Env-mediated capsid release was less pronounced and observed during the whole observation period. Furthermore, in line with their different pH requirements PFV Env but not SFVmac Env containing particles induced strong syncytia formation by a ‘fusion from without’ mechanism on target cells (Picard-Maureau et al., 2003; Stirnnagel et al., 2012).

Glycoprotein-dependent release of infectious FV virions

  1. Top of page
  2. Summary
  3. Introduction
  4. Membrane interactions during FV target cell entry
  5. Glycoprotein-dependent release of infectious FV virions
  6. Glycoprotein-mediated release of subviral particles
  7. Manipulation of FV capsid membrane interactions
  8. Retargeting of PFV Gag to different cellular membranes
  9. Pseudotyping of FV capsids
  10. Conclusions
  11. Acknowledgements
  12. References

The glycoprotein (Env)-dependent budding of viral particles at late stages of the replication cycle is a feature of FVs that is unique among retroviruses but bears homology to hepadnaviruses (Fischer et al., 1998). In general, FV capsid assembly follows a retrovirus type B/D morphotype strategy. This includes a microtubule-dependent transport of capsid (Gag) proteins to the centrosome. The transport is facilitated by a cytoplasmic targeting and retention signal (CTRS) motif within the FV Gag. This CTRS motif shows strong homology to an analogous motif first identified in Mason Pfizer monkey virus (MPMV) (Eastman and Linial, 2001; Yu et al., 2006). FV capsids are preassembled at the centrosome prior to budding across cellular membranes. However, unlike other retroviruses, FVs are unable to release virus-like particles (VLPs) upon Gag expression alone (Fischer et al., 1998). Furthermore, no membrane association of FV capsids is observed in the absence of Env coexpression (Fischer et al., 1998; Pietschmann et al., 1999). This indicates that the FV Gag protein, which is not processed in an orthoretroviral-like manner into matrix (MA), capsid (CA) and nucleocapsid (NC) subunits, lacks a membrane-targeting signal that resides normally in the orthoretroviral MA subunit. Membrane targeting of FV capsids is achieved through a highly specific interaction with the cognate viral glycoproteins during particle morphogenesis (Pietschmann et al., 1999; Lindemann et al., 2001). It involves N-terminal regions of both types of FV structural proteins. The position of the capsid interaction domain within the viral glycoprotein, termed ‘budding domain’, is very unusual because it localizes to the signal or leader peptide (LP). Signal peptides are known for their function in targeting glycoproteins to the secretory pathway. In cases where they are not cotranslationally removed by proteolysis they anchor the protein in the membrane. However, the FV Env protein displays a highly unusual biosynthesis that lacks the typical cotranslational removal of the signal peptide by signal peptidase complex processing observed for other retroviral glycoproteins (Lindemann et al., 2001). As a result a full-length FV Env precursor protein is translated at endoplasmic reticulum (ER) membrane-resident ribosome complexes, and is inserted into the membrane with both N- and C-termini located in the cytoplasm. Only during its transport to the plasma membrane the Env precursor protein is processed into LP (gp18LP), SU (gp80SU) and TM (gp48TM) subunits at two furin or furin-like proteolytic cleavage sites (Duda et al., 2004). As a consequence a tripartite glycoprotein complex, consisting of a LP subunit with type II membrane orientation, an extracellular SU subunit and a type I membrane topology TM subunit, is integral components of released FV particles (Wilk et al., 2000; Lindemann et al., 2001). The Gag interaction (budding) domain of the glycoprotein spans the cytoplasmic N-terminal 15 amino acids and includes two essential and evolutionary highly conserved tryptophane residues. Env presumably interacts with Gag at the trans-Golgi network (Yu et al., 2006). Together with the involvement of the cellular vacuolar protein sorting (VPS) machinery in FV particle release, this is suggestive for an exocytotic budding pathway (Lindemann et al., 2001; Yu et al., 2006). However, there are also clear indications for budding of FVs at the plasma membrane (Lecellier et al., 2002).

The glycoprotein itself contains two trafficking signal motifs, one in each of the cytoplasmic domains (CyD) of the LP and TM subunits. One is a dilysine motif, located near the C-terminus of the rather short (comprising ∼ 13 aa) TM CyD of most FV isolates. It is known to be responsible for the retrieval of glycoproteins to the ER (Goepfert et al., 1999). While this signal can sort Env to the ER, it has only a weak effect on Env intracellular distribution in comparison to the other factors, and it is not required for efficient virus replication (Goepfert et al., 1999). Posttranslational ubiquitination of four of five lysine residues located within the LP subunit N-terminal CyD appears to function as a second, more powerful trafficking signal (Stange et al., 2005; Stanke et al., 2005). It mediates efficient Env removal from the cell surface. Furthermore, Gag coexpression is necessary for the efficient transport of Env to the cell surface, potentially by neutralizing or masking one or both of the two glycoprotein trafficking signals through its interaction with Env (Pietschmann et al., 2000). Summarized, not only is FV Env influencing capsid trafficking by the highly specific interaction but Gag protein coexpression is also altering intracellular distribution and trafficking of the glycoprotein (Pietschmann et al., 2000).

Glycoprotein-mediated release of subviral particles

  1. Top of page
  2. Summary
  3. Introduction
  4. Membrane interactions during FV target cell entry
  5. Glycoprotein-dependent release of infectious FV virions
  6. Glycoprotein-mediated release of subviral particles
  7. Manipulation of FV capsid membrane interactions
  8. Retargeting of PFV Gag to different cellular membranes
  9. Pseudotyping of FV capsids
  10. Conclusions
  11. Acknowledgements
  12. References

The unique requirement of FVs for Gag and Env coexpression to enable budding of virions suggests that unlike other retroviruses both FV structural proteins have an essential function for particle formation. Indeed, in 2003 the formation of so-called subviral particles (SVPs), capsidless particulate structures containing only the viral glycoprotein, was described for FVs (Shaw et al., 2003). Env expression alone was sufficient to induce SVP release. This again is analogous to a similar process observed for hepadnaviruses, which secrete vast amounts of SVPs, the so-called ‘Australia Antigen’. In contrast to hepadnaviruses, however, FVs secrete only small amounts of SVPs. Ubiquitination of the FV Env protein at the CyD of the LP was shown to suppress the intrinsic activity of the FV glycoprotein to induce SVP release. Mutants of the lysine-specifying codons in the LP CyD release large amounts of SVP (Stanke et al., 2005; Stange et al., 2008). Little is known about the mechanisms of FV SVP release. It is not clear if it is a passive process driven mainly by the glycoprotein itself or an active process involving cellular machineries. The latter might be true since SVP release also appears to require the cellular VPS machinery (Stange et al., 2008). However, unlike the release of infectious FV particles, SVP budding was inhibited only by dominant-negative components of late and not by early components of the endosomal sorting complex required for transport (ESCRT) machinery as well (Stange et al., 2005; 2008). Interestingly, the increase of ubiquitination-defective Env protein mutants in SVP release is not at the expense of viral particle release when examined in a replication-competent context (Stanke et al., 2005).

Manipulation of FV capsid membrane interactions

  1. Top of page
  2. Summary
  3. Introduction
  4. Membrane interactions during FV target cell entry
  5. Glycoprotein-dependent release of infectious FV virions
  6. Glycoprotein-mediated release of subviral particles
  7. Manipulation of FV capsid membrane interactions
  8. Retargeting of PFV Gag to different cellular membranes
  9. Pseudotyping of FV capsids
  10. Conclusions
  11. Acknowledgements
  12. References

As a consequence of the unique budding strategy of FVs utilizing this specific interaction of capsid and glycoprotein, pseudotyping of FV vectors with heterologous glycoproteins cannot take place (Pietschmann et al., 1999). This limits the host range of FV vectors to that provided by the cognate FV glycoproteins, which fortunately is extremely broad (Hill et al., 1999; Mergia and Heinkelein, 2003). However, enhancement of naturally poorly permissive target cell populations, by using more favourable heterologous glycoproteins, or even a more selective targeting of specific target cell types, is therefore an attractive challenge. Providing FV vectors with such features requires either a genetic modification of the FV glycoprotein or the development of a FV pseudotyping system.

Since FV capsid proteins appear to lack an authentic membrane-targeting signal, artificial addition of heterologous membrane-targeting domains (MTDs) by genetic fusion to the gag ORF is one approach to potentially overcome this problem. Eastman and Linial (2001) were the first to implement this approach by replacing the N-terminal 10 aa of PFV Gag by the myristoylation signal of the Src kinase protein. Indeed, they could demonstrate a release of viral capsids that was independent of FV Env. However, the released membrane-enveloped capsids were non-infectious even when the cognate viral glycoprotein was coexpressed. Various follow-up studies demonstrated that different MTDs including myristoylation signals of MPMV or human immunodeficiency virus type I (HIV-1) can be used to mediate an Env-independent budding of PFV and feline FV (FFV) capsids (Zhadina et al., 2007; Life et al., 2008; Liu et al., 2011). However, in all cases Env-independent membrane targeting was incompatible with viral replication, resulting in release of non-infectious virions, even when coexpressed with the cognate glycoproteins. Ultrastructural analysis of some MTD-tagged PFV capsids by Life et al. (2008) demonstrated that Myr signal addition does apparently not retarget PFV Gag to the plasma membrane as originally expected. Instead, Gag proteins appear to be still directed to the microtubule organizing centre (MTOC), suggesting that the authentic CTRS signal within Gag is dominant. Furthermore, it revealed that the MTD signals, in particular N-terminal fatty acid addition, interfere with viral morphogenesis. Capsids with aberrant morphologies were observed for all mutants that were myristoylated. Studies for PFV by Life et al. (2008) and for FFV by Liu et al. (2011) also suggested that alterations in the PFV Gag N-terminus interfere with viral replication.

Retargeting of PFV Gag to different cellular membranes

  1. Top of page
  2. Summary
  3. Introduction
  4. Membrane interactions during FV target cell entry
  5. Glycoprotein-dependent release of infectious FV virions
  6. Glycoprotein-mediated release of subviral particles
  7. Manipulation of FV capsid membrane interactions
  8. Retargeting of PFV Gag to different cellular membranes
  9. Pseudotyping of FV capsids
  10. Conclusions
  11. Acknowledgements
  12. References

Research is far away from fully understanding the mechanistic properties of intracellular trafficking and membrane association pathways in FVs. Therefore, we decided to analyse the influence of various MTDs targeting on PFV Gag distribution and on the viral replication cycle. We choose MTDs that are either specific for cellular or endosomal membranes in order to better understand what kind of intracellular trafficking and membrane association pathways are compatible with FV replication. The MTDs used included the complete Gag MA domains of MPMV (MPMV, 107 aa) and HIV-1 (HIV, 136 aa), instead of just their myristoylation signal sequence of about 10 aa employed in the previous studies (Fig. 2A). Further MTDs used for plasma membrane targeting of PFV Gag proteins were the C2 domain (C2, 173 aa) of the ubiquitin ligase WWP1 and as reference to previous studies the v-Src myristoylation signal (SrcM, 10 aa) (Fig. 2A). For targeting to endosomal membranes the Phox homology (PX) domain (PX, 123 aa) of the p40phox subunit of NADPH oxidase and two copies of the Fab1/YOTB/Vac1/EEA 1 (FYVE) domain (FYVE, 78 aa) of Hrs were employed (Fig. 2A). In the previous studies by Eastman and Linial (2001) and Life et al. (2008), the heterologous MTD replaced N-terminal sequences of PFV Gag. Instead, we chose to add the heterologous MTDs to the N-terminus of PFV Gag similar to FFV Gag constructs described recently by Liu et al. (2011). Furthermore, all of our constructs contained a 10 aa PFV Gag p68/p3 cleavage site inserted at the fusion site to allow for potential removal of the N-terminal MTD by FV protease-mediated processing. All MTD fusion constructs also had a haemagglutinin tag and a flexible glycine–serine linker inserted at the fusion site, except the HIV-1 Gag MA.

figure

Figure 2. Manipulation of FV membrane targeting and particle release.

A. Schematic outline of the PFV Gag expression constructs with N-terminal heterologous MTDs. The amino acid positions of the PFV Gag and heterologous MTDs are given. Potential PFV protease cleavage sites are indicated as black arrows, potential HIV cleavage site as grey arrows.

B–F. Biochemical analysis of cellular expression and particle release supported by different PFV Gag packaging constructs. Recombinant viral supernatants were generated essentially as previously described (Lindemann et al., 2001). Briefly, 293T cells were transfected with different Gag packaging constructs in context of a four-component replication-deficient PFV vector system encompassing an EGFP-expressing PFV transfer vector (puc2MD9), as well as Pol (pcziPol) and Env (pczHFVenv EM002) packaging vectors. In samples where the Env packaging vector was omitted pCDNA was added instead. Viral particles were enriched by ultracentrifugation and either subtilisin digested (+ subtilisin) or mock treated (− subtilisin) prior to lysis as previously described (Swiersy et al., 2011). Cell (B and C) and viral (D–F) particle lysates were separated by SDS-PAGE, blotted to nitrocellulose membrane and immunodetected using a mixture of monoclonal antibodies specific for PFV protease and integrase (α-Pol) or Gag and envelope SU subunit (α-Gag + α-Env-SU).

Download figure to PowerPoint

First, we examined the potential of the different Gag fusion proteins to support capsid release in the presence and absence of the cognate glycoprotein. Therefore, 293T cells were transiently transfected with a four-component replication-deficient PFV vector system (Heinkelein et al., 2002), varying the Gag packaging component. All modified Gag constructs were expressed at similar levels in the transfected cells (Fig. 2B). Furthermore, similar Pol and Env expression was observed in all samples examined (Fig. 2B and C). Expression analysis revealed that the introduced N-terminal proteolytic cleavage site was not processed by the viral protease. The potentially bulky PFV PR-RT subunit is unable to access this cleavage site probably due to sterical hindrance. Interestingly, incorporation of a small HIV-PR into HIV-Gag MA capsids, which contain the natural HIV-1 MA/CA-processing site (Fig. 2A), resulted in a removal of the N-terminal MTD tag (data not shown). However, coexpression of HIV-1 PR prevented Env-independent particle release probably as a consequence of premature processing (data not shown).

Analysis of the protein composition of viral particles enriched by ultracentrifugation through sucrose revealed several interesting observations (Fig. 2D–F). First, PFV capsid release was observed only for constructs with MTD targeting the plasma membrane and not endosomal membranes (Fig. 2E). The low amount of PFV capsid protein detected in virus samples of the FYVE-tagged Gag protein appeared to be non-membrane enveloped since it was sensitive to subtilisin digestion (Fig. 2E, compare lanes 13, 14 with lanes 28, 29). In contrast, all other Gag fusion proteins found in particle lysates were subtilisin-resistant (Fig. 2E). Second, in all cases where MTD tagging supported particle export, the release was Env-independent since omission of the PFV Env packaging construct did not abolish capsid protein release in contrast to wild-type PFV Gag (Fig. 2E, lanes 1–10, 16–25). Third, release efficiency of Gag tagged either with the v-Src myristoylation signal (SrcM) or the HIV-1 Gag MA subunit (HIV) appeared superior to wild-type Gag (Fig. 2E, lanes 1–6, 16–21). Particle release of the MPMV Gag MA subunit-tagged (MPMV) or C2 domain-tagged (C2) was only slightly elevated or similar to wild-type PFV Gag (Fig. 2E, lanes 1, 2, 7–10, 16, 17, 22–25). Fourth, in all particle lysates that contained PFV Gag proteins the PFV Pol protein was detected as well (Fig. 2D, lanes 1–10, 16–25). The MPMV Gag MA subunit-tagged construct (MPMV), however, showed very low Pol levels (Fig. 2D, lanes 7, 8, 22, 23). Considering the strong increase in Gag amounts detected in most of the other samples (HIV, SrcM, C2) they also contained reduced relative Pol levels (Fig. 2D, lanes 3–6, 9, 10, 18–21, 24–25). As a result a strongly reduced Gag precursor protein processing at the C-terminus was seen responsible for these tagged Gag proteins (HIV, SrcM, C2) in comparison to wild type, although particle-associated Pol processing appeared normal (Fig. 2D and E, lanes 3–6, 9, 10, 18–21, 24–25). Thus, targeting PFV Gag by heterologous MTDs to the plasma membrane but not to the endosomal membranes is compatible with FV Env-independent budding of FV capsids. Furthermore, addition of plasma membrane MTDs to the N-terminus appears to be incompatible with authentic Gag precursor processing, probably due to a reduced PFV Pol encapsidation efficiency.

The particle morphogenesis of the MTD-tagged mutants expressed in context of the four-component PFV vector system in 293T cells was examined by transmission electron microscopy and exposed several unique phenotypes (Fig. 3). Cells that expressed wild-type Gag and wild-type Env displayed the typical intracellular accumulations of naked capsids and budding structures with prominent Env spikes at different cellular membranes (Fig. 3A and B). No budding structures were detectable in cells expressing endosomal MTD-tagged Gag proteins (Fig. 3K–M). In the cytoplasm of cells expressing the PX-tagged Gag only electron-dense aggregates but no regular capsid structures were seen (Fig. 3K). The electron-dense material may represent aggregated and/or misfolded Gag protein. In contrast, regular shaped capsids were detected between membranous structures in the heavily vacuolized cytoplasm of FYVE-tagged Gag-expressing cells (Fig. 3L and M). However, no budding structures were detectable. In cells expressing the plasma membrane MTD-tagged Gag proteins induction of budding structures with gross changes in their morphology was readily detectable (Fig. 3C–J). V-SrcM-tagged Gag-expressing cells were characterized by the presence of capsid structures with a large-size heterogeneity ranging from wild type-like spherical to tubular structures in the cytoplasm and at cellular membranes (Fig. 3C and D), similarly to previously reported (Life et al., 2008). However, unlike reported we did not observe an accumulation of such structures near the MTOC. Furthermore, underneath the plasma membrane patches of electron-dense material were detected, probably representing oligomerized Gag (Fig. 3D). Some of these patches displayed typical FV Env spike structures on the luminal side (Fig. 3D). On the cell surface, very heterogeneous vesicular structures were observed, displaying mostly dot-like structures on their surface (Fig. 3C). The HIV-1 Gag MA- and MPMV Gag MA-tagged Gag displayed in general similar phenotypes as the v-SrcM-tagged protein (Fig. 3E–H). Differences were the absence of particular structures at intracellular membranes, the even more heterogeneous size of vesicular structures at the cell surface and the absence of the dot-like structures observed in the v-SrcM-tagged samples. The C2-tagged Gag-expressing samples also showed plasma membrane patches of electron-dense material with budding structures of various sizes (Fig. 3I and J). Unique for this mutant was the occurrence of regular shaped FV capsid structures underneath the plasma membrane (Fig. 3J). However, the typical Env spike structures were not detectable, neither on the membrane close to these wild type-like capsid structures nor on more irregular shaped vesicular structures found on the cell surface. Taken together, the ultrastructural analysis indicated that addition of endosomal MTDs to Gag failed to induce budding structures although membrane association is observed. Furthermore, plasma membrane targeting enables induction of budding of mutant PFV particle structures but strongly interferes with normal capsid morphogenesis.

figure

Figure 3. Ultrastructural analysis of particle morphogenesis of MTD-tagged PFV Gag mutants. Electron micrographs showing representative thin sections of transiently transfected 239T cells using different MTD-tagged PFV Gag expression constructs in context of a replication-deficient, four-component PFV vector system as indicated. (A and B) WT-, (C and D) SrcM-, (E and F) HIV-, (G and H) MPMV-, (I and J) C2-, (K) PX- and (L and M) FYVE-tagged PFV Gag. Scale bar: 200 nm.

Download figure to PowerPoint

In a further attempt to characterize the released MTD-tagged PFV particles we examined their infectivity and nucleic acid composition in context of a four-component replication-deficient PFV vector system (Fig. 4). The highest infectivity in combination with the authentic PFV Env protein was detectable in the supernatant of cells expressing the HIV-1 Gag MA-tagged FV capsid protein reaching up to 1% of wild type (Fig. 4, black bars). In supernatants of cells expressing the MPMV Gag MA- or v-SrcM-tagged PFV Gag proteins only very low infectivity levels at 0.05% of wild type were measured. The infectivity values of C2, PX and FYVE domain-tagged Gag-expressing cells were at the detection limit of the assay of about 0.02% of wild type. In contrast, all released MTD-tagged particles packaged similar or even higher relative amounts of viral RNA (Fig. 4, dark grey bars). However, the amount of intraparticle-associated reverse transcribed vDNA was strongly reduced (Fig. 4, light grey bars). Interestingly, SrcM-, HIV-1 Gag MA- and C2-tagged PFV Gag particles displayed similar levels of intraparticle reverse transcription (2–8% of wt), although their relative infectivity varied dramatically (Fig. 4, compare light grey to black bars). For the MPMV Gag MA-tagged particles no intraparticle reverse transcription was detectable.

figure

Figure 4. Infectivity and nucleic acid composition of MTD-tagged PFV Gag mutants. 293T cells were transfected with different Gag packaging constructs in context of the replication-deficient four-component PFV vector system as indicated. Recombinant viral particles were generated as described in legend of Fig. 2. Infectivity (infectivity, black bars) of vector supernatants was examined by an EGFP marker gene transfer assay as previously described (Swiersy et al., 2011). Particle-associated nucleic acids (viral RNA, vRNA dark grey bars; reverse transcript, vDNA light grey bars) were isolated and quantified by Taqman PCR as previously described (Mannigel et al., 2007; Müllers et al., 2011). Mock, uninfected control; ΔEnv, replacement of the PFV Env packaging vector with pCDNA; iRT, replacement of the wild-type PFV Pol packaging vector with a mutant variant thereof with enzymatic inactive reverse transcriptase domain.

Download figure to PowerPoint

Pseudotyping of FV capsids

  1. Top of page
  2. Summary
  3. Introduction
  4. Membrane interactions during FV target cell entry
  5. Glycoprotein-dependent release of infectious FV virions
  6. Glycoprotein-mediated release of subviral particles
  7. Manipulation of FV capsid membrane interactions
  8. Retargeting of PFV Gag to different cellular membranes
  9. Pseudotyping of FV capsids
  10. Conclusions
  11. Acknowledgements
  12. References

The failure of heterologous viral glycoproteins to complement the essential function of the FV Env protein for particle egress is the main reason why FV vectors cannot be pseudotyped (Pietschmann et al., 1999; Lindemann and Rethwilm, 2011). We and others have shown that a FV Env-independent release of FV VLPs can be achieved by fusing heterologous MTDs to FV Gag (see above). However, the permanent membrane attachment apparently interferes strongly with productive infection of target cells, even when the cognate viral Env protein is coexpressed. We recently devised a strategy that replaced the natural FV Gag–Env interaction with a small molecule-controlled heterodimerization system originally developed by Ariad (Ho et al., 2012). The interaction of a mutant PFV Env protein, containing an inactivated LP budding domain, with FV Gag could be restored by this system. Additionally, it enabled a small molecule-dependent release of PFV vector particles with nearly wild-type infectivity. Adaptation of this system by fusing one heterodimerization domain to a MTD, such as the HIV Gag MA subunit, with a second domain fused to the FV Gag protein enabled a very efficient pseudotyping of PFV capsids with a variety of heterologous viral glycoproteins, including vesicular stomatitis virus glycoprotein. Crucial for the success of the system developed is probably the reversibility of FV capsid membrane interaction by using small molecule-controlled heterodimerization domains instead of permanent genetic fusion of a MTD.

Conclusions

  1. Top of page
  2. Summary
  3. Introduction
  4. Membrane interactions during FV target cell entry
  5. Glycoprotein-dependent release of infectious FV virions
  6. Glycoprotein-mediated release of subviral particles
  7. Manipulation of FV capsid membrane interactions
  8. Retargeting of PFV Gag to different cellular membranes
  9. Pseudotyping of FV capsids
  10. Conclusions
  11. Acknowledgements
  12. References

Among retroviruses FVs display a very unique particle egress strategy by the generation of two different types of particle structures that are released from cells. The reason for this phenotype is a result of a special membrane-targeting strategy involving a very specific interaction between the FV capsid and glycoprotein. Unlike other retroviruses the FV glycoprotein appears to have acquired functions that normally reside in the Gag protein. Artificial addition of heterologous MTDs to FV Gag enables Env-independent particle release. However, these virions show strongly aberrant capsid morphologies and are characterized by very low or undetectable infectivity. This might be a consequence of a permanent membrane association due to the very limited processing of FV Gag. In line with this infectious FV particles, with interaction-deficient FV Env or even heterologous glycoproteins, can be generated when a reversible MTD-mediated membrane association is employed instead. Together, these observations suggest that FVs require only a transient interaction of the capsid with cellular membranes during budding and potentially also for uncoating in target cells, which is naturally achieved by the interaction with the cognate viral glycoprotein.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Membrane interactions during FV target cell entry
  5. Glycoprotein-dependent release of infectious FV virions
  6. Glycoprotein-mediated release of subviral particles
  7. Manipulation of FV capsid membrane interactions
  8. Retargeting of PFV Gag to different cellular membranes
  9. Pseudotyping of FV capsids
  10. Conclusions
  11. Acknowledgements
  12. References

This manuscript is dedicated to Hanswalter Zentgraf who passed away on 17 July 2011. We thank B. Hub for excellent technical assistance. This work was supported by grants from the DFG (Li621/3-3, SPP1175 Li621/4-1 + Li621/4-2) to D. L.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Membrane interactions during FV target cell entry
  5. Glycoprotein-dependent release of infectious FV virions
  6. Glycoprotein-mediated release of subviral particles
  7. Manipulation of FV capsid membrane interactions
  8. Retargeting of PFV Gag to different cellular membranes
  9. Pseudotyping of FV capsids
  10. Conclusions
  11. Acknowledgements
  12. References
  • Achong, B.G., Mansell, P.W.A., Epstein, M.A., and Clifford, P. (1971) An unusual virus in cultures from a human nasopharyngeal carcinoma. J Natl Cancer Inst 46: 299307.
  • Duda, A., Stange, A., Lüftenegger, D., Stanke, N., Westphal, D., Pietschmann, T., et al. (2004) Prototype foamy virus envelope glycoprotein leader peptide processing is mediated by a furin-like cellular protease, but cleavage is not essential for viral infectivity. J Virol 78: 1386513870.
  • Duda, A., Lüftenegger, D., Pietschmann, T., and Lindemann, D. (2006) Characterization of the prototype foamy virus envelope glycoprotein receptor-binding domain. J Virol 80: 81588167.
  • Eastman, S.W., and Linial, M.L. (2001) Identification of a conserved residue of foamy virus Gag required for intracellular capsid assembly. J Virol 75: 68576864.
  • Epstein, M.A. (2004) Simian retroviral infections in human beings. Lancet 364: 138139; author reply 139–140.
  • Fischer, N., Heinkelein, M., Lindemann, D., Enssle, J., Baum, C., Werder, E., et al. (1998) Foamy virus particle formation. J Virol 72: 16101615.
  • Goepfert, P.A., Shaw, K., Wang, G., Bansal, A., Edwards, B.H., and Mulligan, M.J. (1999) An endoplasmic reticulum retrieval signal partitions human foamy virus maturation to intracytoplasmic membranes. J Virol 73: 72107217.
  • Heinkelein, M., Dressler, M., Jarmy, G., Rammling, M., Imrich, H., Thurow, J., et al. (2002) Improved primate foamy virus vectors and packaging constructs. J Virol 76: 37743783.
  • Herchenröder, O., Turek, R., Neumann Haefelin, D., Rethwilm, A., and Schneider, J. (1995) Infectious proviral clones of chimpanzee foamy virus (SFVcpz) generated by long PCR reveal close functional relatedness to human foamy virus. Virology 214: 685689.
  • Herchenröder, O., Moosmayer, D., Bock, M., Pietschmann, T., Rethwilm, A., Bieniasz, P.D., et al. (1999) Specific binding of recombinant foamy virus envelope protein to host cells correlates with susceptibility to infection. Virology 255: 228236.
  • Hill, C.L., Bieniasz, P.D., and McClure, M.O. (1999) Properties of human foamy virus relevant to its development as a vector for gene therapy. J Gen Virol 80: 20032009.
  • Ho, Y.P., Schnabel, V., Swiersy, A., Stirnnagel, K., and Lindemann, D. (2012) A small-molecule-controlled system for efficient pseudotyping of prototype foamy virus vectors. Mol Ther 20: 11671176.
  • Lecellier, C.H., Neves, M., Giron, M.L., Tobaly-Tapiero, J., and Saib, A. (2002) Further characterization of equine foamy virus reveals unusual features among the foamy viruses. J Virol 76: 72207227.
  • Lehmann-Che, J., Giron, M.L., Delelis, O., Löchelt, M., Bittoun, P., Tobaly-Tapiero, J., et al. (2005) Protease-dependent uncoating of a complex retrovirus. J Virol 79: 92449253.
  • Lehmann-Che, J., Renault, N., Giron, M.L., Roingeard, P., Clave, E., Tobaly-Tapiero, J., et al. (2007) Centrosomal latency of incoming foamy viruses in resting cells. PLoS Pathog 3: e74.
  • Life, R.B., Lee, E.G., Eastman, S.W., and Linial, M.L. (2008) Mutations in the amino terminus of foamy virus Gag disrupt morphology and infectivity but do not target assembly. J Virol 82: 61096119.
  • Lindemann, D., and Rethwilm, A. (2011) Foamy virus biology and its application for vector development. Viruses 3: 561585.
  • Lindemann, D., Pietschmann, T., Picard-Maureau, M., Berg, A., Heinkelein, M., Thurow, J., et al. (2001) A particle-associated glycoprotein signal peptide essential for virus maturation and infectivity. J Virol 75: 57625771.
  • Linial, M.L., Fan, H., Hahn, B., Löwer, R., Neil, J., Quackenbush, S., et al. (2005) Retroviridae. In Virus Taxonomy. Fauquet, C.M. , Mayo, M.A. , Maniloff, J. , Desselberger, U. , and Ball, L.A. (eds). London: Elsevier Academic Press, pp. 421440.
  • Liu, Y., Kim, Y.B., and Löchelt, M. (2011) N-terminally myristoylated feline foamy virus Gag allows Env-independent budding of sub-viral particles. Viruses 3: 22232237.
  • Mannigel, I., Stange, A., Zentgraf, H., and Lindemann, D. (2007) Correct capsid assembly mediated by a conserved YXXLGL motif in prototype foamy virus Gag is essential for infectivity and reverse transcription of the viral genome. J Virol 81: 33173326.
  • Mergia, A., and Heinkelein, M. (2003) Foamy virus vectors. Curr Top Microbiol Immunol 277: 131159.
  • Müllers, E., Uhlig, T., Stirnnagel, K., Fiebig, U., Zentgraf, H., and Lindemann, D. (2011) Novel functions of prototype foamy virus Gag glycine- arginine-rich boxes in reverse transcription and particle morphogenesis. J Virol 85: 14521463.
  • Nasimuzzaman, M., and Persons, D.A. (2012) Cell Membrane-associated heparan sulfate is a receptor for prototype foamy virus in human, monkey, and rodent cells. Mol Ther 20: 11581166.
  • Picard-Maureau, M., Jarmy, G., Berg, A., Rethwilm, A., and Lindemann, D. (2003) Foamy virus envelope glycoprotein-mediated entry involves a pH-dependent fusion process. J Virol 77: 47224730.
  • Pietschmann, T., Heinkelein, M., Heldmann, M., Zentgraf, H., Rethwilm, A., and Lindemann, D. (1999) Foamy virus capsids require the cognate envelope protein for particle export. J Virol 73: 26132621.
  • Pietschmann, T., Zentgraf, H., Rethwilm, A., and Lindemann, D. (2000) An evolutionarily conserved positively charged amino acid in the putative membrane-spanning domain of the foamy virus envelope protein controls fusion activity. J Virol 74: 44744482.
  • Plochmann, K., Horn, A., Gschmack, E., Armbruster, N., Krieg, J., Wiktorowicz, T., et al. (2012) Heparan sulfate is an attachment factor for foamy virus entry. J Virol 86: 1002810035.
  • Rethwilm, A. (2010) Molecular biology of foamy viruses. Med Microbiol Immunol 199: 197207.
  • Shaw, K.L., Lindemann, D., Mulligan, M.J., and Goepfert, P.A. (2003) Foamy virus envelope glycoprotein is sufficient for particle budding and release. J Virol 77: 23382348.
  • Stange, A., Mannigel, I., Peters, K., Heinkelein, M., Stanke, N., Cartellieri, M., et al. (2005) Characterization of prototype foamy virus gag late assembly domain motifs and their role in particle egress and infectivity. J Virol 79: 54665476.
  • Stange, A., Lüftenegger, D., Reh, J., Weissenhorn, W., and Lindemann, D. (2008) Subviral particle release determinants of prototype foamy virus. J Virol 82: 98589869.
  • Stanke, N., Stange, A., Lüftenegger, D., Zentgraf, H., and Lindemann, D. (2005) Ubiquitination of the prototype foamy virus envelope glycoprotein leader peptide regulates subviral particle release. J Virol 79: 1507415083.
  • Stirnnagel, K., Lüftenegger, D., Stange, A., Swiersy, A., Müllers, E., Reh, J., et al. (2010) Analysis of prototype foamy virus particle-host cell interaction with autofluorescent retroviral particles. Retrovirology 7: 45.
  • Stirnnagel, K., Schupp, D., Dupont, A., Kudryavtsev, V., Reh, J., Müllers, E., et al. (2012) Differential pH-dependent cellular uptake pathways among foamy viruses elucidated using dual-colored fluorescent particles. Retrovirology 9: 71.
  • Swiersy, A., Wiek, C., Reh, J., Zentgraf, H., and Lindemann, D. (2011) Orthoretroviral-like prototype foamy virus Gag-Pol expression is compatible with viral replication. Retrovirology 8: 66.
  • Wilk, T., de Haas, F., Wagner, A., Rutten, T., Fuller, S., Flügel, R.M., and Löchelt, M. (2000) The intact retroviral Env glycoprotein of human foamy virus is a trimer. J Virol 74: 28852887.
  • Yu, S.F., Eastman, S.W., and Linial, M.L. (2006) Foamy virus capsid assembly occurs at a pericentriolar region through a cytoplasmic targeting/retention signal in Gag. Traffic 7: 966977.
  • Zhadina, M., McClure, M.O., Johnson, M.C., and Bieniasz, P.D. (2007) Ubiquitin-dependent virus particle budding without viral protein ubiquitination. Proc Natl Acad Sci USA 104: 2003120036.