Spumaviruses, also known as foamy viruses (FVs), represent the only genus of the retroviral subfamily spumaretrovirinae, and resemble complex retroviruses with respect to their genome structure. The FV replication strategy deviates in many aspects from that of orthoretroviruses [reviewed in ]. Interestingly, many of the unique features of FVs are more reminiscent of another family of reverse transcribing viruses, the hepadnaviridae [reviewed in ]. This includes the expression of Pol as a separate protein, instead of the Gag-Pol fusion proteins typical of orthoretroviruses [reviewed in ]. As a consequence, FVs have a specific strategy to ensure Pol particle incorporation, essential for generation of infectious virions. Both Gag and Pol proteins of FVs bind to full-length genomic viral transcripts. Additionally, protein-protein interactions between Gag and Pol seem to be involved in this assembly process [4–6]. Other aspects of FV assembly are also unique among retroviruses; for example, while FV Gag can preassemble by itself into capsid structures at the cellular microtubule-organizing-center (MTOC) like B/D type orthoretroviruses, it apparently lacks membrane-targeting signals. Therefore, such particles are not released from the cell as virus-like-particles as observed for other retroviruses [reviewed in ]. Similar to Hepatitis B virus (HBV), FV particle budding and release are instead dependent on co-expression of the cognate viral envelope (Env) protein; moreover, this function of FV Env that cannot be complemented by expression of heterologous viral glycoproteins [reviewed in ]. A specific interaction between the cytoplasmic N-terminus of the FV Env glycoprotein, involving the leader peptide (LP) and a conserved W10XXW13 motif, and the N-terminal region of the FV Gag protein, is essential for particle egress. FV Env-independent capsid release can be achieved experimentally by artificial N-terminal fusion of heterologous membrane-targeting signals to the FV Gag. However, these VLPs are non-infectious even when co-expressed with the cognate viral glycoprotein [8–10]. Finally, the structural organization of the FV Gag protein deviates significantly from orthoretroviruses. Unlike orthoretroviral Gag proteins, FV Gag is not processed into separate matrix (MA), capsid (CA) and nucleocapsid (NC) subunits. In fact, only a limited proteolysis is observed during FV particle morphogenesis, resulting in the removal of a C-terminal 3 kD peptide. Both the uncleaved precursor p71Gag and the larger p68Gag cleavage product are incorporated into the FV capsid, where they are found in ratios of 1:1 to 1:4 in released infectious viral particles . Although the FV Gag protein harbors many functional motifs described for other retroviruses (such as an PSAP late assembly (L)-domain, a cytoplasmic targeting and retention signal (CTRS) to mediate assembly at the MTOC, a coiled-coil domain essential for assembly, and a YXXLDL motive important for capsid morphology and reverse transcription), other motifs are either missing from FV Gag or if present, are unique amongst retroviruses [8, 12–15]. This includes the lack of C-terminal Cys-His boxes in Gag implicated in retroviral RNA packaging [reviewed in ]. Instead up to three glycine-arginine-rich sequences (GR-boxes) are found in the C-terminal region of FV Gag. GR-I was reported to bind to nucleic acids and was originally implicated in RNA binding, but this was recently challenged and another function as an interaction motif for the Gag-Pol interaction during Pol particle incorporation was described [4, 16]. GR-II harbors a nuclear localization signal sequence responsible for predominant nuclear targeting of FV Gag at certain time points during viral replication [16, 17]. Furthermore, recently a chromatin-binding site (CBS) within GR-II was identified mediating attachment of FV Gag to host chromosomes .
In recent years, the combination of fluorescently labeled virions with modern imaging techniques has proven to be a powerful tool to study replication in a variety of viral systems. These methods have been particularly useful for dissecting assembly and entry pathways [reviewed in ]. With respect to retroviruses, single virus tracking has revealed that Murine Leukemia Virus (MLV) infection induces establishment of filopodial bridges that enable efficient cell-to-cell transmission; has allowed the quantitation of individual HIV particle genesis in real time; and enabled detailed analysis of the very earliest events during HIV attachment to target cells [20–22].
Further analysis of the FV replication strategy would profit greatly from the availability of functional fluorescent FV particles. For example, the exact cellular location of FV Gag - Env interaction could be determined and examined by time-lapse microscopy. Originally it was thought to occur at the membrane of the endoplasmic reticulum, since FV Env contains an ER retrieval signal and budding seemed to occur at intracellular membranes, which are believed to be the ER. However, Yu et al. reported recently a significant Gag - Env co-localization only in compartments containing Golgi-specific marker proteins, in a study using FV infected fibroblasts and immunostaining of fixed samples . Similarly, the cellular location of the Gag - Pol interaction is currently unknown, and its identification would contribute to the understanding of FV Pol particle incorporation mechanism. Furthermore, very little is known about the sequential events leading to FV entry of target cells, and live imaging of FV uptake could lead to insights into the entry mechanism of these unusual retroviruses.
Currently, it is thought that FV particles bind to a ubiquitous, but as yet unidentified, cellular receptor. This is based largely on the observation that FVs are unique amongst retroviruses in having an extremely broad host range [24, 25]. FV vectors can transduce even bird or reptile cells. Indeed, a species or cell type that is completely resistant to FV Env-mediated transduction has not been reported. After attachment, FV capsids apparently are endocytosed, gaining access to the cytoplasm by a FV Env-mediated pH-dependent fusion process, and seem to migrate to the centrosome by piggybacking on dynein/dynactin motor complexes [26, 27]. There they can reside for long periods of time until disassembling and progressing towards nuclear entry of the FV preintegration complex, induced by yet uncharacterized cellular signals .
A few previous studies have employed enhanced green fluorescent protein (EGFP) tagged FV Gag proteins for cellular assays [9, 18, 26]. Petit et al.  and Tobaly-Tapiero et al.  used different, transiently-expressed N-terminal tagged Gag proteins to characterize the centrosome-targeting and chromatin-binding motifs in PFV Gag. The influence of L-domain mediated Gag ubiquitination on retroviral budding was examined by Zhadina et al.  using artificially membrane-targeted, Env-independently budding PFV Gag protein containing a C-terminal GFP-tag. However, the functional consequences of tagging the FV Gag proteins, compared to untagged wild type FV Gag protein, were not examined in these studies.
In this study, we systematically analyzed the influence of different protein tags on PFV Gag's capacity to support FV replication using recombinant replication-deficient FV vector particles that are capable of single-round infections. We succeeded in identifying for the first time autofluorescent protein (AFP)-tagged PFV Gag constructs that allow generation of fluorescent PFV particles with nearly wild type functionality; these constructs provide a powerful tool for analysis of PFV replication steps by modern imaging techniques. With this tool, a particle-binding assay for target cells was established. In combination with high-titer FV Env containing retroviral vector supernatants, it was used to identify two cell lines that are resistant to PFV Env-mediated marker gene transfer. Interestingly, these cells still displayed retroviral particle attachment in a FV Env-specific manner. Further characterization of the resistance to FV Env-mediated virus entry in these cell lines might ultimately lead to the discovery of currently unknown cellular molecules essential for the early stages of FV infection in target cells.