Murine leukemia virus p12 functions include hitchhiking into the nucleus

When a retrovirus infects a new host cell, its genomic RNA is copied into double-stranded (ds) DNA by the reverse transcriptase enzyme present in the incoming virus particle. This DNA synthesis takes place in the cytoplasm of the cell; the DNA must then make its way into the nucleus before it can be inserted into the chromosomal DNA by the integrase enzyme (which is also carried in the incoming virus). In PNAS, Schneider et al. (1) elegantly dissect some of the complexities of nuclear entry of Moloney murine leukemia virus (MuLV) DNA.

Retroviruses are divided into two subfamilies (i.e., orthoretroviruses and spumaretroviruses). There are six orthoretrovirus genera: alpha-, beta-, gamma-, delta-, epsilon-, and lentiviruses. MuLV is a prototypical gammaretrovirus and has served as the basis for a major class of gene-therapy vectors now under development. A significant difference between gammaretroviruses and many other retroviruses, including the lentivirus HIV-1, is that gammaretroviruses can only infect proliferating cells (2), whereas HIV-1 can infect nondividing cells (3). This ability is a crucial element in HIV-1 biology.

When the dsDNA copy of the retroviral genome enters the nucleus, it is part of the “pre-integration complex” (PIC), which also contains proteins from the virus particle, including integrase (IN) and capsid (CA) (the major structural protein of the mature virus particle), in addition to proteins from the host cell. HIV-1 CA interacts, directly or indirectly, with nuclear import machinery (4); this interaction enables the HIV-1 PIC to enter the interphase nucleus, leading to the successful infection of a nondividing cell.

Unlike HIV-1, gammaretroviral PICs cannot cross the nuclear membrane. This is why gammaretroviral infections require dividing cells, in which the nuclear membrane has broken down. The work by Schneider et al. (1) sheds light on how MuLV ensures that, once the nuclear membrane has reformed at the end of mitosis, the PIC will be within the nucleus.

The basic building block of retrovirus particles is the virus-coded Gag polyprotein. After the particle is released from the virus-producing cell, Gag is cleaved by the viral protease into a series of cleavage products. In all orthoretroviruses, these include matrix (MA), CA, and nucleocapsid. However, in gammaretroviruses, there are two cleavages between MA and CA, leading to the release of a fourth Gag fragment, termed p12. The functions of p12, which is only 84 aa in length, are not well known, despite many years of intensive study of MuLVs (5). Early mutational analysis showed that a short motif within p12, PPPY, is the MuLV “late” domain: binding of PPPY to a cellular protein is required for the release of assembled virus particles from the surface of the virus-producing cell (6). Furthermore, residues near both ends of p12 are important in the infection of new cells by MuLV (7, 8), whereas a central region of p12 is remarkably tolerant of insertions and substitutions (9). The present work by Schneider et al. (1) has used one of the first p12 mutants, termed PM14, in which five consecutive residues near the p12 C terminus have been replaced with alanines; particles containing PM14 p12 are noninfectious (8).

It is now known that p12 is part of the MuLV PIC (10) and tethers the MuLV PIC to the chromosomes in mitotic cells (11). Evidently, this is how the PIC is localized to the interphase nucleus after cell division, enabling integration of the viral DNA into the host DNA. The present work (1) is a detailed dissection of this recently delineated p12 function and adds to our understanding of the question of how gammaretroviruses infect new host cells.

Hitchhiking into the Nucleus

The work by Schneider et al. (1) takes advantage of the fact that domains for tethering to mitotic chromosomes or “chromatin-binding sequences” (CBSs) have previously been described in several proteins of other viruses. These include latency-associated nuclear antigen (LANA) and E2 protein from the DNA viruses Kaposi’s sarcoma herpesvirus (KSHV) (12) and human papillomavirus (HPV) (13), respectively, and a short stretch from the Gag protein of prototype foamy virus (PFV), a spumaretrovirus (14). In KSHV and HPV, the CBSs are used in cells in which the viral genome is present but not directing the production of progeny viruses; tethering to chromosomes ensures that the viral genome, like the chromosomes themselves, is faithfully distributed to the two daughter cells when the cell divides. The PFV CBS seems to function early in infection, in direct analogy to MuLV. In the experiments reported by Schneider et al., CBSs from these viruses were inserted into p12 bearing the PM14 mutation, and the ability of the resulting chimeric virus to infect new cells was measured.

In all three cases, MuLV genomes were obtained in which the heterologous CBSs complemented the PM14 defect. This fundamental result, partially foreshadowed in earlier studies (7, 11), implies that the p12 residues mutated in PM14 contribute to chromatin-binding in MuLV. In addition, however, Schneider et al. examined each case in great detail and were able to extract a remarkable amount of information about MuLV’s requirements from their experiments, including requirements regarding the placement of the added CBS within p12, and regarding apparent incompatibility between certain E2 sequences and a central stretch of p12. Nontethering mutants of LANA and PFV Gag failed to complement PM14 MuLV.

Retrovirus replication is characterized by a high mutation rate. Thus, viruses produced during replication of a poorly infectious virus may include mutants that replicate more efficiently than their parent; these will be rapidly selected during virus passage. The action of natural selection on the broad spectrum of spontaneous mutants is often a rich source of information about the function(s) of viral components. Schneider et al. (1) exploited this phenomenon to learn considerably more about the contribution of the heterologous CBSs to MLV replication.

“Strength” of p12 Tethering to Mitotic Chromosomes

The foregoing discussion has dealt with the ability of MLVs with chimeric p12 proteins to replicate. However, each p12 protein was also characterized with respect to its tendency to associate with mitotic chromosomes. In each case, a fusion of the p12 with green fluorescent protein (GFP) was expressed in mammalian cells, and the degree of colocalization of the fluorescent p12-GFP with the chromosomes was directly measured by confocal microscopy.

As expected, p12-GFP showed a significant tendency to colocalize with the chromosomes, whereas PM14 p12-GFP did not. Addition of a CBS from either PFV, KSHV, or HPV restored targeting of PM14 p12-GFP to mitotic chromosomes. Significantly, however, this restoration was not necessarily correlated with recovery of infectivity in the MuLV containing the same chimeric p12. For example, the p12-GFP fusion with an extended KSHV LANA CBS showed an extremely high association with mitotic chromosomes (far higher than that of wild-type p12-GFP). The corresponding MuLV chimera, however, did not replicate to a detectable level, but generated revertants that did. Strikingly, the chromosomal colocalization of the PM14 p12-GFP containing the revertant tethers, although detectable, was far weaker than that of the parental PM14 p12/KSHV LANA-GFP chimera. In other words, there was strong selection for MuLVs with the optimum, modest level of tethering ability: stronger than that of PM14 MuLV, but weaker than that of PM14 MuLV containing the KSHV LANA insert.

Specificity of Integration of the MuLV Genome into Host Chromatin

Schneider et al. (1) also surveyed the integration-site preferences of their replication-competent chimeric MuLVs. MuLV tends to integrate its genome in the vicinity of transcriptional start sites (15). This targeting is specified by IN (in collaboration with host-cell proteins) and varies between different retroviral genera. Remarkably, all of the chimeras shared this tendency with wild-type MuLV: none of the heterologous CBSs had any influence on the targeting of the PIC at integration. Thus, the p12-mediated tethering to mitotic chromosomes is distinct from the IN-directed association with interphase chromatin leading to integration.

Implications of the Findings

Taken together, these observations suggest the following model (Fig. 1): when an MuLV particle infects a new cell, its RNA is copied into DNA in the cytoplasm. Then, when the cell enters mitosis and its nuclear membrane breaks down, the PIC binds to a chromosome via its p12. After mitosis, the PIC (or part of it), now within the nucleus, is released from the chromatin; finally, the DNA is inserted by IN into interphase chromatin. The tethering to the mitotic chromosome must be modest, presumably because the viral complex must be released before it can be integrated.

Fig. 1.

HIV-1 and MuLV take different routes to the nucleus. Upon infection, both viruses copy their RNA to DNA in the cytoplasm (1). However, in HIV-1, the PIC (DNA plus associated proteins) enters the nucleus (2), and the DNA is integrated into chromosomal DNA (3). In MuLV, the PIC binds to a chromosome during mitosis (2). It is released within the nucleus after mitosis (3), and the DNA is then integrated (4).

The analysis of selected revertants in these experiments will undoubtedly teach us more about chromosomal tethering. Questions that remain unanswered include the significance of placement within p12 for proper tethering function; the specific role of arginine residues, which seem associated with high chromosomal localization, in tethering; and the effect, if any, of serine phosphorylation upon tethering. It will be fascinating to learn more about the “handoff” of viral DNA from its mitotic tether to interphase chromatin after cell division. Finally, it is notable that p12 is the locus for MuLV’s CBS equivalent, as well as for its late domain: it seems likely that the structures of MA, CA, and NC are tightly constrained by the requirements of their primary roles, leaving p12, the remaining product of MuLV Gag cleavage, to carry several short motifs whose function is to interact with specific cellular proteins.