Complementary roles of multiple nuclear targeting signals in the capsid proteins of the parvovirus minute virus of mice during assembly and onset of infection

Abstract

This report describes the distribution of conventional nuclear localization sequences (NLS) and of a beta-stranded so-called nuclear localization motif (NLM) in the two proteins (VP1, 82 kDa; VP2, 63 kDa) forming the T=1 icosahedral capsid of the parvovirus minute virus of mice (MVM) and their functions in viral biogenesis and the onset of infection. The approximately 10 VP1 molecules assembled in the MVM particle harbor in its 142-amino-acid (aa) N-terminal-specific region four clusters of basic amino acids, here called BC1 (aa 6 to 10), BC2 (aa 87 to 90), BC3 (aa 109 to 115), and BC4 (aa 126 to 130), that fit consensus NLS and an NLM placed toward the opposite end of the polypeptide (aa 670 to 680) found to be necessary for VP2 nuclear uptake. Deletions and site-directed mutations constructed in an infectious MVM plasmid showed that BC1, BC2, and NLM are cooperative nuclear transport sequences in singly expressed VP1 subunits and that they conferred nuclear targeting competence on the VP1/VP2 oligomers arising in normal infection, while BC3 and BC4 did not display nuclear transport activity. Notably, VP1 proteins mutated at BC1 and -2, and particularly with BC1 to -4 sequences deleted, induced nuclear and cytoplasmic foci of colocalizing conjugated ubiquitin that could be rescued from the ubiquitin-proteasome degradation pathway by the coexpression of VP2 and NS2 isoforms. These results suggest a role for VP2 in viral morphogenesis by assisting cytoplasmic folding of VP1/VP2 subviral complexes, which is further supported by the capacity of NLM-bearing transport-competent VP2 subunits to recruit VP1 into the nuclear capsid assembly pathway regardless of the BC composition. Instead, all four BC sequences, which are located in the interior of the capsid, were absolutely required by the incoming infectious MVM particle for the onset of infection, suggesting either an important conformational change or a disassembly of the coat for nuclear entry of a VP1-associated viral genome. Therefore, the evolutionarily conserved BC sequences and NLM domains provide complementary nuclear transport functions to distinct supramolecular complexes of capsid proteins during the autonomous parvovirus life cycle.

FIG. 1.

Construction of a VP-1-only MVM genome. (A) Genome organization of the parvovirus MVM in the minor intron region. The P38 promoter and the positions of the two donor (D1 and D2) and the two acceptor (A1 and A2) sites of splicing are indicated using MVMi numbering. VP1 is made from a minor mR3 species that results from the use of the D2 donor (nt 2318) and the A2 acceptor (nt 2399), whereas most viral transcripts are spliced at the D1 donor (nt 2281), thus omitting the VP1 start codon, yielding messengers MR3 and rR3 that use either A1 (nt 2377) or A2 acceptors and express the major VP2 protein. ORF, open reading frame. (B) Inactivation of the minor splicing donor D1 of the MVM genome. The VP1 start codon and the three point genetic changes introduced in the donor splicing site 1 (D1) to produce the mutant plasmid VP1/ΔVP2 are underlined. For simplicity, the alterations that these mutations introduce in the carboxy-terminal region of the NS2 protein isoforms (17) generated from the R2 messengers have not been outlined. (C) Pattern of VP protein expression from the MVM splice donor mutant genome. Protein extracts from NB324K cells transfected with the indicated plasmids and metabolically [³⁵S]Met labeled 16 to 48 h posttransfection were immunoprecipitated with capsid antiserum (MVM) or with serum raised against the VP1-specific N-terminal sequence (VP1). The positions of the VP proteins obtained from [³⁵S]Met-labeled and gradient-purified MVMi capsid (C) are indicated on the left.

FIG. 2.

VP1 nuclear targeting sequences. (A) Distribution of basic amino acid sequences along the VP1 protein. The four clusters of basic sequences (BC1 to BC4) in the VP1-specific N-terminal domain and the NLM in the carboxy-terminal region shared with VP2 are indicated. (B) Subcellular localization of the VP1 mutant proteins. Shown are the nomenclature and genotypes of the genomic MVMi mutants constructed in the VP1 basic sequences and the subcellular distribution of the VP1 mutant proteins in the transfected-cell population. Inactivating mutations in the BC sequences (boxes) and in the NLM (arrows) are represented by crosses. The percentages are the average values from more than 300 stained cells scored 40 to 48 h posttransfection from at least two independent experiments. The phenotypes were examined by epifluorescence with a Zeiss Axiophot microscope and classified in three categories: mostly nuclear (N > C), mixed (N = C), and cytoplasmic (N < C).

FIG. 3.

Sequences involved in the nuclear targeting of VP1/VP2 complexes. Shown are a series of MVMi site-directed and deletion mutants constructed in the VP1-specific BC sequences and in the NLM domain. For the sake of simplicity, other viral genotypes carrying VP1 mutations close to the ΔBC2-3 (89-125 deletion mutant) or the ΔBC1-4 (80-128 deletion mutant) giving similar phenotypes are not depicted. Genomic plasmids carrying the indicated mutations were transfected into NB324K cells, and the subcellular distributions of VP1 and of the VP antigen (VP1 plus VP2) were monitored in the same cells 40 to 48 h afterwards by IF with the α-VP1 and α-VPs antisera, respectively. Phenotypic characterization was performed as for Fig. 2.

FIG. 4.

Capsid formation by VP mutant proteins. Shown is an IF confocal analysis of the subcellular distribution of VP1 and MVMi capsid in cells transfected with the indicated viral plasmids. Double staining was done with a VP1-specific polyclonal antiserum (VP1) and with an MVM capsid MAb (CAPSID). The predominant phenotype for each of the indicated plasmids is shown. Left, VP1-only mutants; right, VP1 mutants with wt or mutant VP2 coexpression.

FIG. 5.

Confocal analysis of Ub conjugation to VP mutant proteins. NB324K cells were processed for IF 40 h posttransfection and costained with an antiserum raised against MVM capsid (MVM; 1/200 dilution) and a MAb recognizing conjugated Ub (FK2; 1/500 dilution). The panels correspond to representative fields of cells transfected with the following plasmids: VP1ΔBC1-2ΔNLM/ΔVP2 (A and B), VP1ΔBC1-2ΔNLM/VP2ΔNLM (C and D), VP1ΔBC1-4ΔNLM /ΔVP2 (E and F), VP1ΔBC1-4ΔNLM/VP2ΔNLM (G and H), and wt MVMi (I and J).

FIG. 6.

Roles of the BC sequences of VP1 in the onset of infection. Plasmid constructs harboring the wt VP2 protein sequence and the indicated genetic changes in the VP1-specific sequence were transfected into NB324K cells, and intracellular DNA-filled virions were harvested and purified by CsCl gradients. (A) Viral DNA amplification. Monolayers of 10⁶ NB324K cells were inoculated with the purified virions, and cell-associated low-molecular-weight DNA was isolated at 0 and 24 h p.i. (lanes 0 and 24) and analyzed by Southern blotting with a ³²P-labeled MVM probe. Exposure was for 48 h with intensifying screening at −70°C. Lanes V, viral genomes extracted from purified particles; lanes M, mock-infected cultures. Purified BC mutant virions contained various amounts of mRF as previously described for ΔVP1 virions (62), the meaning of which is unclear. ss, single-stranded viral genomes. (B) IF analysis of virion infectivity. Cells grown on coverslips were inoculated with normalized numbers of the indicated purified virions and 24 h afterward fixed and stained by IF with VP antiserum. The numbers of scored cells showing VP synthesis are the averages of two independent inoculations.