A New Sendai Virus Vector Deficient in the Matrix Gene Does Not Form Virus Particles and Shows Extensive Cell-to-Cell Spreading

Abstract

A new recombinant Sendai virus vector (SeV/ΔM), in which the gene encoding matrix (M) protein was deleted, was recovered from cDNA and propagated in a packaging cell line expressing M protein by using a Cre/loxP induction system. The titer of SeV/ΔM carrying the enhanced green fluorescent protein gene in place of the M gene was 7 × 10⁷ cell infectious units/ml or more. The new vector showed high levels of infectivity and gene expression, similar to those of wild-type SeV vector, in vitro and in vivo. Virus maturation into a particle was almost completely abolished in cells infected with SeV/ΔM. Instead, SeV/ΔM infection brought about a significant increase of syncytium formation under conditions in which the fusion protein was proteolytically cleaved and activated by trypsin-like protease. This shows that SeV/ΔM spreads markedly to neighboring cells in a cell-to-cell manner, because both hemagglutinin-neuraminidase and active fusion proteins are present at very high levels on the surface of cells infected with SeV/ΔM. Thus, SeV/ΔM is a novel type of vector with the characteristic features of loss of virus particle formation and gain of cell-to-cell spreading via a mechanism dependent on the activation of the fusion protein.

Sendai virus (SeV) belongs to the genus Respirovirus of the family Paramyxoviridae. SeV replicates independently of cellular nuclear functions and does not have a DNA phase during its life cycle, so the possible transformation of cells via the integration of the viral genetic information into the cellular genome is not a concern (25). All members of this family and other families of negative-strand RNA viruses encode a matrix protein (M) that is considered to play a central role in virion formation (12, 38, 46, 47). Virions are formed in so-called lipid rafts on the cell membrane (41), as has been demonstrated for influenza virus (1), measles virus (28), and SeV (2). At these sites, M protein is thought to enhance virion formation by concentrating the envelope protein (also referred to as the spike protein) and viral ribonucleoprotein (RNP). It has been demonstrated that M protein indeed binds to the cytoplasmic tail of the spike proteins of influenza virus (49) and SeV (37) and to the RNPs of influenza virus (35), human parainfluenza virus, and SeV (9). Further, M proteins have been reported to form oligomers with themselves in the cases of SeV (17) and vesicular stomatitis virus (13, 14). Thus, M protein has the capacity to mediate the binding of many of these viral components to lipids. In addition, it has been found that overexpression of the M proteins of SeV (43), human parainfluenza virus (8),and vesicular stomatitis virus (20) causes the budding of virus-like particles (VLPs). These findings indicate that M protein functions as a driving force for virus assembly and budding.

The genome of SeV is a linear, nonsegmented, negative-strand RNA of 15,384 nucleotides. It contains six major genes, which are arranged in tandem on its genome, and it is tightly encapsidated with the nucleoprotein (NP) and is further complexed with the viral phosphoprotein (P) and large protein (L) (the catalytic subunit of the polymerase) (15). This RNP complex constitutes the internal core structure of the virion. The viral envelope contains two spike proteins, hemagglutinin-neuraminidase (HN) and fusion protein (F), which mediate the attachment of virions and the penetration of RNPs into infected cells, respectively. M protein, as noted above, functions in virus assembly and budding. As SeV infects and replicates in most mammalian cells, including human cells, and directs high-level expression of the genes it carries, an SeV vector is considered to be very promising for use in gene therapy (30, 40, 45). In particular, F-deficient SeV (SeV/ΔF) is considered to be safe because it has been modified so that it is nontransmissible due to loss of the F gene in its genomic RNA (26). Considering the nontransmissible VLP formation observed for F-deficient (27) and HN-deficient (42) SeVs, the reduction of nontransmissible VLP formation is one aim to achieve for the next-generation SeV vectors, although it has not been proven that nontransmissible VLP formation causes any kind of adverse effect in vivo. As M protein is considered to play a central role in the virion formation of SeV, deleting the M gene from the SeV genome should be a highly effective way of suppressing VLP formation.

Recovery of M-deficient viruses has been reported for measles virus (6) and rabies virus (33), and the particle-forming capacities of these mutant viruses were markedly inhibited. However, neither of these viruses is suitable for use as a gene transfer vector, because neither virus is produced at high titer. The production process must be considered when a modified virus vector is used industrially for purposes such as gene therapy. In the present study, an M gene-deficient SeV (SeV/ΔM) vector was produced at high titer by using helper cells expressing M proteins. The recovered SeV/ΔM was infectious in vitro and in vivo, and the release of VLPs from cells infected with SeV/ΔM was greatly suppressed. Moreover, cells infected with SeV/ΔM showed marked cell-cell fusion with syncytium formation under conditions in which F was proteolytically cleaved and activated to F₁ and F₂ from the inactive precursor F₀ protein. SeV/ΔM can be used as a new type of SeV vector that has lost virus maturation into particles and gained extensive cell-to-cell spreading.

MATERIALS AND METHODS

Cells and viruses.

Two monkey kidney cell lines, LLC-MK₂ and CV-1, were cultured in Eagle minimal essential medium (MEM) (Gibco-BRL, Rockville, Md.) supplemented with 10% heat-inactivated fetal bovine serum (FBS). A rat myoblast cell line, A-10, was cultured in Dulbecco's modified Eagle MEM (Gibco-BRL) supplemented with 10% heat-inactivated FBS. All cells were cultured in a humidified atmosphere of 5% CO₂-95% air at 37°C. The genome of the attenuated SeV Z strain was used as the starting material for genome modifications in this study. Wild-type SeV carrying the enhanced green fluorescent protein (GFP) gene inserted before the NP open reading frame (SeV + 18GFP) (36) was grown in 10-day-old embryonated chicken eggs (23). F and M gene-deficient SeV vectors were prepared by using stable transformed LLC-MK₂ cells carrying the F gene (LLC-MK₂/F7) (26) and the M (and F) genes (LLC-MK₂/F7/M62) (this report), respectively. An adenovirus vector, AxCANCre (22), expressing Cre recombinase was used for the induction of F protein in LLC-MK₂/F7 cells (cells thus induced are referred to as LLC-MK₂/F7/A) and of M (and F) proteins in LLC-MK₂/F7/M62 cells (cells thus induced are referred to as LLC-MK₂/F7/M62/A). Recombinant vaccinia virus vTF7-3 (11) carrying T7 RNA polymerase was inactivated with psoralen and long-wave UV irradiation and then used for RNP recovery experiments (44).

Antibodies.

A rabbit polyclonal anti-M antibody, N-39F, was raised against three mixed synthesized peptides of SeV M protein (19). Anti-F γ236 (39) and anti-HN HN-2 (34) are mouse monoclonal antibodies. Anti-SeV rabbit polyclonal serum was described previously (23). Goat anti-rabbit immunoglobulin G (IgG) and goat anti-mouse IgG-IgM conjugated with horseradish peroxidase (HRP) were purchased from ICN (Aurola, Ohio) and Biosource International (Camarillo, Calif.), respectively. Goat anti-rabbit IgG conjugated with Alexa Fluor 568 and goat anti-mouse IgG conjugated with Alexa Fluor 488 were purchased from Molecular Probes (Eugene, Oreg.).

Plasmid construction.

For the construction of genomic cDNA of M gene-deficient SeV carrying the GFP gene, full-length genomic cDNA of the M gene-deficient SeV (pSeV18+/ΔM) (4) was utilized. The 2.1-kb BstEII fragment carrying the M-deficient site of pSeV18+/ΔM was subcloned into the BstEII site of pSE280 (Invitrogen, Groningen, The Netherlands) in which the EcoRV recognition sequence at the polylinker region had been deleted previously, to generate pSE-BstEIIfrg. pEGFP, which harbors the GFP gene (TOYOBO, Osaka, Japan), was digested with Acc65I and EcoRI and subcloned, after blunt ending, into the EcoRV site of pSE-BstEIIfrg. The BstEII fragment thus prepared containing the GFP gene was returned to the original pSeV18+/ΔM to generate the M gene-deficient SeV genomic cDNA carrying the GFP gene at the M-deficient site (pSeV18+/ΔM-GFP). For the plasmid expressing M protein under the control of the Cre/loxP induction system (3), a PCR-generated 1.1-kb fragment containing the M gene from pSeV18+ was inserted into the SwaI site of pCALNdLw (3) to generate pCALNdLw/M. Since pCALNdLw/F, which was used for the transfer of the F gene, carries the neomycin resistance gene, the neomycin resistance gene of pCALNdLw/M was replaced with the hygromycin resistance gene. The 4.7-kb HincII-EcoT22I fragment containing the M gene and the 1.8-kb XhoI-HincII fragment, which does not contain the neomycin resistance gene, were prepared from pCALNdLw/M. The hygromycin resistance gene was prepared by performing PCR with pcDNA3.1hygro(+) (Invitrogen) as the template and the primer pair 5′-TCTCGAGTCGCTCGGTACGATGAAAAAGCCTGAACTCACCGCGACGTCTGTCGAG-3′ and 5′-AATGCATGATCAGTAAATTACAATGAACATCGAACCCCAGAGTCCCGCCTATTCCTTTGCCCTCGGACGAGTGCTGGGGCGTC-3′ and then digesting the product with XhoI and EcoT22I. These three fragments were ligated together to generate pCALNdLw-hygroM.

Cloning and analysis of helper cells.

LLC-MK₂/F7 cells were transfected with pCALNdLw-hygroM by using the Superfect transfection reagent (Qiagen, Tokyo, Japan) according to the manufacturer's protocol. Hygromycin (150 μg/ml)-resistant clones that had been propagated from single cells were selected 2 or 3 weeks after the transfection and analyzed for their expression of M protein. Each clone was plated in six-well plates, and when they reached a nearly confluent state, they were infected with AxCANCre at a multiplicity of infection (MOI) of 5 (3, 21). After being cultured for 2 days at 32°C, the cells were recovered and subjected to semiquantitative Western blotting with the anti-M antibody according to a method described previously (19).

Recovery of M gene-deficient SeV from cDNA.

Preparation of cell lysate containing RNPs and primary virions of SeV18+/ΔM-GFP was carried out according to a method described previously (19). Briefly, approximately 10⁷ LLC-MK₂ cells seeded in a 10-cm-diameter dish were infected with psoralen- and long-wave-UV-treated (18) vTF7-3 at an MOI of 2. After 1 h of incubation at room temperature, the cells were washed three times with MEM and transfected with the plasmid mixture described below. For each transfection, plasmids pGEM-NP (4 μg), pGEM-P (2 μg), pGEM-L (4 μg) (23), pGEM-M (4 μg), and pGEM-FHN (4 μg) (27) and the indicated SeV cDNA plasmid (12 μg) were suspended in 200 μl of OptiMEM (Gibco-BRL) with 110 μl of Superfect transfection reagent (Qiagen). After the mixture was allowed to stand at room temperature for 15 min, it was added to the cells in 3 ml of OptiMEM containing 3% FBS, and the cells were further cultured for 3 h. The cells were then washed twice with MEM and cultured for 24 h in MEM containing cytosine β-d-arabinofuranoside (AraC) (40 μg/ml) and trypsin (7.5 μg/ml). Approximately 10⁷ LLC-MK₂/F7/A cells that were expressing F protein after AxCANCre infection were suspended in MEM containing AraC (40 μg/ml) and trypsin (7.5 μg/ml), layered onto the transfected cells (18), and cultured at 37°C for an additional 48 h. The cells were harvested, and the pellet was resuspended in OptiMEM at 10⁷ cells/ml. After three cycles of freezing and thawing, the cell lysate (P0 lysate) was transfected into the selected clones after AxCANCre infection. After that, the cells were cultured at 32°C in MEM containing AraC (40 μg/ml) and trypsin (7.5 μg/ml) for 7 to 16 days. The spread of GFP-expressing cells to neighboring cells was examined by fluorescence microscopy. When viruses could be recovered from culture supernatants, it was considered that the selected clone expressed a sufficient level of functional M protein. The viruses recovered in the culture supernatants were further amplified by several serial infections and culturing at 32°C. The titers of recovered viral vectors were expressed as cell infectious units estimated by determining the proportion of GFP-expressing cells (GFP-CIU) per milliliter (27). As the seed virus for all experiments, we used culture supernatants at the third or fourth passage of the virus, which were stored at −80°C after addition of bovine serum albumin solution (Gibco-BRL) to a final concentration of 1% (wt/vol).

RT-PCR.

Total viral RNA from SeV18+/ΔM-GFP was isolated by using a QIAamp viral RNA minikit (Qiagen). Reverse transcription-PCR (RT-PCR) was performed in a two-step process with the Thermoscript RT-PCR system (Gibco-BRL). RT and PCR amplification were performed with random hexamers and the primer pair 5′-CCAATCTACCATCAGCATCAGC-3′ (forward primer specific for the P gene) and 5′-TTCCCTTCATCGACTATGACC-3′ (reverse primer specific for the F gene).

Detection of viral proteins by Western blotting.

LLC-MK₂ cells (10⁶) grown in six-well plates were infected at an MOI of 3 with SeV18+GFP, SeV18+/ΔF-GFP, or SeV18+/ΔM-GFP and incubated in serum-free MEM at 37°C. Three days after the infection, the culture supernatants were centrifuged at 48,000 × g for 45 min to recover the viral proteins of VLPs. Cells recovered from one well of a six-well plate were stored at −80°C and then thawed in 100 μl of the sample buffer for sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis and Western blotting were carried out according to a method described previously (19). Incubation with the anti-M and anti-SeV primary antibodies was followed by incubation with anti-rabbit IgG conjugated with HRP as the secondary antibody. When anti-F was the primary antibody, anti-mouse IgG conjugated with HRP was used as the secondary antibody. The proteins on the membrane were detected by a chemiluminescence method (ECL Western blotting detection reagents; Amersham Biosciences, Uppsala, Sweden).

Gene transfer to primary cultures of rat cerebral cortex neurons.

Primary cultures of rat cortical neurons were prepared from E17 embryos as described previously (5). Dissociated cells were plated at a density of 4 × 10⁵ cells/well in 24-well culture slides coated with poly-l-lysine (Asahi Technoglass Corp., Tokyo, Japan). The cells were cultured at 37°C in a 5% CO₂ atmosphere for 2 days in neural basal medium supplemented with B27 (Gibco-BRL). The cells were infected with SeV18+/ΔM-GFP at an MOI of 3 and incubated for 36 h. To identify neuronal cells, the cells were fixed with 2% paraformaldehyde at room temperature for 15 min and immunostained with anti-MAP2 monoclonal antibody (Sigma-Aldrich Corp., St. Louis, Mo.), followed by labeling with Alexa Fluor 568-conjugated anti-mouse IgG (Molecular Probes). After being washed, the cells were observed under a DM IRB-SLR fluorescence microscope (Leica, Wetzlar, Germany).

Injection of the M gene-deficient SeV vector into gerbil lateral ventricles.

Adult male Mongolian gerbils (HOS, Saitama, Japan) weighing about 70 g were anesthetized and placed in a stereotaxic frame, and 5 μl of SeV/ΔM-GFP (5 × 10⁶ GFP-CIU) was microinjected into the left lateral ventricle with a 30-gauge 10-μl Hamilton syringe. The injection coordinates relative to the bregma were 1.0 mm laterally to the left side at a depth of 2.0 mm from the cortical surface. GFP expression was assessed by examining coronal sections from the brains of animals sacrificed 3 days after the injection of SeV/ΔM-GFP.

Kinetic analysis of VLP formation.

LLC-MK₂ cells (10⁶) grown in six-well plates were infected at an MOI of 3 with SeV18+/ΔF-GFP or SeV18+/ΔM-GFP and incubated at 37°C in serum-free MEM. The culture supernatants were collected every 24 h, with immediate addition of MEM to the remaining cells. The VLPs were quantified by a hemagglutination (HA) assay according to a method described previously (23).

Quantitative analysis of cytotoxicity.

CV-1 cells (4 × 10⁴) grown in 96-well plates were infected at an MOI of 0.01, 0.03, 0.1, 0.3, 1, 3, 10, or 30 with SeV18+/ΔF-GFP or SeV18+/ΔM-GFP and incubated at 37°C in serum-free MEM. The culture supernatants were recovered 3 days after the infection and assayed with a cytotoxicity detection kit (Roche, Basel, Switzerland) that measures lactate dehydrogenase (LDH) activity released from damaged cells (10).

Cationic liposome-mediated transfection of VLPs.

Culture supernatants (100 μl) collected at the indicated times after infection were mixed with Dosper liposomal transfection reagent (Roche) (12.5 μl), allowed to stand at room temperature for 10 min, and transfected to LLC-MK₂ cells cultured to confluency in six-well plates. The cells were then cultured in serum-free MEM without trypsin for 2 days and observed under a DM IRB-SLR fluorescence microscope (Leica).

Electron microscopy.

Particles obtained by ultracentrifugation at 38,000 × g for 90 min were resuspended in phosphate-buffered saline (PBS), dropped onto microgrids, dried at room temperature, and stained with 4% uranium acetate for 2 min for electron microscopic examination with a JEM-1200EXII instrument (Nippon Denshi, Tokyo, Japan).

Immunofluorescence confocal laser scanning microscopy.

A-10 cells cultured on glass-bottomed dishes (Asahi Technoglass Corp.) were infected at an MOI of 1 with SeV18+GFP or SeV18+/ΔM-GFP and incubated in serum-free Dulbecco's modified Eagle MEM for 24 or 48 h. After being washed with PBS, the cells were fixed with cold methanol at 4°C for 15 min and then washed three times with PBS. The fixed cells were then treated with 0.1% Triton X-100 plus 2% goat serum in PBS. For staining, the cells were incubated with mouse monoclonal anti-F or anti-HN antibody in PBS with 2% goat serum at 37°C for 15 min. Next, they were stained with Alexa Fluor 488-conjugated anti-mouse IgG (Molecular Probes) at 37°C for 30 min. To stain the nuclei, TO_PRO3 (Molecular Probes) was added to the cells at room temperature and left for 15 min. Finally, to prevent quenching, the solution from a Slow Fade antifade kit (Molecular Probes) was substituted for the staining solution, and then the cells were observed under an MRC-1024 confocal laser microscope (Bio-Rad, Hercules, Calif.).

RESULTS

Construction of genomic cDNA of M gene-deficient SeV.

SeV genomic cDNA carrying the GFP gene instead of the M gene (SeV18+/ΔM-GFP) was constructed (Fig. 1A). As GFP expression can be detected in living cells, the use of this cDNA allowed us to confirm the successful recovery and infection of M gene-deficient SeV.

FIG. 1.

Construction of an M gene-deficient SeV vector carrying the GFP gene at the M-deficient site and confirmation of the vector's structure. (A) Structures of recombinant SeV genomes. The open reading frame of the GFP gene was inserted with the SeV end and start signals in the respective positions of the deleted gene(s). The positions of the primers for RT-PCR are shown by arrows. (B) The viral genome structure was confirmed by RT-PCR. The DNA fragment of SeV18+/ΔM-GFP from the 5′-terminal of the P gene to the 3′-terminal of the F gene (containing the GFP gene) was amplified from the plasmid (lane PC) and from cDNA prepared in the presence (lane RT⁺) or absence (lane RT⁻) of reverse transcriptase, and these fragments were compared to the corresponding fragment amplified from the plasmid of SeV18+GFP (lane NC). Amplifications of 1,073- and 1,400-bp DNAs were expected based on the gene structures of SeV18+/ΔM-GFP and SeV18+GFP, respectively. Lane M, markers. (C) Viral proteins were detected by Western blotting. LLC-MK₂ cells were infected with SeV18+GFP (lanes Wild), SeV18+/ΔF-GFP (lanes ΔF), or SeV18+/ΔM-GFP (lanes ΔM) at an MOI of 3. The viral proteins in the cells (Cell) and culture supernatants (Sup) prepared 3 days after the infection were detected by Western blotting with anti-M, anti-F, and anti-SeV (which mainly detects NP protein) antibodies. Lane M, markers.

Establishment of a packaging cell line that expresses SeV M protein.

In order to recover SeV18+/ΔM-GFP from cDNA as virion particles, the missing M gene must be complemented in trans. A transient supply could be provided by cotransfection of an M-expressing plasmid; however, the recovered virus titers would be low. Preferably, an M-expressing packaging cell line should be established in order to achieve production of virions at high titer, and for this purpose a Cre/loxP induction system was employed, as reported for the case of an F-expressing plasmid (26). As we plan in further studies to construct SeV vectors lacking both the M and F genes, simultaneous expression of the M and F proteins in helper cells will be necessary. Therefore, pCALNdLw-hygroM, in which the M gene was inserted into pCALNdLw (3) along with a hygromycin resistance gene, was transfected into LLC-MK₂/F7 cells. At 2 or 3 weeks after the transfection, about 130 hygromycin (150 μg/ml)-resistant clones were thus selected, and their expression of M protein was analyzed after the induction of M (and F) proteins by infection with AxCANCre (3, 21). Semiquantitative Western blotting was carried out with anti-M antibody. Thirty clones were thereby selected (data not shown) for further analysis as described below.

Recovery of M gene-deficient SeV.

Recovery of SeV deficient in the M gene (SeV18+/ΔM-GFP) was carried out in conjunction with evaluation of the clones described above. That is, we examined spreading of GFP-expressing cells (i.e., whether the trans-supply of M protein was achieved) upon the addition of the P0 lysate of SeV18+/ΔM-GFP to each clone. Preparation of P0 lysate was carried out according to a method described previously (19). Meanwhile, clones were plated, infected with AxCANCre at an MOI of 5, and cultured at 32°C for 2 days after the infection. These cells were overlaid with P0 lysate of SeV18+/ΔM-GFP and cultured in serum-free MEM containing 40 μg of AraC per ml and 7.5 μg of trypsin per ml at 32°C. Spreading of GFP-expressing cells induced by SeV18+/ΔM-GFP was observed with a few clones. The spreading was most marked with clone 62, which was used in subsequent experiments. The cells prior to the induction with AxCANCre are referred to as LLC-MK₂/F7/M62, and those after the induction (which expressed M [and F] protein) are referred to as LLC-MK₂/F7/M62/A. M protein expression was maintained at some level during more than 10 passages of LLC-MK₂/F7/M62/A cells (data not shown). We continued to recover SeV18+/ΔM-GFP from LLC-MK₂/F7/M62/A cells; for example, at 6 days after infection at the second vector passage (P2), a viral titer of 9.5 × 10⁷ GFP-CIU was detected, and at 5 days after infection at the fourth vector passage (P4), a viral titer of 3.7 × 10⁷ GFP-CIU was detected.

Structural confirmation of M gene-deficient SeV.

The viral gene structure and the resultant protein expression of SeV18+/ΔM-GFP were analyzed by RT-PCR and Western blotting, respectively. RT-PCR was performed in a two-step process in the presence or absence of reverse transcriptase. As expected from the gene structure of SeV18+/ΔM-GFP, amplification of a 1,073-bp DNA, which corresponded to the fragment containing the GFP gene (from the 5′ terminus of the P gene to the 3′ terminus of the FN gene) (Fig. 1A), was observed (Fig. 1B). When reverse transcriptase was omitted, no amplification of the gene was observed, and when the M gene was inserted instead of the GFP gene (pSeV18+GFP), a 1,400-bp DNA was amplified, which was clearly different in size from the amplified DNA described above, supporting the conclusion that this virus had an M gene-deficient structure.

Confirmation in terms of protein expression was performed by Western blot analysis. LLC-MK₂ cells and their culture supernatants were recovered 3 days after infection with SeV18+/ΔM-GFP, SeV18+/ΔF-GFP, or SeV18+GFP at an MOI of 3 in each case. The culture supernatant was centrifuged to recover viral particles, and the viral proteins were detected with anti-M, anti-F, or anti-SeV (which detects mainly NP protein) antibodies. The fact that M protein was not detected in cells infected with SeV18+/ΔM-GFP while F and NP proteins were detected (Fig. 1C) also confirmed in terms of protein that this virus had the structure of SeV18+/ΔM-GFP. In this case, F protein was not observed in cells infected with SeV18+/ΔF-GFP, while all of the virus proteins examined were detected in cells infected with SeV18+GFP. In addition, regarding the virus proteins in the culture supernatant, very little NP protein was observed in the case of infection with SeV18+/ΔM-GFP, indicating that no or very few VLPs were produced.

Productivity of M gene-deficient SeV from the packaging cell line.

The productivity of SeV18+/ΔM-GFP was examined. LLC-MK₂/F7/M62/A cells were infected with SeV18+/ΔM-GFP at an MOI of 0.5 and cultured at 32°C. The culture supernatant was recovered at various times and replaced with fresh medium. The supernatants thus recovered were assayed for infectivity (GFP-CIU) and HA activity. At 4 to 6 days after the infection, the largest amount of virus was recovered (up to 10⁸ GFP-CIU/ml) (Fig. 2). Although HA activity was maintained even 6 days or more after the infection, strong cytotoxicity was observed at this point (data not shown), suggesting that most of this HA was caused by HN proteins that were not present in virus particles but rather were free or bound to cell debris. Thus, it seemed that for optimal recovery of the virus, the culture supernatant should be collected by the fifth day after the infection.

FIG. 2.

Virus productivity of SeV18+/ΔM-GFP. LLC-MK₂/F7/M62/A cells were infected with SeV18+/ΔM-GFP at an MOI of 0.5 and cultured in serum-free MEM containing 7.5 μg of trypsin per ml at 32°C. The culture supernatants were collected every 24 h, and fresh medium was added immediately to the remaining cells. The infectious ability (GFP-CIU) (bars) and HA activity (HA units [HAU]) (▵) of the recovered supernatants were assayed. The levels of infectious particles were 1 × 10³, 3.9 × 10⁴, 9.5 × 10⁵, 2.9 × 10⁷, 7.1 × 10⁷, 2.2 × 10⁷, 3.8 × 10⁶, and 1.3 × 10⁶ GFP-CIU/ml at 1, 2, 3, 4, 5, 6, 7, and 8 days postinfection (p.i.), respectively.

Infectivity of M gene-deficient SeV in vitro and in vivo.

SeV18+/ΔM-GFP caused efficient gene transfer to a variety of cells, such as LLC-MK₂ cells, CV-1 cells, various lines of cancer cells, and normal human cells (including muscle cells and lung fibroblasts), in vitro (data not shown). To test its infectivity to nondividing cells in vitro, SeV18+/ΔM-GFP was used to infect rat primary cortex neurons at an MOI of 3. Merged microscopic images showed that all of the SeV18+/ΔM-GFP-infected GFP-positive cells corresponding to MAP2-positive neuronal cells (Fig. 3), indicating that M gene-deficient SeV retained the high-level infectivity to wild-type or F gene-deficient type SeV vectors (16, 26) in vitro.

FIG. 3.

(Top panels) Gene transfer of the M gene-deficient SeV vector to nondividing cells in vitro. SeV18+/ΔM-GFP was infected into rat primary neuronal cells derived from the rat cerebral cortex at an MOI of 3. GFP expression was detected 36 h after the infection. The cells were also immunostained with anti-MAP2.

In order to confirm the infectivity of M gene-deficient SeV in vivo, SeV18+/ΔM-GFP was microinjected into the lateral ventricles of gerbils and GFP expression was assessed to verify the gene transfer in vivo. In representative coronal sections from the animals sacrificed 3 days following the injection of SeV/ΔM-GFP, expression of GFP was seen along the bilateral lateral ventricle walls (Fig. 4A). Most of these infected cells were ependymal cells. GFP expression was also observed in the lateral ventricle around the hippocampus and in the third ventricle (Fig. 4B). Efficient infection and gene expression were thus confirmed for M gene-deficient SeV vector in vivo. The expression of GFP was still observed 7 days after the infection but returned to negative 14 days after the injection (data not shown).

FIG. 4.

(Bottom panels) Gene transfer of the M gene-deficient SeV vector to gerbil brain in vivo. SeV18+/ΔM-GFP (5 × 10⁶ GFP-CIU) was microinjected into the left lateral ventricle of an adult Mongolian gerbil. GFP expression was detected throughout the brain of the animal sacrificed 3 days following the injection. Fluorescent light micrographs of coronal sections show a positive GFP reaction in the ependymal cells along the bilateral lateral ventricle walls (A) and also in those around the hippocampus in the lateral ventricle (B).

Quantitative analysis of the level of VLPs of M gene-deficient SeV.

In order to clarify the role of SeV M protein in virion formation, the production of VLPs by cells infected with SeV18+/ΔM-GFP was investigated in comparison with that by cells infected with SeV18+/ΔF-GFP. LLC-MK₂ cells were infected with SeV18+/ΔM-GFP at an MOI of 3, and the culture supernatant was recovered 3 days after the infection, filtered through a 0.45-μm-pore-size filter, and centrifuged according to the conditions for recovery of viral particles, which were then subjected to Western blotting to detect viral proteins semiquantitatively. For comparison, samples that had been similarly prepared from cells infected with SeV18+/ΔF-GFP were used. Serial dilutions of the respective samples were subjected to Western blotting to detect the levels of proteins by using the anti-SeV antibody. The viral protein level in the culture supernatant of cells infected with SeV18+/ΔM-GFP was estimated to be less than 1/100 of that in the supernatant of cells infected with SeV18+/ΔF-GFP (Fig. 5), and this trace amount of viral protein may be explained as being from cell debris (see below).

FIG. 5.

Semiquantitative comparison of viral protein levels in the culture supernatant of cells infected with SeV18+/ΔF-GFP and SeV18+/ΔM-GFP. LLC-MK₂ cells were infected with SeV18+/ΔM-GFP or SeV18+/ΔF-GFP at an MOI of 3 and cultured at 37°C for 3 days. A series of dilutions of the supernatants recovered were assayed by Western blotting with DN-1 antibody, which detects mainly NP protein.

To quantify VLPs by using an assay with a different basis, the culture supernatant of LLC-MK₂ cells infected with SeV18+/ΔM-GFP at an MOI of 3 was collected every 24 h and analyzed by the HA assay. Marked elevation of HA activity was not detected at any time in the case of SeV18+/ΔM-GFP (Fig. 6A). However, at 4 days or more after the infection, HA activity was detectable, although the level was low. Measurement of LDH activity (Fig. 6B), an indicator of cytotoxicity, and fluorescence microscopic observation (Fig. 6C) of the samples both revealed clear cytotoxicity at 4 days or more after infection in the SeV18+/ΔM-GFP-infected cells, strongly suggesting that the elevation of HA activity was due to HN proteins bound to or free from cell debris and probably not due to VLPs. The presence of VLPs in the culture supernatant obtained 5 days after the infection was further examined by using cationic liposomes (Dosper). If there are VLPs carrying RNPs in the supernatant of infected cells, then GFP expression can be detected in the cells transfected with the supernatant plus cationic liposome, since RNPs are able to replicate and express their genes even after forced transfection by agents such as Dosper. Inspection under a fluorescence microscope 2 days after the transfection revealed many GFP-positive cells among cells infected with SeV18+/ΔF-GFP, while almost no GFP-positive cells were observed in the case of SeV18+/ΔM-GFP infection (Fig. 7). These results allows us to estimate that the VLP formation by SeV18+/ΔM-GFP from infected cells was at least less than 1/1,000 of that by SeV18+/ΔF-GFP.

FIG. 6.

(Left panels) Comparison of VLP-forming capabilities of SeV18+/ΔF-GFP and SeV18+/ΔM-GFP. (A) Kinetic quantification of VLPs by HA assay. The culture supernatants of LLC-MK₂ cells infected with SeV18+/ΔM-GFP or SeV18+/ΔF-GFP at an MOI of 3 were recovered at various times (daily), and fresh MEM was added immediately to the remaining cells. VLPs were quantified by the HA assay. HAU, HA units. (B) Cytopathic effect after SeV infection. LDH released from damaged cells was quantified by assaying the same supernatants as analyzed for panel A. Values are relative (A₄₉₀/A₆₅₀). (C) Fluorescence microscopic analysis of the LLC-MK₂ cells described above at 5 days after the infection. Magnification, ×100.

FIG. 7.

(Right panels) Cationic liposome-mediated transfection of VLP. The culture supernatants of LLC-MK₂ cells that had been infected with SeV18+/ΔF-GFP or SeV18+/ΔM-GFP at an MOI of 3 were recovered 5 days after the infection and transfected into newly prepared LLC-MK₂ cells by using cationic liposomes (Dosper). The microscopic observation was carried out 2 days after the transfection. Magnification, ×100.

Electron microscopic observation of particles.

The particles produced from the packaging cell line LLC-MK₂/F7/M62/A transfected with SeV18+/ΔM-GFP were round, with RNP inside and spikes at the surface (Fig. 8A), indicating that SeV M protein was successfully supplied in trans from the packaging cell line. We detected very few particles even in the supernatant of LLC-MK₂ cells, which do not express any SeV M proteins, after infection with SeV18+/ΔM-GFP. Their shape was quite different from the normal SeV shape, and their average size was smaller than that of SeV/ΔM particles from cells expressing M. Moreover, most of them seemed not to bear RNP in the particle (Fig. 8B and C).

FIG. 8.

Electron microscopic ultrastructure of viral particles of SeV/ΔM. The particles were prepared from supernatants of LLC-MK₂/F7/M62/A, an M-expressing packaging cell line, or LLC-MK₂ after infection of SeV18+/ΔM-GFP at an MOI of 3 and subsequent culturing for 3 days at 37°C. The electron microscopic structures of M gene-deficient and M protein-supplied particles of SeV18+/ΔM-GFP (A) and VLPs of SeV18+/ΔM-GFP lacking M protein (B and C) were observed after negative staining. Bars, 50 nm.

The cytopathic effect of M gene-deficient SeV is not enhanced in CV-1 cells.

Infection by SeV18+/ΔM-GFP caused more extensive cytopathic effects in LLC-MK₂ cells than infection by SeV18+/ΔF-GFP (Fig. 6). It is important to know whether this cytopathic effect, which might be due to the suppression of VLP formation by M gene deletion from the SeV vector, occurs in many types of cells. To test this, the cytopathic effects were investigated in CV-1 cells, which are sensitive to SeV infection-dependent cytotoxicity. CV-1 cells plated in 96-well plates were infected with SeV18+/ΔF-GFP or SeV18+/ΔM-GFP and incubated in serum-free MEM for 3 days. The cytotoxicity of SeV18+/ΔM-GFP in these cells was comparable to or lower than that of SeV18+/ΔF-GFP (Fig. 9). Thus, the cytopathic effect of M gene-deficient SeV is not enhanced in CV-1 cells, indicating that the suppression of VLP formation itself by M gene deletion in SeV does not always lead to an increase of SeV infection-dependent cytotoxicity. The relative toxicities of the SeV vectors might differ depending on the cells infected, the culture conditions, and the replication differences of the vectors.

FIG. 9.

Quantitative analysis of SeV infection-dependent cytotoxicity. Cytotoxicity was estimated based on the quantity of LDH released into the cell culture medium. CV-1 cells were infected with SeV18+/ΔF-GFP or SeV18+/ΔM-GFP at an MOI of 0.01, 0.03, 0.1, 0.3, 1, 3, 10, or 30 and cultured in serum-free MEM. The assay was carried out 3 days after the infection. The values are from three experiments. Error bars indicate standard deviations.

Infiltrative spreading of M gene-deficient SeV infection in a cell-to-cell manner.

When LLC-MK₂ cells were infected with M gene-deficient SeV (SeV18+/ΔM-GFP) and cultured in MEM in the presence of trypsin (7.5 μg/ml), they showed marked infiltrative activity with syncytium formation (Fig. 10). The cell fusion activity was significant at a low MOI (Fig. 10A and D) and was not observed in the absence of trypsin (data not shown). In the presence of trypsin, the inactive precursor F₀ protein was proteolytically cleaved and activated to F₁ and F₂. The infection of SeV18+/ΔM-GFP spread to neighboring cells in a cell-to-cell manner even though SeV18+/ΔM-GFP had lost its particle-forming capability in cells in which M protein was not supplied in trans.

FIG. 10.

Spread of M gene-deficient SeV infection in cell-to-cell manner. LLC-MK₂ cells were infected with SeV18+/ΔM-GFP at an MOI of 0.2 (A and D), 1 (B), or 5 (C) and cultured in serum-free MEM containing 7.5 μg of trypsin per ml at 37°C. Microscopic observation was carried out 3 days after the infection. Magnifications, ×100 (A to C) and ×600 (D).

Little alteration of subcellular distribution of F and HN proteins in the absence of M protein.

The subcellular distribution of F and HN proteins in A-10 cells infected with SeV18+/ΔM-GFP was compared with that in cells infected with SeV18+GFP (wild-type SeV) by using confocal laser scanning microscopy. The localization patterns of both the F and HN proteins were not markedly different in cells infected with SeV18+/ΔM-GFP and those infected with SeV18+GFP. Expression of both F and HN was observed mainly around the nucleus, probably near the Golgi complex, at 24 h after infection (Fig. 11). Substantial localization of F and HN proteins to the cell surface was observed 48 h after infection (Fig. 11). The localization of F protein to the cell surface was faster than that of HN protein for both viruses. The time course of the subcellular localization of the F and HN proteins of SeV18+/ΔM-GFP was similar to that for SeV18+GFP. These results indicate that loss of the M protein does not result in significant alteration of the subcellular distribution of the F and HN proteins in cells infected with SeV18+/ΔM-GFP.

FIG. 11.

Subcellular localization of F and HN proteins analyzed by confocal laser scanning microscopy. Stereo three-dimensional images of the subcellular localizations of the F (A) and HN (B) proteins observed under a confocal laser microscope are shown. A-10 cells were cultured at 37°C for 24 or 48 h after infection with SeV18+GFP or SeV18+/ΔM-GFP at an MOI of 1. The image was obtained by immunostaining with anti-F and anti-HN antibodies. The nucleus is shown in blue. Magnification, ×600.

DISCUSSION

A new M gene-deficient SeV vector was produced by using a packaging cell line in which M protein was supplied in trans. For the recovery of cDNA of M gene-deficient SeV, the first vector passage (P1) and subsequent cultures were carried out at 32°C, because this experimental condition enabled good recovery of the viruses. Two factors are considered to account for the enhanced vector recovery at 32°C. One is that the cytotoxicity of AraC, which is added to cultures to inhibit the growth of vaccinia virus, seems to be suppressed by culturing cells at 32°C in comparison with 37°C. The cultures can therefore be continued for 7 to 10 days at 32°C, with the cells remaining intact (data not shown) under the usual SeV recovery conditions (26), whereas the cells are damaged on the third to fourth day of culturing at 37°C, as indicated by increased cell detachment. The second point is that the expression of M protein is maintained in LLC-MK₂/F7/M62/A cells when the cells are cultured at 32°C. M protein expression was well maintained for 2 days at 37°C but decreased thereafter, while it was maintained for at least 8 days at 32°C (data not shown). For the recovery of SeV with inferior transcription or replication efficiency or with poor efficiency of infectious virion formation, such as in the case of attenuated recombinant SeV carrying long foreign genes (36), the culture duration is thought to be directly related to the level of recovery. Virus was recovered at high titers (up to 10⁸ CIU/ml), making concentration procedures unnecessary. This level of virus production allows us to consider the possible industrial use of this vector for purposes such as gene therapy. We constructed and utilized helper cells expressing both M and F proteins (LLC-MK₂/F7/M62/A) for the recovery of SeV/ΔM in this study, as we plan the construction of vectors with deletions of both the M and F genes in the future. We have also established an M (without F)-expressing LLC-MK₂ packaging cell line, and we were able to recover SeV/ΔM at a level similar to that obtained with LLC-MK₂/F7/M62/A (data not shown). Additional expression of F protein in trans in the helper cells may not interfere with or promote the assembly and/or budding of SeV/ΔM, probably because the quantity of F protein expressed from the helper cells is not larger than that produced from the SeV genome (data not shown).

Subcellular transport of the F and HN proteins of the M gene-deficient SeV, SeV18+/ΔM-GFP, was similar to that of those proteins from M-possessing wild-type SeV (Fig. 11). As the M protein of SeV has been shown to be localized on the cell surface by binding to the cytoplasmic tails of both F and HN proteins through the secretory pathway (2, 37), the F and HN proteins are considered to play roles in the transport of M protein. Our results indicate that the loss of M protein does not result in significant alteration of the subcellular distribution of F or HN proteins in cells infected with M gene-deficient SeV. The subcellular transport of F and HN proteins appears to occur in the absence of M protein, and any role of M protein in their transport might be passive rather than active. M protein might be active in viral assembly and budding after the protein transport.

VLP formation by infection with SeV/ΔM was not clearly suggested, even though it is also impossible to exclude that. Trace of positive signals in Western blot analyses (Fig. 5) and HA activity assays (Fig. 6A) might be the results of cell damage (Fig. 6B). Moreover, the particles from the cells infected with SeV/ΔM were small, and most of them seemed not to bear RNP when observed by electron microscopy (Fig. 8B and C). Normal particle formation was confirmed in the case of SeV/ΔM with the M gene deleted and M protein supplied, which was produced by using the packaging cell line (Fig. 8A). In the case of M gene-deficient measles virus, similar particles were detected, but most of them were thought to be cell fragments that had detached because of cell damage due to the infection (6). It was also reported that overexpression of F protein of SeV alone brought about specific particle formation from 293 cells (43). As indicated, we actually observed round particles produced from cells infected with SeV/ΔM; however, most of them seemed not to bear RNP in the particles. M protein might be essential to incorporate RNP in the particle. These results demonstrated the effectiveness of M gene deficiency for suppressing VLP formation and confirmed the critical role of M protein in virion formation of SeV. Some reports on negative-strand RNA viruses have described the effect of the modulation of M protein on virus particle assembly. Especially, rabies virus lacking M protein had reduced particle formation that was less than 1/500,000 of that of wild-type virus (33). Also, measles virus lacking M caused increased cell-cell fusion as a result of severe suppression of virion formation (6). Similar enhanced cell-cell fusion by measles virus was observed in the case of alterations of the cytoplasmic tails of spike proteins (F and H) (7). In addition, the particle formation of spike protein (G)-deficient rabies virus was reduced to 1/30 of that of G-possessing wild-type virus (32). In the case of paramyxovirus, rhabdovirus, and possibly filovirus and orthomyxovirus, virus particle budding is thought to be driven by matrix protein with the assistance of spike proteins (and probably RNP). This property is different from those of other families of viruses. For example, virus particle budding is mediated by both capsid and spike proteins, by capsid or core protein only, and by membrane proteins only in the cases of alphavirus, retrovirus, and coronavirus, respectively (12).

The results reported here are the first example of the recovery of an M gene-deficient negative-strand RNA virus at high titer, and the use of the resultant SeV/ΔM, the so-called M knockout SeV, will enable not only further basic analysis of virion formation but also application as a gene transfer vector expected to have superior features. SeV/ΔM shows extensive cell-to-cell spreading under conditions in which F protein is proteolytically cleaved and activated. Such a high degree of cell-to-cell spreading probably occurs in cells infected with SeV/ΔM because both HN and active F proteins are present at high levels on the surface of cells, from which they are not depleted since there is almost no virion formation. The effect was significant at a low MOI (Fig. 10A) but not at a high MOI (Fig. 10C). Almost all of the sialic acid-containing gangliosides, which are ubiquitously expressed on the cell surface and act as host cell receptors for SeV (29), would be digested at a high MOI by HN protein, and it would take time for these gangliosides to recover. Thus, contact-mediated cell-to-cell spreading resulting in cell fusion results in spreading to noninfected cells at a low MOI but not to previously infected cells at a high MOI. Little or no such cell-to-cell spreading occurred in the absence of trypsin. This indicates that the cell-to-cell spreading of SeV/ΔM can be controlled by the activation of F protein by treatments such as the addition of trypsin-like proteases and suggests that further conversion of the cleavage site of F protein to recognition sequences for other proteases is possible.

A viral vector capable of contact infiltration but incapable of dissemination, due, for example, to a block of particle formation, might be able to overcome a critical disadvantage of generally used advanced gene transfer vectors such as retrovirus, adenovirus, and adeno-associated virus vectors, which are incapable of both particle formation and contact infiltration. In other words, when cells are infected with a virus vector for the purpose of expressing a gene carried by the virus vector, the generally used advanced virus vectors are able to express the encoded protein only in cells directly infected. This is not a problem when cell culture systems are used, in which the cells grow in a layer and can be directly infected with viral vectors. However, when cells are infected in vivo in three-dimensional multilayers, the region and the number of cells directly infected by the virus vectors are limited. For example, it is extremely difficult to directly infect all tumor-forming cells at a tumor site with such virus vectors. By contrast, if virus vectors capable of cell-to-cell spreading are used to infect tumor cells at a tumor site, and even if only a fraction of the cells are directly infected, noninfected cells adjacent to the infected cells can be infected with the vectors through cell-to-cell spreading, resulting in infection of all of the tumor cells. Since the vector is not disseminative, it will not infect other cells in the body through the bloodstream. Thus, SeV/ΔM has the desirable novel feature that it could infect all of the cells in a restricted region or a particular organ without infecting other parts of the body.

SeV infects and replicates in most mammalian cells, including human cells, and directs high-level expression of the genes it carries, and its replication is independent of nuclear functions and does not have a DNA phase during its life cycle, so that the transformation of cells by integration of the genetic information of the virus into the cellular genome is not a concern (25). These properties make SeV vectors very promising for application to gene therapy via the expression of therapeutic genes and vaccine antigens (24, 30, 40, 45, 48). In particular, we plan the clinical application of an SeV vector carrying human fibroblast growth factor-2 for the treatment of peripheral arterial disease. Intramuscular injection of this vector strongly boosted fibroblast growth factor-2, resulting in significant therapeutic effects for limb salvage with increased blood perfusion associated with enhanced expression of endogenous vascular endothelial growth factor in murine models of critical limb ischemia (31). For such use in human gene therapy, SeV/ΔF has been produced and demonstrated to be non-self-transmissible due to loss of the F gene in its genomic RNA (26). In many cases, the expression of genes carried by such SeV vectors continues from several days to several weeks in vivo.

In this study, we first succeeded in the establishment of an M protein-expressing packaging cell line and were thus able to recover SeV/ΔM at high titer. SeV/ΔM might be useful for the same purpose as SeV/ΔF if its F protein produced from the vector in infected cells is not cleaved and activated, as its maturation into a particle from infected cells was almost completely abolished. However, for safety, additional modifications are possible and are now in progress. Another direction for gene therapy with SeV is the treatment of diseases such as cancers. In these cases, a cytolytic vector is acceptable and maybe even preferred. SeV/ΔM, which has lost particle formation and gained extensive cell-to-cell spreading, would fulfill such a need. Further modifications of SeV/ΔM for the treatment of cancers are now in progress. Thus, SeV/ΔM will be one of the key SeV vectors for the development of treatments for diseases requiring reduced cytotoxicity and cytolytic features.