Mumps Virus Is Released from the Apical Surface of Polarized Epithelial Cells, and the Release Is Facilitated by a Rab11-Mediated Transport System

ABSTRACT

Mumps virus (MuV) is an airborne virus that causes a systemic infection in patients. In vivo, the epithelium is a major replication site of MuV, and thus, the mode of MuV infection of epithelial cells is a subject of interest. Our data in the present study showed that MuV entered polarized epithelial cells via both the apical and basolateral surfaces, while progeny viruses were predominantly released from the apical surface. In polarized cells, intracellular transport of viral ribonucleoprotein (vRNP) complexes was dependent on Rab11-positive endosomes, and vRNP complexes were transported to the apical membrane. Expression of a dominant negative form of Rab11 (Rab11S25N) reduced the progeny virus release in polarized cells but not in nonpolarized cells. Although in this way these effects were correlated with cell polarity, Rab11S25N did not modulate the direction of virus release from the apical surface. Therefore, our data suggested that Rab11 is not a regulator of selective apical release of MuV, although it acts as an activator of virus release from polarized epithelial cells. In addition, our data and previous studies on Sendai virus, respiratory syncytial virus, and measles virus suggested that selective apical release from epithelial cells is used by many paramyxoviruses, even though they cause either a systemic infection or a local respiratory infection.

IMPORTANCE Mumps virus (MuV) is the etiological agent of mumps and causes a systemic infection. However, the precise mechanism by which MuV breaks through the epithelial barriers and achieves a systemic infection remains unclear. In the present study, we show that the entry of MuV is bipolar, while the release is predominantly from the apical surface in polarized epithelial cells. In addition, the release of progeny virus was facilitated by a Rab11-positive recycling endosome and microtubule network. Our data provide important insights into the mechanism of transmission and pathogenesis of MuV.

INTRODUCTION

Mumps is a common childhood illness characterized by painful swelling of the parotid glands and is often accompanied by severe complications such as orchitis, aseptic meningitis, pancreatitis, and deafness (1). However, despite the prevalence and seriousness of the disease, the molecular basis of the pathogenesis of mumps is still poorly understood.

Mumps virus (MuV), which belongs to the genus Rubulavirus within the family Paramyxoviridae, is the causative agent of mumps (2). The virus infection of cells is initiated by the binding of the hemagglutinin-neuraminidase (HN) protein to sialic acids of the cell surface (3). After receptor binding, the fusion (F) protein induces pH-independent fusion of the viral envelope with the host plasma membrane, and the viral genomic RNA is released into the cytoplasm (4). The viral genomic RNA encapsidated by the nucleocapsid (N) protein forms an active template for RNA replication and transcription, a viral ribonucleoprotein (vRNP), with viral polymerases composed of the phosphoprotein (P protein) and large (L) protein (5). Viral structural components synthesized in the cytoplasm are transported to the plasma membrane. At the plasma membrane, the matrix (M) protein organizes the assembly of vRNP complexes (N, P, L, and genomic RNA) and envelope proteins (F and HN), leading to efficient budding and release of progeny virions from the infected cells (6).

MuV is transmitted through aerosols or by direct contact with contaminated respiratory secretions (7, 8). Following exposure, the virus initially infects the upper respiratory tract epithelia (8). After breaking through the epithelial barrier, the virus accesses the regional lymph nodes and exploits mononuclear cells for further dissemination (9). During the viremia phase, MuV preferably infects epithelial cells in targeted organs, such as the parotid gland, testis, choroid plexus, pancreas, and kidney, resulting in the various symptoms described above. Although infection of the kidney leads to viruria (10), virus shed in saliva is thought to play a more important role in person-to-person transmission (7). Based on the current knowledge of MuV pathogenesis (11), epithelial cells are the primary targets of MuV in humans.

One of the characteristic features of epithelial cells is the presence of distinct plasma membrane domains, the apical and basolateral membranes, which are separated by tight junctions (12). The apical membrane faces the lumen, whereas the basolateral membrane abuts the underlying stratum of the epithelial cells. Because these membranes are exposed to different physiological environments, they have distinct profiles and distinct machineries for sorting proteins and lipids (13). The specialized properties of epithelial cells can also influence virus infections. While the direction of virus entry is strongly dependent on the distribution of viral receptors, virus release is determined by the localization of the viral membrane proteins and/or the matrix protein (14–20). In the cases of paramyxoviruses, the entry of Nipah virus (NiV) is bipolar, because the entry receptor of NiV, ephrin-B2/-B3, is expressed on both the apical and basolateral surfaces (21). On the other hand, the basolateral expression of nectin-4, which is an epithelial cell receptor for measles virus (MV), restricts entry of the virus to the basolateral surface (22). Unlike the polarity of entry, selective apical releases have been commonly reported in paramyxovirus infections, including those with NiV (21), MV (23), Sendai virus (SeV) (24), and respiratory syncytial virus (RSV) (25). Small GTPase Rab11-mediated transport of the vRNP complex has been reported to facilitate the apical release of paramyxoviruses as well as influenza A virus (IAV) (family Orthomyxoviridae) (26–30).

In order to address the question of how MuV infects polarized epithelial cells, we considered that it would be of interest to analyze the polarity of MuV infection. Here, we present data on directional MuV entry and exit pathways in polarized cells. In addition, we demonstrate that MuV utilizes a Rab11-mediated recycling endosome system for efficient virus release from polarized cells.

MATERIALS AND METHODS

Cells and virus.

Madin-Darby canine kidney (MDCK) II, Calu-3 (human lung epithelial), Vero (African green monkey kidney), and 293T (human kidney) cells were maintained in Dulbecco's modified Eagle's minimal essential medium (DMEM) (Nacalai Tesque, Kyoto, Japan) supplemented with 100 U/ml penicillin, 100 mg/ml streptomycin (P/S), and 10% fetal bovine serum (FBS). Rabbit kidney 13 (RK-13) cells were maintained in minimal essential medium (MEM) (Life Technologies Inc., Rockville, MD) supplemented with P/S and 5% FBS. MDCK II and Vero/hSLAM cells constitutively expressing EGFP-Rab11WT or -Rab11S25N were established in a previous study (29) and cultured in DMEM with 10% FBS and 5 μg/ml puromycin. Calu-3 cells were transduced with a retroviral vector expressing EGFP-Rab11WT or -Rab11S25N as described previously (29). For studies of polarized cells, MDCK and Calu-3 cells were seeded and cultured on permeable Transwell filter membranes with a 0.4-μm or 3.0-μm pore size (Corning Inc., Corning, NY). The formation of an electrically tight monolayer was checked by measuring the transepithelial resistance (TER) with an ERS-2 apparatus (Millipore, Bedford, MA).

The MuV strain Odate, which was isolated from a patient who developed aseptic meningitis (31), and the vaccinia virus strain LC-16 (32) were used in this study.

Plasmids and transfection.

Plasmids encoding enhanced green fluorescent protein (EGFP)-tagged wild-type Rab11 and dominant negative Rab11 (with the mutation S25N) were provided by Tadaki Suzuki. 293T cells were transfected with the plasmids by using TransIT LT1 (Mirus, Madison, WI) and subsequently superinfected with MuV. To assess the transfection efficiency, the number of EGFP-positive cells was counted and expressed as a percentage of the total number of cells that were stained with 4′,6-diamidino-2-phenylindole (DAPI). The percentages of 293T cells expressing EGFP-Rab11WT and -Rab11S25N were >95% for both.

Reagent and antibodies.

Nocodazole was purchased from Sigma (St. Louis, MO). Anti-N (clone 23D), -M (clone 79D), and -HN (clone 78) mouse monoclonal antibodies (MAbs) and anti-MuV N rabbit polyclonal antibody (PAb) were prepared as described previously (33–35). Anti-β-actin (clone AC-15) and anti-α-tubulin (clone DM1A) mouse MAbs were purchased from Sigma. Anti-EGFP mouse MAb and anti-ZO-1 rabbit PAb (ab59720) were purchased from Clontech (Mountain View, CA) and Abcam (Cambridge, United Kingdom), respectively.

Virus titration.

Infectious titers of MuV were determined in triplicate by plaque assay using Vero cells in 12-well plates. After 1 to 2 h of virus adsorption, the cells were cultured in DMEM with 5% FBS and 1% agarose. At 6 days postinoculation, the cells were stained with neutral red solution (Sigma), and the plaque counts were determined. Infectious titers of vaccinia virus were determined in triplicate by plaque assay using RK13 cells in 12-well plates. After 1 h of virus adsorption, the cells were cultured in MEM with 5% FBS and 1.5% carboxymethyl cellulose. At 2 days postinoculation, the cells were stained with crystal violet, and the plaque counts were determined.

Cell extracts and immunoblotting.

For the preparation of cell extracts, cells were washed twice with cold phosphate-buffered saline (PBS) and then lysed in cell lysis buffer (20 mM Tris-HCl [pH 7.5], 135 mM NaCl, 1% Triton X-100, and protease inhibitor cocktail [Complete Mini; Roche, Mannheim, Germany]). For immunoblotting, the cell lysate was boiled in sodium dodecyl sulfate (SDS) sample buffer and subjected to SDS-polyacrylamide gel electrophoresis (SDS-PAGE). The proteins were transferred to polyvinylidene difluoride membranes (Millipore) and incubated with the appropriate antibodies. Each protein was visualized with SuperSignal West Femto maximum-sensitivity substrate (Life Technologies Inc.) and detected by use of an LAS-3000 image analyzer system (Fuji Film, Tokyo, Japan).

Immunofluorescence microscopy.

Cells were fixed in 4% paraformaldehyde in PBS for 15 min at room temperature. Then, the cells were permeabilized with 0.2% Triton X-100 in PBS for 10 min, blocked with PBS containing 2% bovine serum albumin (BSA) for 30 min at room temperature, and incubated with the appropriate antibodies. F-actin and nuclei were stained with Alexa 594-conjugated phalloidin (Life Technologies Inc.) and 4′,6-diamidino-2-phenylindole (DAPI), respectively. The samples were examined under an FV1000D confocal laser-scanning microscope (Olympus, Tokyo, Japan).

Virion purification.

Virions released from cells were purified as described previously (36) with minor modifications. The culture media of MDCK/EGFP-Rab11WT and -Rab11S25N cells infected with MuV were harvested at 24 h postinfection (p.i.) and centrifuged at 7,500 × g for 2 min to remove cell debris. The supernatants were layered on 20% sucrose in TNE (0.1 M NaCl; 0.01 M Tris-HCl, pH 7.5; 0.001 M EDTA) (wt/vol) and centrifuged in an SW41 rotor (Beckman Coulter, Inc., Brea, CA) at 140,000 × g for 1.5 h. Pellets were then resuspended in 0.5 ml of TNE and mixed with 1.3 ml of 80% sucrose in TNE, making 1.8 ml of ∼60% sucrose solution with samples. Layers containing 50% sucrose (1.8 ml) and 10% sucrose (0.6 ml) in TNE were applied to the samples in the ∼60% sucrose solution. The sucrose gradient solutions were then centrifuged in an SW55Ti rotor (Beckman) at 110,000 × g for 3 h. The fractions (1 ml each) were collected from the top of the gradient. MuV virions, which were floated in the second fraction from the top, were precipitated with trichloroacetic acid (TCA) and analyzed by immunoblotting.

RESULTS

MuV entry is bipolar, but release is restricted to the apical surface in polarized epithelial cells.

To assess restriction effects of the pore size for migration of MuV through membrane filters, nonpolarized Vero cells were infected with MuV and grown on 0.4-μm or 3.0-μm Transwell filters, and at 24 h p.i., the virus titers in the apical and basolateral chambers were determined. Virus titers in the basolateral chamber were ∼10 times lower than those in the apical chamber, when 0.4-μm filters were used (Fig. 1A). On the other hand, the difference was less than 3 when 3.0-μm filters were used (Fig. 1A). Thus, 3.0-μm filters were used for this work, unless otherwise noted. To analyze the directional entry and release of MuV in epithelial cells, polarized MDCK cells were infected with MuV at either the apical or basolateral surface, and virus titers in the apical and basolateral media were determined, respectively. As shown in Fig. 1B and C, MuV was predominantly detected in the apical chamber, regardless of the virus entry route. The basolaterally infected cells produced ∼3-fold-lower virus titers than the apically infected cells (Fig. 1C). However, this reduction was likely due to the small restriction of virus migration through the 3.0-μm filters, as shown in Fig. 1A. Therefore, the efficiency of virus entry was comparable between the apical and basolateral infection. MuV infection did not cause significant cytopathic effects in MDCK cells or disrupt the integrity of the polarized cell layer displaying a high TER (>180 Ω/cm²) until 96 h p.i. As in MDCK cells, MuV showed the bipolar entry, the apical release, and little cytopathic effect in another polarized epithelial cell line, Calu-3 (Fig. 1D and E). Analyses by confocal microscopy showed that each viral particle component, i.e., the N (vRNP), M (matrix), and HN (membrane) proteins, was predominantly transported to the apical surface in both polarized MDCK and Calu-3 cells (Fig. 1F and G). Collectively, these data indicate that MuV entry is bipolar, while viral release is restricted to the apical surface in polarized epithelial cells.

FIG 1.

Directional entry and release of MuV from polarized epithelial cells. (A) Vero cells on 0.4-μm or 3.0-μm polycarbonate Transwell filters were infected with MuV at a multiplicity of infection (MOI) of 5.0. Apical and basolateral culture supernatants were collected separately at 24 h p.i., and the infectious titers were determined by plaque assay. (B to E) Polarized MDCK (B and C) and Calu-3 (D and E) cells on 3.0-μm Transwell filters were infected with MuV from the apical or basolateral surface at an MOI of 5.0. Apical and basolateral culture supernatants were collected separately at 24 h p.i., and the infectious titers were determined by plaque assay (C and E). The percentages of total release in the apical and basolateral media are shown in panels B and D. (F and G) Polarized MDCK (F) and Calu-3 (G) cells infected with MuV were immunostained at 24 h p.i. with mouse anti-N, -M, or -HN MAb and AF488-conjugated anti-mouse IgG. Cortical actin and cell nuclei were visualized by AF594-phalloidin (red) and DAPI (blue), respectively. The significance of differences was determined by Student's t test.

Rab11 plays key roles in apical transport of vRNP and efficient virus production in polarized epithelial cells.

Rab11-dependent apical transport has been reported to function in trafficking of the vRNP complex and efficient virus production of many RNA viruses, such as IAV, RSV, SeV, and MV (26-30, 37). To examine the roles of Rab11 in the apical transport of MuV vRNP, the intracellular localizations of MuV proteins in MDCK cells expressing the EGFP-Rab11 wild-type (Rab11WT) or its dominant negative form (Rab11S25N) were used (29). As shown in Fig. 2A, the MuV N protein was colocalized with EGFP-Rab11WT and accumulated at the apical surface, whereas it was concentrated in the cytoplasm of polarized MDCK cells expressing EGFP-Rab11S25N. In EGFP-Rab11WT-expressing MDCK cells, the M protein was accumulated at the apical surface but poorly colocalized with EGFP-Rab11WT (Fig. 2A). On the other hand, the M protein mostly showed a diffuse distribution pattern in the cytoplasm in EGFP-Rab11S25N-expressing MDCK cells (Fig. 2A). Similar distribution patterns of the N and M proteins were also observed in polarized Calu-3 cells expressing either EGFP-Rab11WT or -Rab11S25N (Fig. 2B). In contrast, the N and M proteins were barely localized at the plasma membrane in both EGFP-Rab11WT- and EGFP-Rab11S25N-expressing Vero cells (Fig. 2C). Expression of Rab11S25N did not influence the localization pattern of the HN protein any of the three cell lines (Fig. 2A to C). These results suggested that the vRNP, M, and HN proteins are separately transported to the apical surface and that Rab11 contributes differently to the intracellular transport of the vRNP and M and HN proteins in polarized epithelial cells.

FIG 2.

Rab11 plays a role in apical transport of vRNP in polarized cells. Polarized MDCK (A), Calu-3 (B), or nonpolarized Vero (C) cells expressing EGFP-Rab11WT or -Rab11S25N were infected with MuV at an MOI of 1.0. At 24 h postinfection, the cells were immunostained with mouse anti-MuV N, M or HN MAb and AF633-conjugated anti-mouse IgG (pseudocolored red). Cortical actin was visualized by AF594-phalloidin (pseudocolored blue).

Next, to investigate the effect of altered Rab11 function on MuV propagation, the virus titers released from the cells expressing either EGFP-Rab11WT or -Rab11S25N were determined (Fig. 3A). In a control infection, expression of EGFP-Rab11S25N did not inhibit vaccinia virus replication in polarized MDCK cells (Fig. 3B). In contrast, virus titers in polarized MDCK and Calu-3 cells expressing EGFP-Rab11S25N were ∼20-fold and ∼3-fold lower, respectively, than those in the EGFP-Rab11WT-expressing cells. In the case of nonpolarized cells, the virus titer in the cells expressing EGFP-Rab11S25N was unaffected (293T cells) or was slightly higher than that in the cells expressing EGFP-Rab11WT (Vero cells), as seen in MV infection (29).

FIG 3.

Rab11-mediated transport of vRNP is important for efficient MuV release from the apical surface of polarized cells. (A) Polarized MDCK or Calu-3 or nonpolarized Vero or 293T cells expressing either of the EGFP-Rab11s were infected with MuV at an MOI of 5.0. At 24 h p.i., the supernatants were collected, and the infectious titers were determined. The results shown are from three independent experiments, with the error bars representing the standard deviations. (B) Polarized MDCK cells expressing either of the EGFP-Rab11s were infected with vaccinia virus at an MOI of 10. At 24 h p.i., both the supernatants and cells were collected and sonicated, and then the infectious titers were determined. The results shown are from three independent experiments, with the error bars representing the standard deviations. (C) Polarized MDCK cells expressing either of the EGFP-Rab11s were infected with MuV. At 24 h p.i., the cell lysates were subjected to immunoblotting with the indicated antibodies. The relative band intensities of MuV N were normalized by the β-actin level, and the relative expression of MuV N based on the levels of cells expressing EGFP-Rab11WT is shown, with the error bars representing the standard deviations. (D) Polarized MDCK cells expressing either of the EGFP-Rab11s were infected with MuV. At 24 h p.i., the proteins in the cell lysates and the floated fraction were detected by immunoblotting with the indicated antibodies. The relative band intensities of MuV N and M in the floated fraction were normalized by those in the cell lysate, and the relative expressions of MuV N and M were determined based on the levels of cells expressing EGFP-Rab11WT. (E and F) Polarized MDCK expressing either of the EGFP-Rab11s on 3.0-μm Transwell filters were infected with MuV from the apical surface at an MOI of 5.0. Apical and basolateral culture supernatants were collected separately at 24 h p.i., and the infectious titers were determined by plaque assay. The percentages of total release in the apical and basolateral media are shown in panel E. The significance of differences was determined by Student's t test.

To further characterize the role of Rab11 in MuV propagation, the viral protein synthesis and the physical particle production in EGFP-Rab11WT- and EGFP-Rab11S25N-expressing cells were analyzed. Although viral protein synthesis was slightly decreased in MDCK cells expressing EGFP-Rab11S25N (Fig. 3C), this reduction was not sufficient to explain the inhibition of virus production shown in Fig. 3A. In contrast, physical viral particles purified and detected by immunoblotting of the N and M proteins were barely detectable in the culture media of EGFP-Rab11S25N-expressing cells, indicating the impairment of particle formation in these cells (Fig. 3D). In order to analyze the effect of interruption of Rab11-mediated protein sorting on directional MuV budding, virus titers in both the apical and basolateral media of MDCK cells expressing either EGFP-Rab11WT or -Rab11S25N were determined. In both MDCK/EGFP-Rab11WT and -Rab11S25N cells, the progeny virus was predominantly detected in the apical-chamber media (Fig. 3E and F), indicating that the disruption of Rab11 function did not alter the budding side. Taken together, these findings indicate that Rab11-mediated apical transport is required for efficient virus production in polarized epithelial cells.

Rab11-mediated transport of vRNP is dependent on cell polarity.

Next, we assessed the apical transport of vRNP mediated by Rab11 in the context of polarity. At all measurement time points before 4 days postseeding (p.s.), MDCK cells were not yet polarized, as indicated by the low TER (<180 Ω/cm²) and the lack of distinct ZO-1 staining, which is a marker of the tight junction, at the cell junction (Fig. 4A). At 4 days p.s., MDCK cells had established a polarized monolayer displaying a high TER (>180 Ω/cm²) and clear ZO-1 staining. Both nonpolarized (1 days p.s.) and polarized (5 days p.s.) MDCK cells were infected with MuV, and virus production and vRNP transport were analyzed. Expression of EGFP-Rab11S25N caused a 3-fold reduction in the released progeny virus in nonpolarized MDCK cells but had a much more pronounced effect in polarized MDCK cells, where a >30-fold reduction was observed (Fig. 4B). As shown above, the apical transport of vRNP was not observed in polarized MDCK cells expressing EGFP-Rab11S25N (Fig. 4C). On the other hand, the expression of EGFP-Rab11S25N had a small effect on the surface transport of vRNP in nonpolarized MDCK cells. The data showed that Rab11 is involved in the trafficking of vRNP, especially in completely polarized epithelial cells.

FIG 4.

Rab11-mediated transport of vRNP is dependent on cell polarity. (A) Polarized and nonpolarized MDCK cells were immunostained with rabbit anti-ZO-1 PAb. (B) Polarized or nonpolarized MDCK cells expressing either of the EGFP-Rab11s were infected with MuV at an MOI of 5.0. At 24 h p.i., the supernatants were collected, and the infectious titers were determined. The results shown are from three independent experiments, with the error bars representing the standard deviations. The significance of differences was determined by Student's t test. (C) Polarized or nonpolarized MDCK cells expressing either of the EGFP-Rab11s were infected with MuV. At 24 h p.i., the cells were immunostained with mouse anti-MuV N MAb and AF633-conjugated anti-mouse IgG (pseudocolored red). Cortical actin was visualized by AF594-phalloidin (pseudocolored blue).

Apical transport of vRNP is dependent on the MT network.

Since it has been reported that the vRNP complexes of other RNA viruses are transported along MTs (28–30, 38), we analyzed the effects of nocodazole-induced MT disruption on the vRNP trafficking and virus release in polarized epithelial cells. Nocodazole treatment disrupted the structures of MTs (Fig. 5A) and reduced MuV release from the cells (Fig. 5B). Since viral protein synthesis was not affected by nocodazole treatment (Fig. 5C), the MTs were suggested to play roles in the viral assembly and/or budding steps. Immunofluorescence analysis revealed that the N protein was rarely observed at the apical surface of nocodazole-treated cells (Fig. 5D). Instead, we observed inclusion bodies (IBs) that were larger than those in the absence of the drug. Furthermore, fewer small N-positive dot-like structures, which were thought to be the vRNP en route to the assembly site, were observed in the nocodazole-treated cells. These findings suggest that MuV vRNPs are transported in an MT-dependent manner from IBs to the apical membrane, where viral assembly and budding occur.

FIG 5.

Roles of microtubules in the apical transport of vRNP and viral production. (A) Polarized MDCK cells treated with nocodazole (5 μM) or dimethyl sulfoxide (DMSO) were immunostained with mouse anti-α-tubulin MAb. (B to D) Polarized MDCK cells infected with MuV at an MOI of 5.0 were treated with nocodazole (5 μM) or DMSO. (B) The supernatants were collected at 24 h p.i., and the infectious titers were determined. The results shown are from three independent experiments, with the error bars representing the standard deviations. The significance of differences was determined by Student's t test. (C) At 24 h p.i., cell lysates were collected and subjected to immunoblotting with mouse anti-β-actin MAb and rabbit anti-MuV N PAb. The relative band intensities of MuV N were normalized to the β-actin level, and the relative expression of MuV N is shown based on the levels of DMSO-treated cells. (D) At 24 h p.i., the cells were immunostained with mouse anti-MuV N MAb (green). Cell nuclei were stained with DAPI (blue).

DISCUSSION

In general, virus entry in polarized cells is correlated with the apical and basolateral distribution of the entry receptor (17–21). Because the MuV receptor, sialic acid, is expressed on both the apical and basolateral surfaces, MuV enters the polarized epithelial cells from both sides. The infection from the apical surface has the advantage that it facilitates the transmission of virus to neighboring cells, because virus release occurs via the apical (luminal) side, leading to an efficient regional spread. On the other hand, the basolateral entry is likely used for secondary infection via the bloodstream during the viremic phase. Therefore, the bipolar entry of MuV is thought to be an essential viral strategy for the establishment of local and systemic infections.

In contrast to the bipolar entry, the release of MuV occurs predominantly from the apical surface. The direction of virus release is assumed to determine the spread of infection (39). Basolateral virus release is thought to contribute to systemic spread, while apical virus shedding from the epithelia causes primarily local infections restricted to mucosal surfaces. In keeping with this model, the budding of wild-type SeV is restricted to the apical surface and thus causes a local respiratory infection, while mutation in the M protein has been shown to disrupt the cellular polarity and cause bipolar budding of the virus, resulting in a systemic infection in mice (40, 41). In the case of the MuV infection, apical release supports efficient virus replication in the glandular epithelium, especially the parotid gland, and virus shedding in saliva, leading to person-to-person transmission. Selective apical release from epithelial cells is a common strategy of paramyxoviruses, even though the virus causes either a systemic infection or a local respiratory infection, indicating that the direction of virus release may not be a determinant of paramyxoviral tissue tropisms.

So far, the intracellular trafficking of MuV structural components has not been addressed. In this study, we found that apical transport of the N protein was mediated by a Rab11- and MT-dependent mechanism in polarized epithelial cells. A number of studies have reported that the N protein of paramyxoviruses constitutes vRNP in concert with the P and L proteins and viral genomic RNA (2). Since we previously observed colocalization of the N protein with other vRNP components in infected cells (35), it would appear that the intracellular dynamics of the MuV N protein in the infected cells represents the vRNP trafficking. Intracellular sorting and trafficking of the apical and basolateral membrane components are important to establish and maintain epithelial cell polarity (12). Apical recycling endosomes (ARE) are involved in regulated recycling of specialized apical proteins (42), and Rab11 is a key regulator of ARE and plays roles in the apical vRNP trafficking and particle formation of many RNA viruses in polarized epithelial cells (26–30). However, the disruption of Rab11 function did not affect the apical expression of the HN protein. As the surface glycoproteins are indispensable for the production of infectious MuV, it must be expected that the apical budding is not altered, even though the vRNP and M protein are differently targeted. Collectively, our data suggest that Rab11 acts as an activator of MuV release from polarized epithelial cells but that it is not a regulator of selective apical release.

Rab11 interacts with many effector proteins and regulates various cellular functions, including plasma membrane recycling, endosomal membrane organization, and cytokinesis, and its functions are dependent on the cell type, such as polarized or nonpolarized (43–46). Our data suggested that the Rab11-dependent vRNP transport would be important for efficient virus release from polarized cells, because the number of virus particles released from polarized cells, in which vRNP was transported in a Rab11-mediated manner, was much higher than that from nonpolarized cells, and the efficient release from polarized cells was cancelled by disrupting the Rab11 function (Fig. 4C). Therefore, Rab11 might have additional roles in the late stages, such as assembly and/or budding, at the apical surface of polarized epithelial cells.

The transport of MuV HN protein was not regulated by Rab11. Like other envelope glycoproteins, the HN protein is synthesized in the endoplasmic reticulum (ER) and transported by the secretory pathway. In contrast, the apical accumulation of the M protein was impaired by disrupting the Rab11 functions. Data from various viruses showed that viral matrix proteins themselves may be soluble in the cytoplasm or associate with the cellular membrane via electrostatic forces or specific fatty acid modification (6). The mechanism of intracellular transport of MuV M protein is poorly understood, but our data suggested that the M protein may use a particular transport machinery regulated directly or indirectly by Rab11. Alternatively, the apical accumulation of the M protein may be due to association with a highly stable apical binding partner, such as the N protein, because the interaction of the M protein with the N protein of parainfluenza virus 5, which belongs to the genus Rubulavirus, is important for viral particle assembly (47). The Rab11-mediated vRNP transport could facilitate the interaction between the N and M proteins and the efficient virus budding at the apical surface.

Because epithelial tissues of the respiratory, digestive, and reproductive tracts act as barriers between the body cavities and underlying tissues, viruses that cause systemic infections must find ways to penetrate the barriers. Some viruses, such as HIV (48), Epstein-Barr virus (49), and hepatitis C virus (50), target immune cells and use their innate capacities to cross the epithelia. As another example, NiV has been reported to disrupt the integrity of the epithelial cell layer by syncytium formation, allowing systemic spread into the body (21). In the present study, we showed that MuV is released predominantly from the apical surface. This finding indicates that MuV does not escape the epithelial barrier simply by passing through the epithelial cells from the apical to basolateral side. Furthermore, unlike NiV infection, MuV infection did not cause significant cytopathic effects in either polarized MDCK or Calu-3 cells and did not disrupt the integrity of the polarized cell layer, suggesting that transmigration of apically released virus to the basolateral side through the partly disrupted epithelial barrier does not occur. In the case of MV, which is also transmitted via the respiratory route and causes systemic infections, macrophages and dendritic cells in the lung tissue are considered the main cells targeted for the epithelial crossing (51). Further studies will be needed to clarify how MuV breaks through the epithelial barrier during initial infection.

In conclusion, the present study showed that MuV enters polarized epithelial cells from both the apical and basolateral sides and is predominantly released from the apical surface. We have also demonstrated that the Rab11-positive ARE is an important cellular factor in the process of apical vRNP transport and virus production. These data should provide insights into the mechanism of MuV transmission and pathogenesis.