The Transmembrane Domain of the Respiratory Syncytial Virus F

JOURNAL OF VIROLOGY, Oct. 2005, p. 12528–12535 0022-538X/05/$08.00⫹0 doi:10.1128/JVI.79.19.12528–12535.2005 Copyright © 2005, American Society for Microbiology. All Rights Reserved.

Vol. 79, No. 19

The Transmembrane Domain of the Respiratory Syncytial Virus F Protein Is an Orientation-Independent Apical Plasma Membrane Sorting Sequence Sean C. Brock,1,2 Josh M. Heck,2 Patricia A. McGraw,2 and James E. Crowe, Jr.1,2* Departments of Microbiology and Immunology1 and Pediatrics,2 Vanderbilt University School of Medicine, Nashville, Tennessee 37232 Received 8 February 2005/Accepted 29 April 2005

The processes that facilitate transport of integral membrane proteins though the secretory pathway and subsequently target them to particular cellular membranes are relevant to almost every field of biology. These transport processes involve integration of proteins into the membrane of the endoplasmic reticulum (ER), passage from the ER to the Golgi, and post-Golgi trafficking. The respiratory syncytial virus (RSV) fusion (F) protein is a type I integral membrane protein that is uniformly distributed on the surface of infected nonpolarized cells and localizes to the apical plasma membrane of polarized epithelial cells. We expressed wild-type or altered RSV F proteins to gain a better understanding of secretory transport and plasma membrane targeting of type I membrane proteins in polarized and nonpolarized epithelial cells. Our findings reveal a novel, orientation-independent apical plasma membrane targeting function for the transmembrane domain of the RSV F protein in polarized epithelial cells. This work provides a basis for a more complete understanding of the role of the transmembrane domain and cytoplasmic tail of viral type I integral membrane proteins in secretory transport and plasma membrane targeting in polarized and nonpolarized cells. Respiratory syncytial virus (RSV) is a common pathogen of infancy that infects polarized respiratory epithelial cells by entering and budding from the apical cell membrane (29). The fusion (F) protein of RSV is a type I single-pass integral membrane protein embedded in the cell membrane of infected cells and in the viral envelope after budding (10). Single-pass integral membrane proteins are composed of at least three general domains (ectodomain [ED], transmembrane domain [TMD], and a cytoplasmic tail [CT]); in some cases, they possess an N-terminal endoplasmic reticulum (ER)-translocating signal sequence (SS) (12). These proteins also can be assigned to one of four categories, based on their mode of insertion into the ER membrane: type I, type II, type III, and tail-anchored integral membrane proteins (3, 12). Studies of the trafficking of viral single-pass integral membrane proteins have defined important mechanisms mediating protein processing and transport via the secretory pathway in polarized and nonpolarized mammalian cells (2, 4, 11, 15, 17, 21–23, 28, 33). Several mechanisms underlying apical targeting in polarized cells for type I membrane-spanning proteins have been postulated. Some reports suggested that glycosylation of the ED of type I glycoproteins determines their apical transport to the plasma membrane (13, 19, 24, 30). However, other reports suggest that apical membrane targeting depends on the TMD of type I membrane proteins (14, 17, 28). Finally, some studies have indicated that CT residues are necessary for regulating transport between the Golgi and the plasma membrane in nonpolarized cells (25, 33). In this report, we use RSV F-based * Corresponding author. Mailing address: Vanderbilt University Medical Center, T-2220 Medical Center North, Division of Pediatric Infectious Diseases, 1161 21st Ave. South, Nashville, TN 37232-2905. Phone: (615) 343-8064. Fax: (615) 343-8055. E-mail: james.crowe @vanderbilt.edu.

protein constructs to identify a unique orientation-independent contribution of the TMD to apical plasma membrane targeting in polarized and nonpolarized epithelial cells. MATERIALS AND METHODS Mammalian cell lines and viruses. HEp-2 cells (ATCC CCL-23) were grown in Opti-MEM I medium (Invitrogen) supplemented with 2% fetal bovine serum, 320 ␮g/ml L-glutamine, 2.7 ␮g/ml amphotericin B, and 45 ␮g/ml gentamicin. MDCK cells (type II; ATCC CCL-34) were maintained in Dulbecco’s modified Eagle’s medium and Ham’s F-12 medium (GIBCO Laboratories) supplemented with 7% fetal bovine serum, 320 ␮g/ml L-glutamine, 1% (vol/vol) nonessential amino acids, 2.7 ␮g/ml amphotericin B, and 45 ␮g/ml gentamicin. All cell lines were maintained at 37°C in 5% CO2. The RSV wild-type strain A2 was used for all virus infections and has been described previously (9, 16). Infection and microscopic analysis of RSV-infected cells. Visualization of the RSV F protein in infected cells was carried out by infecting HEp-2 cell monolayers (multiplicity of infection ⫽ 0.5) for a period of 48 h. Infected cells were fixed and permeabilized by incubation at room temperature with phosphatebuffered saline (PBS) supplemented with 3.7% formaldehyde and 0.2% Triton X-100. Fixed and permeabilized specimens were immunostained for the presence of RSV F protein with a mixture of three murine monoclonal antibodies specific for RSV F, followed by secondary labeling with Alexa 568-conjugated murine immunoglobulin-specific goat polyclonal serum, diluted 1:500 (Molecular Probes). Nuclei were labeled by incubation of fixed and permeabilized cells for 30 min at room temperature with TO-PRO-3 iodide diluted 1:1,000 in PBS (Molecular Probes). Samples were examined by confocal microscopy, which was performed using a Zeiss LSM510 confocal microscope. Captured images were visualized using Zeiss LSM Image Browser software, version 3.2.0.70. DNA manipulation and analysis. We generated plasmid DNA constructs encoding chimeric proteins incorporating green fluorescent protein (GFP) and portions of viral fusion proteins. GFP-RSV Ftail and GFP-vesiculostomatitis virus (VSV) Gtail fusion proteins were produced using PCR-amplified regions of cDNA encoding wild-type RSV F or VSV G TM and/or CT. PCR splice overlap extension was used to produce GFP-RSV Ftail(⌬CTp). PCR primer sets for each construct can be found in Table 1. Gel-purified PCR amplicons were cloned into the pGEM T-Easy vector (Promega). Cloned amplicons were extracted using either SalI (5⬘) and BamHI (3⬘), or BamHI (5⬘) and EcoRI (3⬘) restriction sites for cloning into the pEGFP-C1 (Clontech) or pcDNA 3.1 (Invitrogen) mammalian expression vectors, respectively. We also generated full-length or truncated versions of the RSV F protein for expression in mammalian cells using a se-

12528

VOL. 79, 2005

RSV F APICAL PLASMA MEMBRANE SORTING SEQUENCE

12529

TABLE 1. PCR primers used for production of GFP fusion proteins and truncation mutantsa Construct

Forward primer

Reverse primer

GFP-RSV Ftall wt ⌬TM ⌬CT ⌬Ctd ⌬CTp (internal) ⌬CTp (external) TM ⫹ 9 TM ⫹ 6 TM ⫹ 3 TM ⫹ 1

GTCGACTCCACCACAAATATCA GGTGGATCCTATTTAGTTACTAA GTCGACTCCACCACAAATATCA GTCGACTCCACCACAAATATCA CTTATACGTGTCAACTGAGTGGTATAAAT GTCGACTCCACCACAAATATCA GGTGGATCCTATTTAGTTACTAA GGTGGATCCTATTTAGTTACTAA GGTGGATCCTATTTAGTTACTAA GGTGGATCCTATTTAGTTACTAA

GGTGGATCCTATTTAGTTACTAA GTCGACAAGGCCAGAAGCACA GGATCCTTAACAGTATAAGAG GGATCCTTAATCTTTGCTTAGTGT CCACTCAGTTGACAGTATAAGAGCAGTCCAACAGCAAT GGTGGATCCTATTTAGTTACTAA GGATCCTTATAGTGTGACTGGTGTGCTTCT GGATCCTTATGGTGTGCTTCTGGCCTTACAGAATAAGAG GGATCCTTATCTGGCCTTACAGTATAAGAG GGATCCTTACTTACAGTATAAGAGCAGTCC

RSV Fopt(TM ⫹ 1) RSV Fopt(-CT) GFP-VSV Gt(wt)

GAGCTCGGATCCACCATGGAGCTGC GAGCTCGGATCCACCATGGAGCTGC GTCGACCTTGTAGAAGGTTGGTTCAGT

GAATTCTTACTTGCAGTACAGGAGCAGG GAATTCTTAGCAGTACAGGAGCAGGCCC GGATCCTTATCACTTTCCAAGTCG

a

All primers are listed 5⬘ to 3⬘.

quence-optimized RSV F gene (designated Fopt) in pcDNA3.1. We designed a cDNA copy of the RSV F gene that incorporated optimal codon usage for mammalian cells and eliminated predicted stem-loop structures and features that might result in mRNA instability of transcripts. RSV-Fopt constructs were sequenced using a primer hybridizing to the T7 promoter site present in the pcDNA3.1 vector, and GFP-Ftail plasmids were sequenced using the pEGFP-C sequencing primer (Clontech) to ensure that cloned amplicons were inserted in frame and provided the desired coding sequence. DNA and predicted protein sequence analysis was performed using MacVector 7.2 software (Accelrys). All DNA sequences were determined by the Vanderbilt University DNA sequencing facility. Transfection and microscopic analysis of cell cultures. MDCK or HEp-2 cells were grown on glass-bottom tissue culture plates (MatTek Corporation) to allow for microscopic analysis of live or fixed samples. Transient transfections were performed using Effectene (QIAGEN) essentially as specified by the manufacturer, except that the DNA transfection reagent mixture was added to cells while in suspension rather than when adherent. Transiently transfected cells expressing GFP-fused proteins were analyzed as live cells, unless otherwise specified. To observe the cellular distribution of GFP-RSV Ftail constructs relative to resident ER or Golgi markers, transfected HEp-2 cells were fixed and permeabilized with 80% methanol at ⫺20°C. The resident ER chaperone calnexin was immunolabeled with calnexin-specific polyclonal rabbit serum (1:250 dilution) (Stressgene) and visualized using rabbit-specific goat serum conjugated to Alexa Fluor 568 (1:500 dilution) (Molecular Probes). The resident Golgi protein golgin-97 was immunolabeled with a mouse monoclonal antibody (1:250 dilution) (Molecular Probes) and visualized using mouse-specific goat serum conjugated to Alexa Fluor 568 (1:500 dilution). Membrane topology of GFP-RSV Ftail was determined by fixing, but not permeablizing, HEp-2 cells expressing GFP-Ftail by incubation in PBS supplemented with 3.7% formaldehyde at room temperature for 10 min. The CT of GFP-RSV Ftail was labeled using rabbit serum raised against the 13 C-terminal amino acids of the RSV F CT. Labeled cells were visualized using rabbit-specific goat serum conjugated to Alexa Fluor 568. Transiently transfected MDCK cells were fixed, but not permeabilized, to assess cell surface distribution of RSV F(TM ⫹ 1) or (⌬CT). Fixed samples were labeled with a mixture of three mouse monoclonal antibodies specific for the ED of the RSV F protein (1:500 dilution). Labeled F protein was visualized with mousespecific goat serum conjugated to Alexa Fluor 568 (1:500 dilution). Samples were examined by confocal microscopy.

RESULTS Cellular distribution and orientation of RSV F protein and a GFP-RSV Ftail construct in nonpolarized epithelial cells. Following RSV infection, the RSV F protein was processed via the secretory pathway and transported to the cell surface (Fig. 1a) (1, 8, 9, 29). To determine the minimal components needed to mediate RSV F transport to the plasma membrane, we created a GFP fusion protein consisting of the TMD and CT

segments of RSV F fused to the carboxyl terminus of GFP (designated GFP-RSV Ftail). Expression of GFP-RSV Ftail in nonpolarized HEp-2 human epithelial cells caused GFP to be distributed to the Golgi and plasma membrane (Fig. 1c; see also Fig. 3a and b). In contrast, an unaltered form of GFP exhibited diffuse cellular distribution throughout the cytoplasm and nucleus (Fig. 1d). To ensure that the targeting of GFP to the plasma membrane observed for GFP-RSV Ftail was a specific effect of RSV Ftail and not merely due to the nonspecific interaction of a viral TMD with cellular membranes, an analogous fusion protein was created using the TMD and CT from the VSV G protein (designated GFP-VSV Gtail). The cellular distribution of GFP-VSV Gtail differed from that of GFP-RSV Ftail in that it lacked plasma membrane localization (Fig. 1e), indicating that the cellular localization of GFP-RSV Ftail was a result of a specific interaction of RSV Ftail with the secretory pathway. The GFP-RSV Ftail construct does not possess the RSV F N-terminal SS, and GFP does not possess an ER translocation SS. Nonetheless, GFP-RSV Ftail did integrate into the secretory pathway and achieve plasma membrane distribution. Therefore, we sought to determine the membrane orientation of GFP-RSV Ftail. HEp-2 cells transiently transfected with GFP-RSV Ftail were fixed without permeabilization and immunolabeled with antibodies specific for either GFP or the distal region of the CT of RSV F. As illustrated in Fig. 1f, the surface of cells expressing RSV Ftail was immunolabeled with antibodies raised against a peptide derived from the RSV F CT domain. Intracellular GFP-RSV Ftail protein was not immunoreactive, demonstrating that the cell membrane remained intact following fixation. Fixed cell preparations examined using GFP-specific antibodies remained unlabeled; however, GFP immunolabeling was successful when similar cell preparations were fixed and permeabilized prior to staining (data not shown). Taken together, these results demonstrate that GFPRSV Ftail protein assumes a type II membrane topology that is opposite of the orientation of the native full-length RSV F protein in RSV-infected cells. Therefore, the TMD of the RSV F protein contains a unique orientation-independent signal sequence that functions in a hierarchical fashion. In the presence of the F protein N-terminal ER translocation signal se-

12530

BROCK ET AL.

FIG. 1. Cellular distribution and orientation of GFP-RSV Ftail in nonpolarized epithelial cells. (a) HEp-2 cells were infected with RSV for 48 h to determine the cellular distribution of the F protein. RSV F (red) was distributed in a perinuclear manner, consistent with Golgi localization, and at the cell surface (nuclei are pseudocolored blue). (b) Schematic representation of a GFP-tail fusion protein. Both RSV F and VSV G GFP-tail fusion proteins are composed of GFP and a short linker region attached to the TMD and CT of the viral glycoprotein. (c to e) HEp-2 cells were transfected with GFP-RSV Ftail (c), GFP alone (d), or GFP-VSV Gtail (e) to assess the cellular distribution of each protein. Live transfected cells were analyzed by confocal microscopy 48 h after transfection; only GFP-RSV Ftail displayed a cellular distribution similar to that of the RSV F protein. (f) Fixed HEp-2 cells expressing GFP-RSV Ftail were immunolabeled with rabbit sera specific for the CT domain of RSV F to determine the membrane topology of GFP-RSV Ftail protein. Only plasma membrane-localized GFP-RSV Ftail protein was labeled, indicating that the protein was integrated in the plasma membrane in a type II orientation. IF, immunofluorescence; DIC, differential interference contrast.

J. VIROL.

quence, the TMD functions as an ER translocation stop signal, resulting in a type I membrane orientation. In the absence of an N-terminal ER-translocating signal sequence, the TMD acted as a signal anchor sequence, resulting in type II orientation. Previous reports have demonstrated specific roles for protein domains in the process of intracellular transport of integral membrane proteins (7, 18, 23, 25, 32, 33). The data presented in Fig. 1 indicate that features of the RSV F TMD and/or CT are sufficient to direct entry into the secretory pathway and mediate plasma membrane targeting, independent of the RSV F extracellular domain or any other viral proteins. Cellular distribution of GFP-RSV Ftail truncation and deletion mutants. To further dissect the role of the TMD and CT of RSV F in these processes we produced truncation-deletion mutants of the GFP-RSV Ftail construct (Fig. 2a). GFP fusion constructs lacking the entire TMD [GFP-RSV Ftail(⌬TM)] or CT [GFP-RSV Ftail(⌬CT)] of RSV F were produced to determine whether the transport determinants of GFP-RSV Ftail were wholly confined to either protein domain. As expected, expression of GFP-RSV Ftail(⌬TM) in HEp-2 cells revealed a diffuse cellular distribution of GFP that was present throughout the cell and nucleus (Fig. 2c), which is similar to the cellular distribution of unmodified GFP (Fig. 1d), suggesting that this fusion protein lacking a TMD does not specifically associate with membranes. In contrast, GFP-RSV Ftail(⌬CT) exhibited a confined distribution throughout the intracellular space but was excluded from the nucleus, with the exception of occasional nuclear projections (Fig. 2d). Further investigation of the cellular distribution of GFP-RSV Ftail(⌬CT) indicated that it colocalized with the ER chaperone calnexin but did not progress from the ER to the Golgi or plasma membrane, suggesting the lack of the CT domain causes ER retention (Fig. 3c and d). These results indicated that neither the F TMD nor CT alone was sufficient to direct plasma membrane localization. Furthermore, while the TMD itself facilitated association with the ER, the TMD was not sufficient to mediate progression from the ER to Golgi in the context of GFP-RSV Ftail(⌬CT). Therefore, components of both the RSV F TMD and CT are required to accomplish the cellular distribution observed for GFP-RSV Ftail. Next, we sought to identify the minimal components of the F CT necessary to achieve the Golgi-plasma membrane cellular distribution characteristic of GFP-RSV Ftail. The CT of RSV F consists of 24 amino acids; we arbitrarily designated the 12 amino acids immediately following the TMD as TM proximal and the 12 most C-terminal residues as TM distal. Two GFP fusion constructs were designed: one composed of only the TMD and 12 proximal CT residues [GFP-RSV Ftail(⌬CTd)] and one with the TMD and 12 distal CT residues [GFP-RSV Ftail(⌬CTp)] (Fig. 2a). GFP-RSV Ftail(⌬CTp) produced an ER-associated distribution similar to that of GFP-RSV Ftail(⌬CT) (Fig. 2f). However, GFP-RSV Ftail(⌬CTd) exhibited a phenotype analogous to that of GFP-RSV Ftail wild type (Fig. 2b). Four truncated forms of GFP-RSV Ftail(⌬CTd) were created to determine which residues of the proximal region of the CT are required to facilitate plasma membrane distribution. The resulting GFP fusion constructs appended RSV F CT residues 1-9, 1-6, 1-3, or -1 alone to the TMD of GFP-RSV Ftail(⌬CT) construct and were designated GFP-RSV Ftail(TM ⫹ 9), -(TM ⫹ 6), -(TM ⫹ 3), or -(TM ⫹ 1), respectively. The

VOL. 79, 2005


FIG. 2. Cellular distribution of GFP-RSV Ftail truncation and deletion mutants. (a) Amino acid composition of the wild-type and mutant forms of GFP-RSV Ftail. Dashes indicate deleted amino acid residues. The linker region of each construct is composed of 19 amino acids, 4 of which are derived from the C terminus of the predicted ectodomain of RSV F (STTN). The other 15 residues of the linker are encoded in the pEGFP-C1 vector and are not shown. In the CT, the proximal region is shown in plain text; the distal region is underlined. (b to g) Nonpolarized HEp-2 cells were transfected with each GFPRSV Ftail construct noted in the legend to panel a to assess the cellular distribution of each protein. (b) Wild type; (c) ⌬TM; (d) ⌬CT; (e) ⌬CTd; (f) ⌬CTp. Live transfected cells were analyzed by confocal

12531

cellular distribution of these constructs was indistinguishable from one another and from that of GFP-RSV Ftail, as represented by the distribution of GFP-RSV Ftail(TM ⫹ 1) depicted in Fig. 2g. These data indicated that the TMD plus the first amino acid residue of the CT of RSV F are sufficient to direct plasma membrane localization of GFP-RSV Ftail. Cellular distribution of RSV F in nonpolarized epithelial cells in the absence of other viral proteins. A recent report that focused on the type I F protein of the paramyxovirus simian virus 5 showed that all but the first CT residue of the fusion protein could be removed without significantly perturbing virus replication (32). This finding is consistent with the data for the RSV F protein presented in Fig. 2c. However, in contrast to the work with the simian virus 5 F protein constructs, our GFP-RSV Ftail proteins did not contain N-terminal viral ERtranslocating signal sequences or an F protein ectodomain. Furthermore, the usual type I orientation of the TMD of wild-type RSV F protein was reversed for the GFP-RSV Ftail proteins. Therefore, we sought to determine whether truncation of the CT of the full-length RSV F protein in type I orientation would abrogate plasma membrane transport of the full-length protein in a manner analogous to that observed for mutated GFP-RSV Ftail fusion proteins. To pursue this line of investigation, we constructed a DNA nucleotide sequence-optimized form of RSV F cDNA that did not introduce coding changes (designated RSV Fopt), since plasmid-based expression of the RSV F protein from cDNA derived from the native gene sequence is quite poor in mammalian cells. The cellular distribution of full-length RSV Fopt protein that was expressed following transient transfection of HEp-2 cells was indistinguishable from that of F protein produced during RSV infection (Fig. 1a and 3b). To determine the contribution of the CT domain to the cellular localization of RSV Fopt, two truncated forms of the RSV Fopt protein were produced: RSV Fopt(TM ⫹ 1), which possesses only the first CT residue, and RSV Fopt(⌬CT), which does not possess any CT residues (Fig. 4a). When transiently expressed in HEp-2 cells, RSV Fopt(TM ⫹ 1) assumed a Golgi-plasma membrane-localized distribution similar to that of RSV F produced during infection, and that of RSV Fopt, GFP-RSV Ftail, and the GFP-RSV Ftail CT truncation mutants -(TM ⫹ 9), -(TM ⫹ 6), -(TM ⫹ 3), and -(TM ⫹ 1). Furthermore, the RSV Fopt(TM ⫹ 1) protein mediated cell-to-cell fusion, as illustrated by the large multinucleated cell body present in Fig. 4d, which is a hallmark of RSV infection and was also observed following expression of RSV Fopt in mammalian cells. Therefore, not only is the cellular localization of RSV Fopt(TM ⫹ 1) similar to that of the wild-type protein, but this truncated protein also retains the functional characteristics of membrane fusion associated with the RSV F protein. In contrast, expression of RSV Fopt(⌬CT) resulted in perinuclear distribution, consistent with Golgi localization, but was markedly absent from the cell membrane (Fig. 4e). Additionally, Fopt(⌬CT) was not present in the tissue culture su-

microscopy 48 h after transfection. Of the major deletions, only GFPRSV Ftail(⌬CTd) displayed a cellular distribution similar to that of GFP-RSV Ftail and RSV F proteins. (g) Live HEp-2 cells transfected with GFP-RSV Ftail(TM ⫹ 1) exhibited a phenotype indistinguishable from that of GFP-RSV Ftail and RSV F proteins.

12532

BROCK ET AL.

J. VIROL.

FIG. 3. ER and Golgi colocalization of GFP-RSV Ftail constructs in nonpolarized HEp-2 cells. HEp-2 cells were transfected with GFP-RSV Ftail (a and b) or GFP-RSV Ftail(⌬CT) (c and d) to assess the codistribution of each protein with resident ER (a and c) or Golgi (b and d) proteins. Transfected cells were fixed and permeabilized with 80% cold methanol, immunolabeled, and analyzed by confocal microscopy 48 h after transfection. Visualization of the ER or Golgi (pseudocolored red) was achieved by immunolabeling either the ER resident protein calnexin (a and c) or the Golgi-localized protein golgin-97 (b and d). The cellular distribution of GFP-RSV Ftail(⌬CT) was consistent with that of the ER, while the cellular localization of GFP-RSV Ftail was consistent with that of the Golgi and plasma membrane.

pernatant of transfected cells (data not shown), indicating that it is retained in the Golgi or degraded. Taken together, the data in Fig. 4 show that proper cellular localization of the RSV F protein requires principally the presence of the TMD and first CT residue. These data validate the central features of the trafficking determinants identified using GFP-RSV Ftail con-

structs, even though those constructs exhibited an opposite membrane orientation. There was a difference in the location of the block in trafficking between the Fopt(⌬CT) construct, which was found in Golgi (Fig. 4e) and the GFP-RSV Ftail(⌬CT) construct that was retained in the ER (Fig. 2d, 3c, and 3d). This difference suggests an effect of the orientation of

VOL. 79, 2005


12533

FIG. 5. Cellular distribution of RSV Fopt and GFP-RSV Ftail mutants in polarized epithelial cells. MDCK cells were transfected with RSV Fopt(TM ⫹ 1) (a) or RSV Fopt(⌬CT) (b) and allowed to polarize. Transfected, polarized MDCK cells were immunolabeled with antisera specific for RSV F (red) and the basolateral marker protein betacatenin (green). RSV Fopt(TM ⫹ 1) was observed to localize to the apical membrane of polarized MDCK cells (a), while RSV Fopt(⌬CT) was only detected in the intracellular space distributed in a nonpolarized manner (b). MDCK cells were transfected with GFP-RSV Ftail(TM ⫹ 1) (c) or GFP-RSV Ftail(⌬CT) (d) and allowed to polarize. Live transfected cells were analyzed by confocal microscopy 48 h after transfection. GFP-RSV Ftail(TM ⫹ 1) was observed to localize to the apical membrane (c), while GFP-RSV Ftail(⌬CT) assumed a nonpolarized distribution (d).

FIG. 4. Cellular distribution of RSV Fopt in nonpolarized epithelial cells. (a) Schematic representation of full-length and mutant forms of RSV Fopt. (b to e) HEp-2 cells were transfected with each RSV Fopt construct noted in panel a to assess the cellular distribution of each protein. (b) RSV infection; (c) Fopt; (d) Fopt(TM ⫹ 1); (e) Fopt(⌬CT). Transfected cells were fixed, permeabilized, immunolabeled, and analyzed by confocal microscopy 48 h after transfection; RSV-infected HEp-2 cells were analyzed in an analogous manner and are shown for comparison. Only the first residue of the Fopt CT was required to achieve plasma membrane localization.

the TMD on ER-to-Golgi transport or, alternatively, a modulating effect of the F extracellular domain on that process. Cellular distribution of RSV Fopt and GFP-RSV Ftail mutants in polarized epithelial cells. RSV buds preferentially from the apical membrane of infected polarized epithelial cells, and the RSV F protein is transported to the apical surface of polarized epithelial cells during infection (5, 10). Previous reports regarding trafficking of type I glycoproteins in polarized epithelial cells provide differing views as to which

protein domain(s) facilitates apical protein targeting. Initial work in this area suggested that proper glycosylation of the ectodomain of type I proteins drives apical targeting (13, 19, 30). However, more recent work has indicated that the TMD of some of these proteins directs apical targeting (14, 17, 31). In addition, the more recent findings regarding type I protein sorting are supported by studies that demonstrate the ability of the TMD of type II glycoproteins to direct apical membrane targeting of nonapical proteins (14, 15). To define the molecular determinants of apical targeting of RSV F, we considered whether the same components responsible for plasma membrane localization of our RSV F-derived proteins in nonpolarized cells would facilitate apical membrane targeting in polarized epithelial cells. To investigate this issue, we first assessed the cellular distribution of RSV Fopt(TM ⫹ 1) in polarized MDCK cells. The RSV Fopt(TM ⫹ 1) protein localized preferentially to the apical plasma membrane of polarized MDCK cells (Fig. 5a). This finding is consistent with the observed cellular distribution of the wild-type RSV F protein in polarized MDCK cells following RSV infection and demonstrates that apical targeting of RSV F requires neither the 23 carboxylterminal residues of the RSV F CT nor other viral proteins. We next assessed the cellular distribution of RSV Fopt(⌬CT) in polarized MDCK cells. As in nonpolarized cells, RSV Fopt(⌬CT) did not localize to the plasma membrane of polarized epithelial cells but rather exhibited a nonpolarized intracellular distribution that laterally coincided with beta-catenin, a basolaterally localized cellular protein (Fig. 5b). These findings suggest that the TMD and the first residue of the RSV F CT are necessary for proper apical targeting of RSV F in polarized epithelial cells. Furthermore, these data suggest that any trafficking contribution made by the RSV F ectodomain is negligible or secondary to that of TMD and CT residues, since the loss of the first CT residue prevented localization to the

12534

BROCK ET AL.

apical membrane even though the full extracellular domain was present. To further assess whether the ectodomain of RSV F contributed to apical targeting, we determined the cellular localization profile of GFP-RSV Ftail(TM ⫹ 1) and GFP-RSV Ftail(⌬CT) in polarized epithelial cells. When expressed by MDCK cells, GFP-RSV Ftail(TM ⫹ 1) exhibited localization to the apical domain of the cell, while expression of GFP-RSV Ftail(⌬CT) resulted in GFP distribution throughout the intracellular space (Fig. 5c and d, respectively). Both of the GFP fusion proteins, GFP-RSV Ftail(TM ⫹ 1) and GFP-RSV Ftail(⌬CT), displayed polarity profiles similar to those of their RSV Fopt counterparts when expressed in MDCK cells (although a distinction between ER and Golgi localization of the constructs as seen in nonpolarized cells could not be determined at the resolution in the X-Z reconstructions). Therefore, it is unlikely that glycosylation of the ectodomain of RSV F plays a significant role in specifying the apical targeting of the protein. The GFP-RSV Ftail(TM ⫹ 1) construct, which lacks the RSV F ectodomain, is apically targeted, and the truncation of the first CT residue of the RSV F CT domain ablates such targeting, as was observed for RSV Fopt(⌬CT) and GFP-RSV Ftail(⌬CT). DISCUSSION In summary, we have used a panel of truncated forms of RSV F protein and GFP-RSV F fusion proteins to analyze type I integral membrane protein transport in polarized and nonpolarized cells. Our results indicate that the RSV F TMD plus the first residue of the CT are necessary and sufficient to direct plasma membrane localization in nonpolarized cells. Various domains of type I integral membrane proteins have been identified to contribute to plasma cell localization and apical trafficking. By eliminating the extracellular domain of RSV F in our fusion constructs, we showed that glycan interactions of the glycosylated extracellular domain are not necessary for proper trafficking. We also eliminated all but one amino acid of the CT with no apparent effect on plasma cell localization. Remarkably, the same minimal residues required for plasma cell membrane localization also are necessary and sufficient to achieve apical plasma membrane localization in polarized epithelial cells. In addition, we demonstrated that the RSV F TMD acts in a reverse orientation s a signal anchor in the absence of an N-terminal ER-translocating signal sequence, while retaining the ability to localize in a polarized fashion to the same cellular membranes as the wild-type RSV F protein. Our findings are consistent with a hierarchical nature of signal sequences previously observed for related paramyxoviruses (26, 27). These findings identify a novel, orientation-independent plasma membrane targeting function for the TMD of the RSV F protein in polarized and nonpolarized cells. The results suggest that structural determinants in the TMD facilitate apical trafficking, but those determinants function in either orientation. Previous studies have shown that RSV F is found predominantly in detergent-resistant membranes that may represent lipid rafts in cells (6, 20). If the transport of these membrane microdomains controls apical trafficking of proteins localizing to those domains, the apical trafficking of RSV F could be explained by lipid raft trafficking patterns. The molecular de-

J. VIROL.

terminants of incorporation into lipid rafts are not well understood. The orientation-independent nature of the apical trafficking determinants in RSV F protein suggests that there is no sequence-specific interaction of portions of the RSV TMD with the inner or outer leaflets of the plasma cell membrane or with proteins selectively embedded in only one of the leaflets or only one face of a lipid raft domain. Interestingly, the length of the predicted TMD that we studied here was quite long (26 amino acids), which exceeds the length needed to traverse a classical plasma cell membrane. The length could allow an orientation in the membrane other than one of 90 degrees or a sliding insertion in which the borders of the TMD are not fixed in relation to the membrane. Alternatively, the longer length could facilitate segregation into membrane domains that are of greater diameter. Our studies defined the minimal domain and the specific residues at the borders of the TMD and length needed for apical trafficking. These studies do not define the mechanism of incorporation into membranes or lipid rafts, but the studies do give insight into the nature of such interactions because of the orientation-independent features of the TMD function. These results provide a deeper understanding of the minimal requirements for type I protein transport to the plasma membrane and the control of directional trafficking and budding of paramyxoviruses. ACKNOWLEDGMENTS This work was supported by the NIH Respiratory Pathogens Research Unit, grant NO1 AI-65298, and NIH grant R21 DE-014039 (both to J.E.C.); the Viruses, Nucleic Acids, and Cancer Training Grant program, grant T32 CA09385, and a Vanderbilt University Dissertation Enhancement grant (both to S.C.B.). Confocal microscopy studies were performed in part through the use of the VUMC Cell Imaging Shared Resource (supported by NIH grants CA68485, DK20593, and DK58404). DNA sequences were determined by the Vanderbilt DNA Sequencing Facility, supported by NIH grants CA68485, DK20593, and HL65962. We thank Elizabeth M. Johnson for technical assistance and Todd R. Graham, Robert J. Coffey, Anne K. Kenworthy, Timothy R. Peters, and James R. Goldenring for helpful discussions. REFERENCES 1. Anderson, K., E. J. Stott, and G. W. Wertz. 1992. Intracellular processing of the human respiratory syncytial virus fusion glycoprotein: amino acid substitutions affecting folding, transport and cleavage. J. Gen. Virol. 73:1177– 1188. 2. Barman, S., and D. P. Nayak. 2000. Analysis of the transmembrane domain of influenza virus neuraminidase, a type II transmembrane glycoprotein, for apical sorting and raft association. J. Virol. 74:6538–6545. 3. Borgese, N., S. Colombo, and E. Pedrazzini. 2003. The tale of tail-anchored proteins: coming from the cytosol and looking for a membrane. J. Cell Biol. 161:1013–1019. 4. Boulan, E. R., and D. D. Sabatini. 1978. Asymmetric budding of viruses in epithelial monolayers: a model system for study of epithelial polarity. Proc. Natl. Acad. Sci. USA 75:5071–5075. 5. Brock, S. C., J. R. Goldenring, and J. E. Crowe, Jr. 2003. Apical recycling systems regulate directional budding of respiratory syncytial virus from polarized epithelial cells. Proc. Natl. Acad. Sci. USA 100:15143–15148. 6. Brown, G., C. E. Jeffree, T. McDonald, H. W. Rixon, J. D. Aitken, and R. J. Sugrue. 2004. Analysis of the interaction between respiratory syncytial virus and lipid-rafts in Hep2 cells during infection. Virology 327:175–185. 7. Bulbarelli, A., T. Sprocati, M. Barberi, E. Pedrazzini, and N. Borgese. 2002. Trafficking of tail-anchored proteins: transport from the endoplasmic reticulum to the plasma membrane and sorting between surface domains in polarised epithelial cells. J. Cell Sci. 115:1689–1702. 8. Collins, P. L., and G. Mottet. 1991. Post-translational processing and oligomerization of the fusion glycoprotein of human respiratory syncytial virus. J. Gen. Virol. 72:3095–3101. 9. Collins, P. L., R. M. Chanock, and B. R. Murphy. 2001. Respiratory syncytial virus, p. 1443–1485. In D. M. Knipe, P. M. Howley, D. E. Griffin, R. A. Lamb, M. A. Martin, B. Roizman, and S. E. Straus (ed.), Fields virology, 4th ed. Lippincott Williams & Wilkins, Philadelphia, Pa.

VOL. 79, 2005


10. Collins, P. L., Y. T. Huang, and G. W. Wertz. 1984. Nucleotide sequence of the gene encoding the fusion (F) glycoprotein of human respiratory syncytial virus. Proc. Natl. Acad. Sci. USA 81:7683–7687. 11. Compton, T., I. E. Ivanov, T. Gottlieb, M. Rindler, M. Adesnik, and D. D. Sabatini. 1989. A sorting signal for the basolateral delivery of the vesicular stomatitis virus (VSV) G protein lies in its luminal domain: analysis of the targeting of VSV G influenza hemagglutinin chimeras. Proc. Natl. Acad. Sci. USA 86:4112–4116. 12. Goder, V., and M. Spiess. 2001. Topogenesis of membrane proteins: determinants and dynamics. FEBS Lett. 504:87–93. 13. Gut, A., F. Kappeler, N. Hyka, M. S. Balda, H. P. Hauri, and K. Matter. 1998. Carbohydrate-mediated Golgi to cell surface transport and apical targeting of membrane proteins. EMBO J. 17:1919–1929. 14. Huang, X. F., R. W. Compans, S. Chen, R. A. Lamb, and P. Arvan. 1997. Polarized apical targeting directed by the signal/anchor region of simian virus 5 hemagglutinin-neuraminidase. J. Biol. Chem. 272:27598–27604. 15. Kundu, A., R. T. Avalos, C. M. Sanderson, and D. P. Nayak. 1996. Transmembrane domain of influenza virus neuraminidase, a type II protein, possesses an apical sorting signal in polarized MDCK cells. J. Virol. 70:6508– 6515. 16. Lewis, F. A., M. L. Rae, N. I. Lehman, and A. A. Ferris. 1961. A syncytial virus associated with epidemic disease of the lower respiratory tract in infants and young children. Med. J. Aust. 2:932–933. 17. Lin, S., H. Y. Naim, A. C. Rodriguez, and M. G. Roth. 1998. Mutations in the middle of the transmembrane domain reverse the polarity of transport of the influenza virus hemagglutinin in MDCK epithelial cells. J. Cell Biol. 142:51– 57. 18. Lopez, J. A., R. Bustos, A. Portela, B. Garcia-Barreno, and J. A. Melero. 1996. A point mutation in the F1 subunit of human respiratory syncytial virus fusion glycoprotein blocks its cell surface transport at an early stage of the exocytic pathway. J. Gen. Virol. 77:649–660. 19. Matter, K., and I. Mellman. 1994. Mechanisms of cell polarity: sorting and transport in epithelial cells. Curr. Opin. Cell Biol. 6:545–554. 20. McCurdy, L. H., and B. S. Graham. 2003. Role of plasma membrane lipid microdomains in respiratory syncytial virus filament formation. J. Virol. 77:1747–1756. 21. McQueen, N., D. P. Nayak, E. B. Stephens, and R. W. Compans. 1986.

22. 23. 24. 25. 26.

27.

28.

29.

30. 31.

32.

33.

12535

Polarized expression of a chimeric protein in which the transmembrane and cytoplasmic domains of the influenza virus hemagglutinin have been replaced by those of the vesicular stomatitis virus G protein. Proc. Natl. Acad. Sci. USA 83:9318–9322. Mostov, K., T. Su, and M. ter Beest. 2003. Polarized epithelial membrane traffic: conservation and plasticity. Nat. Cell Biol. 5:287–293. Mostov, K. E., M. Verges, and Y. Altschuler. 2000. Membrane traffic in polarized epithelial cells. Curr. Opin. Cell Biol. 12:483–490. Nelson, W. J., and C. Yeaman. 2001. Protein trafficking in the exocytic pathway of polarized epithelial cells. Trends Cell Biol. 11:483–486. Nishimura, N., and W. E. Balch. 1997. A di-acidic signal required for selective export from the endoplasmic reticulum. Science 277:556–558. Parks, G. D., and R. A. Lamb. 1993. Role of NH2-terminal positively charged residues in establishing membrane protein topology. J. Biol. Chem. 268: 19101–19109. Parks, G. D., J. D. Hull, and R. A. Lamb. 1989. Transposition of domains between the M2 and HN viral membrane proteins results in polypeptides which can adopt more than one membrane orientation. J. Cell Biol. 109: 2023–2032. Puddington, L., C. Woodgett, and J. K. Rose. 1987. Replacement of the cytoplasmic domain alters sorting of a viral glycoprotein in polarized cells. Proc. Natl. Acad. Sci. USA 84:2756–2760. Roberts, S. R., R. W. Compans, and G. W. Wertz. 1995. Respiratory syncytial virus matures at the apical surfaces of polarized epithelial cells. J. Virol. 69:2667–2673. Scheiffele, P., J. Peranen, and K. Simons. 1995. N-Glycans as apical sorting signals in epithelial cells. Nature 378:96–98. Scheiffele, P., M. G. Roth, and K. Simons. 1997. Interaction of influenza virus haemagglutinin with sphingolipid-cholesterol membrane domains via its transmembrane domain. EMBO J. 16:5501–5508. Waning, D. L., A. P. Schmitt, G. P. Leser, and R. A. Lamb. 2002. Roles for the cytoplasmic tails of the fusion and hemagglutinin-neuraminidase proteins in budding of the paramyxovirus simian virus 5. J. Virol. 76:9284–9297. Ward, B. M., and B. Moss. 2000. Golgi network targeting and plasma membrane internalization signals in vaccinia virus B5R envelope protein. J. Virol. 74:3771–3780.