INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

The fusion of influenza viruses with target membranes is mediated by the spike membrane glycoprotein, HA. In the viral membrane, HA is organized as a homotrimer, each monomer consisting of two disulfide-linked subunits, HA₁ and HA₂ (14, 38). The HA-mediated fusion is triggered by acidic pH, which causes the HA ectodomain to be converted to a fusogenic conformation, thereby exposing the hydrophobic N terminus of the HA₂ subunit (37), the so-called fusion sequence. X-ray crystallographic studies of a fragment of the ectodomain suggest that the loop of the HA₂ subunit connecting two -helices in the neutral form (39) becomes a part of an extended trimeric coiled coil at low pH (5, 6, 30, 36). Recent reconstruction of the three-dimensional structure of the complete HA ectodomain at low pH by cryoelectron microscopy revealed that the preserved trimeric shape of the ectodomain may serve to guide the orientation of the coiled coil, thereby moving the fusion sequence to the tip of HA toward the target membrane (3). Destabilization of the target membrane by this sequence is then believed to initiate membrane fusion (11, 19).

Expression of HA in the plasma membrane of various cell lines preserves its fusion activity. In studies with these systems, it was revealed that the HA ectodomain and its low-pH-triggered conformational change is sufficient to induce membrane mixing (17, 21-24). Indeed, the HA ectodomain membrane anchored by GPI mediates the merger between the outer layers of the contacting membranes, resulting in a metastable hemifusion intermediate (17, 20, 21, 24). These studies concluded that for full fusion, i.e., merger of the inner leaflet and efficient formation as well as enlargement of pores, TM is required. Recently, it was reported that the fusion sequence of HA₂ plays a role in later fusion steps as well (31).

Previous studies suggested that the specific sequence of the TM of HA-wt is not a requirement for pore formation. Parallel replacement of both the TM and the CT of the HA-wt by related domains of another enveloped virus fusion protein does not inhibit fusion activity (10, 32). However, it was left open whether both domains form a functional entity. Recently, Melikyan et al. (23) addressed this concern by constructing HA chimeras containing the TM and CT of a protein unrelated to fusion, the pIgR. They showed significant inhibition of fusion when only one domain was replaced. Simultaneous replacement did not affect fusion, suggesting that the TM and the CT are not functionally independent. On the other hand, we found that replacement of only one domain of the HA by the corresponding domain from the fusion protein of Sendai virus, the F-protein, did not affect membrane fusion or the formation of small aqueous fusion pores (34). Although this would argue against a functional entity of the two domains, we could not rule out the possibility that the substitution of those domains of HA by related domains from another viral fusion protein preserves the entity. Therefore, in the present work we studied the fusion activity of chimeric HAs containing in a combinatorial fashion the TM and CT of a totally unrelated nonviral and nonfusogenic protein, the T-cell surface glycoprotein CD4.

We expressed HA-wt (subtype H7) and the various chimeric protein constructs in monkey kidney cells (CV-1) using the transient T7-RNA-polymerase vaccinia virus expression system (2, 13, 29, 34) and measured the low-pH-induced fusion of those cells with RBCs and RBC ghosts, respectively, as targets. We further assessed the ability of chimeric HA with the TM and/or CT from Sendai virus F protein (34) to mediate pore widening. For all chimeras tested, we found membrane mixing (redistribution of the lipid analogue R18) and formation of small aqueous fusion pores (redistribution of the low-MW cytoplasmic label calcein) after fusion was triggered; however, subsequent dilation of the pore, allowing the flow of the higher-MW fluorophore TMR-D into the cytoplasm of the transfected cells, was impaired for the chimeras carrying the cytoplasmic domain of CD4.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Abbreviations used in this paper. aa, amino acids; ANOVA, analysis of variance; CT, cytoplasmic tail; DMEM, Dulbecco's modified Eagle's medium; DMEM+, DMEM supplemented with 5% FCS; DMEM, DMEM without FCS; FACS, fluorescence-activated cell sorter; FITC, fluorescein isothiocyanate; FCS, heat-inactivated fetal calf serum; GPI, glycosylphosphatidylinositol; HA, influenza virus hemagglutinin; HA-wt, wild-type HA; MW, molecular weight; PBS, phosphate-buffered saline; PBS++, PBS containing 1 mM CaCl₂ and 1 mM MgCl₂; pIgR, polyimmunoglobulin receptor; RBC, human red blood cell; RIPA, radioimmunoprecipitation assay; RT, room temperature; R18, octadecyl rhodamine B chloride; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis; TM, transmembrane domain; TMR-D, dextran-conjugated tetramethylrhodamine; WGA, wheat germ agglutinin.

Materials. The fluorophores R18, calcein-AM, calcein, and TMR-D (molecular mass, 10 kDa) were purchased from Molecular Probes (Eugene, Oreg.), and Tran³⁵S-label (>1,000 Ci/mmol) was from ICN Radiochemicals (Irvine, Calif.). Enzymes used in molecular cloning were obtained from New England Biolabs (Schwalbach/Taunus, Germany). DMEM, DMEM without L-methionine and without L-glutamine, L-glutamine, 2YT medium, and Lipofectin were purchased from Life Technologies Inc. (Karlsruhe, Germany). AmpliTaq DNA polymerase was obtained from Perkin-Elmer (Vaterstetten, Germany). Oligonucleotide primers used for sequencing and PCR were synthesized by Life Technologies. FCS and 0.05% trypsin-0.02% EDTA solution PBS were obtained from Biochrom KG (Berlin, Germany), and secondary antibody was obtained from Dakopatt (Hamburg, Germany). Protein A-Sepharose, neuraminidase (type V from Clostridium perfringens), trypsin (type XIII from bovine pancreas, tolylsulfonyl phenylalanyl chloromethyl ketone, and WGA (lectin from Triticum vulgaris) were purchased from Sigma (St. Louis, Mo.), and the DNA sequencing kit was from Stratagene (La Jolla, Calif.). Cell culture dishes were purchased from Nunc (Biochrom KG).

Construction of chimeric genes. The various HAs expressed are marked by their composition (ectodomain/transmembrane domain/cytoplasmic tail) as H/H/H (HA-wt subtype 7) and H/H/C, H/C/H, H/C/C, H/H/F, H/F/H, and H/F/F (chimeras). H represents the respective domain of HA-wt, C represents the corresponding sequences from CD4, and F represents those of the domain of the C-tail mutant of Sendai virus glycoprotein F in which cysteine residues replace Ser-530 and Gly-534 (29). The ectodomain of all these proteins is that of HA (sequence not shown). The sequences of the TM region for HA-wt, for CD4 and for the F protein are VILWFSFGASCFLLLAIAMGLVFICV (26 aa), ALIVLGGVAGLLLFIGLGIFFCV (23 aa), and TVITIIVVMVVILVVIIVIIIVL (23 aa), respectively, and those of the CT are KNGNMRCTICI-COOH (11 aa), RCRHRRRQAERMSQIKRLLSEKKTCQCPHRFQKTCSPI-COOH (38 aa), and YRLRRCMLMCNPDERIPRDTOTLEPKIRHMOTNGGFDAMAEKR- COOH (43 aa), respectively.

Cell culture and vaccinia virus-based expression of foreign genes. CV-1 cells were grown at 37°C and 5% CO₂ in DMEM+. Transient expression of HA-wt as well as the chimeric proteins in CV-1 cells was performed as described previously (13, 34). Subconfluent monolayers of CV-1 cells (about 80% confluent, in 3.5 cm dishes) were washed twice with DMEM, infected at a multiplicity of infection of 10 PFU per cell by vaccinia virus vTF7-3 (13), which expresses T7-RNA-polymerase, and incubated in DMEM for 1.5 h at 37°C and 5% CO₂. To adjust the surface expression levels, different amounts of plasmid DNA were used for transfection. A 2-µg (for HA-wt, H/F/F, H/F/H, and H/H/F) or 4-µg (for H/C/C, H/C/H and H/H/C) portion of plasmid vector DNA in 100 µl of DMEM was mixed with preincubated Lipofectin (10 µg in 90 µl of DMEM; 45-min preincubation at room temperature) and incubated for 15 min at RT. DMEM (400 µl) was added, and the mixture was used for transfection of the vaccinia virus-infected CV-1 cells after removal of the virus inoculum.

Analysis of expression levels by immunofluorescence (FACS analysis). CV-1 cells grown in 6-cm dishes (to about 80% confluence) were infected and transfected as described above by using three times the amounts of plasmid DNA, vaccinia virus, Lipofectin, and medium used for 3.5-cm dishes. At 30 min posttransfection, ammonium chloride was added to a final concentration of 10 mM. At 4 h posttransfection, the serum-free medium was replaced by DMEM-10% FCS-10 mM ammonium chloride, and the cells were kept overnight at 32°C and 5% CO₂. They were then washed three times with ice-cold PBS. The cells were removed from the dishes by trypsin treatment (0.05% trypsin-0.02% EDTA in PBS), resuspended, and fixed with 1.5% paraformaldehyde in PBS for 1 h on ice. Immunolabeling was performed with a polyclonal antibody from rabbit (anti-H7) as the primary antibody and an FITC-conjugated porcine anti-rabbit-immunoglobulin as the secondary antibody.

Metabolic labeling experiments and immunoprecipitation. CV-1 cells were infected and transfected as described above but with methionine-deficient DMEM for transfection. At 30 min posttransfection, 3 µl of Tran³⁵S-label (>1,000 Ci/mmol) was added. After radiolabeling for 3.5 h (at 37°C and 5% CO₂), the dishes were transferred to ice. The cells were washed with PBS, lysed with RIPA buffer (1% Triton X-100, 1% deoxycholate, 0.1% SDS, 0.15 M sodium chloride, 20 mM Tris, 10 mM EDTA, 10 mM iodoacetamide) for 5 min, resuspended, and centrifuged in Eppendorf tubes at 1,000 × g for 5 min. The cell debris was removed, and the supernatant was incubated with anti-H7 antibody overnight at 4°C. For precipitation of the expressed HA molecules, the samples were incubated with protein A-Sepharose for 2 h at RT and centrifuged. The Sepharose pellet was washed three times with RIPA buffer. Sample buffer (nonreducing: 50 mM Tris, 2% SDS, 0.1% bromophenol blue, 10% glycerol; the reducing buffer also contained 10 mM mercaptoethanol) was added, and the samples were boiled for 3 min. After centrifugation, the immunoprecipitates were analyzed by SDS-PAGE in a 12% polyacrylamide gel and visualized by autoradiography (1 day) of the vacuum-dried gels with Kodak X-Omat AR films. Quantification of fluorograms was carried out with an Epson GT-9000 scanner and Biometra ScanPack 2.0 software.

Preparation of cells for fusion assays. (i) Preparation of HA-expressing cells. CV-1 cells were treated as described above for expression. At 30 min after transfection, ammonium chloride (final concentration, 10 mM) was added to the medium to prevent inactivation of expressed HA molecules by acidification during intracellular transport (26). As a control, CV-1 cells infected with vaccinia virus and treated by the transfection method but without DNA were used. The cells were incubated for another 3.5 h at 37°C and 5% CO₂. After replacement of the medium by DMEM with 10% FCS and 10 mM ammonium chloride, the cells were kept overnight at 32°C and 5% CO₂. The expressing CV-1 cells were washed with prewarmed DMEM (37°C), treated for 15 min at RT with prewarmed DMEM containing 1 U of neuraminidase per ml and 10 µg of trypsin per ml, and given a final wash with PBS++.

(ii) Preparation and labeling of RBC. Human blood (erythrocyte concentrate) was washed three times with PBS (centrifugation at 2,000 × g for 10 min), and erythrocytes were resuspended in PBS (50% hematocrit). A 20-µl volume of R18-solution (1 mg/ml in ethanol) were added to 200 µl of the RBC suspension in 5 ml of PBS under gentle vortexing. After addition of 10 ml of PBS and incubation of the suspension for 30 min at RT in the dark, 25 ml of DMEM+ was added to the suspension to absorb unbound label. After a further incubation for 20 min at RT in the dark, the RBC suspension was washed in 40 ml of PBS and centrifuged. The RBC pellet was resuspended in 2 ml of PBS, and after addition of the calcein-AM solution (50 µg freshly dissolved in 50 µl of dimethyl sulfoxide), the cells were incubated for 45 min at 37°C in the dark. Subsequently, the suspension was centrifuged and washed in 20 ml of PBS. After resuspension of the pellet in 10 ml of PBS, the cells were incubated at 37°C for 20 min in the dark for cleavage of remaining calcein-AM inside the cell. The suspension of R18- and calcein-labeled RBCs was washed five times in 30 ml of PBS, resuspended in 5 ml of PBS++, and stored on ice until used for binding to the (chimeric) HA-expressing CV-1 cells.

(iii) Preparation and labeling of RBC ghosts. Blood (2 ml) of 50% hematocrit (prepared as described above) was mixed with 30 ml of ice-cold hemolysis buffer (4 mM MgSO₄, 1 mM CaCl₂, 1.2 mM acetic acid, 50 mM vanadate) and pelleted. The unsealed ghosts thus obtained were resuspended in 2 ml of hemolysis buffer containing 1 mg of calcein and 5 mg of TMR-D and incubated on ice for 15 min. Tenfold-concentrated PBS and 3.3 mg of ATP in 100 µl of PBS were added before the suspension was brought to pH 7.3 by titration with sodium hydroxide. The whole procedure was performed on ice. Subsequently, the suspension was incubated for 1 h at 37°C in the dark, washed four times with 30 ml of PBS and, for fusion experiments, resuspended in 5 ml of PBS++.

Fusion assay and fluorescence microscopy. For binding, 1 ml of labeled RBCs or RBC ghosts with a hematocrit of 0.4 and 5%, respectively, was added to the monolayer of HA-expressing cells and incubated for 10 min at RT. The samples were washed with PBS++ to remove unbound RBCs and RBC ghosts, respectively. PBS++ was replaced by sodium acetate buffer (20 mM NaC₂H₃O₂, 150 mM NaCl, 1 mM CaCl₂, 2 mM MgCl₂ [pH 5.0]) to trigger fusion. Fusion was monitored at RT by redistribution of the loaded fluorescent labels from RBCs or RBC ghosts to the membrane and cytoplasm, respectively, of the (chimeric) HA-expressing CV-1 cells. Label redistribution was observed under the fluorescence microscope (Axiovert 100 with a 40× LD Achroplan objective [Zeiss]). Calcein fluorescence was monitored with a "blue" filter set (450-490 excitation filter; lp 520 emission filter). R18 and TMR-D fluorescence was visualized using a "green" filter set (510-560 excitation filter; lp 590 emission filter). Photomicrographs were obtained with an MC 80 microscope camera (Zeiss) on an Ektachrome EPH P1600× color reversal film for push processing (Kodak) at 3,200 ASA. The sodium acetate buffer was removed after 4 min and replaced by PBS++ to prevent cell damage by acidic conditions.

Binding of RBCs to CV-1 cells by lectin and fusion assay. To exclude fusion induction by vaccinia virus proteins, control experiments were performed. Cells expressing only vaccinia virus proteins or additionally HA-wt were incubated for 5 min at RT with WGA (50 µg/ml) in PBS++, and 1 ml of labeled RBCs or labeled RBC ghosts (0.4 and 5% hematocrit, respectively) was added. After further incubation for 10 min at RT in the dark under occasional gentle agitation, unbound cells were removed by five washes with PBS++. Fusion was monitored by fluorescence microscopy as described above.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Expression of HA-wt and the chimeras H/C/C, H/C/H, and H/H/C. HA-wt and the constructs H/C/C, H/C/H, and H/H/C were expressed in CV-1 cells by the transient T7-RNA-polymerase vaccinia virus system, metabolically labeled with [³⁵S]cysteine-[³⁵S]methionine, and immunoprecipitated as described above. The autoradiograph (Fig. 1) of an SDS-polyacrylamide gel obtained under reducing conditions (see Materials and Methods) revealed the biosynthesis of uncleaved precursors of HA₀ and of HA/CD4 chimeras as well as the two cleavage-generated subunits of those precursors. Similar results were obtained for the respective F chimera (not shown), as reported previously (34). Differences in the mobility between HA-wt, H/C/H, H/H/C, and H/C/C indicate different molecular weights based on the various lengths of the CT originating either from HA-wt or from CD4. As expected, this became most evident by comparison between the subunit HA₂ and the corresponding cleavage product of chimeras. Densitometric analysis of the autoradiographs (data not shown) revealed that the efficiency of expression and the cleavage efficiency of HA/CD4 chimeras were similar to those obtained for HA-wt.

View larger version (85K): [in this window] [in a new window]   FIG. 1.   Expression of HA and HA-derived chimeras of CD4 in CV-1 cells. The genes of HA-wt, H/C/C, H/C/H, and H/H/C were cloned into the plasmid vector pTM1 and expressed with the transient T7-RNA-polymerase vaccinia virus expression system (see Materials and Methods). The gene products were metabolically labeled with [35S]cysteine-[35S]methionine. Cell extracts were immunoprecipitated with polyclonal antibody directed against influenza virus HA (H7) and subjected to SDS-PAGE under reducing conditions on a 12% polyacrylamide gel. The upper bands represent the precursor molecules (HA0 and the analogues of the chimeras H/C/C, H/C/H, and H/H/C); the two lower bands represent the subunits HA1 and HA2 (or analogues of the respective chimera) (marked on the left). The exposure time for the autoradiography was 1 day.

Fusion assays. (i) Redistribution of the membrane label R18. We studied membrane fusion between CV-1 cells expressing HA-wt, H/C/C, H/C/H, or H/H/C and R18-labeled RBCs at pH 5.0 (Fig. 2). R18 redistribution for HA-wt as well as for all the chimeras was first observed between 1.5 and 2 min after the pH trigger was set. Between 3 and 4.5 min after lowering of the pH, almost all HA-wt- or chimera-expressing cells with bound RBCs showed R18 labeling, indicating efficient membrane mixing. As revealed by seven independent experiments, no significant difference of the membrane fusion activity between HA-wt and the chimeras was found. This also applies to the chimeras H/F/F, H/F/H, and H/H/F (data not shown), which is in line with our previous study (34).

View larger version (39K): [in this window] [in a new window]   FIG. 2.   Redistribution of the fluorescent label during fusion. HA-wt-, H/C/C-, H/C/H-, H/H/C-, H/F/F-, H/F/H-, and H/H/F-expressing cells were bound to either R18-labeled RBC (top row) or calcein (Cal)- and TMR-D-labeled RBC ghosts (middle and bottom rows, respectively). Fusion was triggered at RT by replacing neutral buffer (PBS++ [pH 7.4]) with acidic buffer (sodium acetate buffer [pH 5.0]). Fluorographs obtained taken 5 min (HA-wt) or 15 min (chimeras) after the pH trigger. For details, see Results.

(ii) Redistribution of the aqueous labels calcein and TMR-D. R18 redistribution between transfected cells and RBCs is indicative of hemifusion of membranes. To address whether our chimeras trigger complete membrane fusion, i.e., fusion of the exoplasmic as well as the cytoplasmic leaflets, and subsequent formation of an aqueous fusion pore, the redistribution of aqueous cytoplasmic fluorescent labels was employed. To this end, we analyzed the fusion of HA-wt- and chimera-expressing CV-1 cells with RBC ghosts doubly labeled with the aqueous labels calcein (MW 623) and TMR-D (MW 10,000). The use of labels with different sizes allowed us to monitor both the formation and the enlargement of the aqueous fusion pore. Based on our observation that binding of two or more ghosts to control CV-1 (i.e., vaccinia virus-infected but not transfected cells) was very rare, we consider that CV-1 cells with two or more bound ghosts are cells expressing HA-wt or chimeric proteins.

View larger version (22K): [in this window] [in a new window]   FIG. 3.   Formation of an aqueous fusion pore, assessed by using fluorescent labels of different MW. Calcein- and TMR-D-labeled RBC ghosts were bound to CV-1 cells expressing HA-wt and the chimeras H/C/C, H/C/H, H/H/C, H/F/F, H/F/H, and H/H/F. Fusion was triggered by replacing neutral buffer (pH 7.4) with acidic buffer (pH 5.0) at RT. The diagrams show the percentage of HA-wt- or chimera-expressing CV-1 cells with calcein redistribution (white columns) and TMR-D redistribution (black columns) (for details, see Results) at 5 min (A) and 20 min (B) after the pH trigger. Each column represents the results of seven independent experiments (mean and standard error of estimate).

(iii) Vaccinia virus-infected cells. To exclude the possibility that vaccinia virus-specific products induce fusion, we studied in parallel whether label redistribution occurred also between labeled RBCs or RBC ghosts and vaccinia virus-infected CV-1 cells (no transfection with vector DNA [see Materials and Methods]). In this experiment, we observed no redistribution of fluorescent labels to CV-1 cells, neither for R18 nor for calcein plus TMR-D, even if RBCs or ghosts were bound to CV-1 cells by WGA before the pH trigger was set. The CV-1 cells and RBCs or RBC ghosts, as well as the labeling of the latter, remained stable and unaffected over a period of 3 h. Since additional control experiments exclude a possible inhibition of fusion by WGA in the medium (see also reference 34), redistribution of the various fluorescent labels between the transfected CV-1 cells and RBCs or RBC ghosts, described above, must be attributed to the fusion activity of HA-wt or chimeric proteins.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Impaired fusion pore enlargement of HA chimeras. In the present study we assessed the significance of the TM and CT of influenza virus HA (subtype H7 from A/fpv/rostock/34) for membrane fusion. While our previous work concentrated on investigating the role of palmitoylation (12, 28) and of the CT and TM of HA by replacement of both domains by those from another fusion protein (Sendai virus F protein) (34), here we have studied the fusion activity of HA chimeras containing the corresponding domains from a totally unrelated, nonviral and nonfusogenic protein, the cellular glycoprotein CD4. We found that the CT and TM of CD4 were tolerated by HA in terms of lipid mixing (redistribution of R18) and the formation of small aqueous fusion pores (transfer of calcein). However, a different picture emerged for these constructs when the enlargement of the aqueous fusion pore was monitored with the fluorescent high-MW dextran TMR-D (MW 10,000). HA-wt readily promoted the formation of a large pore. Extending our previous study, we also found that the HA/F chimeras promoted this late fusion step, although it was delayed with respect to that in HA-wt. We take this as a strong indication that the TM and/or the CT sequences of Sendai virus F protein form a functional entity with the domains of influenza virus HA (H7) during the fusion process, leading to pore formation and growth. The same conclusion can be drawn for the HA/CD4 chimera H/C/H. However, quite differently, the chimeric constructs H/C/C and H/H/C were not able to mediate enlargement of the aqueous pore, although membrane fusion and formation of small pores were not affected (Fig. 2). This suggests that the CT, but not the TM domain of CD4 was responsible for the failure of those two chimeras to mediate pore growth. We noted that the extent of pore formation and growth of H/C/H and HA/F chimeras was lower in comparison to that of HA-wt. This may be caused by their somewhat lower expression in the plasma membrane of CV-1 cells (Table 1). Importantly, the transfection efficiency and surface expression of H/C/C and H/H/C cannot account for their inability to promote pore enlargement, since neither parameter was significantly different from that of H/C/H (Table 1).

Relevance of the TM domain of HA to fusion. The observation that all chimeric proteins, in particular H/C/C and H/H/C, mediate membrane fusion as well as the formation of small aqueous fusion pores indicates that the CT does not affect the conformational change of the HA ectodomain required for initiating fusion as well as steps preceding the enlargement of an aqueous fusion pore. The data demonstrate that the CT affects late steps of the fusion process rather than early intermediates. The observation that the TM of a nonviral, nonfusion integral membrane protein, CD4, can functionally substitute for the corresponding HA domain confirms previous reports by others (23). Similar results have been obtained for other viral fusion proteins. Replacement of the TM of vesicular stomatitis virus G protein with that of CD4, of the viral nonfusion protein gD from herpes simplex virus, or of glycoprotein E3-11.6K from nonenveloped adenovirus preserved the ability of these constructs to mediate syncytium formation (25).

Modification of the CT of HA affects the fusion process. Amino acid residues of the cytoplasmic domain of influenza virus HA are highly conserved among the several HA subtypes, implicating its importance for influenza virus activities (15). So far it is not clear whether the CT plays any functional role in the fusion process. Influenza virus A/Udorn/72 with HA lacking its CT assembled and replicated almost as efficiently as did virions containing HA-wt (15). Deletion of the CT of HA (subtype H3) did not influence syncytium formation (35), and redistribution of a high-MW label (MW 40,000) between transfected CV-1 cells and RBCs was similar to that in HA-wt (22). In contrast, elongation of the CT of HA (subtype H7) by 6 aa caused a fivefold reduction in the extent of syncytium formation between HA-expressing CV-1 cells (27).

Properties of the CT of CD4 that may affect the fusion activity of HA. The most obvious difference between the CTs of HA and CD4 is in their lengths, with the CD4 CT being more than three times as long as the HA CT. In a previous study it was suggested that elongation of the CT of HA-wt may cause a reduction of membrane fusion activity and an impaired fusion pore formation (27). However, if elongation of the CT were the sole cause of the inhibition of pore growth, we would expect a similar inhibition for chimeras containing the CT of the F protein, which is four times as long as that of influenza virus HA-wt. However, this was not the case (Fig. 2 and 3). Furthermore, Melikyan et al. (23) observed the transfer of a large soluble aqueous dye (MW, 40,000) similar to that in HA-wt for a chimera whose TM and CT were replaced by the respective domains of the Env protein of Rous sarcoma virus. The CT of Env is 28 aa, about 2.5 times as long as that of HA-wt. On the other hand, shortening the CT of the pIgR used in that study to 11 aa did not render the respective chimeric HA competent to mediate content mixing (i.e., pore enlargement). Thus, the length of the CT of H/C/C and H/H/C is most probably not the cause of the observed impairment of pore enlargement.

View larger version (45K): [in this window] [in a new window]   FIG. 4.   Hydrophobicity plots of the cytoplasmic domains of influenza virus HA, Sendai virus F protein, vesicular stomatitis virus (VSV) G protein, human immunodeficiency virus type 1 (HIV-1) gp41, herpes simplex virus type 1 (HSV) gD, the deleted CT of pIgR, and human CD4. Hydrophobicity was calculated using the hydrophobicity scale of Abraham and Leo (1) by moving a window of 5 residues along the protein sequence in the direction from the N terminus to the C terminus (18).

How does the hydrophobicity of the cytoplasmic tail promote pore dilation? Based on the observation that GPI-anchored HA causes the formation of a metastable hemifusion diaphragm after triggering the conformational change of the ectodomain (17, 21, 24) the "elastic-coupling" model has been proposed to explain the complete membrane fusion and the subsequent formation of an aqueous fusion pore (21, 23). This model predicts that the stiff linkage between the ectodomain and the TM of HA causes destabilization and eventually rupture of the hemifusion diaphragm, leading to full membrane fusion and pore formation. As we discussed previously (12), an intriguing feature of the model could be an orientational shift of the TM within the viral (plasma) membrane from a vertical to a tilted position. The latter requires at least partial immersion of the CT into the lipid phase. Reduction of the hydrophobicity of the CT either by deacylation (12) or by replacement of the amino acid sequence could hinder its immersion and the tilting of the TM. This would interfere with the fusion process. This hypothesis is also consistent with the observation that fusion is not impaired in the case of HA lacking its CT (15). It is also feasible that modification of the CT may give rise to a strong interaction between HA trimers, forming a small fusion pore, which would inhibit pore enlargement. Future experiments are warranted to test these hypotheses and to determine how the hydrophobicity of the CT influences only the late step of fusion pore growth. Experiments involving recombinant modifications focused on the CT segment close to the TM of fusogenic proteins should contribute to our understanding of pore enlargement during membrane fusion.