![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
Previous Article | Next Article 
Journal of Virology, May 2002, p. 4603-4611, Vol. 76, No. 9
0022-538X/02/$04.00+0 DOI: 10.1128/JVI.76.9.4603-4611.2002
Copyright © 2002, American Society for Microbiology. All Rights Reserved.
Department of Microbiology, Kawasaki Medical School, Kurashiki, Okayama 701-0192, Japan
Received 15 October 2001/ Accepted 17 January 2002
 |
ABSTRACT |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
INTRODUCTION |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The HA1 and HA2 subunits are functionally specialized. HA1 carries receptor-binding activity, and HA2 mediates membrane fusion (4, 24). The membrane fusion process can be divided into three steps: lipid mixing between the outer lipid monolayers of the viral and endosomal membranes (hemifusion), aqueous pore formation by connecting the two inner lipid monolayers, and pore dilation (8). The ectodomain of HA2 is thought to mediate hemifusion, and the transmembrane and cytoplasmic domains are thought to play an important role in pore formation (10, 12, 17). Three cysteine residues in the C-terminal region of HA2 are highly conserved between various subtypes of HA. These residues are posttranslationally modified with palmitic acids through thioester bonds, and these palmitoyl chains bind the cytoplasmic domain to the inner surface of the viral membrane (19, 20, 23, 26). The strict conservation of the cysteine residues implies an important role for the acyl chains in the virus life cycle. However, the role of the acyl chains remains unclear.
It remains a matter of debate whether the palmitylation of HA2 is involved in membrane fusion. To examine this subject, many experiments have been performed with several subtypes of HA in expression and reverse genetics systems. Initially, it was reported that deacylation of the H2 subtype HA impaired syncytium formation, which should be caused by membrane fusion between HA-expressing cells and neighboring cells (19). However, subsequent experiments using H7, H3, and H2 subtypes showed that HA deacylation mutants mediated syncytium formation or membrane fusion between HA-expressing cells and erythrocytes as well as the wild-type HAs (20, 25, 26). Moreover, through reverse genetics experiments with an H3 subtype HA, infectious virus lacking all the acylation sites of HA was recovered (9), indicating that acylation of HA is not essential, at least for H3 HA, for mediating membrane fusion. However, the possibility remains that acylation affects the rate and extent of membrane fusion.
Recently, electrophysiological analysis of HA (H3 subtype) showed that deacylation dramatically decreased the frequency of "flicker," which reflects repetitious opening and closing of small fusion pores in the transition from hemifusion to pore formation (16). Fischer and colleagues reported that deacylation of H7 HA reduced syncytium formation (7). Since syncytium formation requires fusion pore dilation, these results imply that deacylation affects the pore dilation step. The importance of acylation may differ in different subtypes of HA. By reverse genetics experiments with the H3 subtype, deacylation mutants were easily obtained (9), whereas with the H1 subtype only mutants which had lost one acylation site could be obtained (30). Thus, a quantitative analysis is now required to elucidate the role of HA acylation in membrane fusion.
With this aim, we adopted the following strategy: (i) we examined H1 subtype HA because the reverse genetics study suggested that H1 HA was affected most strongly by deacylation as mentioned above; and (ii) we focused on the pore formation stage of the fusion reaction because deacylation of the H3 and H7 HAs seemed to affect flicker and pore dilation, respectively (7, 16). In the sequence of the fusion process, pore formation is located between flicker and pore dilation. In this paper, we set up a hypothesis that deacylation of HA impairs pore formation. To test this hypothesis, we designed cDNAs of H1 subtype HAs lacking acylation sites, expressed these mutants and wild-type HAs in CV-1 cells, and analyzed their membrane fusion activities quantitatively.
|
MATERIALS AND METHODS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Expression of deacylation mutant HAs. The pUCSVL DNAs containing wild-type or mutant HA genes were cut with BamHI to remove the plasmid parts (pUC sequence), recircularized, and then used to transfect CV-1 cells together with the helper DNA as described previously (21). The transfected CV-1 cells were grown in Dulbecco's modified Eagle's medium (DMEM) containing 5% fetal calf serum (FCS). At 7 to 10 days posttransfection, the cells were frozen and thawed three times to harvest recombinant virus stock. The recombinant virus stock was inoculated into fresh CV-1 cells to obtain recombinant virus for experiments. The recombinant viruses were applied to CV-1 cells 48 h before experiments unless otherwise indicated. Before hemadsorption and fusion tests, HA-expressing cells were treated with neuraminidase (Behringwerke) at 5 mU/ml in DMEM for 30 min at 37°C.
Quantification of HA expressed on CV-1 cell surface by flow cytometric analysis. HA-expressing CV-1 cells were detached from culture dishes by EDTA and trypsin treatment and washed once with DMEM containing 5% FCS. The cells were incubated with a fluorescein isothiocyanate (FITC)-conjugated antiserum recognizing the H1 subtype HA for 15 min at room temperature, washed twice with phosphate-buffered saline (PBS), and resuspended in PBS. The fluorescently labeled cell suspension was analyzed using a flow cytometer (Becton Dickinson FACSCalibur 3S).
Palmitylation and intracellular transport assay. To determine palmitylation of HA proteins, CV-1 cells at 42 h postinfection were metabolically labeled with [35S]methionine (1 µCi) in DMEM (1 ml) without methionine or [3H]palmitic acid (1 mCi) in DMEM (1 ml), for 6 h. After cell lysis, HA proteins were immunoprecipitated and analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) as described previously (21).
Hemadsorption assay. After neuraminidase treatment, HA-expressing cells were washed three times with DMEM and incubated with a suspension of freshly prepared guinea pig erythrocytes for 10 min at room temperature. After the incubation, the cells were washed three times with DMEM to remove unbound erythrocytes. To evaluate hemadsorption, the erythrocytes bound to the cells were lysed with distilled water, and the optical density of the lysates was measured at 575 nm to quantify the hemoglobin of cell-bound erythrocytes.
Syncytium formation assay. For the syncytium formation assay, CV-1 cells were grown on gridded culture dishes (Sarstedt) and infected with SV40 recombinant virus. After neuraminidase treatment, HA-expressing cells were incubated with trypsin (20 µg/ml) in DMEM for 10 min at 37°C to cleave HA into HA1 and HA2 subunits. Syncytium formation was induced by exposure of the cells to acidic DMEM (pH 5.0) for 5 min at 37°C. After exposure, the acidic medium was replaced by neutral DMEM (pH 7.2) containing 5% FCS and incubated for 8 h at 37°C. Thereafter, the cells were fixed with ethanol and stained with a 20-fold dilution of Giemsa's solution. To evaluate syncytium formation, the whole area of the bottom of the dish was observed under a microscope, and all syncytia were counted. Nuclei in 40 randomly selected syncytia were also counted. From these data, the number of cell-cell fusion events was calculated from the equation: (number of syncytia) x [(number of nuclei in a syncytium) - 1]. Parallel hemadsorption experiments were performed. Specific syncytium-forming activity was calculated from cell-cell fusion events divided by hemadsorption. Specific activity of HA mutants was normalized to that of wild-type HA.
Erythrocyte labeling for fusion assay. Erythrocytes were labeled with octadecyl rhodamine B (R18) and calcein according to a slightly modified protocol (21). For R18 labeling, a freshly prepared guinea pig erythrocyte suspension (0.25% in PBS, 4 ml) was mixed with 4 µl of R18 solution (2 mM in ethanol) and incubated for 1 h at room temperature with shaking. For calcein labeling, an erythrocyte suspension (1% in PBS, 1 ml) was mixed with 5 µl of calcein AM solution (10 mM in dimethyl sulfoxide) and incubated for 1 h at 37°C. After the labeling incubation, the erythrocytes were washed three times by centrifugation to remove unbound dye. Final pellets were suspended in 10 ml of DMEM.
Cell-erythrocyte fusion assay. For the cell-erythrocyte fusion assay, CV-1 cells were grown on glass base dishes (Iwaki) and infected with SV40 recombinant virus. After neuraminidase treatment, HA-expressing cells were treated with trypsin as above. The cells were incubated with a calcein-labeled erythrocyte suspension or a mixture of calcein- and R18-labeled erythrocytes for 10 min at room temperature and then washed three times with DMEM to remove unbound erythrocytes. Cell-erythrocyte fusion was induced by exposure of the cells to acidic DMEM (pH 4.0 to 5.0) for 5 min at 37°C. After acidic exposure, the medium was replaced by neutral Hanks' balanced salt solution containing 5 mM HEPES (pH 7.0).
The transfer of the dye(s) from erythrocytes to CV-1 cells was observed using a fluorescence microscope, and fluorescent images of the cells were recorded with a digital camera. From the digital images, fusion pore-forming activities of wild-type and mutant HAs were estimated by counting calcein-transferred cells among R18-transferred cells using the public-domain NIH Image program (W. Rasband, available to Internet users at zippy.nimh.nih.gov). In every sample, 50 R18-positive cells were examined.
Calcein leakage assay. HA-expressing CV-1 cells were treated with trypsin, incubated with calcein-labeled erythrocytes, and exposed to acidic medium as described above. After the exposure, the acidic medium was collected. The cells were lysed with distilled water for 1 min. The acidic medium and the lysate were neutralized with HEPES buffer, and calcein fluorescence was measured fluorometrically.
Data processing. For all quantitative analyses, the data from four independent experiments were merged and are presented in Table 1 and Fig. 3 and 6.
|
|
|
|
RESULTS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Metabolic labeling with [35S]methionine showed that the total synthesis of mutant HA proteins did not vary significantly from that of wild-type HA, while labeling with [3H]palmitic acid indicated that these HA mutants were less acylated than wild-type HA (data not shown).
No influence of deacylation on receptor binding of HA. We next examined whether HA mutants could bind to cell receptors as efficiently as wild-type HA. To compare the binding activities of wild-type and mutant HAs with each other, erythrocytes were applied to HA-expressing cells and bound erythrocytes were quantified. Hemadsorption of each HA was normalized to the HA expression level (total fluorescence in Table 1). As shown in Table 1, hemadsorption of HA mutants was not significantly different from that of wild-type HA. However, two HA mutants, m2/1+2 and m2/2+3, showed slightly lower hemadsorption. The lower hemadsorption of the HA mutants m2/1+2 and m2/2+3 is probably a direct effect not of the deacylation but rather of the lower expression level of these mutants.
Deacylation impaired formation of fusion pores by HA. We examined the formation of aqueous fusion pores using a dye transfer assay. Mixtures of R18- and calcein-labeled erythrocytes were bound to HA-expressing CV-1 cells, and the culture medium was replaced with acidic DMEM (pH 5.0) to induce membrane fusion. After the acidic medium exposure, the cells were neutralized and observed with a fluorescence microscope. As shown in Fig. 2, R18 was efficiently transferred to the cells expressing wild-type and mutant HAs. However, calcein transfer into the cells expressing HA mutants was dramatically impaired. We estimated the efficiency of pore formation by wild-type and mutant HAs from the percentage of calcein-transferred cells against R18-transferred cells. As shown in Fig. 3A, the efficiency of pore formation by wild-type HA was 85%, whereas that by HA mutants was between 12% (m2/1+2) and 32% (m1/1).
|
|
HA mutants affected membrane permeability. In addition to the swelling of the erythrocytes, an increase in the fluorescence background was observed in the mutant HA-expressing cell cultures after pH 4.0 treatment (Fig. 4A). To test whether HA mutants induced calcein leakage into the outer medium at pH 4.0, the medium used for acidic treatment was collected, and calcein fluorescence in the medium was measured with a fluorometer. As shown in Fig. 3C, fluorescence in the medium of HA mutant-expressing cells clearly increased after pH 4.0 treatment. The amount of calcein which leaked into medium of HA mutant-expressing cells was three times higher than that of wild-type HA-expressing cells. These data indicate that deacylation of HA made the membrane permeable at pH 4.0.
Deacylation impaired syncytium formation by HA. The results in Fig. 3A demonstrate that deacylation impairs pore formation by H1 HA. Syncytium formation should also therefore be clearly impaired by deacylation of HA. We examined the effect of deacylation of HA on syncytium formation. Cells expressing wild-type and mutant HAs were treated with trypsin, exposed to acidic DMEM (pH 5.0), neutralized, and incubated for 8 h at 37°C. Thereafter, syncytia and nuclei were counted using a light microscope, and cell-cell fusion events were estimated as described in Materials and Methods. As shown in Fig. 5, syncytium formation was dramatically decreased by deacylation of HA. Syncytium-forming activities of all HA mutants were between 1.4% (m1/2) and 12% (m1/1) of wild-type HA activity (Fig. 6).
|
|
DISCUSSION |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Impairment of pore and syncytium formation by HA mutants depends on deacylation and not on HA mutant density. Several researchers have reported that HA density affects membrane fusion efficiency (2, 5, 6, 15). In the present study, five HA mutants, m1/1, m1/2, m3/3, m2/2+3, and m3, were expressed at the same level as wild-type HA, and two HA mutants, m/1+2 and 2+3, were expressed at a slightly lower level than the others (Table 1). However, all HA mutants showed significantly lower activities of pore and syncytium formation than wild-type HA (Fig. 3A and 6). This indicates, therefore, that the observed impairment of pore and syncytium formation by HA mutants is the result of deacylation and not of the density of HA mutants on the cell surface.
Deacylation of HA affects pore formation. Deacylation affected syncytium formation more significantly than pore formation (Fig. 3A and 6). These results imply that deacylation may also affect the fusion pore dilation step, because syncytium formation requires pore dilation. There are two possible interpretations for the deacylation effect on syncytium formation. First, deacylation directly affects pore formation and consequently the restricted pore formation causes poor syncytium formation. Second, deacylation directly affects both steps of pore formation and dilation. The average pore-forming activity of HA mutants was 27% of that by wild-type HA (Fig. 3A), whereas the average syncytium-forming activity of HA mutants was 4% of that of wild-type HA (Fig. 6). The more severe effect of deacylation on syncytium formation appears to support the second interpretation. However, we cannot exclude the first possibility, because the pore-forming activities of HA mutants (27%) may have been overestimated. The efficiency of pore formation was calculated as the number of calcein-transferred cells without considering the amount of transferred calcein in individual cells. As shown in Fig. 2, HA mutant-expressing cells contained obviously lower amounts of calcein than wild-type HA-expressing cells. Presumably, the net efficiency of pore formation of the HA mutants is several times lower than 27% of the wild-type HA level. Thus, the pore formation efficiency of HA mutants is sufficiently low to explain the syncytium formation activities observed in HA mutant-expressing cells. Either way, it is clear that deacylation of HA directly affects fusion pore formation. In other words, the acyl chains in the C-terminal region of HA2 may promote fusion pore formation.
Deacylation of HA induced calcein leakage. We observed soluble dye (calcein) leakage from erythrocytes when the HA mutant-expressing cells were treated at pH 4.0. Several researchers have reported that influenza virus induces hemoglobin leakage (hemolysis) under the acidic conditions which occur during virus-erythrocyte fusion (13, 27, 29). Watanabe and colleagues reported that hemolysis was more evident if the virus membrane was damaged from severe treatments such as repetitious freezing and thawing or sonication (27). The abnormal acidic treatment (pH 4.0) in our experiment might damage the cellular membrane and consequently cause calcein leakage. However, by comparing the mutant HAs with the wild type, it is clear that deacylation contributed significantly to the leakage (see Fig. 3C) and that these observations were not due simply to the acidic treatment. As shown in Fig. 2 and 3C, no significant leakage by HA mutants was observed in the hemifusion stage (at pH 5.0). Furthermore, the calcein leakage and pore formation by HA mutants increased simultaneously at pH 4.0 (Fig. 3B and C). These results imply that the leakage is induced during and/or after pore formation. Presumably, acyl chains maintain lipid bilayer integrity during and/or after fusion pore formation.
Deacylation effects of H1 HA compared with other subtypes of HA. Recently, Fischer and colleagues reported that deacylation of an H7 subtype HA did not affect pore formation but did reduce syncytium formation (7). They suggested that the deacylation of HA might interfere with the dilation of the fusion pore. However, another possibility is that the deacylation also impaired pore formation with the H7 HA to some extent, since in that report pore formation was not analyzed quantitatively. Because the inhibitory effect of deacylation on syncytium formation with H7 HA is not as evident as that seen with H1 HA, it would be difficult to detect the effect of deacylation on pore formation without quantitative evaluation.
In the case of H3 HA, from many studies on the properties of HA lacking acylation sites, the HA mutants seemed to be identical to the wild-type HA in all aspects investigated, such as membrane fusion, receptor binding, biosynthesis, and virus assembly (9, 23, 25). However, in those studies, a quantitative study of pore formation was not performed. Recently, using an electrophysiological analysis, Melikyan and colleagues found that deacylation of H3 HA reduced the frequency of flicker, in which a small fusion pore repetitiously opens and closes in the transition from hemifusion to stable pore formation (16). They also reported that pore formation was not impaired by deacylation. This is inconsistent with our observation of H1 HA. However, it still remains a possibility that the extent of pore formation might be decreased by deacylation of H3 HA, because they examined the kinetics of pore formation up to a maximum level but did not estimate the maximum level itself. Although the biological significance of flicker is not yet known, it may be reasonable to suggest that higher frequency of flicker causes more efficient formation of fusion pores. Since flicker by H1 and H2 HA was also observed before fusion pore formation (14), it remains to be solved whether the low efficiency of pore formation by deacylated H1 HA observed in this study is caused by the restricted flicker.
Influence of impaired pore formation on recombinant virus viability. Reverse genetics studies indicated that the recovery of recombinant virus containing deacylated HA depended on the HA subtype. In H3 HA, the virus containing the triple mutant HA, lacking three acylation sites, was recovered, whereas in H1 HA, viruses with only a single mutant HA were recovered (9, 30). These two subtypes differ in the influence of HA deacylation on fusion pore formation, as discussed above. These observations suggested that the efficiency of pore formation by HA mutants correlates to the viability of recombinant virus. However, the efficiency of pore formation is not sufficient to explain the difference in virus recovery among H1 HA mutants. As shown in Fig. 3, the efficiency of pore formation is not significantly different among HA mutants. By contrast, recombinant virus containing a single mutant HA lacking a first or second acylation site was recovered, but virus containing a third-site single, double, or triple mutant HA was not recovered (30).
Two interpretations for this inconsistency are possible. First, it is possible that a small difference in pore formation among HA mutants might not have been detected by our evaluation, because some factors of pore formation reaction such as the amount of calcein transfer per cell and kinetics was not considered. Second, mutant virus recovery might be affected by factors other than impairment of the HA pore-forming activity. For example, Kawaoka recently found that the 5'- and 3'-terminal regions of viral RNA contain signal sequences to be packed into a viral particle (Y. Kawaoka, Symp. 49th Annu. Meet. Jpn. Soc. Virol. Abstr., p. 19, 2001). The sequence encoding the acylation sites is located close to the 5' terminus of viral RNA, and recombinant viruses containing H1 HA mutants lacking the third site, which is encoded closest to the 5' terminus, have not been recovered (30). Thus, it is possible that the alteration of the nucleotide close to the terminus might destroy the packaging signal for the viral RNA. In any case, further study is required to explain the difference in virus viability among H1 HA mutants.
How could acyl chains of HA enhance pore formation? Conformational change of influenza virus HA molecules in the membrane fusion process was proposed from several studies on the three-dimensional structure of the viral fusion protein (3, 24, 28). Reviewed briefly, under acidic conditions, the HA2 ectodomain is refolded by changing its alpha-helix contents, and the fusion peptide is positioned on the tip of the HA2 ectodomain. After insertion of the fusion peptide into a target membrane, the HA2 ectodomain may become bent like a hairpin by interaction between two alpha-helix domains in the ectodomain, and consequently the fusion peptide may approach the HA2 membrane-anchoring region. This series of conformational changes of the HA2 ectodomain may induce contact between virus and target membranes and then cause hemifusion (Fig. 7A), since even the HA ectodomain bound to the glycosyl phosphatidylinositol membrane anchor could induce hemifusion (10).
|
Another possible function of the acyl chains is to facilitate interaction between the transmembrane domain and the fusion peptide. Recently, Qiao and colleagues found that a point mutation in the fusion peptide arrested fusion at the hemifusion stage (22). This reveals that the fusion peptide also contributes to the transition from hemifusion to pore formation. The fusion peptide should be close to the membrane-anchoring region at the final stage of fusion reaction, and therefore the transmembrane domain and the fusion peptide at the hemifusion stage are thought to interact with each other in the lipid bilayer of the hemifusion area (hemifusion diaphragm), perturb the lipid of the hemifusion diaphragm, and then open the fusion pore (1, 18, 24). Presumably, the acyl chains may facilitate interaction between the transmembrane domain and the fusion peptide in the hemifusion diaphragm and consequently enhance fusion pore opening (Fig. 7).
Conclusion. From analysis of membrane fusion by influenza virus HA (H1 subtype) lacking acylation sites, we conclude that acyl chains facilitate the transition from hemifusion to pore formation.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
|
REFERENCES |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| J. Bacteriol. | Mol. Cell. Biol. | Microbiol. Mol. Biol. Rev. |
|---|
| Clin. Vaccine Immunol. | ALL ASM JOURNALS |
|---|
