Functional Characterization of Heptad Repeat 1 and 2 Mutants of the Spike Protein of Severe Acute Respiratory Syndrome Coronavirus

Cells and antibodies.

Human embryonic kidney 293T and Vero E6 cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% heat-inactivated fetal bovine serum. Hybridoma 183 (clone H12-5C) was described previously (29). Rabbit anti-SP4 was generated by immunizing rabbits with residues 435 to 454 of the S protein (12). Rabbit anti-S protein antiserum was prepared by immunizing rabbits with a bacterially expressed recombinant soluble S protein (residues 511 to 1255) (25). A monoclonal antibody (MAb) directed against ACE2 was purchased from R&D (Minneapolis, MN).

Construction of plasmids.

To construct S plasmids encoding mutations in the HR1 and HR2 sequences, site-directed mutagenesis based on a codon-optimized S gene, pCDNA3.1(−)S, was performed using a QuikChange site-directed mutagenesis kit (Stratagene). Since the S gene was cloned at the XbaI and BamHI sites in pCDNA3.1(−) to facilitate the subcloning steps, the KpnI site in multiple cloning sites of the vector and 3′ to the S gene was first deleted by replacing a 0.15-kb PCR fragment. This DNA fragment, which was flanked by the HindIII and BamHI sites and contained the sequence for transmembrane and cytoplasmic domains of the vesicular stomatitis virus (VSV) glycoprotein G protein, was substituted for the corresponding sequence of pCDNA3.1(−)S. Introduction of this DNA fragment did not affect the coding sequence of the S gene. In order to generate HR2 mutants, the 0.85-kb EcoRI-BamHI fragment of the S gene was subcloned in a pBluescript SKII(+) vector at the corresponding sites, and the plasmid obtained was used as a template to generate HR2 mutations. The mutated fragments were then subcloned back into pCDNA3.1(−)S to generate HR2 mutants. To construct the HR1 mutants, the 3-kb XbaI-EcoRV fragment from pCDNA3.1(−)S was subcloned in the corresponding sites of the pBluescript SKII(+) vector, and the resultant plasmid was used as the template to generate HR1 mutations. The 2-kb KpnI-EcoRI fragments containing HR1 mutations isolated from these mutated plasmids were then subcloned back into the corresponding sites in pCDNA3.1(−)S to yield various HR1 mutants. The following oligonucleotide pairs were used to generate the HR1 and HR2 mutants: 913P, 5′-TTCAACAAGGCCCCTAGCCAGATCCAGGAGAGCCTGACC-3′ (sense) and 5′-CTGGATCTGGCTAGGGGCCTTGTTGAACTGGTTGGC-3′ (antisense); 927P, 5′-ACCAGCACGGCCCCAGGCAAGCTGCAGGACGTGG-3′ (sense) and 5′-CTGCAGCTTGCCTGGGGCCGTGCTGGTCGTGG-3′ (antisense); 941P, 5′-AACGCCCAGGCCCCTAACACCCTGGTGAAGCAGC-3′ (sense) and 5′-CACCAGGGTGTTAGGGGCCTGGGCGTTCTGGTTG-3′ (antisense); 955P, 5′-AACTTCGGCGCCCCAAGCTCCGTGCTGAACGATATC-3 (sense) and 5′-CAGCACGGAGCTTGGGGCGCCGAAGTTGCTGGAC-3′ (antisense); 1151P, 5′-GACCTGGGCGACCCTAGCGGCATCAACGCCAGCG-3′ (sense) and 5′-GTTGATGCCGCTAGGGTCGCCCAGGTCCACGTCC-3′ (antisense); 1165P, 5′-ATCCAGAAGGAGCCAGACCGCCTGAACGAGGTGG-3′ (sense) and 5′-GTTCAGGCGGTCTGGCTCCTTCTGGATGTTCACG-3′ (antisense); and 1179P, 5′-CTGAACGAGAGCCCTATCGACCTCCAGGAGCTGGG-3′ (sense) and 5′-CTGGAGGTCGATAGGGCTCTCGTTCAGGTTCTTGG-3′ (antisense). The underlined nucleotides mark the respective sense and antisense primer Pro substitutions. Primers OptiS2743 (5′-CAACAACGTGTTCCAGACCC-3′) and OptiS3210 (5′-CTTCAACTTCAGCCAGATCC-3′) were used to sequence the DNA fragments containing HR1 mutations, and primer OptiS4154 (5′-TCATCACCACGGACAACACC-3′) was used to sequence the DNA fragments containing HR2 mutations.

pCAGGS-MCS was constructed by replacing the 1-kb PstI-to-EcoRV coding region of pCAGGS-NLS-Cre with a synthetic oligonucleotide linker, 5′-CTGCAGGAATTCTCGAGCTCACGCGTGCATGCGATATC-3′, which contained the PstI/EcoRI/XhoI/SacI/MluI/Sph/EcoRV recognition sequence. The SalI site at the 5′ boundary of the cytomegalovirus immediate-early 1 enhancer was then replaced by an oligonucleotide containing the SalI/NotI/XbaI sites, and the HindIII site at the 3′ boundary of the poly(A) signal was replaced by the NotI site. Therefore, the entire expression cassette could be easily isolated using NotI digestion. ploxP1, derived from pBluescript SKII, contained two tandem repeats of the 34-bp loxP sequence in multiple cloning sites. To generate ploxP-EGFP-loxP, ploxP1 was codigested with SalI and NheI, followed by Klenow fill-in. The 1.15-kb KpnI-to-SspI fragment encoding the enhanced fluorescence green protein (EGFP) and poly(A) sequences of pIRES-EGFP (Clontech) was blunt ended by T4 DNA polymerase and then inserted into the treated ploxP1 vector by blunt-end ligation. The XhoI-EcoRV fragment containing the EGFP gene flanked at both ends by loxP sequences was then isolated and inserted into the same restriction enzyme sites in pCAGGS-MCS to create pCAGGS-loxP-EGFP-loxP. The 0.7-kb BamHI-to-NotI fragment of pDsRed2 N1 (Clontech) was blunt ended by fill-in with the Klenow fragment and was then inserted into the EcoRV site of pCAGGS-loxP-EGFP-loxP to obtain pCAGGS-loxP-EGFP-loxP-DsRed2.

pCDNA3.1(−)ACE2-Ig was generated by replacing the EcoRI-to-KpnI fragment in pCDNA3.1(−)ACE2 with a PCR product encoding the Fc domain of a human immunoglobulin γ1 (Igγ1) heavy chain. The insert was generated using 5′-TGAATTCACCAGCACCTGAACTCCTGG-3′ and 5′-TGGTACCTCATTTACCCGGGGACAGGGAGAGG-3′ as the primers and p2C11-γ1, which encodes a human IgG1 Fc domain, as the template in PCR. The resultant plasmid encoded a fusion protein containing the CH2 and CH3 sequences of the Fc domain, beginning with Pro-Ala-Pro-Glu and ending with Gly-Lys-Val-Asp, attached at the C terminus of Ser-602 within the ectodomain of ACE2. To construct pCAGGS-ACE2-Ig, pCDNA3.1(−)ACE2-Ig was digested with XbaI, blunt ended with the Klenow fragment, and cut by KpnI to obtain the ACE2-CH2-CH3 fusion gene. The pCAGGS-MCS vector was digested by EcoRI and treated with the Klenow enzyme to fill in the staggered end before KpnI digestion. Then, the ACE2-CH2-CH3 DNA fragment was inserted into the treated pCAGGS-MCS vector to obtain pCAGGS-ACE2-Ig.

Transfection of DNA plasmids.

For expression of the SARS S protein, 293T cells grown on 6- or 10-cm petri dishes were transfected with 3 or 10 μg of the codon-optimized wild-type (wt) and mutant pCDNA3.1(−)S constructs by a standard calcium phosphate coprecipitation method (14). For the one-cycle viral replication assay, 293T cells grown in 10-cm petri dishes were cotransfected with 10 μg of pNL4-3R⁻E⁻Luc and 12 μg of either the wt or mutant pCDNA3.1(−)S plasmids. Alternatively, cells were cotransfected with 10 μg of pHXBΔBglCAT and 7.5 μg of either the wt or mutant S plasmids. pCDNA3, which was added to the transfection mixture that did not contain S expression plasmids, was used as a negative control.

Virus preparation and Western blot analysis.

Two days after transfection, cell-free virions were isolated through a 20% sucrose cushion as described previously (13), and equal volumes of cell and viral lysates were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and Western blotting. The membrane blots were incubated with the appropriate primary and secondary antibodies, followed by enhanced chemiluminescence detection.

Assessment of the ACE2-binding ability of the S proteins.

293T cells grown on 6-cm dishes were cotransfected with 2 μg of either the wt or mutant pCDNA3.1(−)S plasmids together with 2 μg of pCAGGS-ACE2-Ig using the Lipofectamine transfection reagent (Invitrogen, Carlsbad, CA) according to the procedures provided by the manufacturer. Two days after transfection, cells were washed with phosphate-buffered saline (PBS) and then lysed in 200 μl of PBS containing 1% 3-[(3-cholamidopropyl)-dimethylammonio]-1-propanesulfonate (CHAPS) and protease inhibitor cocktail (Roche) on ice for 30 min, followed by centrifugation to remove cell debris. Equal volumes of cell lysates were incubated with protein A-Sepharose 4B beads. After extensive washes with PBS containing 0.5% CHAPS, the bound proteins were analyzed by Western blotting using an ACE2 MAb and rabbit anti-S antiserum to detect ACE2 and the S protein, respectively. Alternatively, culture supernatants collected from 293T cells transfected with pCDNA3.1(−)ACE2-ecto were concentrated by Amicon Centriprep YM-10 membranes (Millipore, Bedford, MA). 293T cells transfected with each of the wt and mutant S plasmids were lysed with PBS containing 1% 3-[(3-cholamidopropyl)-dimethylammonio]-2-hydroxy-1-propanesulfonate (CHAPSO) and protease inhibitor cocktail. The cell lysates were then incubated with concentrated soluble ACE2 at 4°C for 2 h, followed by incubation with rabbit anti-S-coated protein A-Sepharose 4B beads at 4°C overnight. After being washed extensively with PBS containing 0.5% CHAPSO, the immune complexes were analyzed by SDS-PAGE, followed by Western blotting using rabbit anti-S and anti-ACE2 MAb to detect precipitated S and ACE2 proteins.

SARS S protein-mediated membrane fusion.

To prepare effector cells for qualitative analysis of S-mediated membrane fusion, 293T cells grown on 6-cm petri dishes were cotransfected by the Lipofectamine transfection method with 3 μg of either the wt or mutant pCDNA3.1(−)S plasmids and 1 μg of pCAGGS-NLS-Cre. For preparation of target cells, 293T cells grown on two 10-cm dishes were cotransfected with 6 μg of pCDNA3.1(−)ACE2 and 4 μg of pCAGGS-loxP-EGFP-loxP-DsRed2. One day after transfection, PBS-washed transfected cells were trypsinized, and 6 × 10⁵ (each) of the effector and target cells were seeded onto 6-cm dishes and then cocultured for 2 days. Cocultured cells were examined under a Zeiss model 135 inverted fluorescence microscope (Carl Zeiss AG, Germany), and green and red fluorescent images were overlaid. For quantitative membrane fusion analysis, effector cells were cotransfected with 4 μg of the wt or mutant S expression plasmids and 2.5 μg of a cytomegalovirus promoter-driven HIV-1 Tat expression plasmid, pCEP4-tat, and target cell were cotransfected with 10 μg of pCDNA3.1(−)ACE2 and 5 μg pIIIexcat, a chloramphenicol acetyltransferase (cat) gene linked to the HIV-1 LTR, by the calcium phosphate coprecipitation method. On the following day, effector and target cells were trypsinized, and 6 × 10⁵ of each kind of cell were cocultured. Two days after coculturing, cells were harvested for CAT activity measurement.

HIV/SARS S pseudotype infection assay.

One day before pseudotype infection, 3 ×10⁴ or 1.5 × 10⁵ Vero E6 cells were seeded onto 24-well plates (for the luciferase assay) or 6-well dishes (for the CAT assay), respectively. Cell-free NL4-3R⁻E⁻Luc or HXBΔBglCAT pseudotypes containing 6 × 10⁴ or 10⁵ cpm of reverse transcriptase (RT) activity in 0.6 or 2 ml, respectively, of Dulbecco's modified Eagle's medium containing 2.5% fetal bovine serum in the presence of 8 μg/ml of Polybrene were added to the wells. After centrifugation at 2,200 rpm for 2 h at room temperature, the cultures were incubated at 37°C. Two days after transfection, cells were harvested and prepared for the luciferase or CAT assay.

RT, luciferase, and CAT assays.

Cell-free, SARS S protein-pseudotyped viruses were assayed for virion-associated RT activity as described previously (13, 14). Cell lysates (25 μl) prepared from NL4-3R⁻E⁻Luc pseudotype infection were assayed for firefly luciferase activity using 50 μl of luciferase assay substrate reagent (Promega). Cell extracts prepared from an HXBΔBglCAT pseudotype infection were assayed for CAT activity using a Packard Instant Imager (model A202401; Packard Instrument Company, Meriden, CT) as described previously (14).

Binding of HIV/SARS S pseudotypes to ACE2-expressing cells.

The wt and mutant pseudotypes produced as described above were concentrated by an Amicon Ultra 15 filter with a nominal molecular mass limit of 100 kDa (Millipore), and comparable amounts of wt and mutant pseudotypes (containing 10⁶ cpm of RT activity) were incubated with 293T cells transfected with pCDNA3.1(−)ACE2 at 4°C for 1 h. After being washed three times with PBS containing 0.5% bovine serum albumin (BSA), cells were blocked with PBS containing 2% BSA on ice for 30 min. The cells were successively incubated with rabbit anti-S and fluorescein isothiocyanate-conjugated anti-rabbit IgG on ice. After each antibody incubation, the cells were washed three times with ice-cold PBS containing 0.5% BSA. After fixation with 4% paraformaldehyde for 30 min, the immunostained cells were quantitated by fluorescence-activated cell sorter analysis according to the procedure described previously (10).

Pulse-chase and immunoprecipitation.

Two days after transfection, 293T cells expressing the wt or 913P mutant S proteins were metabolically labeled with [³⁵S]methionine for 30 min and chased with excess cold methionine as described previously (10, 29) for different time periods. Equal amounts of cell lysates were incubated with rabbit anti-S preadsorbed onto protein A-Sepharose 4B beads at 4°C overnight, extensively washed, and separated by SDS-PAGE followed by fluorography.

Cell surface biotinylation of SARS spike proteins.

Transfected 293T cells grown in 6-cm petri dishes were metabolically labeled with [³⁵S]methionine overnight, and labeled cells were surface biotinylated with 0.25 mg/ml of membrane-impermeable sulfo-N-hydroxysuccinimide-biotin on ice for 30 min as described previously (10, 11, 29). Cell lysates were immunoprecipitated with rabbit anti-S preadsorbed onto protein A-Sepharose 4B beads. The precipitated S proteins were released from the beads by being boiled in sample loading buffer without bromophenol blue. A fraction of samples was directly resolved by SDS-PAGE, and another fraction of samples was precipitated with neutravidin-agarose beads (Pierce) prior to SDS-PAGE.

Construction and expression of the HR1 and HR2 mutants.

To understand the functional involvement of the two HR1 and HR2 motifs in the SARS-CoV life cycle, the HR residues Ile-913, Leu-927, Leu-941, and Ile-955 within HR1 and Ile-1151, Ile-1165, and Leu-1179 within HR2 (Fig. 1A) were individually replaced by a Pro residue via site-directed mutagenesis based on pCDNA3.1(−)S, a codon-optimized S expression plasmid (31, 36). The relative positions of these substituted residues in the α-helical, hairpin structure are indicated in Fig. 1B. Pro is believed to drastically disrupt the α-helical structures of HR motifs in envelope glycoproteins. pCDNA3.1(−)S encodes the signal sequence of CD5, the entire TOR2 strain S protein but without its signal sequence (residues 12 to 1255), and a nine-residue C-terminal tag, TETSQVAPA, derived from the C terminus of bovine rhodopsin.

To assess the expression of S proteins, 293T cells were transfected with each of the wt and mutant constructs, and cell lysates prepared at 24 h after transfection were analyzed by Western blotting using a rabbit monospecific anti-SP4 polyclonal antibody, which is directed against residues 435 to 454 of the S protein. The expressed wt S protein migrated as a tightly spaced doublet, marked as S_L and S_H, around 180 to 200 kDa in SDS-PAGE, with the faster-migrating, lower-molecular-mass S_L form being predominant over the slower-migrating, higher-molecular-mass S_H form (Fig. 2A, lane 2). Similar results were obtained when a rabbit polyclonal antibody directed against a recombinant soluble S protein was used in Western blotting (data not shown). Consistent with other results (55), no significant cleavage products of the S protein were observed in addition to these two forms. Furthermore, a species migrating between the separating and stacking gels was noted (Fig. 2A, lane 2). This high-molecular-weight species represented aggregates or high-ordered multimeric structures of the S protein that were not dissociated by treatment with SDS and the reducing agent during the boiling of samples (2, 43). When cell lysates prepared at 48 h posttransfection were analyzed, the S_H form of the wt S protein was more predominant than the S_L form (Fig. 2B, lane 2). These results together suggest that S_L is converted to S_H during biosynthesis and maturation of the S protein. The nature of this doublet may represent differential glycosylation of the wt S protein (2, 43). An endoglycosidase H digestion study also showed that the faster-migrating, lower-molecular-weight species contains high-mannose N-glycans while the slower-migrating, higher-molecular-weight species contains the complex-type N-glycans (41).

When the expressions of the mutant proteins were analyzed, with the exception of the 913P mutant, all mutants produced S_L and S_H levels comparable to those produced by the wt S construct (Fig. 2A and B). Nevertheless, the 913P mutant produced a smaller amount of the S_H form than did the wt and other mutants regardless of whether lysates prepared at 24 or 48 h were analyzed (Fig. 2A and B, lane 3). Moreover, high-ordered multimeric structures of all mutants were also observed (Fig. 2A and B).

Altered processing of the 913P mutant S protein.

To understand the relationship between the S_L and S_H forms of the wt and 913P mutant proteins, a pulse-chase study was performed. Cell lysates were precipitated with rabbit anti-S antiserum, and the precipitated proteins were resolved by SDS-PAGE. The intensity of the wt S_L form decreased while that of the S_H form increased as the chase proceeded (Fig. 3, top panel), indicating that wt S_L is converted to S_H during protein maturation. Although the intensity of the 913P mutant S_L gradually decreased at a rate slower than that of wt S_L as the chase proceeded, the S_H form of this mutant could not be efficiently detected even after 8 h of chasing (Fig. 3, bottom panel). These results indicate that a mutation at Ile-913 may affect the glycosylation process of the S protein.

Cell surface expression of the HR1 and HR2 mutants.

To understand whether the S_L and S_H forms of the wt and mutants are expressed on the cell surface, 293T cells transfected with the wt or each of the mutant S expression plasmids were metabolically labeled with [³⁵S]methionine, and cell surface biotinylation using membrane-impermeable sulfo-N-hydroxysuccinimide-biotin was performed. Cell lysates were first immunoprecipitated with rabbit anti-S, and the immune complexes were directly separated by SDS-PAGE or precipitated with neutravidin-agarose beads prior to SDS-PAGE. As shown by the Western blot analyses (Fig. 2), with the exception of the 913P mutant in which only the S_L form was observed, all mutants produced comparable or even slightly greater amounts of the S_L and S_H forms than did the wt construct when the total protein expression was assessed (Fig. 4A, top panel). The 913P mutant showed only the S_L form on the cell surface, whereas all of the other mutants produced similar or even slightly greater amounts of the S_L and S_H forms on the cell surface than the wt protein (Fig. 4A, bottom panel). This result indicated that mutations in the HR1 and HR2 sequences do not affect cell surface expression of the S protein.

ACE2-binding ability of the HR1 and HR2 mutants.

The ACE2-binding site in the S protein has been mapped to residues 270 to 510 (1) or 318 to 510 (54), as shown by the binding of S fragments to Vero E6 cells, followed by flow cytometry or by immunoprecipitation of ³⁵S-labeled ACE2 with S fragments fused to the Fc domain of human IgG1, respectively. It was expected that mutations in the HR1 or HR2 motif would not affect the receptor-binding ability of the S protein. To confirm this notion, 293T cells were cotransfected with or without the wt or mutant S expression plasmids in the presence or absence of a plasmid encoding soluble ACE2 fused to the Fc domain of human IgG1. Cell lysates were precipitated with protein A-Sepharose beads, and the bound proteins were analyzed by Western blotting using rabbit anti-S and MAb anti-ACE2. In the absence of ACE2-Ig coexpression, the wt S protein did not bind to the protein A-Sepharose 4B beads (Fig. 4B, lane 2). Comparison of the mobility of the precipitated S protein to that of the S_L and S_H forms shown in Fig. 2, 3, and 4A revealed that S_L but not S_H was bound to ACE2-Ig (Fig. 4B, lane 4). The wt S and ACE2-Ig may rapidly interact with each other in the endoplasmic reticulum after biosynthesis, resulting in coprecipitation of the wt S_L form with ACE2-Ig-coated protein A-Sepharose beads. However, the wt S molecules which do not bind to ACE2 are transported to the Golgi for further processing into S_H. A similar or even slightly greater amount of mutant S_L proteins than that of wt S_L was bound to ACE2-Ig (Fig. 4B, lanes 4 to 11).

To further determine the ACE2-binding ability of these S mutants, concentrated soluble ACE2 collected from culture supernatants of 293T cells transfected with pCDNA3.1(−)ACE2-ecto (31) was incubated with lysates of cells transfected with each of the wt and mutant S plasmids. After incubation with rabbit anti-S-coated protein A-Sepharose 4B beads, the precipitated complexes were resolved by SDS-PAGE, followed by Western blotting using rabbit anti-S and ACE2 MAb. This assay determined the specific binding of wt S with ACE2 since incubation with rabbit anti-S, but not with a rabbit control serum, coprecipitated ACE2 (Fig. 4C, left panel). Levels of soluble ACE2 coprecipitated with mutant proteins were comparable to or even greater than that of ACE2 coprecipitated with the wt protein (Fig. 4C, right panel). Also, the levels of soluble ACE2 coprecipitated paralleled the levels of mutant proteins precipitated (Fig. 4C, right panel). These results taken together indicated that mutations in the HR1 or HR2 motif do not affect the receptor-binding ability of the S protein.

Assessment of the membrane fusion abilities of the HR1 and HR2 mutants.

To examine whether these S mutants expressed on the cell surface are able to mediate membrane fusion, we first developed a sensitive cell-to-cell fusion assay based on the loxP-Cre recombination system of bacteriophage P1 (27, 48). Effector cells cotransfected with the S- and Cre-expressing plasmids were cocultured with target cells cotransfected with the loxP-EGFP-loxP-DsRed2 gene and an ACE2 expression plasmid. If membrane fusion between the effector and target cells occurred, Cre recombinase would catalyze the reciprocal site-specific recombination between the two specific 34-bp sites of loxP. The intramolecular recombination of the two loxP sites positioned head to tail resulted in deletion of the intervening EGFP sequences, and an intermolecular recombination caused integration at the loxP site, converting the target gene from EGFP expression to DsRed2 expression. The wt S-mediated membrane fusion was strictly dependent on the S protein expressed on effector cells and ACE2 on target cells (Fig. 5A). The 913P, 927P, 941P, 955P, and 1165P mutants displayed significantly impaired membrane fusion phenotypes compared to the wt S protein (Fig. 5B). However, the 1151P and 1179P mutants still exhibited effective membrane fusion (Fig. 5B).

To quantify membrane fusion mediated by the wt and mutant S proteins, 293T cells cotransfected with wt or mutant S and HIV-1 Tat expression plasmids were cocultured with 293T cells cotransfected with the ACE2 expression plasmid along with pIIIexcat, in which a cat gene is driven by the HIV-1 LTR. This assay is based on the ability of Tat expressed in S-synthesizing effector cells to transactivate HIV-1 LTR-linked cat gene expression in ACE2-expressing target cells upon membrane fusion. Consistent with the membrane fusion determined by the Cre-loxP system (Fig. 5B), the 1151P and 1179P mutant proteins effectively mediated membrane fusion as did the wt S protein, whereas the membrane fusion abilities of the other five mutants were significantly reduced (Fig. 5C). When the results from three individual experiments were averaged, the 1151P and 1179P mutants were found to slightly enhance membrane fusion whereas the 913P, 927P, 941P, 955P, and 1165P mutants strikingly decreased membrane fusion, compared to the wt S protein (Fig. 5D).

Incorporation of the HR1 and HR2 mutant proteins into the HIV-1 pseudotypes.

The SARS-CoV S pseudotypes based on the vector particles of VSV, murine leukemia virus, HIV, and SIV (19, 21, 24, 30, 35, 36, 41, 59) have been demonstrated, and infection with these pseudotypes is restricted to target cells expressing ACE2 (30, 31, 36). These retroviral and VSV pseudotyping systems therefore provide a quantitative assay for monitoring SARS-CoV S protein-mediated viral infection, while concomitantly avoiding the need for specialized biosafety containment for handling the live SARS-CoV. Since all of the mutant S proteins are expressed on the cell surface, we then determined whether these mutant S proteins could be assembled and incorporated into HIV-1-like particles using an env-defective, luciferase gene-containing HIV-1 reporter viral expression system. 293T cells were cotransfected with pNL4-3R⁻E⁻Luc (15) and each of the wt and mutant S plasmids. Forty-eight hours after transfection, lysates of cells and virions were subjected to SDS-PAGE, followed by Western blotting using MAb 183, which specifically recognizes HIV-1 capsid protein p24, and rabbit anti-S. Similar amounts of the Gag precursor, Pr55, and its cleaved products, p41 and p24, were detected in cells as well as virions obtained from each transfection (Fig. 6A). Consistent with the Western blotting results for cell lysates obtained 48 h after transfection (Fig. 2B), the slower-migrating S_H form of the wt S protein was present at a level higher than that of the faster-migrating S_L form (Fig. 6A, left panel, lane 2). With the exception of the 913P mutant, all mutant S constructs produced comparable or even slightly greater amounts of the cell-associated S_H form than the wt construct (Fig. 6A, left panel). The S protein of the 913P mutant was present mainly in S_L, with S_H expressed at a level lower than those in the wt and other mutants (Fig. 6A, left panel), confirming that the processing of S_L to S_H in the 913P mutant is impaired. When virions were analyzed for S protein incorporation, the S_H forms of all mutants except for the 913P mutant were sufficiently incorporated into the virions compared to wt S_H (Fig. 6A, right panel). Moreover, the S_L forms of the wt and all mutants were less efficiently incorporated into the virions compared to their S_H counterparts (Fig. 6A, right panel).

Binding of pseudotypes to ACE2-expressing cells.

To further determine the specific interactions of mutant proteins with ACE2, binding of wt and mutant pseudotypes to ACE2-expressing 293T cells was determined by fluorescence-activated cell sorter analysis using rabbit anti-S. With the exception of the 913P mutant, which did not show effective ACE2 binding, mutant protein levels similar to or even greater than that of the wt S protein bound to the surface of ACE2-expressing target cells (Fig. 6B), indicating that mutations in HR1 and HR2 motifs do not affect the binding of the S protein present on the viral envelope to ACE2.

Effects of mutations in the HR1 and HR2 motifs on viral entry.

Since all of the mutant S_H proteins, except the 913P mutant, were effectively incorporated into the HIV-1-like viral particles, we then determined whether these mutant S proteins were functional for viral entry by an env trans-complementation assay based on pNL4-3R⁻E⁻Luc, which measures viral replication in a single-cycle mode. Pseudotyped viruses produced from each of the four HR1 mutants expressed significantly reduced luciferase activity upon challenge to target Vero E6 cells compared to the wt S pseudotype (Fig. 7A). Among the HR2 mutants, the 1151P mutant mediated one-cycle viral transmission as effectively as did the wt S protein, whereas the 1179P mutant still mediated a low degree of viral entry (Fig. 7A). In contrast, viral entry mediated by the 1165P mutant was significantly impaired (Fig. 7A).

To confirm that the 1179P mutant still mediates a low level of viral entry activity, the pseudotyping system based on another env-deficient, cat gene-containing pHXBΔBglCAT reporter virus (22) was employed. As shown in a representative analysis, the 1179P mutant still mediated one-cycle viral infectivity compared to the other HR1 and HR2 mutants; nevertheless, viral infectivity mediated by this mutant was strikingly reduced compared to that mediated by the wt S protein (Fig. 7B). Quantitation from three individual studies clearly demonstrated that virus-to-cell transmission mediated by the 1179P mutant S protein was severely impaired but not completely blocked (Fig. 7C).