Previous Article | Next Article ![]()
Journal of Virology, May 2005, p. 5900-5906, Vol. 79, No. 10
0022-538X/05/$08.00+0 doi:10.1128/JVI.79.10.5900-5906.2005
Copyright © 2005, American Society for Microbiology. All Rights Reserved.
Department of Cancer Immunology and AIDS, Dana-Farber Cancer Institute, Department of Medicine,1 Partners AIDS Research Center, Brigham and Women's Hospital, Department of Medicine (Microbiology and Molecular Genetics), Harvard Medical School, Boston, Massachusetts 02115,2 Laboratory of Infectious Diseases, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, Maryland 208923
Received 29 July 2004/ Accepted 20 December 2004
|
|
|---|
|
|
|---|
SARS-CoV is readily transmissible by close contact within households and the health care environment. Evidence of airborne transmission is now well documented (17, 34). However, in numerous human cases the exact mode of transmission cannot be established (29) and the reasons for heterogeneity of transmission, particularly superspreader events, remain unknown (13, 22). The role of wild animals as reservoirs for SARS-CoV has been suggested by the detection of SARS-like coronaviruses (SARS-like-CoVs) in the Himalayan palm civet (civet cat) and raccoon dog, which were tested from markets selling wild animals for human consumption. In addition, antibody studies in people working in these markets have shown that some have had prior infection with SARS-like-CoV (5). Moreover, a recent study confirmed that all new human cases in the resurgence of SARS in 2003/2004 in Guangdong were caused by independent and multiple interspecies transmissions from animals to humans (4). Although the mass culling of civet cats in Guangdong likely provided a temporary break in this chain, the virus reservoir has almost certainly not been eliminated.
Currently, prevention of SARS has largely relied on improved awareness, surveillance, and institution of local, regional and international public health care measures (23). Significant efforts in the area of SARS vaccine research have been initiated, and several recent reports have documented that transfer of immune serum from mice with prior SARS-CoV infection, or from mice vaccinated with a DNA plasmid encoding SARS S protein or a vaccinia virus expressing the S protein, can prevent virus replication in the lungs and upper respiratory tract (1, 24, 31). In addition, in SARS-CoV infection of humans, decreasing virus titers from nasopharyngeal aspirates, serum, urine, and stool have been observed to be coincident with the development of neutralizing antibodies (9, 19). Treatment of SARS with convalescent plasma has been reported (2, 27). These studies support the importance of humoral immunity in protection against SARS-CoV and suggest that a specific and effective human monoclonal antibody (MAb) should be developed to provide a prophylaxis and early treatment against SARS in the event that episodic or even widespread reemergence into the human population occurs.
We have recently isolated a high-affinity recombinant human MAb (80R) against the S1 domain of the SARS-CoV S protein, which acts as a viral entry inhibitor (25). Through blocking the association of S protein to its receptor ACE2 (11), 80R potently neutralizes SARS-CoV infection in vitro. In the present study, we further investigated the prophylactic effectiveness of 80R immunoglobulin G1 (IgG1) in vivo in a mouse model of SARS. We also defined the unique features of the 80R neutralizing epitope, and an extended panel of 80R sensitivity and resistance markers has been established that will serve as a useful tool to guide the potential prophylactic use of this antibody in an outbreak setting.
|
|
|---|
Production of 80R scFv and whole human IgG1. 80R single-chain antibody (scFv) and IgG1 were expressed and purified as previously described (25). Briefly, six-His-tagged 80R scFv was expressed in Escherichia coli XL1-Blue (Stratagene) in the prokaryotic expression vector pSyn1 and purified from the periplasmic fraction by immobilized metal affinity chromatography. Whole human IgG1 of 80R was expressed in 293T cells by transient transfection and purified by protein A-Sepharose affinity chromatography. The S1-binding activities of purified soluble 80R scFv and IgG1 of 80R were confirmed by enzyme-linked immunosorbent assay.
Construction of full-length spike, S1-Ig, truncation variants, and mutants. The plasmid encoding a codon-optimized form of the S1 gene (residues 12 to 672 were defined as the S1 domain of the SARS-CoV S protein), fused with the Fc portion of human IgG1 (S1-Ig), was previously described (10, 11, 26). Plasmids encoding residues 318 to 510, 327 to 510, 318 to 490, and other truncation variants of S1 were generated by PCR using S1-Ig as a template. Mutations within S1-Ig or within S1(318-510)-Ig were generated by site-directed mutagenesis using the QuikChange method (Stratagene). S1 or full-length spike genes of SARS-CoV Tor2 (accession number AY274119), GD03T (AY525636) and SARS-like-CoV SZ3 (AY304486) were generated de novo by recursive PCR. Full-length spike proteins for immunoprecipitation were fused with a carboxyl-terminal nine-amino-acid (C9) tag (11, 15). All variants and mutations were confirmed by DNA sequencing.
Radioimmunoprecipitation of S1, truncation variants, and mutations. S1-Ig, S1(318-510)-Ig, and other truncations and mutations were expressed in 293T cells, secreted in culture supernatants, and metabolically labeled for 24 h with [35S]cysteine and [35S]methionine (NEN Life Science). One microgram of 80R scFv was conjugated to 20 µl of anti-His6 agarose beads in PBS buffer by incubating 2 h at 4°C and followed by washing two times with 1 ml of PBS. The 80R scFv beads were used to precipitate Fc-tagged S1 or its derivates. Five hundred microliters of culture supernatants which contains metabolically labeled S1 or its derivates was incubated with beads for 4 h at 4°C, and then beads were washed three times with PBS containing 0.25% Nonidet P-40. Full-length S protein was also expressed in 293T cells and metabolically labeled, and cells were lysed with 1% 3-[(3-cholamidopropyl)dimethylammonio]-2-hydroxy-1-propanesulfonate (CHAPSO)/PBS. One microgram of 80R IgG1 or anti-C9 antibody 1D4 was used to precipitate S protein from cellular lysate. Bound proteins were eluted in reducing Laemmli sample buffer at 100°C for 5 min. Proteins were subjected to 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis, visualized by phosphorimaging, and quantified using ImageQuant software.
80R IgG1 inhibition of infection by S protein-pseudotyped viruses.
S protein-pseudotyped lentiviruses expressing a luciferase reporter gene were produced as described previously (16, 32). Briefly, 293T cells were cotransfected with a plasmid encoding S protein variants with a modified carboxyl-terminal ht2 (15), a plasmid pCMV
R8.2 encoding human immunodeficiency virus type 1 (HIV-1) Gag-Pol, and a plasmid (pHIV-Luc) encoding the firefly luciferase reporter gene under control of the HIV-1 long terminal repeat. Forty-eight hours posttransfection, viral supernatants were harvested and 5 µl of S protein-pseudotyped virus was used for infection of 6,000 ACE2-expressing 293T cells in a 96-well plate (human ACE2-expression cells for Tor2, GD03T; civet ACE2-expression cells for SZ3). Infection efficiency was quantitated by measuring the luciferase activity in the target cells with an EG&G Berthold Microplate Luminometer (LB 96V).
|
|
|---|
12.5 mg/kg of body weight), 4/4 mice had a more-than-4-log reduction in viral load (to below the assay limit), whereas the equivalent amount of human IgG1 had no effect. At a dose of 50 µg/mouse, 1 of 4 mice again showed a viral load reduction to below the limit of detection and 3 of 4 mice showed a nearly 4-log reduction in viral titer. At the lowest dose (1:25 dilution, 10 µg/mouse), 4/4 mice became infected and the virus load was reduced about 10-fold. This excellent level of protection is comparable to that seen when the animals were injected with convalescent-phase sera from previously infected mice (24). These results indicate that the prophylactic administration of 80R can efficiently protect mice from SARS-CoV infection. |
View this table: [in a new window] |
TABLE 1. Protection from virus replication in the lower respiratory tracts of mice following passive transfer of anti-SARS MAb 80R
|
![]() View larger version (34K): [in a new window] |
FIG. 1. Truncations and point mutations of S1(318-510) were analyzed to define the 80R antibody epitope. S1 residues 318 to 510 fused to the Fc region of human IgG1 and truncation or mutation variants of S1(318-510) containing the indicated residues were metabolically labeled and precipitated by protein A or 80R scFv. (a) S1(324-503) was the smallest fragment bound to 80R. Either N-terminal or C-terminal truncation variants slightly smaller than that either had decreased expression or lost binding activity to 80R. (b) Critical residues for the 80R epitope were observed. Individual alanine substitution of glutamic acid 452 or aspartic acids 454 and 480 in the S1(318-510) fragment impaired or abolished binding to 80R. IP, immunoprecipitation.
|
10% and 100% of wild-type S1(318-510)-Ig by 80R scFv, respectively. D454A and D480A variants were not detectable on precipitation with 80R scFv. The same result was obtained for variants in S1-Ig (data not shown). These data demonstrated that E452 and D454 contribute to the association of S protein with 80R antibody similarly to that of S1 with ACE2. D463 does not affect 80R binding to S1 but affects the association of ACE2 to S1. D480 plays no significant role in S protein association with ACE2 but is critically important for 80R-antibody binding.
80R-binding characteristics of variant S proteins with amino acid substitutions in the 80R-binding domain that occur in SARS-like-CoV from civet cats and that evolved during the 2002/2003 outbreak and in a 2003/2004 Guangdong index patient.
The S1 domains of spike proteins contain neutralizing epitopes for most known coronaviruses and therefore are likely to encode determinants for host tropism, neutralizing antibody, and viral virulence. Recent molecular epidemiological studies of SARS-CoVs have shown that amino acid changes are mainly located in the S1 region, which contains its receptor (ACE2)-binding domain (3). We compared 78 reported human SARS-CoV sequences and SARS-like-CoV sequences from the Himalayan palm civets for amino acid sequence variations in the ACE2- and 80R-binding domains (residues 318 to 510) and summarized all variations in Table 2 (3, 5, 30, 33). A total of six differences are observed at positions 344, 360, 472, 479, 480, and 487 in this region. Among the human SARS-CoVs isolated from the 2002/2003 epidemic, the amino acid arginine or lysine was present at position 344 in the early isolates (K344K/R) but lysine was present in the middle and late isolates (3). The amino acids at positions 472 and 480 were lysine and aspartic acid in all reported SARS-like-CoV S sequences and human SARS-CoV S sequences but were proline and glycine in the S sequence of the 2003/2004 Guangdong index patient (GD03T0013, referred as GD03T) (L472P and D480G). Major amino acid variations in the spike between human SARS-CoVs of 2002/2003 and SARS-like-CoVs are located at positions 360, 479, and 487 (F360S, N479K, T487S, SARS-CoV/SARS-like-CoV). The above sequence variations in the 80R antibody-binding domain could be associated with significant changes in the binding activity of 80R. The 80R antibody was originally screened against the S1 domain of S protein of a late-phase human SARS-CoV isolate (Tor2, NC_004718). For substitution analysis, each amino acid in Tor2 was individually replaced with a corresponding changed amino acid in order to examine the effect of these residues on 80R antibody binding. For example, K (Tor2) at position 344 was replaced with R (Civet SARS-like-CoV, isolate SZ3). We also replaced each amino acid with alanine to investigate whether these residues normally contribute to form the 80R epitope. As shown in Fig. 2a, no effect on the binding of 80R was found for the variants of either F360S and L472P or F360A and L472A in S1(318-510) and no significant changes in binding were observed with the variants K344R and T487S. However, mutation to alanine with the variants R344A and T487A resulted in
20% and
50% reductions in binding, respectively. At position 479, N479K substitution resulted in an
50% decrease in 80R binding, whereas N479A substitution resulted in only
20% reduction. These results implied that lysine 344, asparagine 479, and threonine 487 normally contribute to some degree to the binding of 80R to the S1 domain, either by forming part of the 80R-binding site or by facilitating correct folding of the protein. As shown in Fig. 1b and 2a, D480A substitution completely abolished binding to 80R. The same result was observed when aspartic acid was changed to glycine, which is found in the sequence of the 2003/2004 Guangdong index patient GD03T.
|
View this table: [in a new window] |
TABLE 2. Amino acid changes in SARS-like-CoVs and SARS-CoVs from human cases
|
![]() View larger version (36K): [in a new window] |
FIG. 2. Effects on 80R binding of variant amino acid substitutions of S protein that occur in animal SARS-like-CoVs and human SARS-CoVs. (a) Indicated amino acid residues in S1(318-510) of Tor2 were individually replaced with corresponding variant amino acids found in SARS-like-CoVs or other human SARS-CoVs. These residues also were replaced with alanine. Alterations of K344A, N479A, and T487A affected the binding to 80R to some degree. N479K substitution resulted in an 50% decrease in 80R binding, and D480G substitution totally abolished binding to 80R. (b) Multiple substitutions with the amino acids of civet SZ3 virus (344R/360S/479K/487S) in the S1(318-510)-Ig construct of Tor2 had no effect on 80R binding, as well as full-length S1(12-672) of SZ3, which was de novo synthesized. Multiple substitutions with the amino acids of human GD03T virus (344R/360S/472P/480G/487S) in the S1(318-510)-Ig construct of Tor2 and full-length S1(12-672) of GD03T completely lost binding to 80R scFv. (c) Full-length S protein of Tor2 and variants containing amino acid substitutions of isolates SZ3 or GD03T were precipitated by 1D4, which recognizes a C9 tag present at the carboxyl terminus of each S protein, or by 80R IgG1 and analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. The binding activities of these full-length S proteins to 80R IgG1 were consistent with those of their receptor-binding domains (318 to 510) or S1 domains (12 to 672) to 80R scFv. WT, wild type; IP, immunoprecipitation.
|
80R IgG1 neutralization of pseudoviral infection mediated by full-length S protein variants. The neutralization activity of 80R IgG1 against variant SARS-CoV was evaluated by using a pseudovirus system. Single-round HIV luciferase reporter viruses pseudotyped by S protein variants of Tor2, SZ3, or GD03T were tested for neutralization sensitivity to increasing concentrations of 80R IgG1 or nonrelevant human IgG1. As expected, Tor2 is very sensitive to neutralization of 80R IgG1, with a 90% inhibitory concentration around 2 µg/ml (Fig. 3a); 80R IgG1 could also efficiently block SZ3 pseudoviral infection (Fig. 3b); in contrast, GD03T is essentially resistant to neutralization by 80R IgG1 in the concentration range assayed (Fig. 3c).
![]() View larger version (10K): [in a new window] |
FIG. 3. 80R IgG1 neutralization of pseudoviral infection mediated by full-length SARS-CoV spike variants. HIVs pseudotyped with the S protein from Tor2, SZ3, or GD03T isolates were incubated with the indicated concentration of 80R IgG1 (solid line, diamonds) or nonrelevant human IgG1 (solid line, squares) for 1 h prior to infection. At 48 h after infection, luciferase activities in target cells were measured and relative viral inhibition was calculated as the ratio of luciferase activity in the presence to absence of 80R IgG1 or nonrelevant human IgG1. (a) 80R IgG1 efficiently blocked Tor2 S protein-pseudotyped HIV infection, with a 90% inhibitory concentration around 2 µg/ml. (b) 80R IgG1 also efficiently neutralized SZ3 S protein pseudoviral infection. (c) In contrast to Tor2 and SZ3, GD03T S protein-pseudotyped virus was essentially resistant to the neutralization of 80R IgG1 with a concentration up to 50 µg/ml. These results are representative of two experiments with similar results.
|
|
|
|---|
One surprise of this study was the absolute conformational dependence of the neutralizing epitope of 80R, since we showed previously that 80R can recognize the denatured S1 protein (25). However, binding of 80R was not detected when a complete series of 18-mer S1 peptides with 10-amino-acid overlap, or some larger linear and conformationally constrained peptides that centered around essential residues for 80R binding, were examined (data not shown). The 80R neutralizing epitope was shown to be located in the same region as the ACE2 receptor-binding domain (residues 318 to 510). In addition, the 80R epitope could be further reduced but minimally required a 180-amino-acid fragment lying between 324 and 503 of the S1 domain. Interestingly, Zeng et al. recently showed that a plasmid DNA vaccine encoding residues 18 to 495 elicited high antibody titers in mice but no SARS-CoV neutralizing activity (35). Our data are consistent with this observation, since the carboxy-terminal deletion prior to amino acid 503 in S1 likely failed to maintain the conformationally sensitive neutralizing epitope. Thus, this 180-amino-acid region encompassing the 80R epitope can now be considered an essential core region of the S protein for subunit vaccines aimed at eliciting potent neutralizing antibody responses. Whether other neutralizing epitopes exist on the S1 protein remains to be determined.
Some acidic residues previously analyzed for ACE2 receptor binding were tested by alanine substitution for their effect on the binding of 80R in this study. Glutamic acid 452 and aspartic acid 454 contribute to both the association of S protein with 80R and the association of S1 with ACE2. Aspartic acid 463 reduces the association of ACE2 with S1 but has no effect in 80R binding to S1. Substitution of aspartic acid 480 abolishes the binding to 80R but has no effect on the binding to ACE2. These data strongly suggest that the S determinants involved in the binding of receptor and of neutralizing antibodies are in part overlapping and are likely to result from both common and unique contact residues. Further studies of 80R antibody binding through an analysis of amino acid variants between civet SARS-like CoVs and human SARS-CoVs found that lysine 344, asparagine 479, and threonine 487 contribute to the binding of 80R with the S1 domain. Asparagine 479 of human SARS-CoV S1 changed to lysine of civet SARS-like-CoV S1 resulted in partial resistance to 80R antibody. However, this resistance was not seen in the context of the complete civet SZ3 S protein. Glycine at position 480 of S1 of SARS-CoV isolated from the 2003/2004 Guangdong index patient conferred complete resistance to 80R. Critical amino acids for the 80R epitope were not found to vary among all the human SARS-CoVs from the 2002/2003 epidemic. Therefore, the vast majority of SARS-CoVs, including the well-adapted and more pathogenic human SARS-CoVs from the 2002/2003 epidemic and SARS-like CoVs isolated to date, are likely to be sensitive to 80R.
For passive immunotherapy to be effective in a SARS outbreak, 80R IgG1 would have to be active against a plethora of viral isolates that may emerge. We propose that testing for susceptibility to 80R neutralization can be managed in an outbreak by rapid and early genotyping of the DNA fragment encoding the 180 amino acids of the 80R epitope from the outbreak strains. For example, in some outbreaks, such as the one which recently occurred as a result of a laboratory exposure at the National Institute of Virology in Beijing, it should be possible to predict the susceptibility of the SARS-CoV to 80R by comparing the amino acid sequences of the S1 protein of the laboratory strain with the 80R neutralizing epitope elucidated by this study. However, in a natural reemergence such as that which occurred in 2003/2004 in Guangdong, it would not be possible to know a priori whether the new viruses would be susceptible to 80R. The resistance to 80R neutralization of the SARS-CoV from the 2003/2004 Guangdong index patient because of the D480G mutation provides one such sobering example. Although the GD03T virus appeared to be only weakly pathogenic and no secondary infections with SARS-CoVs of this genotype have been reported, other mutations that have different properties on viral tropism and replication kinetics may emerge in future outbreaks. Variations in the genomes of SARS-like CoVs in the animal reservoirs will likely continue to occur because of the high mutation rate of RNA viruses. Indeed, recent surveillance of the known animal reservoirs of SARS-like-CoVs have confirmed that the 2003/2004 viruses are different from those that circulated in animals and humans in the outbreak of 2002/2003 (4). Thus, our study cautions that any prophylaxis strategy based on neutralizing antibody responses, whether by passive immunotherapy or active immunization, should be monitored closely by rapid and early genotyping of the target neutralizing epitope.
At present, management of potentially SARS-CoV-exposed individuals is still based on observation and quarantine, even on a large scale (http://www.who.int/csr/don/2004_04_28/en/). We propose that 80R IgG1 may be useful for emergency prophylaxis of potentially SARS-CoV-exposed individuals by treatment with a single intravenous or intramuscular dose (half-life of IgG1, circa 21 days) sufficient to achieve a serum level that would result in
99% reduction in the expected virus titers in the serum or tissue of humans or animals based on previously published studies (18, 20).
The results of these 80R studies also provide further validation of the results from several recent vaccine studies that have demonstrated the importance of the spike protein in eliciting neutralizing antibody responses (1, 31). However, the pros and cons of passive immunoprophylaxis versus vaccination requires additional consideration. Since heightened awareness and public health care measures that are now in place, it is uncertain if a worldwide epidemic or even a regional epidemic of the magnitude seen in 2002/2003 will reoccur (13). Therefore, the cost effectiveness of vaccinating a susceptible population with a low incidence of infection versus employing passive immunotherapy in an outbreak setting must be considered.
In summary, we propose that 80R IgG1 may be useful for the emergency immunoprophylaxis of SARS. The vast majority of SARS-CoVs isolated so far remain sensitive to 80R. In an outbreak setting, early and rapid genotyping of the 540-bp S1 gene fragment encoding the 80R epitope from index cases may provide an accurate guide to susceptibility or resistance to 80R prophylaxis. Our study also cautions that for any prophylaxis strategy based on neutralizing antibody responses, whether by passive or active immunization, a genotyping monitor will be necessary for effective use. Therefore, there should be ongoing surveillance in place to monitor closely, on a regular basis, the SARS-like-CoV genotypes in the known animal reservoirs, as is now being routinely performed with influenza virus (7).
|
|
|---|
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Copyright © 2009 by the American Society for Microbiology. For an alternate route to Journals.ASM.org, visit: http://intl-journals.asm.org | More Info»